.\"Copyright (C) 2011, Kaz Kylheku . .\"All rights reserved. .\" .\"BSD License: .\" .\"Redistribution and use in source and binary forms, with or without .\"modification, are permitted provided that the following conditions .\"are met: .\" .\" 1. Redistributions of source code must retain the above copyright .\" notice, this list of conditions and the following disclaimer. .\" 2. Redistributions in binary form must reproduce the above copyright .\" notice, this list of conditions and the following disclaimer in .\" the documentation and/or other materials provided with the .\" distribution. .\" 3. The name of the author may not be used to endorse or promote .\" products derived from this software without specific prior .\" written permission. .\" .\"THIS SOFTWARE IS PROVIDED ``AS IS'' AND WITHOUT ANY EXPRESS OR .\"IMPLIED WARRANTIES, INCLUDING, WITHOUT LIMITATION, THE IMPLIED .\"WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE. .TH "TXR" 1 2012-04-20 "Utility Commands" "Txr Text Processing Language" "Kaz Kylheku" .SH NAME txr \- text processing language (version 65) .SH SYNOPSIS .B txr [ options ] query-file { data-file }* .sp .SH DESCRIPTION .B TXR is a language oriented toward processing text from files or streams, using multiple programming paradigms. A .B TXR script is called a query, and it specifies a pattern which matches (a prefix of) an entire file, or multiple files. Patterns can consists of large chunks of multi-line freeform text, which is matched literally against material in the input sources. Free variables occurring in the pattern (denoted by the @ symbol) are bound to the pieces of text occurring in the corresponding positions. If the overall match is successful, then .B TXR can do one of two things: it can report the list of variables which were bound, in the form of a set of variable assignments which can be evaluated by the .B eval command of the POSIX shell language, or generate a custom report according to special directives in the query. Patterns can be arbitrarily complex, and can be broken down into named pattern functions, which may be mutually recursive. TXR patterns can work horizontally (characters within a line) or vertically (spanning multiple lines). Multiple lines can be treated as a single line. In addition to embedded variables which implicitly match text, the .B TXR query language supports a number of directives, for matching text using regular expressions, for continuing a match in another file, for searching through a file for the place where an entire sub-query matches, for collecting lists, and for combining sub-queries using logical conjunction, disjunction and negation, and numerous others. Furthermore, embedded within TXR is a powerful Lisp dialect. TXR Lisp supports functional and imperative programming, and provides data types such as symbols, strings, vectors, hash tables with weak reference support, lazy lists, and arbitrary-precision (bignum integers). .SH ARGUMENTS AND OPTIONS Options which don't take an argument may be combined together. The -v and -q options are mutually exclusive. Of these two, the one which occurs in the rightmost position in the argument list dominates. The -c and -f options are also mutually exclusive; if both are specified, it is a fatal error. .IP -Dvar=value Bind the variable .IR var to the value .IR value prior to processing the query. The name is in scope over the entire query, so that all occurrence of the variable are substituted and match the equivalent text. If the value contains commas, these are interpreted as separators, which give rise to a list value. For instance -Da,b,c creates a list of the strings "a", "b" and "c". (See Collect Directive bellow). List variables provide a multiple match. That is to say, if a list variable occurs in a query, a successful match occurs if any of its values matches the text. If more than one value matches the text, the first one is taken. .IP -Dvar Binds the variable .IR var to an empty string value prior to processing the query. .IP -q Quiet operation during matching. Certain error messages are not reported on the standard error device (but the if the situations occur, they still fail the query). This option does not suppress error generation during the parsing of the query, only during its execution. .IP -d Invoke the interactive txr debugger. See the DEBUGGER section. .IP -v Verbose operation. Detailed logging is enabled. .IP -b Suppresses the printing of variable bindings for a successful query, and the word . IR false for a failed query. The program still sets an appropriate termination status. Bindings are implicitly suppressed if the TXR query performs an output operation on any file stream other than *stddebug*. (Internal streams like string streams do not count as output.) .IP -B Force the printing of variable bindings for a successful query, and the word . IR false for a failed query, even if the program produced output. .IP "-l or --lisp-bindings" Print the variable bindings in Lisp syntax instead of shell syntax. .IP "-a num" The decimal integer argument specifies the maximum number of array dimensions to use for variables arising out of collect. The default is 1. Additional dimensions are expressed using numeric suffixes in the generated variable names. For instance, consider the three-dimensional list arising out of a triply nested collect: ((("a" "b") ("c" "d")) (("e" "f") ("g" "h"))). Suppose this is bound to a variable V. With -a 1, this will be reported as: V_0_0[0]="a" V_0_1[0]="b" V_1_0[0]="c" V_1_1[0]="d" V_0_0[1]="e" V_0_1[1]="f" V_1_0[1]="g" V_1_1[1]="h" The leftmost bracketed index is the most major index. That is to say, the dimension order is: NAME_m_m+1_..._n[1][2]...[m-1]. .IP "-c query" Specifies the query in the form of a command line argument. If this option is used, the query-file argument is omitted. The first non-option argument, if there is one, now specifies the first input source rather than a query. Unlike queries read from a file, (non-empty) queries specified as arguments using -c do not have to properly end in a newline. Internally, TXR adds the missing newline before parsing the query. Thus -c "@a" is a valid query which matches a line. Example: # read two lines "1" and "2" from standard input, # binding them to variables a and b. Standard # input is specified as - and the data # comes from shell "here document" redirection. txr -c "@a @b" - < twoline.txr #!/usr/bin/txr @a @b [Ctrl-D] $ chmod a+x twoline.txr $ ./twoline.txr - 1 2 [Ctrl-D] a=1 b=2 A script written in this manner will not pass options to txr. For instance, if the above script is invoked like this ./twoline.txr -Da=42 the -D option isn't passed down to txr; -Da=42 is an ordinary argument (which the script will try to open as an input file). This behavior is useful if the script author wants not to expose the txr options to the user of the script. However, if the hash bang line can use the -f option: #!/usr/bin/txr -f Now, the name of the script is passed as an argument to the -f option, and TXR will look for more options after that. .SS Whitespace Outside of directives, whitespace is significant in TXR queries, and represents a pattern match for whitespace in the input. An extent of text consisting of an undivided mixture of tabs and spaces is a whitespace token. Whitespace tokens match a precisely identical piece of whitespace in the input, with one exception: a whitespace token consisting of precisely one space has a special meaning. It is equivalent to the regular expression @/[ ]+/: match an extent of one or more spaces (but not tabs!) Thus, the query line "a b" (one space) matches texts like "a b", "a b", et cetera (arbitrary number of tabs and spaces between a and b). However "a b" (two spaces) matches only "a b" (two spaces). For matching a single space, the syntax @\ can be used (backslash-escaped space). It is more often necessary to match multiple spaces, than to exactly match one space, so this rule simplifies many queries and adds inconvenience to only few. In output clauses, string and character literals and quasiliterals, a space token denotes a space. .SS Text Query material which is not escaped by the special character @ is literal text, which matches input character for character. Text which occurs at the beginning of a line matches the beginning of a line. Text which starts in the middle of a line, other than following a variable, must match exactly at the current position, where the previous match left off. Moreover, if the text is the last element in the line, its match is anchored to the end of the line. An empty query line matches an empty line in the input. Note that an empty input stream does not contain any lines, and therefore is not matched by an empty line. An empty line in the input is represented by a newline character which is either the first character of the file, or follows a previous newline-terminated line. Input streams which end without terminating their last line with a newline are tolerated, and are treated as if they had the terminator. Text which follows a variable has special semantics, discussed in the section Variables below. A query may not leave unmatched material in a line which is covered by the query. However, a query may leave unmatched lines. In the following example, the query matches the text, even though the text has an extra line. Query: Four score and seven years ago our Text: Four score and seven years ago our forefathers In the following example, the query .B fails to match the text, because the text has extra material on one line. Query: I can carry nearly eighty gigs in my head Text: I can carry nearly eighty gigs of data in my head Needless to say, if the text has insufficient material relative to the query, that is a failure also. To match arbitrary material from the current position to the end of a line, the "match any sequence of characters, including empty" regular expression @/.*/ can be used. Example: Query: I can carry nearly eighty gigs@/.*/ Text: I can carry nearly eighty gigs of data In this example, the query matches, since the regular expression matches the string "of data". (See Regular Expressions section below). Another way to do this is: Query: I can carry nearly eighty gigs@(skip) .SS Special Characters in Text Control characters may be embedded directly in a query (with the exception of newline characters). An alternative to embedding is to use escape syntax. The following escapes are supported: .IP @\e A backslash immediately followed by a newline introduces a physical line break without breaking up the logical line. Material following this sequence continues to be interpreted as a continuation of the previous line, so that indentation can be introduced to show the continuation without appearing in the data. .IP @\e A backslash followed by a space encodes a space. This is useful in line continuations when it is necessary for leading spaces to be preserved. For instance the two line sequence abcd@\ @\ efg is equivalent to the line abcd efg The two spaces before the @\ in the second line are consumed. The spaces after are preserved. .IP @\ea Alert character (ASCII 7, BEL). .IP @\eb Backspace (ASCII 8, BS). .IP @\et Horizontal tab (ASCII 9, HT). .IP @\en Line feed (ASCII 10, LF). Serves as abstract newline on POSIX systems. .IP @\ev Vertical tab (ASCII 11, VT). .IP @\ef Form feed (ASCII 12, FF). This character clears the screen on many kinds of terminals, or ejects a page of text from a line printer. .IP @\er Carriage return (ASCII 13, CR). .IP @\ee Escape (ASCII 27, ESC) .IP @\exHEX A @\ex followed by a sequence of hex digits is interpreted as a hexadecimal numeric character code. For instance @\ex41 is the ASCII character A. .IP @\eOCTAL A @\e followed by a sequence of octal digits (0 through 7) is interpreted as an octal character code. For instance @\e010 is character 8, same as @\eb. .PP Note that if a newline is embedded into a query line with @\en, this does not split the line into two; it's embedded into the line and thus cannot match anything. However, @\en may be useful in the @(cat) directive and in @(output). .SS International Characters .B TXR represents text internally using wide characters, which are used to represent Unicode code points. The query language, as well as all data sources, are assumed to be in the UTF-8 encoding. In the query language, extended characters can be used directly in comments, literal text, string literals, quasiliterals and regular expressions. Extended characters can also be expressed indirectly using hexadecimal or octal escapes. On some platforms, wide characters may be restricted to 16 bits, so that .B TXR can only work with characters in the BMP (Basic Multilingual Plane) subset of Unicode. .B TXR does not use the localization features of the system library; its handling of extended characters is not affected by environment variables like LANG and L_CTYPE. The program reads and writes only the UTF-8 encoding. If .B TXR encounters an invalid bytes in the UTF-8 input, what happens depends on the context in which this occurs. In a query, comments are read without regard for encoding, so invalid encoding bytes in comments are not detected. A comment is simply a sequence of bytes terminated by a newline. In lexical elements which represent text, such as string literals, invalid or unexpected encoding bytes are treated as syntax errors. The scanner issues an error message, then discards a byte and resumes scanning. Certain sequences pass through the scanner without triggering an error, namely some UTF-8 overlong sequences. These are caught when when the lexeme is subject to UTF-8 decoding, and treated in the same manner as other UTF-8 data, described in the following paragraph. Invalid bytes in data are treated as follows. When an invalid byte is encountered in the middle of a multibyte character, or if the input ends in the middle of a multibyte character, or if a character is extracted which is encoded as an overlong form, the UTF-8 decoder returns to the starting byte of the ill-formed multibyte character, and extracts just that byte, mapping it to the Unicode character range U+DC00 through U+DCFF. The decoding resumes afresh at the following byte, expecting that byte to be the start of a UTF-8 code. .SS Regular Expression Directives In place of a piece of text (see section Text above), a regular expression directive may be used, which has the following syntax: @/RE/ where the RE part enclosed in slashes represents regular expression syntax (described in the section Regular Expressions below). Long regular expressions can be broken into multiple lines using a backslash-newline sequence. Whitespace before the sequence or after the sequence is not significant, so the following two are equivalent: @/reg \e ular/ @/regular/ There may not be whitespace between the backslash and newline. Whereas literal text simply represents itself, regular expression denotes a (potentially infinite) set of texts. The regular expression directive matches the longest piece of text (possibly empty) which belongs to the set denoted by the regular expression. The match is anchored to the current position; thus if the directive is the first element of a line, the match is anchored to the start of a line. If the regular expression directive is the last element of a line, it is anchored to the end of the line also: the regular expression must match the text from the current position to the end of the line. Even if the regular expression matches the empty string, the match will fail if the input is empty, or has run out of data. For instance suppose the third line of the query is the regular expression @/.*/, but the input is a file which has only two lines. This will fail: the data has line for the regular expression to match. A line containing no characters is not the same thing as the absence of a line, even though both abstractions imply an absence of characters. Like text which follows a variable, a regular expression directive which follows a variable has special semantics, discussed in the section Variables below. .SS Variables Much of the query syntax consists of arbitrary text, which matches file data character for character. Embedded within the query may be variables and directives which are introduced by a @ character. Two consecutive @@ characters encode a literal @. A variable matching or substitution directive is written in one of several ways: @NAME @{NAME} @*NAME @*{NAME} @{NAME /RE/} @{NAME (FUN [ ARGS ... ])} @{NAME NUMBER} The forms with an * indicate a long match, see Longest Match below. The last two forms with the embedded regexp /RE/ or number have special semantics, see Positive Match below. The identifier t cannot be used as a name; it is a reserved symbol which denotes the value true. An attempt to use the variable @t will result in an exception. The symbol nil can be used as a variable name, but it has special semantics, described in a section below. When the @NAME form is used, the name itself may consist of any combination of one or more letters, numbers, and underscores. It may not look like a number, so that for instance 123 is not a valid name, but 12A is valid. Case is sensitive, so that @FOO is different from @foo, which is different from @Foo. The braces around a name can be used when material which follows would otherwise be interpreted as being part of the name. When a name is enclosed in braces, the following additional characters may be used as part of the name: ! $ % & * + - < = > ? \e ^ _ ~ The rule holds that a name cannot look like a number so +123 is not a name, but these are valid names: a->b, *xyz*, foo-bar. The syntax @FOO_bar introduces the name "FOO_bar", whereas @{FOO}_bar means the variable named "FOO" followed by the text "_bar". There may be whitespace between the @ and the name, or opening brace. Whitespace is also allowed in the interior of the braces. It is not significant. If a variable has no prior binding, then it specifies a match. The match is determined from some current position in the data: the character which immediately follows all that has been matched previously. If a variable occurs at the start of a line, it matches some text at the start of the line. If it occurs at the end of a line, it matches everything from the current position to the end of the line. .SS Negative Match If a variable is one of the plain forms @NAME, @{NAME}, @*NAME or @*{NAME}, then this is a "negative match". The extent of the matched text (the text bound to the variable) is determined by looking at what follows the variable, and ranges from the current position to some position where the following material finds a match. This is why this is called a "negative match": the spanned text which ends up bound to the variable is that in which the match for the trailing material did not occur. A variable may be followed by a piece of text, a regular expression directive, a function call, a directive, another variable, or nothing (i.e. occurs at the end of a line). These cases are discussed in detail below. .SS Variable Followed by Nothing If the variable is followed by nothing, the negative match extends from the current position in the data, to the end of the line. Example: pattern: "a b c @FOO" data: "a b c defghijk" result: FOO="defghijk" .SS Variable Followed by Text For the purposes of determining the negative match, text is defined as a sequence of literal text and regular expressions, not divided by a directive. So for instance in this example: @a:@/foo/bcd e@(maybe)f@(end) the variable @a is considered to be followed by ":@/foo/bcd e". If a variable is followed by text, then the extent of the negative match is determined by searching for the first occurrence of that text within the line, starting at the current position. The variable matches everything between the current position and the matching position (not including the matching position). Any whitespace which follows the variable (and is not enclosed inside braces that surround the variable name) is part of the text. For example: pattern: "a b @FOO e f" data: "a b c d e f" result: FOO="c d" In the above example, the pattern text "a b " matches the data "a b ". So when the @FOO variable is processed, the data being matched is the remaining "c d e f". The text which follows @FOO is " e f". This is found within the data "c d e f" at position 3 (counting from 0). So positions 0-2 ("c d") constitute the matching text which is bound to FOO. .SS Variable Followed by a Function Call or Directive If the variable is followed by a function call, or a directive, the extent is determined by scanning the text for the first position where a match occurs for the regular expression, call or directive. (For a description of functions, see FUNCTIONS.) Note that the given variable and the function or directive are considered in isolation. This means, for instance, that @var@(skip)text is a degenerate form. The @(skip) will be processed alone, without regard for the trailing text and so consume the input to the end of the line. The right way to express the most probable intent of this is @{var}text. Another degenerate case is @var@(bind ...), or in general, a variable followed by some directive not used for matching text. Watch out for the following pitfall: @a @b@(bind have_ab "y") The intent here is that the variable b captures everything after the space to the end of the line, and then the variable have_ab is set to "y". But since @(bind) always succeeds, b captures an empty string, and then the whole line fails if there is any material after the space. The right way to do this is: @a @b@(eol)@(bind have_ab "y") That is to say, match an explicit @(eol) after the variable. This will search for the end of the lne and capture the spanning text into b, as intended. The bind then happens afterward. .SS Consecutive Variables If an unbound variable specified a fixed-width match or a regular expression, then the issue of consecutive variables does not arise. Such a variable consumes text regardless of any context which follows it. However, what if an unbound variable with no modifier is followed by another variable? The behavior depends on the nature of the other variable. If the other variable also has no modifier, this is a semantic error which will cause the query to fail. A diagnostic message will be issued, unless operating in quiet mode via -q. The reason is that there is no way to bind two consecutive variables to an extent of text; this is an ambiguous situation, since there is no matching criterion for dividing the text between two variables. (In theory, a repetition of the same variable, like @FOO@FOO, could find a solution by dividing the match extent in half, which would work only in the case when it contains an even number of characters. This behavior seems to have dubious value). An unbound variable may be followed by one which is bound. The bound variable is replaced by the text which it denotes, and the logic proceeds accordingly. Variables are never bound to regular expressions, so the regular expression match does not arise in this case. The @* syntax for longest match is available. Example: pattern: "@FOO:@BAR@FOO" data: "xyz:defxyz" result: FOO=xyz, BAR=def Here, FOO is matched with "xyz", based on the delimiting around the colon. The colon in the pattern then matches the colon in the data, so that BAR is considered for matching against "defxyz". BAR is followed by FOO, which is already bound to "xyz". Thus "xyz" is located in the "defxyz" data following "def", and so BAR is bound to "def". If an unbound variable is followed by a variable which is bound to a list, or nested list, then each character string in the list is tried in turn to produce a match. The first match is taken. An unbound variable may be followed by another unbound variable which specifies a regular expression or function call match. This is a special case called a "double variable match". What happens is that the text is searched using the regular expression or function. If the search fails, than neither variable is bound: it is a matching failure. If the search succeeds, than the first variable is bound to the text which is skipped by the search. The second variable is bound to the text matched by the regular expression or function. Examples: pattern: "@foo@{bar /abc/}" data: "xyz@#abc" result: foo="xyz@#", BAR="abc" .SS Consecutive Variables Via Directive Two variables can be de-facto consecutive in a manner shown in the following example: @var1@(all)@var2@(end) The @(all) directive does nothing other than assert that all clauses must match. It has only one clause, @var2. So this is equivalent to just @var1@var2, except that if both variables are unbound, no semantic error is identified in this situation. The situation is handled as a variable followed by a directive. Of course @var2 matches any position current position, and so @var1 ends up with nothing. Example 1: b matches at position 0 and a gets nothing: pattern: "@a@(all)@b@(end)" data: "abc" result: a="" b="abc" Example 2: *a specifies longest match (see Longest Match below), and so a gets everything: pattern: "@*a@(all)@b@(end)" data: "abc" result: a="abc" b="" .SS Longest Match The closest-match behavior for the negative match can be overridden to longest match behavior. A special syntax is provided for this: an asterisk between the @ and the variable, e.g: pattern: "a @*{FOO}cd" data: "a b cdcdcdcd" result: FOO="b cdcdcd" pattern: "a @{FOO}cd" data: "a b cdcdcd" result: FOO="b " In the former example, the match extends to the rightmost occurrence of "cd", and so FOO receives "b cdcdcd". In the latter example, the * syntax isn't used, and so a leftmost match takes place. The extent covers only the "b ", stopping at the first "cd" occurrence. .SS Positive Match There are syntax variants of variable syntax which have an embedded expression enclosed with the variable in braces: @{NAME /RE/} @{NAME (FUN [ARGS ...])} @{NAME NUMBER} These specify a variable binding that is driven by a positive match derived from a regular expression, function or character count, rather than from trailing material (which is regarded as a "negative" match, since the variable is bound to material which is .B skipped in order to match the trailing material). In the /RE/ form, the match extends over all characters from the current position which match the regular expression RE. (see Regular Expressions section below). In the (FUN [ARGS ...]) form, the match extends over characters which are matched by the call to the function, if the call succeeds. Thus @{x (y z w)} is just like @(y z w), except that the region of text skipped over by @(y z w) is also bound to the variable x. See FUNCTIONS below. In the NUMBER form, the match processes a field of text which consists of the specified number of characters, which must be nonnegative number. If the data line doesn't have that many characters starting at the current position, the match fails. A match for zero characters produces an empty string. The text which is actually matched by this construct is all text within the specified field, but excluding leading and trailing whitespace. If the field contains only spaces, then an empty string is extracted. This syntax is processed without consideration of what other syntax follows. A positive match may be directly followed by an unbound variable. .SS Symbol nil as a Variable If the symbol nil is used as a variable, it behaves like a variable which has no binding. Furthermore, no binding is created. @nil allows the variable matching syntax to be used to skip material, in ways different from and complementary to @(skip). .SS Regular Expressions Regular expressions are a language for specifying sets of character strings. Through the use of pattern matching elements, regular expression is able to denote an infinite set of texts. .B TXR contains an original implementation of regular expressions, which supports the following syntax: .IP . (period) is a "wildcard" that matches any character. .IP [] Character class: matches a single character, from the set specified by special syntax written between the square brackets. Supports basic regexp character class syntax; no POSIX notation like [:digit:]. The regex tokens \es, \ed and \ew are permitted in character classes, but not their complementing counterparts. These tokens simply contribute their characters to the class. The class [a-zA-Z] means match an uppercase or lowercase letter; the class [0-9a-f] means match a digit or a lowercase letter; the class [^0-9] means match a non-digit, et cetera. The class [\ed.] means match a digit or the period character. A ] or - can be used within a character class, but must be escaped with a backslash. A ^ in the first position denotes a complemented class, unless it is escaped by backslash. In any other position, it denotes itself. Two backslashes code for one backslash. So for instance [\e[\e-] means match a [ or - character, [^^] means match any character other than ^, and [\e^\e\e] means match either a ^ or a backslash. Regex operators such as *, + and & appearing in a character class represent ordinary characters. The characters -, ] and ^ occurring outside of a character class are ordinary. Unescaped / characters can appear within a character class. The empty character class [] matches no character at all, and its complement [^] matches any character, and is treated as a synonym for the . (period) wildcard operator. .IP "\es, \ew and \ed" These regex tokens each match a single character. The \es regex token matches a wide variety of ASCII whitespace characters and Unicode spaces. The \ew token matches alphabetic word characters; it is equivalent to the character class [A-Za-z_]. The \ed token matches a digit, and is equivalent to [0-9]. .IP "\eS, \eW and \eD" These regex tokens are the complemented counterparts of \es, \ew and \ed. The \eS token matches all those characters which \es does not match, \eW matches all characters that \ew does not match and \eD matches nondigits. .IP empty An empty expression is a regular expression. It represents the set of strings consisting of the empty string; i.e. it matches just the empty string. The empty regex can appear alone as a full regular expression (for instance the .B TXR syntax @// with nothing between the slashes) and can also be passed as a subexpression to operators, though this may require the use of parentheses to make the empty regex explicit. For example, the expression a| means: match either a, or nothing. The forms * and (*) are syntax errors; though not useful, the correct way to match the empty expression zero or more times is the syntax ()*. .IP nomatch The nomatch regular expression represents the empty set: it matches no strings at all, not even the empty string. There is no dedicated syntax to directly express nomatch in the regex language. However, the empty character class [] is equivalent to nomatch, and may be considered to be a notation for it. Other representations of nomatch are possible: for instance, the regex ~.* which is the complement of the regex that denotes the set of all possible strings, and thus denotes the empty set. A nomatch has uses; for instance, it can be used to temporarily "comment out" regular expressions. The regex ([]abc|xyz) is equivalent to (xyz), since the []abc branch cannot match anything. Using [] to "block" a subexpression allows you to leave it in place, then enable it later by removing the "block". .IP (R) If R is a regular expression, then so is (R). The contents of parentheses denote one regular expression unit, so that for instance in (RE)*, the * operator applies to the entire parenthesized group. The syntax () is valid and equivalent to the empty regular expression. .IP R? optionally match the preceding regular expression R. .IP R* match the expression R zero or more times. This operator is sometimes called the "Kleene star", or "Kleene closure". The Kleene closure favors the longest match. Roughly speaking, if there are two or more ways in which R1*R2 can match, than that match occurs in which R1* matches the longest possible text. .IP R+ match the preceding expression R one or more times. Like R*, this favors the longest possible match: R+ is equivalent to RR*. .IP R1%R2 match R1 zero or more times, then match R2. If this match can occur in more than one way, then it occurs such that R1 is matched the fewest number of times, which is opposite from the behavior of R1*R2. Repetitions of R1 terminate at the earliest point in the text where a non-empty match for R2 occurs. Because it favors shorter matches, % is termed a non-greedy operator. If R2 is the empty expression, or equivalent to it, then R1%R2 reduces to R1*. So for instance (R%) is equivalent to (R*), since the missing right operand is interpreted as the empty regex. Note that whereas the expression (R1*R2) is equivalent to (R1*)R2, the expression (R1%R2) is .B not equivalent to (R1%)R2. .IP ~R match the opposite of the following expression R; i.e. match exactly those texts that R does not match. This operator is called complement, or logical not. .IP R1R2 Two consecutive regular expressions denote catenation: the left expression must match, and then the right. .IP R1|R2 match either the expression R1 or R2. This operator is known by a number of names: union, logical or, disjunction, branch, or alternative. .IP R1&R2 match both the expression R1 and R2 simultaneously; i.e. the matching text must be one of the texts which are in the intersection of the set of texts matched by R1 and the set matched by R2. This operator is called intersection, logical and, or conjunction. .PP Any escaped character which does not fall into the above escaping conventions, or any unescaped character which is not a regular expression operator, denotes one-position match of that character itself. Any of the special characters, including the delimiting /, can be escaped with a backslash to suppress its meaning and denote the character itself. Furthermore, all of the same escapes as are described in the section Special Characters in Text above are supported---the difference is that in regular expressions, the @ character is not required, so for example a tab is coded as \et rather than @\e\t. Octal and hex character escapes can be optionally terminated by a semicolon, which is useful if the following characters are octal or hex digits not intended to be part of the escape. Precedence table, highest to lowest: .TS tab(!); l l l. operators!class!associativity (R) []!primary! R? R+ R* R%...!postfix!left-to-right R1R2!catenation!left-to-right ~R ...%R!unary!right-to-left R1&R2!intersection!left-to-right R1|R2!union!left-to-right .TE The % operator is like a postfix operator with respect to its left operand, but like a unary operator with respect to its right operand. Thus a~b%c~d is a(~(b%(c(~d)))), demonstrating right-to-left associativity, where all of b% may be regarded as a unary operator being applied to c~d. Similarly, a?*+%b means (((a?)*)+)%b, where the trailing %b behaves like a postfix operator. In .B TXR, regular expression matches do not span multiple lines. The regex language has no feature for multi-line matching. However, the @(freeform) directive allows the remaining portion of the input to be treated as one string in which line terminators appear as explicit characters. Regular expressions may freely match through this sequence. It's possible for a regular expression to match an empty string. For instance, if the next input character is z, facing a the regular expression /a?/, there is a zero-character match: the regular expression's state machine can reach an acceptance state without consuming any characters. Examples: pattern: @A@/a?/@/.*/ data: zzzzz result: A="" pattern: @{A /a?/}@B data: zzzzz result: A="", B="zzzz" pattern: @*A@/a?/ data: zzzzz result: A="zzzzz" In the first example, variable @A is followed by a regular expression which can match an empty string. The expression faces the letter "z" at position 0 in the data line. A zero-character match occurs there, therefore the variable A takes on the empty string. The @/.*/ regular expression then consumes the line. Similarly, in the second example, the /a?/ regular expression faces a "z", and thus yields an empty string which is bound to A. Variable @B consumes the entire line. The third example requests the longest match for the variable binding. Thus, a search takes place for the rightmost position where the regular expression matches. The regular expression matches anywhere, including the empty string after the last character, which is the rightmost place. Thus variable A fetches the entire line. For additional information about the advanced regular expression operators, NOTES ON EXOTIC REGULAR EXPRESSIONS below. .SS Directives The general syntax of a directive is: @EXPR where expr is a parenthesized list of subexpressions. A subexpression is an symbol, number, string literal, character literal, quasiliteral, regular expression, or a parenthesized expression. So, examples of syntactically valid directives are: @(banana) @(a b c (d e f)) @( a (b (c d) (e ) )) @("apple" #\eb #\espace 3) @(a /[a-z]*/ b) @(_ `@file.txt`) A symbol is lexically the same thing as a variable name (the type enclosed in braces in the @{NAME} syntax) and the same rules apply: it can consist of all the same characters, and must not look like a number. Tokens that look like numbers are treated as numbers. .SS Special Symbols Just like in the programming language Lisp, the names nil and t cannot be used as variables. They always represent themselves, and have many uses, internal to the program as well as externally visible. The nil symbol stands for the empty list object, an object which marks the end of a list, and boolean false. It is synonymous with the syntax () which may be used interchangeably with nil in most constructs. .SS Keyword Symbols Names whose names begin with the : character are keyword symbols. These also may not be used as variables either and stand for themselves. Keywords are useful for labeling information and situations. .SS Character Literals Character literals are introduced by the #\e syntax, which is either followed by a character name, the letter x followed by hex digits, the letter o followed by octal digits, or a single character. Valid character names are: nul, alarm, backspace, tab, linefeed, newline, vtab, page, return, esc, space. This convention for character literals is similar to that of the Scheme language. Note that #\elinefeed and #\enewline are the same character. .SS String Literals String literals are delimited by double respectively, and may not span multiple lines. A double quote within a string literal is encoded using \e" and a backslash is encoded as \e\e. Backslash escapes like \en and \et are recognized, as are hexadecimal escapes like \exFF or \exxabc and octal escapes like \e123. Ambiguity between an escape and subsequent text can be resolved by using trailing semicolon delimiter: "\exabc;d" is a string consisting of the character U+0ABC followed by "d". The semicolon delimiter disappears. To write a literal semicolon immediately after a hex or octal escape, write two semicolons, the first of which will be interpreted as a delimiter. Thus, "\ex21;;" represents "!;". .SS String Quasiliterals Quasiliterals are similar to string literals, except that they may contain variable references denoted by the usual @ syntax. The quasiliteral represents a string formed by substituting the values of those variables into the literal template. If a is bound to "apple" and b to "banana", the quasiliteral `one@a and two @{b}s` represents the string "one apple and two bananas". A backquote escaped by a backslash represents itself, and two consecutive @ characters code for a literal @. There is no \e@ escape. Quasiliterals support the full output variable syntax. Expressions within variables substitutions follow the evaluation rules of TXR Lisp when the quasiliteral occurs in TXR Lisp, and the rules of the TXR pattern language when the quasiliteral occurs in the pattern language. .SS Numbers TXR supports integers and floating-point numbers. An integer constant is made up of digits 0 through 9, optionally preceded by a + or - sign. Examples: 123 -34 +0 -0 +234483527304983792384729384723234 An integer constant can also be specified in hexadecimal using the prefix #x followed by an optional sign, followed by hexadecimal digits: 0 through 9 and the upper or lower case letters A through F: #xFF ;; 255 #x-ABC ;; -2748 A floating-point constant is marked by the inclusion of a decimal point, the exponential "e notation", or both. It is an optional sign, followed by a mantissa consisting of digits, a decimal point, more digits, and then an optional exponential notation consisting of the letter "e" or "E", an optional "+" or "-" sign, and then digits indicating the exponent value. In the mantissa, the digits are not optional. At least one digit must either precede the decimal point or follow. That is to say, a decimal point by itself is not a floating-point constant. Examples: .123 123. 1E-3 20E40 .9E1 9.E19 -.5 +3E+3 Examples which are not floating-point constant tokens: . (consing dot) 123E (the symbol 123E) 1.0E- (floating point 1.0 followed by symbol E-) .e (consing dot followed by symbol e) In TXR there is a special "dotdot" token consisting of two consecutive periods. An integer constant followed immediately by dotdot is recognized as such; it is not treated as a floating constant followed by a dot. That is to say, 123.. does not mean 123. . (floating point 123.0 value followed by dot token). It means 123 .. (integer 123 followed by .. token). Dialect note: unlike in Common Lisp, 123. is not an integer, but the same as 123.0. .SS Comments Comments of the form @; were already covered. Inside directives, comments are introduced just by a ; character. Example: @(foo ; this is a comment bar ; this is another comment ) This is equivalent to @(foo bar). .SS Directives-driven Syntax Some directives not only denote an expression, but are also involved in surrounding syntax. For instance, the directive @(collect) not only denotes an expression, but it also introduces a syntactic phrase which requires a matching @(end) directive. So in other words, @(collect) is not only an expression, but serves as a kind of token in a higher level phrase structure grammar. Usually if this type of "syntactic directive" occurs alone in a line, not preceded or followed by other material, it is involved in a "vertical" (or line oriented) syntax. If such a directive is embedded in a line (has preceding or trailing material) then it is in a horizontal syntactic and semantic context (character-oriented). There is an exception: the a definition of a horizontal function looks like this: @(define name (arg))body material@(end) Yet, this is considered one vertical item, which means that it does not match a line of data. (This is necessary because all horizontal syntax matches something within a line of data.) Many directives have a horizontal and vertical syntax, with different but closely related semantics. A few are still "vertical only", and some are horizontal only but in future releases, these exceptions will be minimized. A summary of the available directives follows: .IP @(eof) Explicitly match the end of file. Fails if unmatched data remains in the input stream. .IP @(eol) Explicitly match the end of line. Fails if the the current position is not the end of a line. Also Fails if no data remains (there is no current line). .IP @(next) Continue matching in another file or other data source. .IP @(block) Groups to gether a sequence of directives into a logical name block, which can be explicitly terminated from within using the @(accept) and @(fail) directives. Blocks are discussed in the section BLOCKS below. .IP @(skip) Treat the remaining query as a subquery unit, and search the lines (or characters) of the input file until that subquery matches somewhere. A skip is also an anonymous block. .IP @(trailer) Treat the remaining query or subquery as a match for a trailing context. That is to say, if the remainder matches, the data position is not advanced. .IP @(freeform) Treat the remainder of the input as one big string, and apply the following query line to that string. The newline characters (or custom separators) appear explicitly in that string. .IP @(fuzz) The fuzz directive, inspired by the patch utility, specifies a partial match for some lines. .IP @(some) Multiple clauses are each applied to the same input. Succeeds if at least one of the clauses matches the input. The bindings established by earlier successful clauses are visible to the later clauses. .IP @(all) Multiple clauses are applied to the same input. Succeeds if and only if each one of the clauses matches. The clauses are applied in sequence, and evaluation stops on the first failure. The bindings established by earlier successful clauses are visible to the later clauses. .IP @(none) Multiple clauses are applied to the same input. Succeeds if and only if none of them match. The clauses are applied in sequence, and evaluation stops on the first success. No bindings are ever produced by this construct. .IP @(maybe) Multiple clauses are applied to the same input. No failure occurs if none of them match. The bindings established by earlier successful clauses are visible to the later clauses. .IP @(cases) Multiple clauses are applied to the same input. Evaluation stops on the first successful clause. .IP @(choose) Multiple clauses are applied to the same input. The one whose effect persists is the one which maximizes or minimizes the length of a particular variable. .IP "@(define NAME ( ARGUMENTS ...))" Introduces a function. Functions are discussed in the FUNCTIONS section below. .IP @(gather) Searches text for matches for multiple clauses which may occur in arbitrary order. For convenience, lines of the first clause are treated as separate clauses. .IP @(collect) Search the data for multiple matches of a clause. Collect the bindings in the clause into lists, which are output as array variables. The @(collect) directive is line oriented. It works with a multi-line pattern and scans line by line. A similar directive called @(coll) works within one line. A collect is an anonymous block. .IP @(and) Separator of clauses for @(some), @(all), @(none), @(maybe) and @(cases). Equivalent to @(or). Choice is stylistic. .IP @(or) Separator of clauses for @(some), @(all), @(none), @(maybe) and @(cases). Equivalent to @(and). Choice is stylistic. .IP @(end) Required terminator for @(some), @(all), @(none), @(maybe), @(cases), @(collect), @(coll), @(output), and @(repeat). .IP @(fail) Terminate the processing of a block, as if it were a failed match. Blocks are discussed in the section BLOCKS below. .IP @(accept) Terminate the processing of a block, as if it were a successful match. What bindings emerge may depend on the kind of block: collect has special semantics. Blocks are discussed in the section BLOCKS below. .IP @(try) Indicates the start of a try block, which is related to exception handling, discussed in the EXCEPTIONS section below. .IP "@(catch), @(finally)" Special clauses within @(try). See EXCEPTIONS below. .IP "@(defex), @(throw)" Define custom exception types; throw an exception. See EXCEPTIONS below. .IP @(flatten) Normalizes a set of specified variables to one-dimensional lists. Those variables which have scalar value are reduced to lists of that value. Those which are lists of lists (to an arbitrary level of nesting) are converted to flat lists of their leaf values. .IP @(merge) Binds a new variable which is the result of merging two or more other variables. Merging has somewhat complicated semantics. .IP @(cat) Decimates a list (any number of dimensions) to a string, by catenating its constituent strings, with an optional separator string between all of the values. .IP @(bind) Binds one or more variables against a value using a structural pattern match. A limited form of unification takes place which can cause a match to fail. .IP @(set) Destructively assigns one or more existing variables using a structural pattern, using syntax similar to bind. Assignment to unbound variables triggers an error. .IP @(rebind) Evaluates an expression in the current binding environment, and then creates new bindings for the variables in the structural pattern. Useful for temporarily overriding variable values in a scope. .IP @(output) A directive which encloses an output clause in the query. An output section does not match text, but produces text. The directives above are not understood in an output clause. .IP @(repeat) A directive understood within an @(output) section, for repeating multi-line text, with successive substitutions pulled from lists. The directive @(rep) produces iteration over lists horizontally within one line. .IP @(deffilter) This directive is used for defining named filters, which are useful for filtering variable substitutions in output blocks. Filters are useful when data must be translated between different representations that have different special characters or other syntax, requiring escaping or similar treatment. Note that it is also possible to use a function as a filter. See Function Filters below. .IP @(filter) The filter directive passes one or more variables through a given filter or chain or filters, updating them with the filtered values. .IP @(load) The load directive loads another TXR file and interprets its contents. .PP .SH INPUT SCANNING AND DATA MANIPULATION .SS The Next Directive The next directive indicates that the remainder of the query is to be applied to a new input source. It can only occur by itself as the only element in a query line, and takes various arguments, according to these possiblities: @(next) @(next SOURCE) @(next SOURCE :nothrow) @(next :args) @(next :env) @(next :list EXPR) @(next :string EXPR) The lone @(next) without arguments switches to the next file in the argument list which was passed to the .B TXR utility. However, "switch to the next file" means in a pattern matching way, not in an imperative way. It is possible for the pattern matching logic to implicitly backtrack to the previous file. If SOURCE is given, it must be text-valued expression which denotes an input source; it may be a string literal, quasiliteral or a variable. For instance, if variable A contains the text "data", then @(next A) means switch to the file called "data", and @(next `@A.txt`) means to switch to the file "data.txt". If the input source cannot be opened for whatever reason, .B TXR throws an exception (see EXCEPTIONS below). An unhandled exception will terminate the program. Often, such a drastic measure is inconvenient; if @(next) is invoked with the :nothrow keyword, then if the input source cannot be opened, the situation is treated as a simple match failure. The variant @(next :args) means that the remaining command line arguments are to be treated as a data source. For this purpose, each argument is considered to be a line of text. If an argument is currently being processed as an input source, that argument is included. Note that if the first entry in the argument list is not intended to name an input source, then the query should begin with @(next :args) or some other form of next directive, to prevent an attempt to open the input source named by that argument. If the very first directive of a query is any variant of the next directive, then .B TXR avoids opening the first input source, but it does open the input source for any other directive, even one which does not consume any data. The variant @(next :env) means that the list of process environment variables is treated as a source of data. It looks like a text file stream consisting of lines of the form "name=value". If this feature is not available on a given platform, an exception is thrown. The syntax @(next :list EXPR) treats the expression as a source of text. The value of the expression is flattened to a list in a way similar to the @(flatten) directive. The resulting list is treated as if it were the lines of a text file: each element of the list is a line. If the lines happen contain embedded newline characters, they are a visible constituent of the line, and do not act as line separators. The syntax @(next :string EXPR) treats the expression as a source of text. The value of the expression must be a string. Newlines in the string are interpreted as line terminators. A string which is not terminated by a newline is tolerated, so that: @(next :string "abc") @a binds a to "abc". Likewise, this is also the case with input files and other streams whose last line is not terminated by a newline. However, watch out for empty strings, which are analogous to a correctly formed empty file which contains no lines: @(next :string "") @a This will not bind a to ""; it is a matching failure. The behavior of :list is different. The query @(next :list "") @a binds a to "". The reason is that under :list the string "" is flattened to the list ("") which is not an empty input stream, but a stream consisting of one empty line. Note that "remainder of the query" which is applied to the stream opened by @(next) refers to the subquery in which the next directive appears, not necessarily the entire query. For example, the following query looks for the line starting with "xyz" at the top of the file "foo.txt", within a some directive. After the @(end) which terminates the @(some), the "abc" is matched in the previous file again. @(some) @(next "foo.txt") xyz@suffix @(end) abc However, if the @(some) subquery successfully matched "xyz@suffix" within the file foo.text, there is now a binding for the suffix variable, which is visible to the remainder of the entire query. The variable bindings survive beyond the clause, but the data stream does not. The @(next) directive supports the file name conventions as the command line. The name - means standard input. Text which starts with a ! is interpreted as a shell command whose output is read like a file. These interpretations are applied after variable substitution. If the file is specified as @a, but the variable a expands to "!echo foo", then the output of the "echo foo" command will be processed. .SS The Skip Directive The skip directive considers the remainder of the query as a search pattern. The remainder is no longer required to strictly match at the current line in the current file. Rather, the current file is searched, starting with the current line, for the first line where the entire remainder of the query will successfully match. If no such line is found, the skip directive fails. If a matching position is found, the remainder of the query is understood to be processed there. Of course, the remainder of the query can itself contain skip directives. Each such directive performs a recursive subsearch. Skip comes in vertical and horizontal flavors. For instance, skip and match the last line: @(skip) @last @(eof) Skip and match the last character of the line: @(skip)@{last 1}@(eol) The skip directive has two optional arguments. If the first argument is a number, its value limits the range of lines scanned for a match. Judicious use of this feature can improve the performance of queries. Example: scan until "size: @SIZE" matches, which must happen within the next 15 lines: @(skip 15) size: @SIZE Without the range limitation skip will keep searching until it consumes the entire input source. While sometimes this is what is intended, often it is not. Sometimes a skip is nested within a collect, or following another skip. For instance, consider: @(collect) begin @BEG_SYMBOL @(skip) end @BEG_SYMBOL @(end) The collect iterates over the entire input. But, potentially, so does the skip. Suppose that "begin x" is matched, but the data has no matching "end x". The skip will search in vain all the way to the end of the data, and then the collect will try another iteration back at the beginning, just one line down from the original starting point. If it is a reasonable expectation that an "end x" occurs 15 lines of a "begin x", this can be written instead: @(collect) begin @BEG_SYMBOL @(skip 15) end @BEG_SYMBOL @(end) If the symbol nil is used in place of a number, it means to scan an unlimited range of lines; thus, @(skip nil) is equivalent to @(skip). If the symbol :greedy is used, it changes the semantics of the skip to longest match semantics, like the regular expression * operator. For instance, match the last three space-separated tokens of the line: @(skip :greedy) @a @b @c Without :greedy, the variable @c will can match multiple tokens, and end up with spaces in it, because nothing follows @c and so it matches from any position which follows a space to the end of the line. Also note the space in front of @a. Without this space, @a will get an empty string. A line oriented example of greedy skip: match the last line without using @eof: @(skip :greedy) @last_line There may be a second numeric argument. This specifies a minimum number of lines to skip before looking for a match. For instance, skip 15 lines and then search indefinitely for "begin ...": @(skip nil 15) begin @BEG_SYMBOL The two arguments may be used together. For instance, the following matches if, and only if, the 15th line of input starts with "begin ": @(skip 1 15) begin @BEG_SYMBOL Essentially, @(skip 1 ) means "hard skip by " lines, then match the query without scanning. @(skip 1 0) is the same as @(skip 1), which is a noop, because it means: "the remainder of the query must match starting on the very next line", or, more briefly, "skip exactly zero lines", which is the behavior if the skip directive is omitted altogether. Here is one trick for grabbing the fourth line from the bottom of the input: @(skip) @fourth_from_bottom @(skip 1 3) @(eof) Or using greedy skip: @(skip :greedy) @fourth_from_bottom @(skip 1 3) Nongreedy skip with the @(eof) has a slight advantage because the greedy skip will keep scanning even though it has found the correct match, then backtrack to the last good match once it runs out of data. The regular skip with explicit @(eof) will stop when the @(eof) matches. .SS Reducing Backtracking with Blocks Skip can consider considerable CPU time when multiple skips are nested. Consider: @(skip) A @(skip) B @(skip) C This is actually nesting: the second a third skips occur within the body of the first one, and thus this creates nested iteration. TXR is searching for the combination of skips which find match the pattern of lines A, B and C, with backtracking behavior. The outermost skip marches through the data until it finds A, followed by a pattern match for the second skip. The second skip iterates within to find B, followed by the third skip, and the third skip iterates to find C. If there is only one line A, and one B, then this is reasonably fast. But suppose there are many lines matching A and B, giving rise to a large number combinations of skips which match A and B, and yet no match for C, triggering backtracking. The nested stepping which tries the combinations of A and B can give rise to a considerable running time. One way to deal with the problem is to unravel the nesting with the help of blocks. For example: @(block) @ (skip) A @(end) @(block) @ (skip) B @(end) @(skip) C Now the scope of each skip is just the remainder of the block in which it occurs. The first skip finds A, and then the block ends. Control passes to the next block, and backtracking will not take place to a block which completed (unless all these blocks are enclosed in some larger construct which backtracks, causing the blocks to be re-executed. Of course, this rewrite is not equivalent, and cannot be used for instance in backreferencing situations such as: @; @; Find some three lines which are the same. @; @(skip) @line @(skip) @line @(skip) @line This example depends on the nested search-within-search semantics. .SS The Trailer Directive The trailer directive introduces a trailing portion of a query or subquery which matches input material normally, but in the event of a successful match, does not advance the current position. This can be used, for instance, to cause @(collect) to match partially overlapping regions. Example: @(collect) @line @(trailer) @(skip) @line @(end) This script collects each line which has a duplicate somewhere later in the input. Without the @(trailer) directive, this does not work properly for inputs like: 111 222 111 222 Without @(trailer), the first duplicate pair constitutes a match which spans over the 222. After that pair is found, the matching continues after the second 111. With the @(trailer) directive in place, the collect body, on each iteration, only consumes the lines matched prior to @(trailer). .SS The Freeform Directive The freeform directive provides a useful alternative to .B TXR's line-oriented matching discipline. The freeform directive treats all remaining input from the current input source as one big line. The query line which immediately follows freeform is applied to that line. The syntax variations are: @(freeform) ... query line .. @(freeform NUMBER) ... query line .. @(freeform STRING) ... query line .. @(freeform NUMBER STRING) ... query line .. The string and numeric arguments, if both are present, may be given in either order. If a numeric argument is given, it limits the range of lines which are combined together. For instance @(freeform 5) means to only consider the next five lines to to be one big line. Without a numeric argument, freeform is "bottomless". It can match the entire file, which creates the risk of allocating a large amount of memory. If a string argument is given, it specifies a custom line terminator. The default terminator is "\en". The terminator does not have to be one character long. Freeform does not convert the entire remainder of the input into one big line all at once, but does so in a dynamic, lazy fashion, which takes place as the data is accessed. So at any time, only some prefix of the data exists as a flat line in which newlines are replaced by the terminator string, and the remainder of the data still remains as a list of lines. After the subquery is applied to the virtual line, the unmatched remainder of that line is broken up into multiple lines again, by looking for and removing all occurences of the terminator string within the flattened portion. Care must be taken if the terminator is other than the default "\en". All occurences of the terminator string are treated as line terminators in the flattened portion of the data, so extra line breaks may be introduced. Likewise, in the yet unflattened portion, no breaking takes place, even if the text contains occurences of the terminator string. The extent of data which is flattened, and the amount of it which remains, depends entirely on the query line underneath @(flatten). In the following example, lines of data are flattened using $ as the line terminator. Query: @(freeform "$") @a$@b: @c @d Data: 1 2:3 4 Output: a="1" b="2" c="3" d="4" The data is turned into the virtual line 1$2:3$4$. The @a$@b: subquery matches the 1$2: portion, binding a to 1, and b to 2. The remaining portion 3$4$ is then split into separate lines again according to the line terminator $: 3 4 Thus the remainder of the query @c @d faces these lines, binding c to 3 and d to 4. Note that since the data does not contain dollar signs, there is no ambiguity; the meaning may be understood in terms of the entire data being flattened and split again. In the following example, freeform is used to solve a tokenizing problem. The Unix password file has fields separated by colons. Some fields may be empty. Using freeform, we can join the password file using ":" as a terminator. By restricting freeform to one line, we can obtain each line of the password file with a terminating ":", allowing for a simple tokenization, because now the fields are colon-terminated rather than colon-separated. Example: @(next "/etc/passwd") @(collect) @(freeform 1 ":") @(coll)@{token /[^:]*/}:@(end) @(end) .SS The Fuzz Directive The fuzz directive allows for an imperfect match spanning a set number of lines. It takes two arguments, both expressions that should evaluate to integers: @(fuzz m n) ... This expresses that over the next n query lines, the matching strictness is relaxed a little bit. Only m out of those n lines have to match. Afterward, the rest of the query follows normal, strict processing. In the degenerate situation that there are fewer than n query lines following the fuzz directive, then m of them must succeed nevertheless. (If there are fewer than m, then this is impossible.) .SS The Some, All, None, Maybe, Cases and Choose directives These directives, called the parallel directives, combine multiple subqueries, which are applied at the same input position, rather than to consecutive input. They come in vertical (line mode) and horizontal (character mode) flavors. In horizontal mode, the current position is understood to be a character position in the line being processed. The clauses advance this character position by moving it to the right. In vertical mode, the current position is understood to be a line of text within the stream. A clause advances the position by some whole number of lines. The syntax of these parallel directives follows this example: @(some) subquery1 . . . @(and) subquery2 . . . @(and) subquery3 . . . @(end) And in horizontal mode: @(some)subquery1...@(and)subquery2...@(and)subquery3...@(end) Long horizontal lines can be broken up with line continuations, allowing the above example to be written like this, which is considered a single logical line: @(some)@\ subquery1...@\ @(and)@\ subquery2...@\ @(and)@\ subquery3...@\ @(end) The @(some), @(all), @(none), @(maybe), @(cases) or @(choose) must be followed by at least one subquery clause, and be terminated by @(end). If there are two or more subqueries, these additional clauses are indicated by @(and) or @(or), which are interchangeable. The separator and terminator directives also must appear as the only element in a query line. The choose directive requires keyword arguments. See below. The syntax supports arbitrary nesting. For example: QUERY: SYNTAX TREE: @(all) all -+ @ (skip) +- skip -+ @ (some) | +- some -+ it | | +- TEXT @ (and) | | +- and @ (none) | | +- none -+ was | | | +- TEXT @ (end) | | | +- end @ (end) | | +- end a dark | +- TEXT @(end) *- end nesting can be indicated using whitespace between @ and the directive expression. Thus, the above is an @(all) query containing a @(skip) clause which applies to a @(some) that is followed by the the text line "a dark". The @(some) clause combines the text line "it", and a @(none) clause which contains just one clause consisting of the line "was". The semantics of the parallel directives is: .IP @(all) Each of the clauses is matched at the current position. If any of the clauses fails to match, the directive fails (and thus does not produce any variable bindings). Clauses following the failed directive are not evaluated. Bindings extracted by a successful clause are visible the clauses which follow, and if the directive succeeds, all of the combined bindings emerge. .IP "@(some [ :resolve (vars ...) ])" Each of the clauses is matched at the current position. If any of the clauses succeed, the directive succeeds, retaining the bindings accumulated by the successully matching clauses. Evaluation does not stop on the first successful clause. Bindings extracted by a successful clause are visible to the clauses which follow. The :resolve parameter is for situations when the @(some) directive has multiple clauses that need to bind some common variables to different values: for instance, output parameters in functions. Resolve takes a list of variable name symbols as an argument. This is called the resolve set. If the clauses of @(some) bind variables in the resolve set, those bindings are not visible to later clauses. However, those bindings do emerge out of the @(some) directive as a whole. This creates a conflict: what if two or more clauses introduce non-matching bindings for a variable in the resolve set? This is why it is called the resolve set: conflicts for variables in the resolve set are automatically resolved in favor of later directives. Example: @(some :resolve (x)) @ (bind a "a") @ (bind x "x1") @(or) @ (bind b "b") @ (bind x "x2") @(end) Here, the two clauses both introduce a binding for x. Without the :resolve parameter, this would mean that the second clause fails, because x comes in with the value "x1", which does not bind with "x2". But because x is placed into the resolve set, the second clause does not see the "x1" binding. Both clauses establish their bindings independently creating a conflict over x. The conflict is resolved in favor of the second clause, and so the bindings which emerge from the directive are: a="a" b="b" x="x2" .IP @(none) Each of the clauses is matched at the current position. The directive succeeds only if all of the clauses fail. If any clause succeeds, the directive fails, and subsequent clauses are not evaluated. Thus, this directive never produces variable bindings, only matching success or failure. .IP @(maybe) Each of the clauses is matched at the current position. The directive always succeeds, even if all of the clauses fail. Whatever bindings are found in any of the clauses are retained. Bindings extracted by any successful clause are visible the clauses which follow. .IP @(cases) Each of the clauses is matched at the current position. The The clauses are matched, in order, at the current position. If any clause matches, the matching stops and the bindings collected from that clause are retained. Any remaining clauses after that one are not processed. If no clause matches, the directive fails, and produces no bindings. .IP "@(choose [ :longest | :shortest ])" Each of the clauses is matched at the current position in order. In this construct, bindings established an earlier clause are not visible to later clauses. Although any or all of the clauses can potentially match, the clause which succeeds is the one which maximizes or minimizes the length of the text bound to the specified variable. The other clauses have no effect. For all of the parallel directives other than @(none) and @(choose), the query advances the input position by the greatest number of lines that match in any of the successfully matching subclauses that are evaluated. The @(none) directive does not advance the input position. For instance if there are two subclauses, and one of them matches three lines, but the other one matches five lines, then the overall clause is considered to have made a five line match at its position. If more directives follow, they begin matching five lines down from that position. .SS The Gather Directive Sometimes text is structured as items that can appear in an arbitrary order. When multiple matches need to be extracted, there is a combinatorial explosion of possible orders, making it impractical to write pattern matches for all the possible orders. The gather directive is for these situations. It specifies multiple clauses which all have to match somewhere in the data, but in any order. For further convenience, the lines of the first clause of the gather directive are implicitly treated as separate clauses. The syntax follows this pattern @(gather) one-line-query1 one-line-query2 . . . one-line-queryN @(and) multi line query1 . . . @(and) multi line query2 . . . @(end) Of course the multi-line clauses are optional. The gather directive takes keyword parameters, see below. Similarly to @(collect), @(gather) has an optional until/last clause: @(gather) ... @(until) ... @(end) How gather works is that the text is searched for matches for the single line and multi-line queries. The clauses are applied in the order in which they appear. Whenever one of the clauses matches, any bindings it produces are retained and it is removed from further consideration. Multiple clauses can match at the same text position. The position advances by the longest match from among the clauses which matched. If no clauses match, the position advances by one line. The search stops when all clauses are eliminated, and then the cumulative bindings are produced. If the data runs out, but unmatched clauses remain, the directive fails. Example: extract several environment variables, which do not appear in a particular order: @(next :env) @(gather) USER=@USER HOME=@HOME SHELL=@SHELL @(end) If the until or last clause is present and a match occurs, then the matches from the other clauses are discarded and the gather terminates. The difference between until and last is that any bindings bindings established in last are retained, and the input position is advanced past the matched material. The until/last clause has visibility to bindings established in the previous clauses in that same iteration, even though those bindings end up thrown away. .SS Gather Keyword Parameters The gather diretive accepts the keyword parameter :vars. The argument to vars is a list of required and optional variables. Optional variables are denoted by the specification of a default value. Example: @(gather :vars (a b c (d "foo"))) ... @(end) Here, a, b, c and e are required variables, and d is optional. Variable e is required because its default value is the empty list (), same as the symbol nil. The presence of vars changes the behavior in three ways. Firstly, even if all the clauses in the gather match successfully and are eliminated, the directive will fail if the required variables do not have bindings. It doesn't matter whether the bindings are existing, or whether they are established by the gather. Secondly, if some of the clauses of the gather did not match, but all of the required variables have bindings, then the directive succeeds. Without the presence of :vars, it would fail in this situation. Thirdly, if the the gather succeeds (all required variables have bindings), then all of the optional variables which do not have bindings are given bindings to their default values. .SS The Collect Directive The syntax of the collect directive is: @(collect) ... lines of subquery @(end) or with an until or last clause: @(collect) ... lines of subquery: main clause @(until) ... lines of subquery: until clause @(end) @(collect) ... lines of subquery: main clause @(last) ... lines of subquery: last clause @(end) The repeat symbol may be specified instead of collect, which changes the meaning, see below: @(repeat) ... lines of subquery @(end) The subquery is matched repeatedly, starting at the current line. If it fails to match, it is tried starting at the subsequent line. If it matches successfully, it is tried at the line following the entire extent of matched data, if there is one. Thus, the collected regions do not overlap. (Overlapping behavior can be obtained: see the @(trailer) directive). Unless certain keywords are specified, or unless the collect is explicitly failed with @(fail), it always succeeds, even if it collects nothing, and even if the until/last clause never finds a match. If no until/last clause is specified, and the collect is not limited using parameters, the collect is unbounded. It consumes the entire data file. If any query material follows such the collect clause, it will fail if it tries to match anything in the current file; but of course, it is possible to continue matching in another file by means of @(next). If an until/last clause is specified, the collection stops when that clause matches at the current position. If an until clause terminates collect, no bindings are collected at that position, even if the main clause matches at that position also. Moreover, the position is not advanced. The remainder of the query begins matching at that position. If a last clause terminates collect, the behavior is different. Any bindings captured by the main clause are thrown away, just like with the until clause. However, the bindings in the last clause itself survive, and the position is advanced to skip over that material. Example: Query: @(collect) @a @(until) 42 @b @(end) @c Data: 1 2 3 42 5 6 Output: a[0]="1" a[1]="2" a[2]="3" c="42" The line 42 is not collected, even though it matches @a. Furthermore, the until does not advance the position, so variable c takes 42. If the @(until) is changed to @(last) the output will be different: Output: a[0]="1" a[1]="2" a[2]="3" b=5 c=6 The 42 is not collected into the a list, just like before. But now the binding captured by @b emerges. Furthermore, the position advances so variable now takes 6. The binding variables within the clause of a collect are treated specially. The multiple matches for each variable are collected into lists, which then appear as array variables in the final output. Example: Query: @(collect) @a:@b:@c @(end) Data: John:Doe:101 Mary:Jane:202 Bob:Coder:313 Output: a[0]="John" a[1]="Mary" a[2]="Bob" b[0]="Doe" b[1]="Jane" b[2]="Coder" c[0]="101" c[1]="202" c[2]="313" The query matches the data in three places, so each variable becomes a list of three elements, reported as an array. Variables with list bindings may be referenced in a query. They denote a multiple match. The -D command line option can establish a one-dimensional list binding. Collect clauses may be nested. Variable matches collated into lists in an inner collect, are again collated into nested lists in the outer collect. Thus an unbound variable wrapped in N nestings of @(collect) will be an N-dimensional list. A one dimensional list is a list of strings; a two dimensional list is a list of lists of strings, etc. It is important to note that the variables which are bound within the main clause of a collect---i.e. the variables which are subject to collection---appear, within the collect, as normal one-value bindings. The collation into lists happens outside of the collect. So for instance in the query: @(collect) @x=@x @(end) The left @x establishes a binding for some material preceding an equal sign. The right @x refers to that binding. The value of @x is different in each iteration, and these values are collected. What finally comes out of the collect clause is list variable called x which holds each value that was ever instantiated under that name within the collect clause. Also note that the until clause has visibility over the bindings established in the main clause. This is true even in the terminating case when the until clause matches, and the bindings of the main clause are discarded. .SS Collect Keyword Parameters By default, collect searches the rest of the input indefinitely, or until the @(until) clause matches. It skips arbitrary amounts of nonmatching material before the first match, and between matches. Within the @(collect) syntax, it is possible to specify some useful keyword parameters for additional control of the behavior. For instance @(collect :maxgap 5) means that the collect will terminate if it does not find a match within five lines of the starting position, or if more than five lines are skipped since any successful match. A :maxgap of 0 means that the collected regions must be adjacent. For instance: @(collect :maxgap 0) M @a @(end) means: from here, collect consecutive lines of the form "M ...". This will not search for the first such line, nor will it skip lines which do not match this form. Other keywords are :mingap, and :gap. The :mingap keyword specifies a minimum gap between matches, but has no effect on the distance to the first match. The :gap keyword specifies :mingap and :maxgap at the same time, and can only be used if these other two are not used. Thus: @(collect :gap 1) @a @(end) means collect every other line starting with the current line. Several other supported keywords are :times, :mintimes, :maxtimes and lines. The shorthand :times N means the same thing as :mintimes N :maxtimes N. These specify how many matches should be collected. If there are fewer than mintimes matches, the collect fails. If maxtimes matches are collected, collect stops collecting immediately. Example: @(collect :times 3) @a @b @(end) This will collect a match for "@a @b" exactly three times. If three matches are not found, it will fail. The :lines parameter specifies the upper bound on how many lines should be scanned by collect, measuring from the starting position. The extent of the collect body is not counted. Example: @(collect :lines 2) foo: @a bar: @b baz: @c @(end) The above collect will look for a match only twice: at the current position, and one line down. There is one more keyword, :vars, discussed in the following section. .SS Specifying Variables in Collect Normally, any variable for which a new binding occurs in a collect is collected. A collect clause may be sloppy: it can neglect to collect some variables on some iterations, or bind some variables which are intended to behave like local temporaries, but end up collated into lists. Another issue is that the collect clause might not match anything at all, and then none of the variables are bound. The :vars keyword allows the query writer to add discipline the collect body. The argument to :vars is a list of variable specs. A variable spec is either a symbol, or a ( ) pair, where the expression specifies a default value. When a :vars list is specified, it means that only the given variables can emerge from the successful collect. Any newly introduced bindings for other variables do not propagate. Furthermore, for any variable which is not specified with a default value, the collect body, whenever it matches successfully, must bind that variable. If it neglects to bind the variable, an exception of type query_error is thrown. (If a collect body matches successfully, but produces no new bindings, then this error is suppressed.) For any variable which does have a default value, if the collect body neglects to bind that variable, the behavior is as if the collect did bind that variable to that default value. The default values are expressions, and so can be quasiliterals. Lastly, if in the event that the collect does not match anything, the variables specified in vars (whether or not they have a default value) are all bound to empty lists. (These bindings are established after the processing of the until/last clause, if present.) Example: @(collect :vars (a b (c "foo"))) @a @c @(end) Here, if the body "@a @c" matches, an error will be thrown because one of the mandatory variables is b, and the body neglects to produce a binding for b. Example: @(collect :vars (a (c "foo"))) @a @b @(end) Here, if "@a @b" matches, only a will be collected, but not b, because b is not in the variable list. Furthermore, because there is no binding for c in the body, a binding is created with the value "foo", exactly as if c matched such a piece of text. In the following example, the assumption is that THIS NEVER MATCHES is not found anywhere in the input but the line THIS DOES MATCH is found and has a successor which is bound to a. Because the body did not match, the :vars a and b should be bound to empty lists. But a is bound by the last clause to some text, so this takes precedence. Only b is bound to a an empty list. @(collect :vars (a b)) THIS NEVER MATCHES @(last) THIS DOES MATCH @a @(end) The following means: do not allow any variables to propagate out of any iteration of the collect and therefore collect nothing: @(collect :vars nil) ... @(end) Instead of writing @(collect :vars nil), it is possible to write @(repeat). @(repeat) takes all collect keywords, except for :vars. There is a repeat directive used in @(output) clauses; that is a different repeat directive. .SS The Coll Directive The coll directive is a kind of miniature version of the collect directive. Whereas the collect directive works with multi-line clauses on line-oriented material, coll works within a single line. With coll, it is possible to recognize repeating regularities within a line and collect lists. Regular-expression based Positive Match variables work well with coll. Example: collect a comma-separated list, terminated by a space. pattern: @(coll)@{A /[^, ]+/}@(until) @(end)@B data: foo,bar,xyzzy blorch result: A[0]="foo" A[1]="bar" A[2]="xyzzy" B=blorch Here, the variable A is bound to tokens which match the regular expression /[^, ]+/: non-empty sequence of characters other than commas or spaces. Like its big cousin, the coll directive searches for matches. If no match occurs at the current character position, it tries at the next character position. Whenever a match occurs, it continues at the character position which follows the last character of the match, if such a position exists. If not bounded by an until clause, it will exhaust the entire line. If the until clause matches, then the collection stops at that position, and any bindings from that iteration are discarded. Like collect, coll also supports a last clause, which propagates varaible bindings and advances the position. Coll clauses nest, and variables bound within a coll are available to within the rest of the coll clause, including the until clause, and appear as single values. The final list aggregation is only visible after the coll clause. The behavior of coll is troublesome, when delimited variables are used, because in text file formats, the material which separates items is not repeated after the last item. For instance, a comma-separated list usually not appear as "a,b,c," but rather "a,b,c". There might not be any explicit termination---the last item might be at the very end of the line. So for instance, the following result is not satisfactory: pattern: @(coll)@a @(end) data: 1 2 3 4 5 result: a[0]="1" a[1]="2" a[2]="3" a[3]="4" What happened to the 5? After matching "4 ", coll continues to look for matches. It tries "5", which does not match, because it is not followed by a space. Then the line is consumed. So in this sequence, a valid item is either followed by a space, or by nothing. So it is tempting to try this: pattern: @(coll)@a@/ ?/@(end) data: 1 2 3 4 5 result: a[0]="" a[1]="" a[2]="" a[3]="" a[4]="" a[5]="" a[6]="" a[7]="" a[8]="" however, the problem is that the regular expression / ?/ (match either a space or nothing), matches at any position. So when it is used as a variable delimiter, it matches at the current position, which binds the empty string to the variable, the extent of the match being zero. In this situation, the coll directive proceeds character by character. The solution is to use positive matching: specify the regular expression which matches the item, rather than a trying to match whatever follows. The collect directive will recognize all items which match the regular expression. pattern: @(coll)@{a /[^ ]+/}@(end) data: 1 2 3 4 5 result: a[0]="1" a[1]="2" a[2]="3" a[3]="4" a[4]="5" The until clause can specify a pattern which, when recognized, terminates the collection. So for instance, suppose that the list of items may or may not be terminated by a semicolon. We must exclude the semicolon from being a valid character inside an item, and add an until clause which recognizes a semicolon: pattern: @(coll)@{a /[^ ;]+/}@(until);@(end); data: 1 2 3 4 5; result: a[0]="1" a[1]="2" a[2]="3" a[3]="4" a[4]="5" data: 1 2 3 4 5; result: a[0]="1" a[1]="2" a[2]="3" a[3]="4" a[4]="5" Semicolon or not, the items are collected properly. Note that the @(end) is followed by a semicolon. That's because when the @(until) clause meets a match, the matching material is not consumed. Instead of regular expression hacks, this problem can be nicely solved with cases: pattern: @(coll)@(cases)@a @(or)@a@(end)@(end) data: 1 2 3 4 5 result: a[0]="1" a[1]="2" a[2]="3" a[3]="4" a[4]="5" .SS Coll Keyword Parameters The @(coll) directive takes most of the same parameters as @(collect). See the section Collect Keyword Parameters above. So for instance @(coll :gap 0) means that the collects must be consecutive, and @(coll :maxtimes 2) means that at most two matches will be collected. The :lines keyword does not exist, but there is an analogous :chars keyword. .SS The Flatten Directive. The flatten directive can be used to convert variables to one dimensional lists. Variables which have a scalar value are converted to lists containing that value. Variables which are multidimensional lists are flattened to one-dimensional lists. Example (without @(flatten)) pattern: @b @(collect) @(collect) @a @(end) @(end) data: 0 1 2 3 4 5 result: b="0" a_0[0]="1" a_1[0]="2" a_2[0]="3" a_3[0]="4" a_4[0]="5" Example (with flatten): pattern: @b @(collect) @(collect) @a @(end) @(end) @(flatten a b) data: 0 1 2 3 4 5 result: b[0]="0" a[0]="1" a[1]="2" a[2]="3" a[3]="4" a[4]="5" .SS The Merge Directive The merge directive provides a way of combining two or more variables in a somewhat complicated but very useful way. To understand what merge does we first have to define a property called depth. The depth of an atom such as a string is defined as 1. The depth of an empty list is 0. The depth of a nonempty list is one plus the depth of its deepest element. So for instance "foo" has depth 1, ("foo") has depth 2, and ("foo" ("bar")) has depth three. We can now define the binary (two argument) merge operation as follows. (merge A B) first normalizes the values A and B such that they have equal depth. 1. A value which has depth zero is put into a one element list. 2. If either value has a smaller depth than the other, it is wrapped in a list as many times as needed to give it equal depth. Finally, the values are appended together. Merge takes more than two arguments. These are merged by a left reduction. The leftmost two values are merged, and then this result is merged with the third value, and so on. Merge is useful for combining the results from collects at different levels of nesting such that elements are at the appropriate depth. .SS The Cat Directive The @(cat) directive converts a list variable into a single piece of text. The syntax is: @(cat VAR [ SEP ]) The SEP argument specifies a separating piece of text. If no separator is specified, then a single space is used. Example: pattern: @(coll)@{a /[^ ]+/}@(end) @(cat a ":") data: 1 2 3 4 5 result: a="1:2:3:4:5" .SS The Bind Directive The syntax of the @(bind) directive is: @(bind pattern expression { keyword value }*) The @(bind) directive is a kind of pattern match, which matches one or more variables on in the left hand side pattern to the value of a variable on the right hand side. The right hand side variable must have a binding, or else the directive fails. Any variables on the left hand side which are unbound receive a matching piece of the right hand side value. Any variables on the left which are already bound must match their corresponding value, or the bind fails. Any variables which are already bound and which do match their corresponding value remain unchanged (the match can be inexact). The simplest bind is of one variable against itself, for instance bind A against A: @(bind A A) This will fail if A is not bound, (and complain loudly). If A is bound, it succeeds, since A matches A. The next simplest bind binds one variable to another: @(bind A B) Here, if A is unbound, it takes on the same value as B. If A is bound, it has to match B, or the bind fails. Matching means that either .IP - A and B are the same text .IP - A is text, B is a list, and A occurs within B. .IP - vice versa: B is text, A is a list, and B occurs within A. .IP - A and B are lists and are either identical, or one is found as substructure within the other. .PP The right hand side does not have to be a variable. It may be some other object, like a string, quasiliteral, regexp, or list of strings, et cetera. For instance @(bind A "ab\tc") will bind the string "ab\tc" (the letter a, b, a tab character, and c) to the variable A if A is unbound. If A is bound, this will fail unless A already contains an identical string. However, the right hand side of cannot be an unbound variable, nor a complex expression that contains unbound variables. The left hand side of a bind can be a nested list pattern containing variables. The last item of a list at any nesting level can be preceded by a dot, which means that the variable matches the rest of the list from that position. Example: suppose that the list A contains ("now" "now" "brown" "cow"). Then the directive @(bind (H N . C) A), assuming that H, N and C are unbound variables, will bind H to "how", N to "now", and C to the remainder of the list ("brown" "cow"). Example: suppose that the list A is nested to two dimensions and contains (("how" "now") ("brown" "cow")). Then @(bind ((H N) (B C)) A) binds H to "how", N to "now", B to "brown" and C to "cow". The dot notation may be used at any nesting level. it must be preceded and followed by a symbol: the forms (.) (. X) and (X .) are invalid. The number of items in a left pattern match must match the number of items in the corresponding right side object. So the pattern () only matches an empty list. The notation () and nil means exactly the same thing. The symbols nil, t and keyword symbols may be used on either side. They represent themselves. For example @(bind :foo :bar) fails, but @(bind :foo :foo) succeeds since the two sides denote the same keyword symbol object. .SS Keywords in The Bind Directive The Bind directive accepts these keywords: .IP :lfilt The argument to :lfilt is a filter specification. When the left side pattern contains a binding which is therefore matched against its counterpart from the right side expression, the left side is filtered through the filter specified by :lfilt for the purposes of the comparison. For example: @(bind "a" "A" :lfilt :upcase) produces a match, since the left side is the same as the right after filtering through the :upcase filter. .IP :rfilt The argument to :rfilt is a filter specification. The specified filter is applied to the right hand side material prior to matching it against the left side. The filter is not applied if the left side is a variable with no binding. It is only applied to determine a match. Binding takes place the unmodified right hand side object. Example, the following produces a match: @(bind "A" "a" :rfilt :upcase) .IP :filter This keyword is a shorthand to specify both filters to the same value. So for instance :filter :upcase is equivalent to :lfilt :upcase :rfilt :upcase. For a description of filters, see Output Filtering below. Of course, compound filters like (:from_html :upcase) are supported with all these keywords. The filters apply across arbitrary patterns and nested data. Example: @(bind (a b c) ("A" "B" "C")) @(bind (a b c) (("z" "a") "b" "c") :rfilt :upcase) Here, the first bind establishes the values for a, b and c, and the second bind succeeds, because the value of a matches the second element of the list ("z" "a") if it is upcased, and likewise b matches "b" and c matches "c" if these are upcased. .SS The Set Directive The @(set) directive resembles bind, but is not a pattern match. It overwrites the previous values of variables with new values from the right hand side. Each variable that is assigned must have an existing binding. Examples follow. Store the value of A back into A, achieving nothing: @(set A A) Exchange the values of A and B: @(set (A B) (B A)) Store a string into A: @(set A "text") Store a list into A: @(set A ("line1" "line2")) Destructuring assignment. D assumed to contain the list @(bind D ("A" ("B1" "B2") "C1" "C2")) @(bind (A B C) (() () ())) @(set (A B . C) D) A ends up with "A", B ends up with ("B1" "B2") and C gets ("C1" and "C2"). .SS The Rebind Directive The @(rebind) directive resembles @(set) but it is not an assignment. It combines the semanticss of @(local), @(bind) and @(set). The expression on the right hand side is evaluated in the current environment. Then the variables in the pattern on the left are introduced as new bindings, whose values come from the pattern. Rebind makes it easy to create temporary bindings based on existing bindings. @(define pattern-function (arg)) @;; inside a pattern function: @(rebind recursion-level @(+ recursion-level 1)) @;; ... @(end) When the function terminates, the previous value of recursion-level is restored. The effect is like the following, but much easier to write and faster to execute: @(define pattern-function (arg)) @;; inside a pattern function: @(local temp) @(set temp recursion-level) @(local recursion-level) @(set recursion-level @(+ temp 1)) @;; ... @(end) .SH BLOCKS .SS Introduction Blocks are sections of a query which are denoted by a name. Blocks denoted by the name nil are understood as anonymous. The @(block NAME) directive introduces a named block, except when the name is the word nil. The @(block) directive introduces an unnamed block, equivalent to @(block nil). The @(skip) and @(collect) directives introduce implicit anonymous blocks, as do function bodies. Blocks are useful for terminating parts of a pattern matchin search prematurely, and escaping to a higher level. This makes blocks not only useful for simplifying the semantics of certain pattern matches, but also an optimization tool. Judicious use of blocks and escapes can reduce or eliminate the amount backtracking that TXR performs. .SS Block Scope The names of blocks are in a distinct namespace from the variable binding space. So @(block foo) has no interaction with the variable @foo. A block extends from the @(block ...) directive which introduces it, until the matching @(end), and may be empty. For instance: @(some) abc @(block foo) xyz @(end) @(end) Here, the block foo occurs in a @(some) clause, and so it extends to the @(end) which terminates the block. After that @(end), the name foo is not associated with a block (is not "in scope"). The second @(end) terminates the @(some) block. The implicit anonymous block introduced by @(skip) has the same scope as the @(skip): it extends over all of the material which follows the skip, to the end of the containing subquery. .SS Block Nesting Blocks may nest, and nested blocks may have the same names as blocks in which they are nested. For instance: @(block) @(block) ... @(end) @(end) is a nesting of two anonymous blocks, and @(block foo) @(block foo) @(end) @(end) is a nesting of two named blocks which happen to have the same name. When a nested block has the same name as an outer block, it creates a block scope in which the outer block is "shadowed"; that is to say, directives which refer to that block name within the nested block refer to the inner block, and not to the outer one. .SS Block Semantics A block normally does nothing. The query material in the block is evaluated normally. However, a block serves as a termination point for @(fail) and @(accept) directives which are in scope of that block and refer to it. The precise meaning of these directives is: .IP "@(fail NAME)" Immediately terminate the enclosing query block called NAME, as if that block failed to match anything. If more than one block by that name encloses the directive, the inner-most block is terminated. No bindings emerge from a failed block. .IP @(fail) Immediately terminate the innermost enclosing anonymous block, as if that block failed to match. If the implicit block introduced by @(skip) is terminated in this manner, this has the effect of causing the skip itself to fail. I.e. the behavior is as if skip search did not find a match for the trailing material, except that it takes place prematurely (before the end of the available data source is reached). If the implicit block associated with a @(collect) is terminated this way, then the entire collect fails. This is a special behavior, because a collect normally does not fail, even if it matches and collects nothing! To prematurely terminate a collect by means of its anonymous block, without failing it, use @(accept). .IP "@(accept NAME)" Immediately terminate the enclosing query block called NAME, as if that block successfully matched. If more than one block by that name encloses the directive, the inner-most block is terminated. Any bindings established within that block until this point emerge from that block. .IP @(accept) Immediately terminate the innermost enclosing anonymous block, as if that block successfully matched. Any bindings established within that block until this point emerge from that block. If the implicit block introduced by @(skip) is terminated in this manner, this has the effect of causing the skip itself to succeed, as if all of the trailing material successfully matched. If the implicit block associated with a @(collect) is terminated this way, then the collection stops. All bindings collected in the current iteration of the collect are discarded. Bindings collected in previous iterations are retained, and collated into lists in accordance with the semantics of collect. Example: alternative way to @(until) termination: @(collect) @ (maybe) --- @ (accept) @ (end) @LINE @(end) This query will collect entire lines into a list called LINE. However, if the line --- is matched (by the embedded @(maybe)), the collection is terminated. Only the lines up to, and not including the --- line, are collected. The effect is identical to: @(collect) @LINE @(until) --- @(end) The difference (not relevant in these examples) is that the until clause has visibility into the bindings set up by the main clause. However, the following example has a different meaning: @(collect) @LINE @ (maybe) --- @ (accept) @ (end) @(end) Now, lines are collected until the end of the data source, or until a line is found which is followed by a --- line. If such a line is found, the collection stops, and that line is not included in the collection! The @(accept) terminates the process of the collect body, and so the action of collecting the last @LINE binding into the list is not performed. .SS Data Extent of Terminated Blocks A query block may have matched some material prior to being terminated by accept. In that case, it is deemed to have only matched that material, and not any material which follows. This may matter, depending on the context in which the block occurs. Example: Query: @(some) @(block foo) @first @(accept foo) @ignored @(end) @second Data: 1 2 3 Output: first="1" second="2" At the point where the accept occurs, the foo block has matched the first line, bound the text "1" to the variable @first. The block is then terminated. Not only does the @first binding emerge from this terminated block, but what also emerges is that the block advanced the data past the first line to the second line. So next, the @(some) directive ends, and propagates the bindings and position. Thus the @second which follows then matches the second line and takes the text "2". In the following query, the foo block occurs inside a maybe clause. Inside the foo block there is a @(some) clause. Its first subclause matches variable @first and then terminates block foo. Since block foo is outside of the @(some) directive, this has the effect of terminating the @(some) clause: Query: @(maybe) @(block foo) @ (some) @first @ (accept foo) @ (or) @one @two @three @four @ (end) @(end) @second Data: 1 2 3 4 5 Output: first="1" second="2" The second clause of the @(some) directive, namely: @one @two @three @four is never processed. The reason is that subclauses are processed in top to bottom order, but the processing was aborted within the first clause the @(accept foo). The @(some) construct never had the opportunity to match four lines. If the @(accept foo) line is removed from the above query, the output is different: Query: @(maybe) @(block foo) @ (some) @first @# <-- @(accept foo) removed from here!!! @ (or) @one @two @three @four @ (end) @(end) @second Data: 1 2 3 4 5 Output: first="1" one="1" two="2" three="3" four="4" second="5" Now, all clauses of the @(some) directive have the opportunity to match. The second clause grabs four lines, which is the longest match. And so, the next line of input available for matching is 5, which goes to the @second variable. .SS Interaction between Trailer and Accept Directives If one of the clauses which follow a @(trailer) request a successful termination to an outer block via @(accept), then @(trailer) intercepts the transfer and adjusts the data extent to the position that it was given. Example: Query: @(block) @(trailer) @line1 @line2 @(accept) @(end) @line3 Data: 1 2 3 Output: line1="1" line2="2" line3="1" The variable line3 is bound to 1 because although the @(accept) yields a data position which is advanced to the third line, this is intercepted by @(trailer) and adjusted back to the first line. Directives other than @(trailer) have no such special interaction with accept. .SH FUNCTIONS .SS Introduction .B TXR functions allow a query to be structured to avoid repetition. On a theoretical note, because .B TXR functions support recursion, functions enable TXR to match some kinds of patterns which exhibit self-embedding, or nesting, and thus cannot be matched by a regular language. Functions in .B TXR are not exactly like functions in mathematics or functional languages, and are not like procedures in imperative programming languages. They are not exactly like macros either. What it means for a .B TXR function to take arguments and produce a result is different from the conventional notion of a function. A .B TXR function may have one or more parameters. When such a function is invoked, an argument must be specified for each parameter. However, a special behavior is at play here. Namely, some or all of the argument expressions may be unbound variables. In that case, the corresponding parameters behave like unbound variables also. Thus .B TXR function calls can transmit the "unbound" state from argument to parameter. It should be mentioned that functions have access to all bindings that are visible in the caller; functions may refer to variables which are not mentioned in their parameter list. With regard to returning, .B TXR functions are also unconventional. If the function fails, then the function call is considered to have failed. The function call behaves like a kind of match; if the function fails, then the call is like a failed match. When a function call succeeds, then the bindings emanating from that function are processed specially. Firstly, any bindings for variables which do not correspond to one of the function's parameters are thrown away. Functions may internally bind arbitrary variables in order to get their job done, but only those variables which are named in the function argument list may propagate out of the function call. Thus, a function with no arguments can only indicate matching success or failure, but not produce any bindings. Secondly, variables do not propagate out of the function directly, but undergo a renaming. For each parameter which went into the function as an unbound variable (because its corresponding argument was an unbound variable), if that parameter now has a value, that value is bound onto the corresponding argument. Example: @(define collect_words (list)) @(coll)@{list /[^ \t]/}@(end) @(end) The above function "collect_words" contains a query which collects words from a line (sequences of characters other than space or tab), into the list variable called "list". This variable is named in the parameter list of the function, therefore, its value, if it has one, is permitted to escape from the function call. Suppose the input data is: Fine summer day and the function is called like this: @(collect_words wordlist) The result is: wordlist[0]=Fine wordlist[1]=summer wordlist[1]=day How it works is that in the function call @(collect_words wordlist), "wordlist" is an unbound variable. The parameter corresponding to that unbound variable is the parameter "list". Therefore, that parameter is unbound over the body of the function. The function body collects the words of "Fine summer day" into the variable "list", and then yields the that binding. Then the function call completes by noticing that the function parameter "list" now has a binding, and that the corresponding argument "wordlist" has no binding. The binding is thus transferred to the "wordlist" variable. After that, the bindings produced by the function are thrown away. The only enduring effects are: .IP - the function matched and consumed some input; and .IP - the function succeeded; and .IP - the wordlist variable now has a binding. .PP Another way to understand the parameter behavior is that function parameters behave like proxies which represent their arguments. If an argument is an established value, such as a character string or bound variable, the parameter is a proxy for that value and behaves just like that value. If an argument is an unbound variable, the function parameter acts as a proxy representing that unbound variable. The effect of binding the proxy is that the variable becomes bound, an effect which is settled when the function goes out of scope. Within the function, both the original variable and the proxy are visible simultaneously, and are independent. What if a function binds both of them? Suppose a function has a parameter called P, which is called with an argument A, and then in the function @A and @P are bound. This is permitted, and they can even be bound to different values. However, when the function terminates, the local binding of A simply disappears (because, remember, the symbol A is not a member of the list of parameters). Only the value bound to P emerges, and is bound to A, which still appears unbound at that point. .SS Definition Syntax Function definition syntax comes in two flavors: vertical and horizontal. Horizontal definitions actually come in two forms, the distinction between which is hardly noticeable, and the need for which is made clear below. A function definition begins with a @(define ...) directive. For vertical functions, this is the only element in a line. The define symbol must be followed by a symbol, which is the name of the function being defined. After the symbol, there is a parenthesized optional argument list. If there is no such list, or if the list is specified as () or the symbol "nil" then the function has no parameters. Examples of valid define syntax are: @(define foo) @(define bar ()) @(define match (a b c)) If the define directive is followed by more material on the same line, then it it defines a horizontal function: @(define match_x)x@(end) If the define is the sole element in a line, then it is a vertical function, and the function definition continues below: @(define match_x) x @(end) The difference between the two is that a horizontal function matches characters within a line, whereas a vertical function matches lines within a stream. The former match_x matches the character x, advancing to the next character position. The latter match_x matches a line consisting of the character x, advancing to the next line. Material between @(define) and @(end) is the function body. The define directive may be followed directly by the @(end) directive, in which case the function has an empty body. Functions may be nested within function bodies. Such local functions have dynamic scope. They are visible in the function body in which they are defined, and in any functions invoked from that body. The body of a function is an anonymous block. (See BLOCKS above). .SS Two Forms of The Horizontal Function If a horizontal function is defined as the only element of a line, it may not be followed by additional material. The following construct is erroneous: @(define horiz (x))@foo:@bar@(end)lalala This kind of definition is actually considered to be in the vertical context, and like other directives that have special effects and that do not match anything, it does not consume a line of input. If the above syntax were allowed, it would mean that the line would not only define a function but also match "lalala". This would, in turn, would mean that the @(define)...@(end) is actually in horizontal mode, and so it matches a span of zero characters within a line (which means that is would require a line of input to match: a nasty thing for a non-matching directive to do!) A horizontal function can be defined in an actual horizontal context. This occurs if its is in a line where it is preceded by other material. For instance: X@(define fun)...@(end)Y This is a query line which must match the text XY. It also defines the function fun. The main use of this form is for nested horizontal functions: @(define fun)@(define local_fun)...@(end)@(end) .SS Vertical-Horizontal Overloading A function of the same name may be defined as both vertical and horizontal. Both functions are available at the same time. Which one is used by a call is resolved by context. See the section Vertical Versus Horizontal Calls below. .SS Call Syntax A function is invoked by compound directive whose first symbol is the name of that function. Additional elements in the directive are the arguments. Arguments may be symbols, or other objects like string and character literals, quasiliterals ore regular expressions. Example: Query: @(define pair (a b)) @a @b @(end) @(pair first second) @(pair "ice" cream) Data: one two ice milk Output: first="one" second="two" cream="milk" The first call to the function takes the line "one two". The parameter "a" takes "one" and parameter b takes "two". These are rebound to the arguments first and second. The second call to the function binds the a parameter to the word "ice", and the b is unbound, because the corresponding argument "cream" is unbound. Thus inside the function, @a is forced to match "ice". Then a space is matched and @b collects the text "milk". When the function returns, the unbound "cream" variable gets this value. If a symbol occurs multiple times in the argument list, it constrains both parameters to bind to the same value. That is to say, all parameters which, in the body of the function, bind a value, and which are all derived from the same argument symbol must bind to the same value. This is settled when the function terminates, not while it is matching. Example: Query: @(define pair (a b)) @a @b @(end) @(pair same same) Data: one two Output: [query fails, prints "false"] .SS Vertical Versus Horizontal Calls A function call which is the only element of the query line in which it occurs is ambiguous. It can go either to a vertical function or to the horizontal one. If both are defined, then it goes to the vertical one. Example: Query: @(define which (x))@(bind x "horizontal")@(end) @(define which (x)) @(bind x "vertical") @(end) @(which fun) Output: fun="vertical" Not only does this call go to the vertical function, but it is in a vertical context. If only a horizontal function is defined, then that is the one which is called, even if the call is the only element in the line. This takes place in a horizontal character-matching context, which requires a line of input which can be traversed: Example: Query: @(define which (x))@(bind x "horizontal")@(end) @(which fun) Data: ABC Output: false The query failed. Why? Because since @(which fun) is in horizontal mode, it matches characters in a line. Since the function body consists of @(bind ...) which doesn't match any characters, the function call requires an empty line to match. The line ABC is not empty, and so there is a matching failure. The following example corrects this: Example: Query: @(define which (x))@(bind x "horizontal")@(end) @(which fun) Data: Output: fun="horizontal" A call made in a clearly horizontal context will prefer the horizontal function, and only fall back on the vertical one if the horizontal one doesn't exist. (In this fall-back case, the vertical function is called with empty data; it is useful for calling vertical functions which process arguments and produce values.) In the next example, the call is followed by trailing material, placing it in a horizontal context. Leading material will do the same thing: Example: Query: @(define which (x))@(bind x "horizontal")@(end) @(define which (x)) @(bind x "vertical") @(end) @(which fun)B Data: B Output: fun="horizontal" .SS Nested Functions Function definitions may appear in a function. Such definitions are visible in all functions which are invoked from the body (and not necessarily enclosed in the body). In other words, the scope is dynamic, not lexical. Inner definitions shadow outer definitions. This means that a caller can redirect the function calls that take place in a callee, by defining local functions which capture the references. Example: Query: @(define which) @ (fun) @(end) @(define fun) @ (output) toplevel fun! @ (end) @(end) @(define callee) @ (define fun) @ (output) local fun! @ (end) @ (end) @ (which) @(end) @(callee) @(which) Output: local fun! toplevel fun! Here, the function "which" is defined which calls "fun". A toplevel definition of "fun" is introduced which outputs "toplevel fun!". The function "callee" provides its own local definition of "fun" which outputs "local fun!" before calling "which". When callee is invoked, it calls @(which), whose @(fun) call is routed to callee's local definition. When @(which) is called directly from the top level, its @(fun) call goes to the toplevel definition. .SH MODULARIZATION .SS The Load Directive The syntx of the load directive is: @(load EXPR) Where EXPR evaluates to a string giving the path of the file to load. Unless the path is absolute, it is interpreted relative to the directory of the source file from which the @(load) form was read. If there was no such source file (for instance, the script was read from standard input), then it is resolved relative to the current working directory. If the file cannot be opened, then the .txr suffix is added and another attempt is made. Thus load expressions need not refer to the suffix. In the future, additional suffixes may be searched (compiled versions of a file). Loading is performed at evaluation time; it is not a source file inclusion mechanism. A TXR script is read from beginning to end and parsed prior to being evaluated. See also: the *self-path* variable in TXR Lisp. .SH OUTPUT .SS Introduction A .B TXR query may perform custom output. Output is performed by @(output) clauses, which may be embedded anywhere in the query, or placed at the end. Output occurs as a side effect of producing a part of a query which contains an @(output) directive, and is executed even if that part of the query ultimately fails to find a match. Thus output can be useful for debugging. An output clause specifies that its output goes to a file, pipe, or (by default) standard output. If any output clause is executed whose destination is standard output, .B TXR makes a note of this, and later, just prior to termination, suppresses the usual printing of the variable bindings or the word false. .SS The Output Directive The syntax of the @(output) directive is: @(output [ DESTINATION ] { bool-keyword | keyword value }* ) . . one or more output directives or lines . @(end) The optional destination is a filename, the special name, - which redirects to standard output, or a shell command preceded by the ! symbol. In the first form, the destination may be specified as a variable which holds text, a string literal or a quasiliteral The keyword list consists of a mixture of boolean keywords which do not have an argument, or keywords with arguments. The following boolean keywords are supported: .IP :nothrow The output directive throws an exception if the output destination cannot be opened, unless the :nothrow keyword is present, in which case the situation is treated as a match failure. Note that since command pipes are processes that report errors asynchronously, a failing command will not throw an immediate exception that can be suppressed with :nothrow. This is for synchronous errors, like trying to open a destination file, but not having permissions, etc. .IP :append This keyword is meaningful for files, specifying append mode: the output is to be added to the end of the file rather than overwriting the file. The following value keywords are supported: .IP :filter The argument can be a symbol, which specifies a filter to be applied to the variable substitutions occuring within the output clause. The argument can also be a list of filter symbols, which specifies that multiple filters are to be applied, in left to right order. See the later sections Output Filtering below, and The Deffilter Directive. .IP :into The argument of :into is a symbol which denotes a variable. The output will go into that variable. If the variable is unbound, it will be created. Otherwise, its contents are overwritten unless the :append keyword is used. If :append is used, then the new content will be appened to the previous content of the variable, after flattening the content to a list, as if by the @(flatten) directive. .SS Output Text Text in an output clause is not matched against anything, but is output verbatim to the destination file, device or command pipe. .SS Output Variables Variables occurring in an output clause do not match anything, but instead their contents are output. A variable being output can be any object. If it is of a type other than a string, it will be converted to a string as if by the tostring function in TXR Lisp. A list is converted to a string in a special way: the elements are individually converted to a string and then they are catenated together. The default separator string is a single space: an alternate separation can be specified as an argument in the brace substitution syntax. Lists may be output within @(repeat) or @(rep) clauses. Each nesting of these constructs removes one level of nesting from the list variables that it contains. In an output clause, the @{NAME NUMBER} variable syntax generates fixed-width field, which contains the variable's text. The absolute value of the number specifies the field width. For instance -20 and 20 both specify a field width of twenty. If the text is longer than the field, then it overflows the field. If the text is shorter than the field, then it is left-adjusted within that field, if the width is specified as a positive number, and right-adjusted if the width is specified as negative. An output variable may specify a filter which overrides any filter established for the output clause. The syntax for this is @(NAME :filter }. The filter specification syntax is the same as in the output clause. See Output Filtering below. .SS Output Variables: Indexing Additional syntax is supported in output variables that is does not appear in pattern matching variables. A square bracket index notation may be used to extract elements or ranges from a variable, which works with strings, vectors and lists. Elements are indexed from zero. This notation is only available in brace-enclosed syntax, and looks like this: .IP @{NAME[expr]} Extract the element at the position given by expr. .IP @{NAME[expr1..expr2]} Extract a range of elements from the position given by expr1, up to one position less than the position given by expr2. If the variable is a list, it is treated as a list substitution, exactly as if it were the value of an unsubscripted list variable. The elements of the list are converted to strings and catenated together wit ha separator string between them, the default one being a single space. An alternate character may be given as a string argument in the brace notation. .TP Example: @(bind a ("a" "b" "c" "d")) @(output) @{a[1..3] "," 10} @(end) The above produces the text "b,c" in a field 10 spaces wide. The [1..3] argument extracts a range of a; the "," argument specifies an alternate separator string, and 10 specifies the field width. .SS Output Substitutions The brace syntax has another syntactic and semantic extension. In place of the symbol, an expression may appear. The value of that expression is substituted. Example: @(bind a "foo") @(output) @{`@a:` -10} Here, the quasiliteral expression `@a:` is evaluated, producing the string "foo:". This string is printed right-adjusted in a 10 character field. .SS The Repeat Directive The repeat directive is generates repeated text from a ``boilerplate'', by taking successive elements from lists. The syntax of repeat is like this: @(repeat) . . main clause material, required . . special clauses, optional . . @(end) Repeat has four types of special clauses, any of which may be specified with empty contents, or omitted entirely. They are described below. Repeat takes arguments, also described below. All of the material in the main clause and optional clauses is examined for the presence of variables. If none of the variables hold lists which contain at least one item, then no output is performed, (unless the repeat specifies an @(empty) clause, see below). Otherwise, among those variables which contain non-empty lists, repeat finds the length of the longest list. This length of this list determines the number of repetitions, R. If the repeat contains only a main clause, then the lines of this clause is output R times. Over the first repetition, all of the variables which, outside of the repeat, contain lists are locally rebound to just their first item. Over the second repetition, all of the list variables are bound to their second item, and so forth. Any variables which hold shorter lists than the longest list eventually end up with empty values over some repetitions. Example: if the list A holds "1", "2" and "3"; the list B holds "A", "B"; and the variable C holds "X", then @(repeat) >> @C >> @A @B @(end) will produce three repetitions (since there are two lists, the longest of which has three items). The output is: >> X >> 1 A >> X >> 2 B >> X >> 3 The last line has a trailing space, since it is produced by "@A @B", where @B has an empty value. Since C is not a list variable, it produces the same value in each repetition. The special clauses are: .IP @(single) If the repeat produces exactly one repetition, then the contents of this clause are processed for that one and only repetition, instead of the main clause or any other clause which would otherwise be processed. .IP @(first) The body of this clause specifies an alternative body to be used for the first repetition, instead of the material from the main clause. .IP @(last) The body of this clause is used instead of the main clause for the last repetition. .IP @(empty) If the repeat produces no repetitions, then the body of this clause is output. If this clause is absent or empty, the repeat produces no output. .IP "@(mod n m)" The forms n and m are expressions that evaluate to integers. The value of m should be nonzero. The clause denoted this way is active if the repetition modulo m is equal to n. The first repetition is numbered zero. For instance the clause headed by @(mod 0 2) will be used on repetitions 0, 2, 4, 6, ... and @(mod 1 2) will be used on repetitions 1, 3, 5, 7, ... .IP "@(modlast n m)" The meaning of n and m is the same as in @(mod n m), but one more condition is imposed. This clause is used if the repetition modulo m is equal to n, and if it is the last repetition. .PP The precedence among the clauses which take an iteration is: single > first > mod > modlast > last > main. That is if two or more of these clauses can apply to a repetition, then the leftmost one in this precedence list applies. For instance, if there is just a single repetition, then any of these special clause types can apply to that repetition, since it is the only repetition, as well as the first and last one. In this situation, if there is a @(single) clause present, then the repetition is processed using that clause. Otherwise, if there is a @(first) clause present, that clause is used. Failing that, @(mod) is used if there is such a clause and its numeric conditions are satisfied. If not then @(modlast) clauses are considered, and if there are none, or none of them activate, then @(last) is considered. If none of those clauses are present or apply, then the repetition is processed using the main clause. Repeat supports an optional keyword argument: @(repeat [:counter ]) This designates a symbol which will behave as an integer variable over the scope of the clauses inside the repeat. The variable provides access to the reptition count, starting at zero, incrementing with each repetition. .SS Nested Repeats If a repeat clause encloses variables which holds multidimensional lists, those lists require additional nesting levels of repeat (or rep). It is an error to attempt to output a list variable which has not been decimated into primary elements via a repeat construct. Suppose that a variable X is two-dimensional (contains a list of lists). X must be twice nested in a repeat. The outer repeat will walk over the lists contained in X. The inner repeat will walk over the elements of each of these lists. A nested repeat may be embedded in any of the clauses of a repeat, not only the main clause. .SS The Rep Directive The @(rep) directive is similar to @(repeat), but whereas @(repeat) is line oriented, @(rep) generates material within a line. It has all the same clauses, but everything is specified within one line: @(rep)... main material ... .... special clauses ...@(end) More than one @(rep) can occur within a line, mixed with other material. A @(rep) can be nested within a @(repeat) or within another @(rep). .SS Repeat and Rep Examples Example 1: show the list L in parentheses, with spaces between the elements, or the symbol NIL if the list is empty: @(output) @(rep)@L @(single)(@L)@(first)(@L @(last)@L)@(empty)NIL@(end) @(end) Here, the @(empty) clause specifies NIL. So if there are no repetitions, the text NIL is produced. If there is a single item in the list L, then @(single)(@L) produces that item between parentheses. Otherwise if there are two or more items, the first item is produced with a leading parenthesis followed by a space by @(first)(@L , and the last item is produced with a closing parenthesis: @(last)@L). All items in between are emitted with a trailing space by the main clause: @(rep)@L . Example 2: show the list L like Example 1 above, but the empty list is (). @(output) (@(rep)@L @(last)@L@(end)) @(end) This is simpler. The parentheses are part of the text which surrounds the @(rep) construct, produced unconditionally. If the list L is empty, then @(rep) produces no output, resulting in (). If the list L has one or more items, then they are produced with spaces each one, except the last which has no space. If the list has exactly one item, then the @(last) applies to it instead of the main clause: it is produced with no trailing space. .SS Output Filtering Often it is necessary to transform the output to preserve its meaning under the convention of a given data format. For instance, if a piece of text contains the characters < or >, then if that text is being substituted into HTML, these should be replaced by < and >. This is what filtering is for. Filtering is applied to the contents of output variables, not to any template text. .B TXR implements named filters. Built-in filters are named by keywords, given below. User-defined filters are possible, however. See notes on the deffilter directive below. Instead of a filter name, the syntax (fun NAME) can be used. This denotes that the function called NAME is to be used as a filter. This is discussed in the next section Function Filters below. Built-in filters named by keywords: .IP :to_html Filter text to HTML, representing special characters using HTML ampersand sequences. For instance '>' is replaced by '>'. .IP :from_html Filter text with HTML codes into text in which the codes are replaced by the corresponding characters. For instance '>' is replaced by '>'. .IP :upcase Convert the 26 lower case letters of the English alphabet to upper case. .IP :downcase Convert the 26 upper case letters of the English alphabet to lower case. .IP :frompercent Decode percent-encoded text. Character triplets consisting of the % character followed by a pair of hexadecimal digits (case insensitive) are are converted to bytes having the value represented by the hexadecimal digits (most significant nybble first). Sequences of one or more such bytes are treated as UTF-8 data and decoded to characters. .IP :topercent Convert to percent encoding according to RFC 3986. The text is first converted to UTF-8 bytes. The bytes are then converted back to text as follows. Bytes in the range 0 to 32, and 127 to 255 (note: including the ASCII DEL), bytes whose values correspond to ASCII characters which are listed by RFC 3986 as being in the "reserved set", and the byte value corresponding to the ASCII % character are encoded as a three-character sequence consisting of the % character followed by two hexadecimal digits derived from the byte value (most significant nybble first, upper case). All other bytes are converted directly to characters of the same value without any such encoding. .IP :fromurl Decode from URL encoding, which is like percent encoding, except that if the unencoded + character occurs, it is decoded to a space character. Of course %20 still decodes to space, and %2B to the + character. .IP :tourl Encode to URL encoding, which is like percent encoding except that a space maps to + rather than %20. The + character, being in the reserved set, encodes to %2B. .PP :tonumber Converts strings to numbers. Strings that contain a period, e or E are converted to floating point as if by the function flo-str. Otherwise they are converted to integer as if using int-str with a radix of 10. Non-numeric junk results in the object nil. .PP :tointeger Converts strings to integers as if using int-str with a radix of 10. Non-numeric junk results in the object nil. .PP :tofloat Converts strings to floating-point values as if using the function flo-str. Non-numeric junk results in the object nil. .PP :hextoint Converts strings to integers as if using int-str with a radix of 16. Non-numeric junk results in the object nil. .PP .TP Examples To escape HTML characters in all variable substitutions occuring in an output clause, specify :filter :to_html in the directive: @(output :filter :to_html) ... @(end) To filter an individual variable, add the syntax to the variable spec: @(output) @{x :filter :to_html} @(end) Multiple filters can be applied at the same time. For instance: @(output) @{x :filter (:upcase :to_html)} @(end) This will fold the contents of x to upper case, and then encode any special characters into HTML. Beware of combinations that do not make sense. For instance, suppose the original text is HTML, containing codes like '"'. The compound filter (:upcase :from_html) will not work because '"' will turn to '"' which no longer be recognized by the :from_html filter, because the entity names in HTML codes are case-sensitive. Capture some numeric variables and convert to numbers: @date @time @temperature @pressure @(filter :tofloat temperature pressure) @;; temperature and pressure can now be used in calculations .SS Function Filters A function can be used as a filter. For this to be possible, the function must conform to certain rules: .IP 1. The function must take two special arguments, which may be followed by additional arguments. .IP 2. When the function is called, the first argument will be bound to a string, and the second argument will be unbound. The function must produce a value by binding it to the second argument. If the filter is to be used as the final filter in a chain, it must produce a string. For instance, the following is a valid filter function: @(define foo_to_bar (in out)) @ (next :string in) @ (cases) foo @ (bind out "bar") @ (or) @ (bind out in) @ (end) @(end) This function binds the out parameter to "bar" if the in parameter is "foo", otherwise it binds the out parameter to a copy of the in parameter. This is a simple filter. To use the filter, use the syntax (:fun foo_to_bar) in place of a filter name. For instance in the bind directive: @(bind "foo" "bar" :lfilt (:fun foo_to_bar)) The above should succeed since the left side is filtered from "foo" to "bar", so that there is a match. Of course, function filters can be used in a chain: @(output :filter (:downcase (:fun foo_to_bar) :upcase)) ... @(end) Here is a split function which takes an extra argument. @(define split (in out sep)) @ (next :list in) @ (coll)@(maybe)@token@sep@(or)@token@(end)@(end) @ (bind out token) @(end) Furthermore, note that it produces a list rather than a string. This function separates the argument in into tokens according to the separator text sep. Here is another function, join, which catenates a list: @(define join (in out sep)) @ (output :into out) @ (rep)@in@sep@(last)@in@(end) @ (end) @(end) Now here is these two being used in a chain: @(bind text "how,are,you") @(output :filter (:fun split ",") (:fun join "-")) @text @(end) Output: how-are-you When the filter invokes a function, it generates the first two arguments internally to pass in the input value and capture the output. The remaining arguments from the (:fun ...) construct are also passed to the function. Thus the "," and "-" are passed as the sep argument to split and join. Note that split puts out a list, which join accepts. So the overall filter chain operates on a string: a string goes into split, and a string comes out of join. .SS The Deffilter Directive The deffilter directive allows a query to define a custom filter, which can then be used in @(output) clauses to transform substituted data. This directive's syntax is illustrated in this example: Query: @(deffilter rot13 ("a" "n") ("b" "o") ("c" "p") ("d" "q") ("e" "r") ("f" "s") ("g" "t") ("h" "u") ("i" "v") ("j" "w") ("k" "x") ("l" "y") ("m" "z") ("n" "a") ("o" "b") ("p" "c") ("q" "d") ("r" "e") ("s" "f") ("t" "g") ("u" "h") ("v" "i") ("w" "j") ("x" "k") ("y" "l") ("z" "m")) @(collect) @line @(end) @(output :filter rot13) @(repeat) @line @(end) @(end) Input: hey there! Output: url gurer! The deffilter symbol must be followed by the name of the filter to be defined, followed by forms which evaluate to lists of strings. Each list must be at least two elements long and specifies one or more texts which are mapped to a replacement text. For instance, the following specifies a telephone keypad mapping from upper case letters to digits. @(deffilter alpha_to_phone ("E" "0") ("J" "N" "Q" "1") ("R" "W" "X" "2") ("D" "S" "Y" "3") ("F" "T" "4") ("A" "M" "5") ("C" "I" "V" "6") ("B" "K" "U" "7") ("L" "O" "P" "8") ("G" "H" "Z" "9")) @(deffilter foo (`@a` `@b`) ("c" `->@d`)) @(bind x ("from" "to")) @(bind y ("---" "+++")) @(deffilter sub x y) The last deffilter above equivalent to @(deffilter sub ("from" "to") ("---" "+++")). Filtering works using a longest match algorithm. The input is scanned from left to right, and the longest piece of text is identified at every character position which matches a string on the left hand side, and that text is replaced with its associated replacement text. The scanning then continues at the first character after the matched text. If none of the strings matches at a given character position, then that character is passed through the filter untranslated, and the scan continues at the next character in the input. Filtering is not in-place but rather instantiates a new text, and so replacement text is not re-scanned for more replacements. If a filter definition accidentally contains two or more repetitions of the same left hand string with different right hand translations, the later ones take precedence. No warning is issued. .SS The Filter Directive The syntax of the filter directive is: @(filter FILTER { VAR }+ ) A filter is specified, followed by one or more variables whose values are filtered and stored back into each variable. Example: convert a, b, and c to upper case and HTML encode: @(filter (:upcase :to_html) a b c) .SH EXCEPTIONS .SS Introduction The exceptions mechanism in .B TXR is another disciplined form of non-local transfer, in addition to the blocks mechanism (see BLOCKS above). Like blocks, exceptions provide a construct which serves as the target for a dynamic exit. Both blocks and exceptions can be used to bail out of deep nesting when some condition occurs. However, exceptions provide more complexity. Exceptions are useful for error handling, and TXR in fact maps certain error situations to exception control transfers. However, exceptions are not inherently an error-handling mechanism; they are a structured dynamic control transfer mechanism, one of whose applications is error handling. An exception control transfer (simply called an exception) is always identified by a symbol, which is its type. Types are organized in a subtype-supertype hierarchy. For instance, the file_error exception type is a subtype of the error type. This means that a file error is a kind of error. An exception handling block which catches exceptions of type error will catch exceptions of type file_error, but a block which catches file_error will not catch all exceptions of type error. A query_error is a kind of error, but not a kind of file_error. The symbol t is the supertype of every type: every exception type is considered to be a kind of t. (Mnemonic: t stands for type, as in any type). Exceptions are handled using @(catch) clauses within a @(try) directive. In addition to being useful for exception handling, the @(try) directive also provides unwind protection by means of a @(finally) clause, which specifies query material to be executed unconditionally when the try clause terminates, no matter how it terminates. .SS The Try Directive The general syntax of the try directive is @(try) ... main clause, required ... ... optional catch clauses ... ... optional finally clause @(end) A catch clause looks like: @(catch TYPE) . . . and also the this form, equivalent to @(catch (t)): @(catch) . . . which catches all exceptions. A finally clause looks like: @(finally) ... . . The main clause may not be empty, but the catch and finally may be. A try clause is surrounded by an implicit anonymous block (see BLOCKS section above). So for instance, the following is a no-op (an operation with no effect, other than successful execution): @(try) @(accept) @(end) The @(accept) causes a successful termination of the implicit anonymous block. Execution resumes with query lines or directives which follow, if any. Try clauses and blocks interact. For instance, a block accept from within a try clause invokes a finally. Query: @(block foo) @ (try) @ (accept foo) @ (finally) @ (output) bye! @ (end) @ (end) Output: bye! How this works: the try block's main clause is @(accept foo). This causes the enclosing block named foo to terminate, as a successful match. Since the try is nested within this block, it too must terminate in order for the block to terminate. But the try has a finally clause, which executes unconditionally, no matter how the try block terminates. The finally clause performs some output, which is seen. .SS The Finally Clause A try directive can terminate in one of three ways. The main clause may match successfully, and possibly yield some new variable bindings. The main clause may fail to match. Or the main clause may be terminated by a non-local control transfer, like an exception being thrown or a block return (like the block foo example in the previous section). No matter how the try clause terminates, the finally clause is processed. Now, the finally clause is itself a query which binds variables, which leads to the question: what happens to such variables? What if the finally block fails as a query? Another question is: what if a finally clause itself initiates a control transfer? Answers follow. Firstly, a finally clause will contribute variable bindings only if the main clause terminates normally (either as a successful or failed match). If the main clause successfully matches, then the finally block continues matching at the next position in the data, and contributes bindings. If the main clause fails, then the finally block matches at the same position. The overall try directive succeeds as a match if either the main clause or the finally clause succeed. If both fail, then the try directive is a failed match. The subquery in which it is located fails, et cetera. Example: Query: @(try) @a @(finally) @b @(end) @c Data: 1 2 3 Output: a=1 b=2 c=3 In this example, the main clause of the try captures line "1" of the data as variable a, then the finally clause captures "2" as b, and then the query continues with the @c variable after try block, and captures "3". Example: Query: @(try) hello @a @(finally) @b @(end) @c Data: 1 2 Output: b=1 c=2 In this example, the main clause of the try fails to match, because the input is not prefixed with "hello ". However, the finally clause matches, binding b to "1". This means that the try block is a successful match, and so processing continues with @c which captures "2". When finally clauses are processed during a non-local return, they have no externally visible effect if they do not bind variables. However, their execution makes itself known if they perform side effects, such as output. A finally clause guards only the main clause and the catch clauses. It does not guard itself. Once the finally clause is executing, the try block is no longer guarded. This means if a nonlocal transfer, such as a block accept or exception, is initiated within the finally clause, it will not re-execute the finally clause. The finally clause is simply abandoned. The disestablishment of blocks and try clauses is properly interleaved with the execution of finally clauses. This means that all surrounding exit points are visible in a finally clause, even if the finally clause is being invoked as part of a transfer to a distant exit point. The finally clause can make a control transfer to an exit point which is more near than the original one, thereby "hijacking" the control transfer. Also, the anonymous block established by the try directive is visible in the finally clause. Example: @(try) @ (try) @ (next "nonexistent-file") @ (finally) @ (accept) @ (end) @(catch file_error) @ (output) file error caught @ (end) @(end) In this example, the @(next) directive throws an exception of type file_error, because the given file does not exist. The exit point for this exception is the @(catch file_error) clause in the outer-most try block. The inner block is not eligible because it contains no catch clauses at all. However, the inner try block has a finally clause, and so during the processing of this exception which is headed for the @(catch file_error), the finally clause performs an anonymous accept. The exit point for the accept is the anonymous block surrounding the inner try. So the original transfer to the catch clause is forgotten. The inner try terminates successfully, and since it constitutes the main clause of the outer try, that also terminates successfully. The "file error caught" message is never printed. .SS Catch Clauses Catch clauses establish a try block as a potential exit point for an exception-induced control transfer (called a ``throw''). A catch clause specifies an optional list of symbols which represent the exception types which it catches. The catch clause will catch exceptions which are a subtype of any one of those exception types. If a try block has more than one catch clause which can match a given exception, the first one will be invoked. When a catch is invoked, it is of course understood that the main clause did not terminate normally, and so the main clause could not have produced any bindings. Catches are processed prior to finally. If a catch clause itself throws an exception, that exception cannot be caught by that same clause or its siblings in the same try block. The catches of that block are no longer visible at that point. Nevertheless, the catch clauses are still protected by the finally block. If a catch clause throws, the finally block is still processed. If a finally block throws an exception, then it is simply aborted; the remaining directives in that block are not processed. So the success or failure of the try block depends on the behavior of the catch clause or the finally, if there is one. If either of them succeed, then the try block is considered a successful match. Example: Query: @(try) @ (next "nonexistent-file") @ x @ (catch file_error) @a @(finally) @b @(end) @c Data: 1 2 3 Output: a=1 b=2 c=3 Here, the try block's main clause is terminated abruptly by a file_error exception from the @(next) directive. This is handled by the catch clause, which binds variable a to the input line "1". Then the finally clause executes, binding b to "2". The try block then terminates successfully, and so @c takes "3". .SS Catch Clauses with Parameters A catch may have parameters following the type name, like this: @(catch pair (a b)) To write a catch-all with parameters, explicitly write the master supertype t: @(catch t (arg ...)) Parameters are useful in conjunction with throw. The built-in error exceptions generate one argument, which is a string containing the error message. Using throw, arbitrary parameters can be passed from the throw site to the catches. .SS The Throw Directive The throw directive generates an exception. A type must be specified, followed by optional arguments. For example, @(throw pair "a" `@file.txt`) throws an exception of type pair, with two arguments, being "a" and the expansion of the quasiliteral `@file.txt`. The selection of the target catch is performed purely using the type name; the parameters are not involved in the selection. Binding takes place between the arguments given in throw, and the target catch. If any catch parameter, for which a throw argument is given, is a bound variable, it has to be identical to the argument, otherwise the catch fails. (Control still passes to the catch, but the catch is a failed match). Query: @(bind a "apple") @(try) @(throw e "banana") @(catch e (a)) @(end) Output: false If any argument is an unbound variable, the corresponding parameter in the catch is left alone: if it is an unbound variable, it remains unbound, and if it is bound, it stays as is. Query: @(try) @(trow e "honda" unbound) @(catch e (car1 car2)) @car1 @car2 @(end) Data: honda toyota Output: car1="honda" car2="toyota" If a catch has fewer parameters than there are throw arguments, the excess arguments are ignored. Query: @(try) @(throw e "banana" "apple" "pear") @(catch e (fruit)) @(end) Output: fruit="banana" If a catch has more parameters than there are throw arguments, the excess parameters are left alone. They may be bound or unbound variables. Query: @(try) @(trow e "honda") @(catch e (car1 car2)) @car1 @car2 @(end) Data: honda toyota Output: car1="honda" car2="toyota" A throw argument passing a value to a catch parameter which is unbound causes that parameter to be bound to that value. Throw arguments are evaluated in the context of the throw, and the bindings which are available there. Consideration of what parameters are bound is done in the context of the catch. Query: @(bind c "c") @(try) @(forget c) @(bind (a c) ("a" "lc")) @(throw e a c) @(catch e (b a)) @(end) Output: c="c" b="a" a="lc" In the above example, c has a toplevel binding to the string "c", but is then unbound within the try construct, and rebound to the value "c". Since the try construct is terminated by a throw, these modifications of the binding environment are discarded. Hence, at the end of the query, variable c ends up bound to the original value "c". The throw still takes place within the scope of the bindings set up by the try clause, so the values of a and c that are thrown are "a" and "lc". However, at the catch site, variable a does not have a binding. At that point, the binding to "a" established in the try has disappeared already. Being unbound, the catch parameter a can take whatever value the corresponding throw argument provides, so it ends up with "lc". .SS The Defex Directive The defex directive allows the query writer to invent custom exception types, which are arranged in a type hierarchy (meaning that some exception types are considered subtypes of other types). Subtyping means that if an exception type B is a subtype of A, then every exception of type B is also considered to be of type A. So a catch for type A will also catch exceptions of type B. Every type is a supertype of itself: an A is a kind of A. This of course implies that ever type is a subtype of itself also. Furthermore, every type is a subtype of the type t, which has no supertype other than itself. Type nil is is a subtype of every type, including itself. The subtyping relationship is transitive also. If A is a subtype of B, and B is a subtype of C, then A is a subtype of C. Defex may be invoked with no arguments, in which case it does nothing: @(defex) It may be invoked with one argument, which must be a symbol. This introduces a new exception type. Strictly speaking, such an introduction is not necessary; any symbol may be used as an exception type without being introduced by @(defex): @(defex a) Therefore, this also does nothing, other than document the intent to use a as an exception. If two or more argument symbols are given, the symbols are all introduced as types, engaged in a subtype-supertype relationship from left to right. That is to say, the first (leftmost) symbol is a subtype of the next one, which is a subtype of the next one and so on. The last symbol, if it had not been already defined as a subtype of some type, becomes a direct subtype of the master supertype t. Example: @(defex d e) @(defex a b c d) The fist directive defines d as a subtype of e, and e as a subtype of t. The second defines a as a subtype of b, b as a subtype of c, and c as a subtype of d, which is already defined as a subtype of e. Thus a is now a subtype of e. It should be obvious that the above could be condensed to: @(defex a b c d e) Example: Query: @(defex gorilla ape primate) @(defex monkey primate) @(defex human primate) @(collect) @(try) @(skip) @(cases) gorilla @name @(throw gorilla name) @(or) monkey @name @(throw monkey name) @(or) human @name @(throw human name) @(end)@#cases @(catch primate (name)) @kind @name @(output) we have a primate @name of kind @kind @(end)@#output @(end)@#try @(end)@#collect Input: gorilla joe human bob monkey alice Output: we have a primate joe of kind gorilla we have a primate bob of kind human we have a primate alice of kind monkey Exception types have a pervasive scope. Once a type relationship is introduced, it is visible everywhere. Moreover, the defex directive is destructive, meaning that the supertype of a type can be redefined. This is necessary so that something like the following works right. @(defex gorilla ape) @(defex ape primate) These directives are evaluated in sequence. So after the first one, the ape type has the type t as its immediate supertype. But in the second directive, ape appears again, and is assigned the primate supertype, while retaining gorilla as a subtype. This situation could instead be diagnosed as an error, forcing the programmer to reorder the statements, but instead TXR obliges. However, there are limitations. It is an error to define a subtype-supertype relationship between two types if they are already connected by such a relationship, directly or transitively. So the following definitions are in error: @(defex a b) @(defex b c) @(defex a c)@# error: a is already a subtype of c, through b @(defex x y) @(defex y x)@# error: circularity; y is already a supertype of x. .SH TXR LISP The TXR language contains an embedded Lisp dialect called TXR Lisp. This language is exposed in TXR in two ways. Firstly, in any situation that calls for an expression, a Lisp compound expression can be used, if it is preceded by the @ symbol. The Lisp expression is evaluated and its value becomes the value of that expression. Thus, TXR directives are embedded in literal text using @, and Lisp expressions are embedded in directives using @ also. Secondly, the @(do) directive can be used for evaluating one or more Lisp forms, such that their value is thrown away. This is useful for evaluating some Lisp code for the sake of its side effect, such as defining a variable, updating a hash table, et cetera. Examples: Bind variable a to the integer 4: @(bind a @(+ 2 2)) Define several Lisp functions using @(do): @(do (defun add (x y) (+ x y)) (defun occurs (item list) (cond ((null list) nil) ((atom list) (eql item list)) (t (or (eq (first list) item) (occurs item (rest list))))))) .SS Overview TXR Lisp is a small and simple dialect, like Scheme, but much more similar to Common Lisp than Scheme. It has separate value and function binding namespaces, like Common Lisp (and thus is a Lisp-2 type dialect), and represents boolean true and false with the symbols t and nil (but note the case sensitivity of identifiers denoting symbols!) Furthermore, the symbol nil is also the empty list, which terminates nonempty lists. Function and variable bindings are dynamically scoped in TXR Lisp. However, closures do capture variables. .SS Additional Syntax Much of the TXR Lisp syntax has been introduced in the previous sections of the manual, since directive forms are based on it. There is some additional syntax that is useful in TXR Lisp programming. .SS Quoting/Unquoting .IP 'form The quote character in front of a form is used for suppressing evaluation, which is useful for forms that evaluate to something other than themselves. For instance if '(+ 2 2) is evaluated, the value is the three-element list (+ 2 2), wheras if (+ 2 2) is evaluated, the value is 4. Similarly, the value of 'a is the symbol a itself, whereas the value of a is the value of the variable a. Note that TXR Lisp does not have a distinct quote and backquote syntax. There is only one quote, which supports unquoting. A quoted form which contains no unquotes codifies an ordinary quote. A quoted form which contains unquotes expresses a quasiquote. (This should not be confused with a string quasiliteral, although it is a related concept.) .IP ,form Thes comma character is used within a quoted list to denote an unquote. Wheras the quote suppresses evaluation, the comma introduces an exception: an element of a form which is evaluated. For example, list '(a b c ,(+ 2 2) (+ 2 2)) is the list (a b c 4 (+ 2 2)). Everything in the quote stands for itself, except for the ,(+ 2 2) which is evaluated. The presence of ,(+ 2 2) in the quoted list turns it into a quasiquote. .IP ,*form The comma-star operator is used within a quoted list to denote a splicing unquote. Wheras the quote suppresses evaluation, the comma introduces an exception: the form which follows ,* must evaluate to a list. That list is spliced into the quoted list. For example: '(a b c ,*(list (+ 3 3) (+ 4 4) d)) evaluates to (a b c 6 8 d). The expression (list (+ 3 3) (+ 4 4)) is evaluated to produce the list (6 8), and this list is spliced into the quoted template. .IP ,'form The comma-quote combination has a special meaning: the quote always behaves as a regular quote and not a quasiquote, even if form contains unquotes. Therefore, it does not "capture" these unquotes: they cannot "belong" to this quote. The comma and quote "cancel out", so the only effect of comma-quote is to add one level of unquoting. So for instance, whereas in '(a b c '(,d)), the subsitution of d belongs to the inner quote (it is unquoted by the leftmost comma which belongs to the innermost quote) by contrast, in '(a b c '(,',d)) the d is now one comma removed from the leftmost comma and thus the substitution of d belongs to the outer quote. In other dialects of Lisp, this would be written `(a b c `(,',d)), making it explicit which kind of quote is being specified. TXR Lisp works out which kind of quote to use internally. .IP ,*'form The comma-splice form is analogous to comma-quote (see above). Like in the ,' combination, in the ,*' combination, the quote behaves as a regular quote and not a quasiquote. .SS Quasiquoting non-List Objects Quasiquoting is supported over hash table and vector literals (see Vectors and Hashes below). A hash table or vector literal can be quoted, like any object, for instance: '#(1 2 3) ; value is #(1 2 3) If unquotes occur in the vector, it is a quasivector. (let ((a 42)) '#(1 ,a 3)) ; value is #(1 42 3) The vector in the following example is also a quasivector. It contains unquotes, and is though the quote is not directly applied to it, it is surrounded in a quote. (let ((a 42)) '(a b c #(d ,a))) ; value is (a b c #(d 42)) Hash table literals have two parts: the list of hash construction arguments and the key-value pairs. For instance: #H((:equal-based) (a 1) (b 2)) where (:equal-based) is the list of arguments and the pairs are (a 1) and (b 2). In quasiquoting, the arguments and pairs are treated as separate syntax; it is not one big list. So the following is not a possible way to express the above hash: ;; not supported: splicing across the entire syntax (let ((hash-syntax '((:equal-based) (a 1) (b 2)))) '#H(,*hash-syntax)) This is correct: ;; fine: splicing hash arguments and contents separatly (let ((hash-args '(:equal-based)) (hash-contents '((a 1) (b 2)))) '#H(,hash-args ,*hash-contents)) .SS Vectors .IP "#(...)" A hash token followed by a list denotes a vector. For example #(1 2 a) is a three-element vector containing the numbers 1 and 2, and the symbol a. .SS Hashes .IP "#H((*) ( )*)" The notation #H followed by a nested list syntax denotes a hash table literal. The first item in the syntax is a list of keywords. These are the same keywords as are used when calling the function hash to construct a hash table. Allowed keywords are: :equal-based, :weak-keys, :weak-values. An empty list can be specified as nil or (), which defaults to a hash table basd on the eq function, with no weak semantics. .SS Nested Quotes Quotes can be nested. What if it is necessary to unquote something in the nested list? The following will not work in TXR Lisp like it does in Common Lisp or Scheme: '(1 2 3 '(4 5 6 ,(+ 1 2))). This is because the quote is also "active" as a quasiquote, and so the ,(+ 1 2) belongs to the inner quote, which protects it from evaluation. To get the (+ 1 2) value "through" to the inner quote, the unquote syntax must also be nested using multiple commas, like this: '(1 2 3 '(4 5 6 ,',(+ 1 2))). The leftmost comma goes with the innermost quote. The quote between the commas protects the (+ 1 2) from repeated evaluations: the two unquotes call for two evaluations, but we only want (+ 1 2) to be evaluated once. .SS The .. notation In TXR Lisp, there is a special "dotdot" notation consiting of a pair of dots. This can be written between successive atoms or compound expressions, and is a shorthand for cons. That is to say, A .. B translates to (cons A B), and so for instance (a b .. (c d) e .. f . g) means (a (cons b (c d)) (cons e f) . g). This is a syntactic sugar useful in certain situations in which a cons is used to represent a pair of numbers or other objects. For instance, if L is a list, then [L 1 .. 3] computes a sublist of L consisting of elements 1 through 2 (counting from zero). .SS The DWIM Brackets TXR Lisp has a square bracket notation. The syntax [...] is a shorthand way of writing (dwim ...). The [] syntax is useful for situations where the expressive style of a Lisp-1 dialect is useful. For instance if foo is a variable which holds a function object, then [foo 3] can be used to call it, instead of (call foo 3). If foo is a vector, then [foo 3] retrieves the fourth element, like (vecref foo 3). Indexing over lists, strings and hash tables is possible, and the notation is assignable. Furthermore, any arguments enclosed in [] which are symbols are treated according to a modified namespace lookup rule. More details are given in the documentation for the dwim operator. .SS Regular Expressions In TXR Lisp, the / character can occur in symbol names, and the / token is a symbol. Therefore the /regex/ syntax is absent, replaced with the #/regex/ syntax. .SH CONTROL FLOW AND SEQUENCING When the first element of a compound expression is an operator symbol, the interpretation of the meaning of that form is under the complete control of that operator. The following sections list all of the operators available in TXR Lisp. In these sections Syntax is indicated using these conventions: .TP A symbol in angle brackets denotes some syntactic unit: it may be a symbol or compound form. The syntactic unit is explained in the Description section. .TP {syntax}* * This indicates a repetition of zero or more of the given syntax enclosed in the braces or syntactic unit. .TP [syntax] [] Square brackets indicate optional syntax. .TP alternative1 | alternative2 | ... | alternativeN Multiple syntactic variations allowed in one place are indicated as bar-separated items. .SS Operators progn and prog1 .TP Syntax: (progn
*) (prog1 *) .TP Description The progn operator evaluates forms in order, and returns the value of the last form. The return value of (progn) is nil. The prog1 operator evaluates forms in order, and returns the value of the first form. The return value of (prog1) is nil. Various other operators such as let also arrange for the evaluation of a body of forms, the value of the last of which is returned. These operators are said to feature an "implicit progn". .SS Operator cond .TP Syntax: (cond {( {form}*)}*) .TP Description: The cond operator provides a multi-branching conditional evaluation of forms. Enclosed in the cond form are groups of forms expressed as lists. Each group must be a list of at least one form. The forms are processed from left to right as follows: the first form, , in each group is evaluated. If it evaluates true, then the remaining forms in that group, if any, are also evaluated. Processing then terminates and the result of the last form in the group is taken as the result of cond. If is the only form in the group, then result of is taken as the result of cond. If the first form of a group yields nil, then processing continues with the next group, if any. If all form groups yield nil, then the cond form yields nil. This holds in the case that the syntax is empty: (cond) yields nil. .SS Operator if .TP Syntax: (if []) .TP Description: The if operator provides a simple two-way-selective evaluation control. The form is evaluated. If it yields true then is evaluated, and that form's return value becomes the return value of the if. If yields false, then is evaluated and its return value is taken to be that of if. If is omitted, then the behavior is as if were specified as nil. .SS Operator and .TP Syntax: (and {}*) .TP Description: The and operator provides three functionalities in one. It computes the logical "and" function over several forms. It controls evaluation (a.k.a. "short-circuiting"). It also allows the convenient substitution of an arbitrary true value in the true case. The and operator evaluates as follows. First, a return value is established and initialized to the value t. The forms, if any, are evaluated from left to right. The return value is overwritten with the result of each form. Evaluation stops when all forms are exhausted, or when any of them yields nil. When evaluation stops, the operator yields the return value. .TP Examples: (and) -> t (and (> 10 5) (stringp "foo")) -> t (and 1 2 3) -> 3 .SS Operator or .TP Syntax: (or {}*) .TP Description: The and operator provides three functionalities in one. It computes the logical "or" function over several forms. It controls evaluation (a.k.a. "short-circuiting"). The behavior of or also provides for a simplified selection of the first non-nil value from a sequence of forms. The or operator evaluates as follows. First, a return value is established and initialized to the value nil. The forms, if any, are evaluated from left to right. The return value is overwritten with the result of each form. Evaluation stops when all forms are exhausted, or when a form yields a true value. When evaluation stops, the operator yields the return value. .TP Examples: (or) -> nil (or 1 2) -> 1 (or nil 2) -> 2 (or (> 10 20) (stringp "foo")) -> t .SS Operator unwind-protect .TP Syntax: (unwind-protect *) .TP Description: The unwind-protect operator evaluates in such a way that no matter how the execution of terminates, the -s will be executed. The cleanup forms, however, are not protected. If a cleanup form terminates via some non-local jump, the subsequent cleanup forms are not evaluated. Cleanup forms themselves can "hijack" a non-local control transfer such as an exception. If a cleanup form is evaluated during the processing of a dynamic control transfer such as an exception, and that cleanup form initiats its own dynamic control transfer, the original control transfer is aborted and replaced with the new one. .TP Example: (block foo (unwind-protect (progn (return-from foo 42) (format t "not reached!\en")) (format t "cleanup!\n"))) In this example, the protected progn form terminates by returning from block foo. Therefore the form does not complete and so the output "not reached!" is not produced. However, the cleanup form excecutes, producing the output "cleanup!". .SS Operator block .TP Syntax: (block *) .TP Description: The block operator introduces a named block around the execution of some forms. The argument must be a symbol. Since a block name is not a variable binding, keyword symbols are permitted, and so are the symbols t and nil. A block named by the symbol nil is slighlty special: it is understood to be an anonymous block. Blocks in TXR Lisp have dynamic scope. This means that the following situation is allowed: (defun func () (return-from foo 42)) (block foo (func)) The function can return from the foo block even though the foo block does not lexically surround foo. Thus blocks in TXR Lisp provide dynamic non-local returns, as well as returns out of lexical nesting. .TP Dialect Note: In Common Lisp, blocks are lexical. A separate mechanism consisting of catch and throw operators performs non-local transfer based on symbols. The TXR Lisp example: (defun func () (return-from foo 42)) (block foo (func)) is not allowed in Common Lisp, but can be transliterated to: (defun func () (throw 'foo 42)) (catch 'foo (func)) Note that foo is quoted in CL. This underscores the dynamic nature of the construct. THROW itself is a function and not an operator. .SS Operators return, return-from .TP Syntax: (return []) (return-from []) .TP Description: The return operator must be dynamically enclosed within an anonymous block (a block named by the symbol nil). It immediately terminates the evaluation of the innermost anonyous block which encloses it, causing it to return the specified value. If the value is omitted, the anonymous block returns nil. The return-from operator must be dynamically enclosed within a named block whose name matches the argument. It immediately terminates the evaluation of the innermost such block, causing it to return the specified value. If the value is omitted, that block returns nil. .TP Example: (block foo (let ((a "abc\n") (b "def\n")) (pprint a *stdout*) (return-from foo 42) (pprint b *stdout*))) Here, the output produced is "abc". The value of b is not printed because the return-from terminates block foo, and so the second pprint form is not evaluated. .SH EVALUATION .SS Operator dwim .TP Syntax: (dwim *) [*] .TP Description: The dwim operator's name is an acronym: DWIM may be taken to mean "Do What I Mean", or alternatively, "Dispatch, in a Way that is Intelligent and Meaningful". The notation [...] is a shorthand equivalent to (dwim ...) and is the preferred way for writing dwim expressions. The dwim operator takes a variable number of arguments, which are all evaluated in the same way: the first argument is not evaluated different from the remaining arguments. Furthermore, the evaluation of symbols is done differently: all of the enclosing scopes are considered as if the function and variable namespaces are collapsed into a single namespace, in which variable names take precedence over functions if there exist mutiple bindings. This allows functions to be referenced without the fun operator. All forms which are not symbols are evaluated using the normal evaluation rules. The first argument may not be an operator such as let, et cetera. How many are required depends on the type of object to which the first argument expression evaluates: of the first argument. The possibilities are: .IP "[ *]" Call the given the function object to the given arguments. .IP "[ *]" If the first expression evaluates to a symbol, that symbol is resolved in the function namespace, and then the resulting function, if found, is called with the given arguments. .IP "[ ]" Retrieve the specified element from the specified list. Index zero refers to the first element. Indexed list access does not throw exceptions. Negative indices yield nil, and indices beyond the end of a list yield nil. (However assignment to a nonexistent list element throws.) .IP "[ ..]" Retrieve the specified range of elements, exactly as if using (sub-list ). The range of elements is specified in the car and cdr fields of a cons cell, for which the .. (dotdot) syntactic sugar is useful. See the section on Indexing below. .IP "[ ]" Retrieve the specified element of a vector. This is equivalent to (vecref ). .IP "[ ..]" Retrieve the specified range of elements, exactly as if using (sub-vec ). The range of elements is specified in the car and cdr fields of a cons cell, for which the .. (dotdot) syntactic sugar is useful. See the section on Range Indexing below. .IP "[ ]" Retrieve the specified element of a string. This is equivalent to (chr-str ). .IP "[ ..]" Retrieve the specified range of characters from the string, exactly as if using (sub-str ). The range of elements is specified in the car and cdr fields of a cons cell, for which the .. (dotdot) syntactic sugar is useful. See the section on Indexing below. .IP "[ ]" Retrieve a value from the hash table corresponding to , or if there is no such entry. The first argument may not be an operator such as let, only a function. The places denoted by the dwim operator can be assigned. There are some restrictions. List, string and vector ranges can only be replaced using the set operator. The other operators like push do not apply. Characters in a string can only be assigned with set or incremented with inc and dec. The source of a range assignment can be a string, vector or list, regardless of whether the target is a string, vector or list. If the target is a string, the replacement sequence must be a string, or a list or vector of characters. .TP Range Indexing Vector and list range indexing is based from zero. The first element element zero. Furthermore, the value -1 refers to the last element of the vector or list, and -2 to the second last and so forth. So the range 1 .. -2 means "everything except for the first element and the last two". The symbol t represents the position one past the end of the vector, string or list, so 0 .. t denotes the entire list or vector, and the range t .. t represents the empty range just beyond the last element. It is possible to assign to t .. t. For instance: (defvar list '(1 2 3)) (set [list t .. t] '(4)) ;; list is now (1 2 3 4) Either end of the range can also be specified as nil. If the start is specified as nil, it means zero. If the end is specified as nil, it means one element past the end. Thus nil .. nil spans all of the elements. The value zero has a "floating" behavior when used as the end of a range. If the start of the range is a negative value, and the end of the range is zero, the zero is interpreted as being the position past the end of the sequence, rather than the first element. For instance the range -1..0 means the same thing as -1..t or -1..nil. Zero at the start of a range always means the first element, so that 0..-1 refers to all the elements except for the last one. .TP Notes: The dwim operator allows for a Lisp-1 flavor of programming in TXR Lisp, which is normally Lisp-2, with some useful extensions. A Lisp-1 dialect is one in which an expression like (a b) treats both a and b as expressions with the same evaluation rules. The symbols a and b are looked up in a variable namespace. A function call occurs if the value of variable a is a function object. Thus in a Lisp-1, named functions do not exist as such: they are just variable bindings. In a Lisp-1 (car 1 2) means that there is a variable called car, which holds a function. In a Lisp-2 (car 1 2) means that there is a function called car, and so (car car car) is possible, because there can be also a variable called car. The Lisp-1 design has certain disadvantages, which are avoided in TXR Lisp by confining the Lisp-1 expressivity inside the [...] notation, in which operators are not allowed. When round parentheses are used, the normal Lisp-2 rules apply. A "best of both worlds" situation is achieved. The square brackets are just as convenient as parentheses and at the same time visually distinct, making it clear that different rules apply. The Lisp-1 is useful for functional programming, because it eliminates occurences of the call and fun operators. For instance: ;; regular notation (funcall foo (fun second) '((1 a) (2 b))) ;; [] notation [foo second '((1 a) (2 b))] Lisp-1 dialects can also provide useful extensions by giving a meaning to objects other than functions in the first position of a form, and the dwim/[...] syntax does exactly this. However, unlike Lisp-1 dialects, the [] syntax does not allow operators. It .B is an operator: (dwim ...). .SS Function identity .TP Syntax: (identity ) .TP Description: The identity function returns its argument. .TP Notes: The identify function is useful as a functional argument, when a transformation function is required, but no transformation is actually desired. .SS Function eval .TP Syntax: (eval ) .TP Description: The eval function treats the object as a Lisp expression, which is evaluated. The side effects implied by the form are performed, and the value which it produces is returned. The object specifies an environment for resolving the function and variable references encountered in the expression. The object nil can be specified as an environment, in which case the evaluation takes place in the top-level environment. .SH MUTATION .SS Operators inc, dec, set, push, pop, flip and del .TP Syntax: (inc []) (dec []) (set ) (push ) (pop ) (flip ) (del ) .TP Description: These destructive operators update the value of a place. A place is a storage location which is denoted by a form. Place forms are identical to value accessing forms. That is to say, any form recognized as a place by these operators can be evaluated by itself to retrieve the value of the storage location. However, the converse is false: not all forms which access storage location are recognized as places. With are exceptions noted below, it is an error if a place does not exist. For instance, a variable being assigned must exist. Literal objects which are directly specified in the source code are considered part of the program body. Modifying parts of these objects therefore gives rise to self-modifying code. The behavior of self-modifying code is not specified. The inc and dec update the place by adding or subtracting, respectively, a displacement to or from that number. If the expression is specified, then it is evaluated and its value is used as the increment. Otherwise, a default increment of 1 is used. The prior value of the place and the delta must be suitable operands for the + and - functions. (inc x) is equivalent to (set x (+ 1 x)), except that expression x is evaluated only once to determine the storage location. The inc and dec operators return the new value that was stored. The set operator overwrites the previous value of a place with a new value, and also returns that value. The push and pop operators operate on a place which holds a list. The push operator updates the list by replacing it with a new list which has a new item at the front, followed by the previous list. The item is returned. The pop operator performs the reverse operation: it removes the first item from the list and returns it. (push y x) is similar to (let ((temp y)) (set x (cons temp x)) temp) except that x is evaluated only once to determine the storage place, and no such temporary variable is visible to the program. Similarly, (pop x) is much like (let ((temp (car x))) (set x (cdr x)) temp) except that x is evaluated only once, and no such temporary variable is visible to the program. The flip operator toggles a place between true and false. If the place contains a value other than nil, then its value is replaced with nil. If it contains nil, it is replaced with t. The del operator does not modify the value of a place, but rather deletes the place itself. Index values and ranges of lists denoted using the dwim operator indexing notation can be subject to a deletion, as can hash table entries denoted using dwim or gethash. It is an error to try to delete other kinds of places such as simple variables. The del operator returns the value of the place that was deleted. Deleting from a sequence means removing the element or elements. Deleting a hash place means removing the corresponding entry from the hash table. Currently, these forms are recognized as places: (car ) (cdr ) (gethash ) (vecref ) (dwim ...) [ ...] ;; equivalent to (dwim ...) A place denotes a variable. If the variable does not exist, it is an error. The (car ) and (cdr ) places denote the corresponding slots of a cons cell. The form must be an expression which evaluates to a cons. The gethash place denotes a value stored in a hash table. The form must evaluate to a hash table. If the place does not exist in the hash table under the given key, then the destructive operation will create it. In that case, the form is evaluated to determine the initial value of the place. Otherwise it is ignored. The vecref place denotes a vector element, allowing vector elements to be treated as assignment places. The dwim/[] place denotes a vector element, list element, string, or hash table, depending on the type of obj. .SH BINDING AND ITERATION .SS Operators let and let* .TP Syntax: (let ({ | ( )}*) *) (let* ({ | ( )}*) *) .TP Description: The let and let* operators introduce a new scope with variables and evaluate forms in that scope. The operator symbol, either let or let*, is followed by a list which can contain any mixture of variable name symbols, or ( ) pairs. A symbol denotes the name of variable to be instantiated and initialized to the value nil. A symbol specified with an init-form denotes a variable which is intialized from the value of the init-form. The symbols t and nil may not be used as variables, and neither can be keyword symbols: symbols denoted by a leading colon. The difference between let and let* is that in let*, later init-forms have visibility over the variables established by earlier variables in the same let* construct. In plain let, the variables are not visible to any of the init-forms. When the variables are established, then the body forms are evaluated in order. The value of the last form becomes the return value of the let. If the forms are omitted, then the return value nil is produced. The variable list may be empty. .TP Examples: (let ((a 1) (b 2)) (list a b)) -> (1 2) (let* ((a 1) (b (+ a 1))) (list a b (+ a b))) -> (1 2 3) (let ()) -> nil (let (:a nil)) -> error, :a and nil can't be used as variables .SS Operators for and for* .TP Syntax: ({for | for*} ({ | ( )}*) ( *) (*) *) .TP Description: The for and for* operators combine variable binding with loop iteration. The first argument is a list of variables with optional initializers, exactly the same as in the let and let* operators. Furthermore, the difference between for and for* is like that between let and let* with regard to this list of variables. The for operators execute these steps: 1. Establish bindings for the specified variables similarly to let and let*. The variable bindings are visible over the , each , each and each . 2. Establish an anonymous block over the remaining forms, allowing the return operator to be used to terminate the loop. 3. Evaluate . If yields nil, then each is evaluated, and the vale of the last of these forms is is the result value of the for loop. If there are no such forms then the return value is nil. 4. Otherwise, if yields non-nil, then each is evaluated in turn. Then, each is evaluated in turn and processing resumes at step 2. Furthermore, the for operators establish an anonymous block, allowing the return operator to be used to terminate at any point. .SS Operators each, each*, collect-each and collect-each* .TP Syntax: (each ({( )}*) *) (each* ({( )}*) *) (collect-each ({( )}*) *) (collect-each* ({( )}*) *) .TP Description: These operator establish a loop for iterating over the elements of one or more lists. Each must evaluate to a list. The lists are then iterated in parallel over repeated evaluations of the -s, which each variable being assigned to successive elements of its list. The shortest list determines the number of iterations, so if any of the -s evaluate to an empty list, the body is not executed. The body forms are enclosed in an anonymous block, allowing the return operator to terminate the looop prematurely and optionally specify the return value. The collect-each and collect-each* variants are like each and each*, except that for each iteration, the resulting value of the body is collected into a list. When the iteration terminates, the return value is this collection. The alternate forms denoted by the adorned symbols each* and collect-each* variants differ from each and collect-each in the following way. The plain forms evaluate the -s in an environment in which none of the variables are yet visible. By contrast, the alternate forms evaluate each in an environment in which bindings for the previous variables are visible. In this phase of evaluation, variables are list-valued: one by one they are each bound to the list object emanating from their corresponding . Just before the first loop iteration, however, the variables are assigned the first item from each of their lists. .TP Examples: ;; print numbers from 1 to 10 and whether they are even or odd (each* ((n (range 1 10)) (even (collect-each ((n m)) (evenp m)))) ;; n is a list here (format t "~s is ~s\n" n (if even "even" "odd"))) ;; n is an item here Output: 1 is odd 2 is even 3 is odd 4 is even 5 is odd 6 is even 7 is odd 8 is even 9 is odd 10 is even .SH FUNCTION OBJECTS AND NAMED FUNCTIONS .SS Operator defun .TP Syntax: (defun ({ [. ]}*) *) Description: The defun operator introduces a new function in the global function namespace. The function is similar to a lambda, and has the same parameter syntax and semantics as the lambda operator. Unlike in lambda, the -s of a defun are surrounded by a block. The name of this block is the same as the name of the function, making it possible to terminate the function and return a value using (return-from ). For more information, see the definition of the block operator. A function may call itself by name, allowing for recursion. .SS Operator lambda .TP Syntax: (lambda ({}*[. ]) {}*) (lambda {}*) .TP Description: The lambda operator produces a value which is a function. Like in most other Lisps, functions are objects in TXR Lisp. They can be passed to functions as arguments, returned from functions, aggregated into lists, stored in variables, et cetera. The first argument of lambda is the list of parameters for the function. It may be empty, and it may also be an improper list (dot notation) where the terminating atom is a symbol other than nil. It can also be a single symbol. The second and subsequent arguments are the forms making up the function body. The body may be empty. When a function is called, the parameters are instantiated as variables that are visible to the body forms. The variables are initialized from the values of the argument expressions appearing in the function call. The dotted notation can be used to write a function that accepts a variable number of arguments. To write a function that accepts variable arguments only, with no required arguments, use a single symbol. The keyword symbol : can appear in the parameter list. It is not an argument, but a separator between required parameters and optional parameters. When the function is called, optional parameter for which arguments are not supplied take on the value nil. Functions created by lambda capture the surrounding variable bindings. .TP Examples: Counting function. This function, which takes no arguments, captures the variable "counter". Whenever this object is called, it increments the counter by 1 and returns the incremented value. (let ((counter 0)) (lambda () (inc counter))) Function that takes two or more arguments. The third and subsequent arguments are aggregated into a list passed as the single parameter z: (lambda (x y . z) (list 'my-arguments-are x y z)) Variadic funcion: (lambda args (list 'my-list-of-arguments args)) Optional arguments: [(lambda (x : y) (list x y)) 1] -> (1 nil) [(lambda (x : y) (list x y)) 1 2] -> (1 2) .SS Operator call .TP Syntax: (call {}*) .TP Description: The call operator invokes a function. must evaluate to a function. Each is evaluated in left to right order and the resulting values are passed to the function as arguments. The return value of the (call ...) expression is that of the function applied to those arguments. The may be any Lisp form that produces a function as its value: a symbol denoting a variable in which a function is stored, a lambda expression, a function call which returns a function, or (fun ...) expression. .TP Examples: Apply arguments 1 2 to a lambda which adds them to produce 3: (call (lambda (a b) (+ a b)) 1 2) -> 3 Useless use of call on a named function; equivalent to (list 1 2): (call (fun list) 1 2) -> (1 2) .SS Operator fun .TP Syntax: (fun ) .TP Description: The fun operator retrieves the function object corresponding to a named function. . The is a symbol denoting a named function: a built in function, or one defined by defun. .TP Dialect Note: A lambda expression is not a function name in TXR Lisp. The syntax (fun (lambda ...)) is invalid. .SS Function symbol-function .TP Syntax: (symbol-function ) .TP Description: The symbol-function retrieves the toplevel function binding of the given symbol if it has one. If the symbol has no toplevel function binding, the value nil is returned. .TP Dialect note: The symbol-function form is currently not an assignable place. Only the defun operator defines functions. .SS Function func-get-form .TP Syntax: (func-get-form ) .TP Description: The func-get-form function retrieves a source code form of , which must be an interpreted function. The source code form has the syntax ( {}*). .SS Function func-get-env .TP Syntax: (func-get-env ) .TP Description: The func-get-env function retrieves the environment object associated with function . The environment object holds the captured bindings of a lexical closure. .SS Function functionp .TP Syntax: (functionp ) .TP Description: The functionp function returns t if is a function, otherwise it returns nil. .SS Function interp-fun-p .TP Syntax: (interp-fun-p ) .TP Description: The interp-fun-p function returns t if is an interpreted function, otherwise it returns nil. .SH OBJECT TYPE AND EQUIVALENCE .SS Function typeof .TP Syntax: (typeof ) .TP Description The typeof function returns a symbol representing the type of . .RS The core types are identified by the following symbols. .IP cons A cons cell. .IP str String. .IP lit A literal string embedded in the TXR executable image. .IP chr Character. .IP fixnum Fixnum integer. An integer that fits into the value word, not having to be heap allocated. .IP sym Symbol. .IP pkg Symbol package. .IP fun Function. .IP vec Vector. .IP lcons Lazy cons. .IP lstr Lazy string. .IP env Function/variable binding environment. .IP bignum A bignum integer: arbitrary precision integer that is heap-allocated. .PP There are additional kinds of objects, such as streams. .SS Functions null and not .TP Syntax: (null ) (not ) .TP Description: The null and not functions are synonyms. They tests whether is the object nil. They return t if this is the case, nil otherwise. .TP Examples: (null '()) -> t (null nil) -> t (null ()) -> t (if (null x) (format t "x is nil!")) (let ((list '(b c d))) (if (not (memq 'a list)) (format t "list ~s does not contain the symbol a\en"))) .SS Functions eq, eql and equal .TP Syntax: (eq ) (eql ) (equal ) .TP Description: The principal equality test functions eq, eql and equal test whether two objects are equivalent, using different criteria. They return t if the objects are equivalent, and nil otherwise. The eq function uses the strictest equivalence test, called implementation equality. The eq function returns t if, and only if, and are actually the same object. The eq test is is implemented by comparing the raw bit pattern of the value, whether or not it is an immediate value or a pointer to a heaped object, including its type tags. Consequently, two objects of different type are never equal, two character values are eq if they are the same character, and two fixnum integers are eq if they have the same value. All other objects kinds are actually represented as pointers, and are eq if they point to the same object in memory. So two bignum integers might not be eq even if they have the same numeric value, two lists might not be eq even if all their corresponding elements are eq, two strings might not be eq even if they hold identical text, etc. The eql function is slightly less strict than eq. The difference between eql and eq is that if and are bignums which have the same numeric value, eql returns t, even if they are different objects. For all other objects, eql behaves like eq. The equal function is less strict than eql. In general, it recurses into some kinds of aggregate objects to perform a structural equivalence. If and are eql then they are also equal. If the two objects are both cons cells, then they are equal if their "car" fields are equal and their "cdr" fields are equal. If two objects are vectors, they are equal if they have the same length, and their corresponding elements are equal. If two objects are strings, they are equal if they are textually identical. If two objects are functions, they are equal if they have equal environments, and if they have equal functions. Two compiled functions are the same if they are the same function. Two interpreted functions are equal if their list structure is equal. For some aggregate objects, there is no special semantics. Two hashes, symbols, packages, or streams are equal if they are the same hash. Certain object types have a custom equal function. .SH BASIC LIST LIBRARY When the first element of a compound form is a symbol denoting a function, the evaluation takes place as follows. The remaining forms, if any, denote the arguments to the function. They are evaluated in left to right order to produce the argument values, and passed to the function. An exception is thrown if there are not enough arguments, or too many. Programs can define named functions with the defun operator The following are Lisp functions and variables built-in to TXR. .SS Function cons .TP Syntax: (cons ) .TP Description: The cons function allocates, intializes and returns a single cons cell. A cons has two fields called "car" and "cdr", which are accessed by functions of the same name, or by the functions "first" and "rest", which are alternative spellings. Lists are made up of conses. A (proper) list is either the symbol nil denoting an empty list, or a cons cell which holds the first item of the list in its "car", and the list of the remaining items in "cdr". (cons 1 nil) allocates a one element list denoted (1). The "cdr" is nil, so there are no additional items. A cons cell with a "cdr" other than nil is printed with the dotted pair notation. For example (cons 1 2) yields (1 . 2). The notation (1 . nil) is valid as input into the machine, but is printed as (1). A list terminated by an atom other than nil is called an improper list, and the dot notation is extended to cover improper lists. For instance (1 2 . 3) is an improper list of two elements, terminated by 3, and can be constructed using (cons 1 (cons 2 3)). Another notation for this list is (1 . (2 . 3)) The list (1 2) is (1 . (2 . nil)). .SS Function atom .TP Syntax: (atom ) .TP Description: The atom function tests whether is an atom. It returns t if this is the case, nil otherwise. All values which are not cons cells are atoms. (atom x) is equivalent to (not (consp x)). .TP Examples: (atom 3) -> t (atom (cons 1 2)) -> nil (atom "abc") -> t (atom '(3)) -> nil .SS Function consp .TP Syntax: (consp ) .TP Description: The atom function tests whether is a cons. It returns t if this is the case, nil otherwise. (consp x) is equivalent to (not (atom x)). Non-empty lists test positive under consp because a list is represented as a reference to the first cons in a chain of one or more conses. .TP Examples: (consp 3) -> nil (consp (cons 1 2)) -> t (consp "abc") -> nil (consp '(3)) -> t .SS Functions car and first .TP Syntax: (car ) (first ) .TP Description: The functions car and first are synonyms. They retrieve the "car" field of a cons cell. (car (cons 1 2)) yields 1. For programming convenience, (car nil) is allowed, and returns nil, even though nil isn't a cons and doesn't have a "car" field. .SS Functions cdr and rest .TP Syntax: (cdr ) (rest ) .TP Description: The functions cdr and rest are synonyms. They retrieve the "cdr" field of a cons cell. (cdr (cons 1 2)) yields 2. For programming convenience, (cdr nil) is allowed, and returns nil, even though nil isn't a cons and doesn't have a "cdr" field. .TP Example: Walk every element of the list (1 2 3): (for ((i '(1 2 3))) (i) ((set i (cdr i))) (print (car i) *stdout*) (print #\enewline *stdout*)) The variable i marches over the cons cells which make up the "backbone" of the list. The elements are retrieved using the car function. Advancing to the next cell is achieved using (cdr i). If i is the last cell in a (proper) list, (cdr i) yields nil. The guard expression i fails and the loop terminates. .SS Functions rplaca and rplacd .TP Syntax: (rplaca ) (rplacd ) .TP Description: The rplaca and rplacd functions assign new values into the "car" and "cdr" fields of the cell . Note that (rplaca x y) is the same as the more generic (set (car x) y), and likewise (rplacd x y) can be written as (set (cdr x) y). It is an error if is not a cons or lazy cons. In particular, whereas (car nil) is correct, (rplaca nil ...) is erroneous. .SS Functions second, third, fourth, fifth and sixth .TP Syntax: (first ) (second ) (third ) (fourth ) (fifth ) (sixth ) .TP Description: These functions access the elements of a proper list by position. If the list is shorter than implied, these functions return nil. .TP Examples: (third '(1 2)) -> nil (second '(1 2)) -> 2 (third '(1 2 . 3)) -> **error** .SS Functions append and append* .TP Syntax: (append [* ]) (append* [* ]) .TP Description: The append function creates a new list which is a catenation of the arguments. All arguments are optional, such that (append) produces the empty list. If a single argument is specified, then append simply returns the value of that argument. It may be any kind of object. If N arguments are specified, where N > 1, then the first N-1 arguments must be proper lists. Copies of these lists are catenated together. The last argument N, shown in the above syntax as , may be any kind of object. It is installed into the cdr field of the last cons cell of the resulting list. Thus, if argument N is also a list, it is catenated onto the resulting list, but without being copied. Argument N may be an atom other than nil; in that case append produces an improper list. The append* function works like append, but returns a lazy list which produces the catenation of the lists on demand. If some of the arguments are themselves lazy lists which are infinite, then append* can return immediately, whereas append will get caught in an infinite loop trying to produce a catenation and eventually exhaust available memory. (However, the last argument to append may be an infinite lazy list, because append does not traverse the last argument.) .TP Examples: ;; An atom is returned. (append 3) -> 3 ;; A list is also just returned: no copying takes place. ;; The eq function can verify that the same object emerges ;; from append that went in. (let ((list '(1 2 3))) (eq (append list) list)) -> t (append '(1 2 3) '(4 5 6) 7) -> '(1 2 3 4 5 6 . 7)) ;; the (4 5 6) tail of the resulting list is the original ;; (4 5 6) object, shared with that list. (append '(1 2 3) '(4 5 6)) -> '(1 2 3 4 5 6) (append nil) -> nil ;; (1 2 3) is copied: it is not the last argument (append '(1 2 3) nil) -> (1 2 3) ;; empty lists disappear (append nil '(1 2 3) nil '(4 5 6)) -> (1 2 3 4 5 6) (append nil nil nil) -> nil ;; atoms and improper lists other than in the last position ;; are erroneous (append '(a . b) 3 '(1 2 3)) -> **error** .SS Function list .TP Syntax: (list *) .TP Description: The list function creates a new list, whose elements are the argument values. .TP Examples: (list) -> nil (list 1) -> (1) (list 'a 'b) -> (a b) .SS Function sub-list .TP Syntax: (sub-list [ []]) .TP Description: The sub-list function extracts a sublist from . It is exactly like the more generic function sub, except that it operates only on lists. For a description of the arguments and semantics, refer to the sub function. .SS Function replace-list .TP Syntax: (replace-list [ []]) .TP Description: The replace-list function replaces a subrange of with items from the item-sequence argument, which may be any kind of sequence (list, vector or string). It is like the replace function, except that the first argument must be a list. For a description of the arguments and semantics, refer to the replace function. .SS Functions listp and proper-listp .TP Syntax: (listp ) (proper-listp ) .TP Description: The listp and proper-listp functions test, respectively, whether is a list, or a proper list, and return t or nil accordingly. The listp test is weaker, and executes without having to traverse the object. (listp x) is equivalent to (or (null x) (consp x)). The empty list is a list, and a cons cell is a list. The proper-listp function returns t only for proper lists. A proper list is either nil, or a cons whose cdr is a proper list. proper-listp traverses the list, and its execution will not terminate if the list is circular. .SS Function length-list .TP Syntax: (length-list ) .TP Description: The length-list function returns the length of , which may be a proper or improper list. The length of a list is the number of conses in that list. .SS Function copy-list .TP Syntax: (copy-list ) .TP Description: The copy-list function which returns a list similar to , but with a newly allocated cons cell structure. If is an atom, it is simply returned. Otherwise, is a cons cell, and copy-list returns (cons (car ) (copy-list (cdr ))) except that recursion is not used. .TP Dialect Note: Common Lisp does not allow the argument to be an atom, except for the empty list nil. .SS Function copy-cons .TP Syntax: (copy-cons ) .TP Description: This function creates a fresh cons cell, whose car and cdr fields are copied from . .SS Functions reverse, nreverse .TP Syntax: (reverse ) (nreverse ) .TP Description: The functions reverse and nreverse produce an object which contains the same items as proper list , but in reverse order. If is nil, then both functions return nil. The reverse function is non-destructive: it creates a new list. The nreverse function creates the structure of the reversed list out of the cons cells of the input list, thereby destructively altering it (if it contains more than one element). How nreverse uses the material from the original list is unspecified. It may rearrange the cons cells into a reverse order, or it may keep the structure intact, but transfer the "car" values among cons cells into reverse order. Other approaches are possible. .SS Function ldiff .TP Syntax: (ldiff ) .TP Description: The values and are proper lists. The ldiff function determines whether is a structural suffix of (meaning that it actually is a suffix, and is not merely equal to one). This is true if and are the same object, or else, recursively, if is a suffix of (cdr ). The object nil is the sublist of every list, including itself. The ldiff function returns a new list consisting of the elements of the prefix of which come before the suffix. The elements are in the same order as in . If is not a suffix of , then a copy of is returned. .TP Examples: ;;; unspecified: the compiler could make '(2 3) a suffix of '(1 2 3), ;;; or they could be separate objects. (ldiff '(1 2 3) '(2 3)) -> either (1) or (1 2 3) ;; b is the (1 2) suffix of a, so the ldiff is (1) (let ((a '(1 2 3)) (b (cdr a))) (ldiff a b)) -> (1) .SS Functions flatten, flatten* .TP Syntax: (flatten ) (flatten* ) .TP Description: The flatten function produces a list whose elements are all of the non-nil atoms contained in the structure of . The flatten* function works like flatten except that flatten creates and returns a complete flattened list, whereas flatten* produces a lazy list which is instantiated on demand. This is particularly useful when the input structure is itself lazy. .TP Examples: (flatten '(1 2 () (3 4))) -> (1 2 3 4) ;; precisely equivalent to previous example! nil is the same thing as () (flatten '(1 2 nil (3 4))) -> (1 2 3 4) (flatten nil) -> nil (flatten '(((()) ()))) -> nil .SS Functions memq, memql and memqual .TP Syntax: (memq ) (memql ) (memqual ) .TP Description: The memq, memql and memqual functions search the for a member which is, respectively, eq, eql or equal to . (See the eq, eql and equal functions below.) If no such element found, nil is returned. Otherwise, that tail of the list is returned whose first element is the matching object. .SS Functions remq, remql and remqual .TP Syntax: (remq ) (remql ) (remqual ) .TP Description The remq, remql and remqual functions produce a new list based on , removing the the items which are eq, eql or equal to . The input is unmodified, but the returned list may share substructure with it. If no items are removed, it is possible that the return value is itself. .SS Functions remq*, remql* and remqual* .TP Syntax: (remq* ) (remql* ) (remqual* ) .TP Description: The remq*, remql* and remqual* functions are lazy versions of remq, remql and remqual. Rather than computing the entire new list prior to returning, these functions return a lazy list. Caution: these functions can still get into infinite looping behavior. For instance, in (remql* 0 (repeat '(0))), remql will keep consuming the 0 values coming out of the infinite list, looking for the first item that does not have to be deleted, in order to instantiate the first lazy value. .TP Examples: ;; Return a list of all the natural numbers, excluding 13, ;; then take the first 100 of these. ;; If remql is used, it will loop until memory is exhausted, ;; because (range 1) is an infinite list. [(remql* 13 (range 1)) 0..100] .SH APPLICATIVE LIST PROCESSING .SS Functions remove-if, keep-if, remove-if* and keep-if* .TP Syntax: (remove-if : ) (keep-if : ) (remove-if* : ) (keep-if* : ) .TP Description The remove-if function produces a list whose contents are those of but with those elements removed which satisfy . Those elements which are not removed appear in the same order. The result list may share substructure with the input list, and may even be the same list object if no items are removed. The optional specifies how each element from the is transformed to an argument to . If this argument is omitted or specified as nil, then the predicate function is applied to the elements directly, a behavior which is identical to being (fun identity). The keep-if function is exactly like remove-if, except the sense of the predicate is inverted. The function keep-if retains those items which remove-if will delete, and removes those that remove-if will preserve. The remove-if* and keep-if* are like remove-if and keep-if, but produce lazy lists. .TP Examples: ;; remove any element numerically equal to 3. (remove-if (op = 3) '(1 2 3 4 3.0 5)) -> (1 2 4 5) ;; remove those pairs whose first element begins with "abc" [remove-if (op equal [@1 0..3] "abc") '(("abcd" 4) ("defg" 5)) car] -> (("defg 5)) ;; equivalent, without test function (remove-if (op equal [(car @1) 0..3] "abc") '(("abcd" 4) ("defg" 5))) -> (("defg 5)) .SS Function tree-find .TP Syntax: (tree-find ) .TP Description: The tree-find function searches for an occurence of . Tree can be any atom, or a cons. If it is a cons, it is understood to be a proper list whose elements are also trees. The equivalence test is performed by which must take two arguments, and has conventions similar to eq, eql or equal. tree-find works as follows. If is equivalent to under , then t is returned to announce a successful finding. If this test fails, and is an atom, nil is returned immediately to indicate that the find failed. Otherwise, is taken to be a proper list, and tree-find is recursively applied to each element of the list in turn, using the same and arguments, stopping at the first element which returns non-nil. .SS Functions find and find-if .TP Syntax: (find [ []]) (find-if []) .TP Description: The find and find-if functions search through a list for an item which matches a key, or satisfies a predicate function, respectively. The keyfun argument specifies a function which is applied to the elements of the list to produce the comparison key. If this argument is omitted, then the untransformed elements of the list themselves are searched. The find function's testfun argument specifies the test function which is used to compare the comparison keys from the list to the search key. If this argument is omitted, then the equal function is used. The first element from the list whose comparison key (as retrieved by the key function) matches the search (under the test function) is returned. If no such element is found, nil is returned. The find-if function's predfun argument specifies a predicate function which is applied to the successive comparison keys pulled from the list by applying the key function to successive elements. The first element for which the predicate function yields true is returned. If no such element is found, nil is returned. .SS Function set-diff .TP Syntax: (set_diff [ []]) .TP Description: The set-diff function treats the lists and as if they were sets and computes the set difference: a list which contains those elements in which do not occur in . Element equivalence is determined by a combination of testfun and keyfun. Elements are compared pairwise, and each element of a pair is passed through the keyfun function to produce a comparison value. The comparison values are compared with the testfun function. If keyfun is omitted, then the untransformed elements themselves are compared, and if testfun is omitted, then the equal function is used. If contains duplicate elements which do not occur in list2 (and thus are preserved in the set difference) then these duplicates appear in the resulting list. Furthermore, the order of the items from list1 is preserved. .SS Functions mapcar, mappend, mapcar* and mappend* .TP Syntax: (mapcar *) (mappend *) (mapcar* *) (mappend* *) .TP Description: When given three arguments, the mapcar function processes applies to the elements of and returns a list of the resulting values. Essentially, the list is filtered through the function. When additional lists are given as arguments, this filtering behavior is generalized in the following way: mapcar traverses the lists in parallel, taking a value from each list as an argument to the function. If there are two lists, the function is called with two arguments and so forth. The process is limited by the length of the shortest list. The return values of the function are collected into a new list which is returned. The mappend function works like mapcar, with the following difference. Rather than accumulating the values returned by the function into a list, mappend expects the items returned by the function to be lists which are catenated with append, and the resulting list is returned. That is to say, (mappend f a b c) is equivalent to (apply (fun append) (mapcar f a b c)). The mapcar* and mappend* functions work like mapcar and mappend, respectively. However, they return lazy lists rather than generating the entire output list prior to returning. .TP Caveats: Like mappend, mappend* must "consume" empty lists. For instance, if the function being mapped puts out a sequence of nil values, then the result must be the empty list nil, because (append nil nil nil nil ...) is nil. Suppose that mappend* is used on inputs which are infinite lazy lists, such that the function returns nil values indefinitely. For instance: ;; Danger: infinite loop!!! (mappend* (fun identity) (repeat '(nil))) The mappend* function is caught in a loop trying to consume and squash an infinite stream of nil values. .TP Examples: ;; multiply every element by two (mapcar (lambda (item) (* 2 item)) '(1 2 3)) -> (4 6 8) ;; "zipper" two lists together (mapcar (lambda (le ri) (list le ri)) '(1 2 3) '(a b c)) '((1 a) (2 b) (3 c))) ;; like append, mappend allows a lone atom or a trailing atom: (mappend (fun identity) 3) -> (3) (mappend (fun identity) '((1) 2)) -> (1 . 2) ;; take just the even numbers (mappend (lambda (item) (if (evenp x) (list x))) '(1 2 3 4 5)) -> (2 4) .SS Function apply .TP Syntax: (apply ) .TP Description: The apply function converts a list of values into individual arguments which are passed to . The return value of the apply invocation is that of . .TP Examples: ;; '(1 2 3) becomes arguments to list, thus (list 1 2 3). (apply (fun list) '(1 2 3)) -> (1 2 3) .TP Dialect note: TXR Lisp apply does not take additional arguments before the list. In Common Lisp we can write (apply #'list 1 2 (list 3 4 5)) which yields (1 2 3 4 5). In TXR Lisp, this usage can be simulated using (apply (fun list) (list 1 2 (list 3 4 5))) or (apply (fun list) '(,1 ,2 ,*(list 3 4 5))) . .SS Functions reduce-left and reduce-right .TP Syntax: (reduce-left ) (reduce-right ) .TP Description: The reduce-left and reduce-right functions reduce lists of operands specified in to a single value by the repeated application of . First, both functions initialize an internal accumulator with . Under reduce-left, the list is processed left to right. If elements remain to be processed, the is repeatedly called with two arguments: the accumulator and the next element from the list. After each call, the return value of the function replaces the accumulator. When no more items remain, the accumulator is returned. Under reduce-right, the list is processed right to left. If elements remain to be processed, the is repeatedly called with two arguments: the next element from the list and the accumulator. After each call, the return value of the function replaces the accumulator. When no more items remain, the accumulator is returned. The specifies how each element from the is converted to an argument to . The value nil is equivalent to (fun identity), which means that each list element is taken as the value itself. .TP Examples: ;;; list is empty, so 1 is just returned: (reduce-left (fun +) () 1 nil) -> 1 ;;; computes (- (- (- 0 1) 2) 3) (reduce-left (fun -) '(1 2 3) 0 nil) -> -6 ;;; computes (- 1 (- 2 (- 3 0))) (reduce-right (fun -) '(1 2 3) 0 nil) -> 2 ;;; computes (* 1 2 3) (reduce-left (fun *) '((1) (2) (3)) 1 (fun first)) -> 6 .SS Function some, all and none .TP Syntax: (some ) (all ) (none ) .TP Description The some, all and none functions apply a predicate test over a list of elements. The elements of are reduced to values using , which is a one-argument function. If is specified as nil, then (fun identity) is substituted, and thus the values of the list are taken as they are. These functions have short-circuiting semantics and return conventions similar to the and and or operators. The some function applies to successive values produced by retrieving elements of and processing them through . If the list is empty, it returns nil. Otherwise it returns the first non-nil return value returned by a call to and stops evaluating more elements. If returns nil for all elements, it returns nil. The all function applies to successive values produced by retrieving elements of and processing them through . If the list is empty, it returns t. Otherwise, if yields nil for any value, the all function immediately returns without invoking on any more elements. If all the elements are processed, then the all function returns the value which yielded for the last element. The none function applies to successive values produced by retrieving elements of and processing them through . If the list is empty, it returns t. Otherwise, if yields non-nil for any value, the none function immediately returns nil. If yields nil for all values, the none function returns t. .TP Examples: ;; some of the integers are odd (some (fun oddp) '(2 4 6 9) nil) -> t ;; none of the integers are even (none (fun evenp) '(1 3 4 7) nil) -> t .SH ASSOCIATION LISTS Association lists are ordinary lists formed according to a special convention. Firstly, any empty list is a valid association list. A non-empty association list contains only cons cells as the key elements. These cons cells are understood to represent key/value associations, hence the name "association list". .SS Function assoc .TP Syntax: (assoc ) .TP Description: The assoc function searches an association list for a cons cell whose car field is equivalent to (with equality determined by the equal function). The first such cons is returned. If no such cons is found, nil is returned. .SS Function assq .TP Syntax: (assql ) .TP Description: The assql function is just like assoc, except that the equality test is determined using the eql function rather than equal. .SS Function acons .TP Syntax: (acons ) .TP Description: The acons function constructs a new alist by consing a new cons to the front of . The following equivalence holds: (acons car cdr alist) <--> (cons (cons car cdr) alist) .SS Function acons-new .TP Syntax: (acons-new ) .TP Description: The acons-new function searches , as if using the assoc function, for an existing cell which matches the key provided by the car argument. If such a cell exists, then its cdr field is overwritten with the argument, and then the list is returned. If no such cell exists, then a new list is returned by adding a new cell to the input list consisting of the and values, as if by the acons function. .SS Function aconsql-new .TP Syntax: (aconsql-new ) .TP Description: This function is like acons-new, except that the eql function is used for equality testing. Thus, the list is searched for an existing cell as if using the assql function rather than assoc. .SS Function alist-remove .TP Syntax: (alist-remove ) .TP Description: The alist-remove function takes association list and produces a duplicate from which cells matching the specified keys have been removed. The argument is a list of the keys not to appear in the output list. .SS Function alist-nremove .TP Syntax: (alist-nremove ) .TP Description: The alist-nremove function is like alist-remove, but potentially destructive. The input list may be destroyed and its structural material re-used to form the output list. The application should not retain references to the input list. .SS Function copy-alist .TP Syntax: (copy-alist ) .TP Description: The copy-alist function duplicates . Unlike copy-list, which only duplicates list structure, copy-alist also duplicates each cons cell of the input alist. That is to say, each element of the output list is produced as if by the copy-cons function applied to the corresponding element of the input list. .SH PROPERTY LISTS .SS Function prop .TP Syntax: (prop ) .TP Description: A property list a flat list of even length consisting of interleaved pairs of property names (usually symbols) and their values (arbitrary objects). An example property list is (:a 1 :b "two") which contains two properties, :a having value 1, and :b having value "two". The prop function searches property list for key . If the key is found, then the value next to it is returned. Otherwise nil is returned. It is ambiguous whether nil is returned due to the property not being found, or due to the property being present with a nil value. .SH LIST SORTING .SS Function merge .TP Syntax: (merge ) .TP Description: The merge function merges two sorted lists and into a single sorted list. The semantics and defaulting behavior of the and arguments are the same as those of the sort function. The input lists are assumed to be sorted according to these functions. This function is destructive. The application should not retain references to the input lists, since the output list is formed out of the structure of the input lists. .SS Function multi-sort .TP Syntax: (multi-sort []) .TP Description: The multi-sort function regards a list of lists to be the columns of a database. The corresponding elements from each list constitute a record. These records are to be sorted, producing a new list of lists. The argument supplies the list of lists which comprise the columns of the database. The lists should ideally be of the same length. If the lists are of different lengths, then the shortest list is taken to be the length of the database. Excess elements in the longer lists are ignored, and do not appear in the sorted output. The argument supplies a list of comparison functions which are applied to the columns. Successive functions correspond to successive columns. If is an empty list, then the sorted database will emerge in the original order. If contains exactly one function, then the rows of the database is sorted according to the first column. The remaining columns simply follow their row. If contains more than one function, then additional columns are taken into consideration if the items in the previous columns compare equal. For instance if two elements from column one compare equal, then the corresponding second column elements are compared using the second column comparison function. The optional argument supplies transformation functions through which column entries are converted to comparison keys, similarly to the single key function used in the sort function and others. If there are more key functions than less functions, the excess key functions are ignored. .SH LAZY LISTS AND LAZY EVALUATION .SS Function make-lazy-cons .TP Syntax: (make-lazy-cons ) .TP Description: The function make-lazy-cons makes a special kind of cons cell called a lazy cons, or lcons. Lazy conses are useful for implementing lazy lists. Lazy lists are lists which are not allocated all at once. Rather, their elements materialize when they are accessed, like magic stepping stones appearing under one's feet out of thin air. A lazy cons has "car" and "cdr" fields like a regular cons, and those fields are initialized to nil when the lazy cons is created. A lazy cons also has an update function, the one which is provided as the argument to make-lazy-cons. When either the car and cdr fields of a cons are accessed for the first time, the function is automatically invoked first. That function has the opportunity to initialize the car and cdr fields. Once the function is called, it is removed from the lazy cons: the lazy cons no longer has an update function. To continue a lazy list, the function can make another call to make-lazy-cons and install the resulting cons as the cdr of the lazy cons. .TP Example: ;;; lazy list of integers between min and max (defun integer-range (min max) (let ((counter min)) ;; min is greater than max; just return empty list, ;; otherwise return a lazy list (if (> min max) nil (make-lazy-cons (lambda (lcons) ;; install next number into car (rplaca lcons counter) ;; now deal wit cdr field (cond ;; max reached, terminate list with nil! ((eql counter max) (rplacd lcons nil)) ;; max not reached: increment counter ;; and extend with another lazy cons (t (inc counter) (rplacd lcons (make-lazy-cons (lcons-fun lcons)))))))))) .SS Function lcons-fun .TP Syntax: (lcons-fun ) .TP Description: The lcons-fun function retrieves the update function of a lazy cons. Once a lazy cons has been accessed, it no longer has an update function and lcons-fun returns nil. While the update function of a lazy cons is executing, it is still accessible. This allows the update function to retrieve a reference to itself and propagate itself into another lazy cons (as in the example under make-lazy-cons). .SS Function generate .TP Syntax: (generate ) .TP Description: The generate function produces a lazy list which dynamically produces items according to the following logic. The arguments to generate are functions which do not take any arguments. The return value of generate is a lazy list. When the lazy list is accessed, for instance with the functions car and cdr, it produces items on demand. Prior to producing each item, is called. If it returns a true boolean value (any value other than nil), then the function is called, and its return value is incorporated as the next item of the lazy list. But if yields nil, then the lazy list immediately terminates. Prior to returning the lazy list, generate invokes the one time. If while-fun yields nil, then generate returns the empty list nil instead of a lazy list. Otherwise, it instantiates a lazy list, and invokes the gen-func to populate it with the first item. .SS Function repeat .TP Syntax: (repeat {}*) .TP Description: The repeat function produces an infinite lazy list formed by the repeatedly cycled catenation of the argument lists. .SS Operator gen .TP Syntax: (gen ) .TP Description: The gen operator produces a lazy list, in a manner similar to the generate function. Whereas the generate function takes functional arguments, the gen operator takes two expressions, which is often more convenient. The return value of gen is a lazy list. When the lazy list is accessed, for instance with the functions car and cdr, it produces items on demand. Prior to producing each item, the is evaluated, in its original lexical scope. If the expression yields a true value (non-nil), then is evaluated, and its return value is incorporated as the next item of the lazy list. If the expression yields nil, then the lazy list immediately terminates. The gen operator itself immediately evaluates before producing the lazy list. If the expression yields nil, then the operator returns the empty list nil. Otherwise, it instantiates the lazy list and invokes the to force the first item. .TP Example: @(do ;; Make a lazy list of integers up to 1000 ;; access and print the first three. (let* ((counter 0) (list (gen (< counter 1000) (inc counter)))) (format t "~s ~s ~s\en" (pop list) (pop list) (pop list)))) Output: 1 2 3 .SS Operator delay .TP Syntax: (delay ) .TP Description: The delay operator arranges for the delayed (or "lazy") evaluation of . This means that the expression is not evaluated immediately. Rather, the delay expression produces a promise object. The promise object can later be passed to the force function (described later in this document). The force function will trigger the evaluation of the expression and retrieve the value. The expression is evaluated in the original scope, no mater where the force takes place. The expression is evaluated at most once, by the first call to force. Additional calls to force only retrieve a cached value. .TP Example: @(do ;; list is popped only once: the value is computed ;; just once when force is called on a given promise ;; for the first time. (defun get-it (promise) (format t "*list* is ~s\en" *list*) (format t "item is ~s\en" (force promise)) (format t "item is ~s\en" (force promise)) (format t "*list* is ~s\en" *list*)) (defvar *list* '(1 2 3)) (get-it (delay (pop *list*)))) Output: *list* is (1 2 3) item is 1 item is 1 *list* is (2 3) .SS Function force .TP Syntax: (force ) .TP Description: The force function accepts a promise object produced by the delay function. The first time force is invoked , the promise expression is evaluated (in its original lexical environment, regardless of where in the program the force call takes place). The value of the expression is cached inside and returned, becoming the return value of the force function call. If the force function is invoked additional times on the same promise, the cached value is retrieved. .SH CHARACTERS AND STRINGS .SS Function mkstring .TP Syntax: (mkstring ) .TP Description: The mkstring function constructs a string object of a length specified by the parameter. Every position in the string is initialized with , which must be a character value. .SS Function copy-str .TP Syntax: (copy-str ) .TP Description: The copy-str function constructs a new string whose contents are identical to . .SS Function upcase-str .TP Syntax: (upcase-str ) .TP Description: The upcase-str function produces a copy of such that all lower-case characters of the English alphabet are mapped to their upper case counterparts. .SS Function downcase-str .TP Syntax: (downcase-str ) .TP Description: The downcase-str function produces a copy of such that all upper case characters of the English alphabet are mapped to their lower case counterparts. .SS Function string-extend .TP Syntax: (string-extend ) .TP Description: The string-extend function destructively increases the length of , which must be an ordinary dynamic string. It is an error to invoke this function on a literal string or a lazy string. The argument can be a character, string or integer. If it is a string or character, it specifies material which is to be added to the end of the string: either a single character or a sequence of characters. If it is an integer, it specifies the number of characters to be added to the string. If is an integer, the newly added characters have indeterminate contents. The string appears to be the original one because of an internal terminating null character remains in place, but the characters beyond the terminating zero are indeterminate. .SS Function stringp .TP Syntax: (stringp ) .TP Description: The stringp function returns t if is one of the several kinds of strings. Otherwise it returns nil. .SS Function lazy-stringp .TP Syntax: (lazy-stringp ) .TP Description: The lazy-stringp function returns t if is a lazy string. Otherwise it returns nil. .SS Function length-str .TP Syntax: (length-str ) .TP Description: The length-str function returns the length in characters. The argument must be a string. .SS Function search-str .TP Syntax: (search-str [ []]) .TP Description: The search-str function finds an occurrence of the string inside the string and returns its position. If no such occurrence exists, it returns nil. If a argument is specified, it gives the starting index for the search. If the argument is specified and is non-nil, it means that the search is conducted right-to-left. If multiple matches are possible, it will find the rightmost one rather than the leftmost one. .SS Function search-str-tree .TP Syntax: (search-str-tree [ []]) .TP Description: The search-str-tree function is similar to search-str, except that instead of searching for the occurence of a single needle string, it searches for the occurence of numerous strings at the same time. These search strings are specified, via the argument, as an arbitrarily structured tree whose leaves are strings. The function finds the earliest possible match, in the given search direction, from among all of the needle strings. If is a single string, the semantics is equivalent to search-str. .SS Function match-str .TP Syntax: (match-str []) .TP Description: The match-str function determines how many characters of match a prefix of . If the argument is specified, then the function tests how many characters of match a prefix of that portion of which starts at the given position. .SS Function match-str-tree .TP Syntax: (match-str-tree []) .TP Description: The match-str-tree function is a generalization of match-str which matches multiple test strings against at the same time. The value reported is the longest match from among any of the strings. The strings are specified as an arbitrarily shaped tree structure which has strings at the leaves. If is a single string atom, then the function behaves exactly like match-str. .SS Function sub-str .TP Syntax: (sub-str [ []]) .TP Description: The sub-str function extracts a substring from . It is exactly like the more generic function sub, except that it operates only on strings. For a description of the arguments and semantics, refer to the sub function. .SS Function replace-str .TP Syntax: (replace-str [ []]) .TP Description: The replace-str function replaces a substring of with items from , which may be any kind of sequence (list, vector or string) provided that it, if it is nonempty, it contains only characters. It is like the replace function, except that the first argument must be a list. For a description of the arguments and semantics, refer to the replace function. .SS Function cat-str .TP Syntax: (cat-str []) .TP Description: The cat-str function catenates a list of strings given by into a single string. The optional argument specifies a separator string which is interposed between the catenated strings. .SS Function split-str .TP Syntax: (split-str ) .TP Description: The split-str function breaks the into pieces, returing a list thereof. The argument must be a string. It specifies the separator character sequence within . All non-overlapping occurences of within are identified in left to right order, and are removed from . The string is broken into pieces according to the gaps left behind by the removed separators. Adjacent occurrences of within are considered to be separate gaps which come between empty strings. This operation is nondestructive: is not modified in any way. .SS Function split-str-set .TP Syntax: (split-str ) .TP Description: The split-str function breaks the into pieces, returing a list thereof. The argument must be a string. It specifies a set of characters. All occurences of any of these characters within are identified, and are removed from . The string is broken into pieces according to the gaps left behind by the removed separators. Adjacent occurrences of characters from within are considered to be separate gaps which come between empty strings. This operation is nondestructive: is not modified in any way. .SS Function list-str .TP Syntax: (list-str ) .TP Description: The list-str function converts a string into a list of characters. .SS Function trim-str .TP Syntax: (trim-str ) .TP Description: The trim-str function produces a copy of from which leading and trailing whitespace is removed. Whitespace consists of spaces, tabs, carriage returns, linefeeds, vertical tabs and form feeds. .SS Function string-lt .TP Syntax: (string-lt ) .TP Description: The string-lt function returns t if is lexicographically prior to . The behavior does not depend on any kind of locale. Note that this function forces (fully instantiates) any lazy string arguments, even if doing is is not necessary. .SS Function chrp .TP Syntax: (chrp ) .TP Description: Returns t if is a character, otherwise nil. .SS Function chr-isalnum .TP Syntax: (chr-isalnum ) .TP Description: Returns t if is an alpha-numeric character, otherwise nil. Alpha-numeric means one of the upper or lower case letters of the English alphabet found in ASCII, or an ASCII digit. This function is not affected by locale. .SS Function chr-isalpha .TP Syntax: (chr-isalpha ) .TP Description: Returns t if is an alphabetic character, otherwise nil. Alphabetic means one of the upper or lower case letters of the English alphabet found in ASCII. This function is not affected by locale. .SS Function chr-isascii .TP Syntax: (chr-isalpha ) .TP Description: This function returns t if the code of character is in the range 0 to 127, inclusive. For characters outside of this range, it returns nil. .SS Function chr-iscntrl .TP Syntax: (chr-iscntrl ) .TP Description: This function returns t if the character is a character whose code ranges from 0 to 31, or is 127. In other words, any non-printable ASCII character. For other characters, it returns nil. .SS Function chr-isdigit .TP Syntax: (chr-isdigit ) .TP Description: This function returns t if the character is is an ASCII digit. Otherwise, it returns nil. .SS Function chr-isgraph .TP Syntax: (chr-isgraph ) .TP Description: This function returns t if is a non-space printable ASCII character. It returns nil if it is a space or control character. It also returns nil for non-ASCII characters: Unicode characters with a code above 127. .SS Function chr-islower .TP Syntax: (chr-islower ) .TP Description: This function returns t if is an ASCII lower case letter. Otherwise it returns nil. .SS Function chr-isprint .TP Syntax: (chr-isprint ) .TP Description: This function returns t if is an ASCII character which is not a control character. It also returns nil for all non-ASCII characters: Unicode characters with a code above 127. .SS Function chr-ispunct .TP Syntax: (chr-ispunct ) .TP Description: This function returns t if is an ASCII character which is not a control character. It also returns nil for all non-ASCII characters: Unicode characters with a code above 127. .SS Function chr-isspace .TP Syntax: (chr-isspace ) .TP Description: This function returns t if is an ASCII whitespace character: any of the characters in the set #\espace, #\etab, #\elinefeed, #\enewline, #\ereturn, #\evtab, and #\epage. For all other characters, it returns nil. .SS Function chr-isupper .TP Syntax: (chr-isupper ) .TP Description: This function returns t if is an ASCII upper case letter. Otherwise it returns nil. .SS Function chr-isxdigit .TP Syntax: (chr-isxdigit ) .TP Description: This function returns t if is a hexadecimal digit. One of the ASCII letters A through F, or their lower-case equivalents, or an ASCII digit 0 through 9. .SS Function chr-toupper .TP Syntax: (chr-toupper ) .TP Description: If character is a lower case ASCII letter character, this function returns the upper case equivalent character. If it is some other character, then it just returns . .SS Function chr-tolower .TP Syntax: (chr-tolower ) .TP Description: If character is an upper case ASCII letter character, this function returns the lower case equivalent character. If it is some other character, then it just returns . .SS Functions num-chr and chr-num .TP Syntax: (num-chr ) (chr-num ) .TP Description: The argument must be a character. The num-chr function returns that character's Unicode code point value as an integer. The argument must be a fixnum integer in the range 0 to #\e10FFFF. The argument is taken to be a Unicode code point value and the corresponding character object is returned. .SS Function chr-str .TP Syntax: (chr-str ) .TP Description: The chr-str function performs random access on string to retrieve the character whose position is given by integer , which must be within range of the string. The index value 0 corresponds to the first (leftmost) character of the string and so non-negative values up to one less than the length are possible. Negative index values are also allowed, such that -1 corresponds to the last (rightmost) character of the string, and so negative values down to the additive inverse of the string length are possible. An empty string cannot be indexed. A string of length one supports index 0 and index -1. A string of length two is indexed left to right by the values 0 and 1, and from right to left by -1 and -2. .TP Notes: Direct use of chr-str is equivalent to the DWIM bracket notation except that must be a string. The following relation holds: (chr-str s i) --> [s i] since [s i] <--> (ref s i), this also holds: (chr-str s i) --> (ref s i) .SS Function chr-str-set .TP Syntax: (chr-str-set ) .TP Description: The chr-str function performs random access on string to overwrite the character whose position is given by integer , which must be within range of the string. The character at is overwritten with character . The argument works exactly as in chr-str. The argument must be a modifiable string. .TP Notes: Direct use of chr-str is equivalent to the DWIM bracket notation except that must be a string. The following relation holds: (chr-str-set s i c) --> (set [s i] c) since (set [s i] c) <--> (refset s i c), this also holds: (chr-str s i) --> (refset s i c) .SS Function span-str .TP Syntax: (span-str ) .TP Description: The span-str function determines the longest prefix of string which consists only of the characters in string , in any combination. .SS Function compl-span-str .TP Syntax: (compl-span-str ) .TP Description: The compl-span-str function determines the longest prefix of string which consists only of the characters which do not appear in , in any combination. .SS Function break-str .TP Syntax: (break-str ) .TP Description: The break-str function returns an integer which represents the position of the first character in string which appears in string . If there is no such character, then nil is returned. .SH VECTORS .SS Function vector .TP Syntax: (vector ) .TP Description: The vector function creates and returns a vector object of the specified length. The elements of the vector are initialized to nil. .SS Function vectorp .TP Syntax: (vectorp ) .TP Description: The vectorp function returns t if is a vector, otherwise it returns nil. .SS Function vec-set-length .TP Syntax: (vec-set-length ) .TP Description: The vec-set-length modifies the length of , making it longer or shorter. If the vector is made longer, then the newly added elements are initialized to nil. The argument must be nonnegative. The return value is . .SS Function vecref .TP Syntax: (vecref ) .TP Description: The vecref function performs indexing into a vector. It retrieves an element of at position , counted from zero. The value must range from 0 to one less than the length of the vector. The specified element is returned. .SS Function vec-push .TP Syntax: (vec-push ) .TP Description: The vec-push function extends the length of a vector by one element, and sets the new element to the value . The previous length of the vector (which is also the position of ) is returned. This function performs similarly to the generic function ref, except that the first argument must be a vector. .SS Function length-vec .TP Syntax: (length-vec ) .TP Description: The length-vec function returns the length of vector . It performs similarly to the generic length function, except that the argument must be a vector. .SS Function size-vec .TP Syntax: (size-vec ) .TP Description: The size-vec function returns the number of elements for which storage is reserved in the vector vec. .TP Notes: The length of the vector can be extended up to this size without any memory allocation operations having to be performed. .SS Function vector-list .TP Syntax: (vector-list ) .TP Description: This function returns a vector which contains all of the same elements and in the same order as list . .SS Function list-vector .TP Syntax: (list-vector ) .TP Description: The list-vector function returns a list of the elements of vector . .SS Function copy-vec .TP Syntax: (copy-vec ) .TP Description: The copy-vec function returns a new vector object of the same length as and containing the same elements in the same order. .SS Function sub-vec .TP Syntax: (sub-vec [ []]) .TP Description: The sub-vec function extracts a subvector from . It is exactly like the more generic function sub, except that it operates only on vectors. For a description of the arguments and semantics, refer to the sub function. .SS Function replace-vec .TP Syntax: (replace-vec [ []]) .TP Description: The replace-vec function replaces a subrange of with items from the item-sequence argument, which may be any kind of sequence (list, vector or string). It is like the replace function, except that the first argument must be a vector. For a description of the arguments and semantics, refer to the replace function. .SS Function cat-vec .TP Syntax: (cat-vec ) .TP Description: The argument is a list of vectors. The cat-vec function produces a catenation of the vectors listed in . It returns a single large vector formed by catenating those vectors together in order. .SH GENERIC SEQUENCE OPERATIONS .SS Function length .SS Function sub .TP Syntax: (sub [ []]) Description: The sub function extracts a slice from input sequence . The slice is a sequence of the same type as . If the parameter is omitted, the behavior is as if it were specified as nil. Likewise, if the parameter is omitted, the behavior is as if nil were specified. Thus (sub a) means (sub a nil nil). The following equivalence holds between the sub function and the DWIM-bracket syntax: (sub seq from to) <--> [seq from..to] The the description of the dwim operator---in particular, the section on Range Indexing---explains the semantics of the range specification. If the sequence is a list, the output sequence may share substructure with the input sequence. .SS Function replace .TP Syntax: (replace [ []]) .TP Description: The replace function replaces a subsequence of the with . The replaced subsequence may be empty, in which case an insertion is performed. If is empty (for example, the empty list nil), then a deletion is performed. If the and parameters are omitted, their values default to nil. The following equivalence holds between assignment to a place denoted by DWIM bracket syntax and the replace function: (set seq (replace seq new from to)) <--> (set [seq from..to] new) The the description of the dwim operator---in particular, the section on Range Indexing---explains the semantics of the range specification. This operation is destructive: it may work "in place" by modifying the original sequence. The caller should retain the return value and stop relying on the original input sequence. .SS Functions ref and refset .TP Syntax: (ref ) (refset ) .TP Description: The ref and refset functions perform array-like indexing into sequences. The ref function retrieves an element of , whereas refset overwrites an element of with a new value. The argument is based from zero, and negative values are permitted, with a special meaning as described in the Range Indexing section under the description of the dwim operator. The refset function returns the new value. The following equivalences hold between ref and refset, and the DWIM bracket syntax: (ref seq idx) <--> [seq idx] (refset seq idx new) <--> (set [seq idx] new) The difference is that ref and refset are first class functions which can be used in functional programming as higher order functions, whereas the bracket notation is syntactic sugar, and set is an operator, not a function. Therefore the brackets cannot replace all uses of ref and refset. .SS Function sort .TP Syntax: (sort []) .TP Description: The sort function destructively sorts , producing a sequence which is sorted according to the and arguments. The argument specifies a function which is applied to elements of the sequence to obtain the key values which are then compared using the lessfun. If is omitted, the identity function is used by default: the sequence elements themselves are their own sort keys. The argument specifies the comparison function which determines the sorting order. It must be a binary function which can be invoked on pairs of keys as produced by the key function. It must return a non-nil value if the left argument is considered to be lesser than the right argument. For instance, if the numeric function < is used on numeric keys, it produces an ascending sorted order. If the function > is used, then a descending sort is produced. The sort function is stable for sequences which are lists. This means that the original order of items which are considered identical is preserved. For strings and vectors, the sort is not stable. .SH MATH LIBRARY .SS Arithmetic functions +, - .TP Syntax: (+ *) (- *) (* *) .TP Description: The +, - and * functions perform addition, subtraction and multiplication, respectively. Additionally, the - function performs additive inverse. The + function requires zero or more arguments. When called with no arguments, it produces 0 (the identity element for adddition), otherwise it produces the sum over all of the arguments. Similarly, the * function requires zero or more arguments. When called with no arguments, it produces 1 (the identity element for multiplication). Otherwise it produces the product of all the arguments. The semantics of - changes from subtraction to additive inverse when there is only one argument. The argument is treated as a subtrahend, against an implicit minuend of zero. When there are two or more argument, the first one is the minuend, and the remaining are subtrahends. When there are three or more operands, these operations are performed as if by binary operations, in a left-associative way. That is to say, (+ a b c) means (+ (+ a b) c). The sum of a b is computed first, and then this is added to c. Similarly (- a b c) means (- (- a b) c). First, b is subtracted from a, and then c is subtracted from that result. The arithmetic inverse is performed as if it were subtraction from integer 0. That is, (- x) means the same thing as (- 0 x). The operands of +, - and * can be characters, integers (fixnum and bignum), and floats, in nearly any combination. If two operands have different types, then one of them is converted to the type of the one with the higher rank, according to this ranking: character < integer < float. For instance if one operand is integer, and the other float, the integer is converted to a float. .TP Restrictions: Characters are not considered numbers, and participate in these operations in limited ways. Subtraction can be used to computed the displacement between the Unicode values of characters, and an integer displacement can be added to a character, or subtracted from a character. For instance (- #\e9 #\e0) is 9. The Unicode value of a character C can be found using (- C #\ex0): the displacement from the NUL character. The rules can be stated as a set of restrictions: Two characters may not be added together. A character may not be subtracted from an integer (which also rules out the possibility of computing the additive inverse of a character). A character operand may not be opposite to a floating point operand in any operation. A character may not be an operand of multiplication. .SS Functions /, trunc, mod .TP Syntax: (/ ) (trunc ) (mod ) Description: The arguments to these functions are numbers. Characters are not permitted. The / function performs floating-point division. Each operands is first converted to floating-point type, if necessary. The trunc function performs a division of by whose result is truncated to integer toward zero. If both operands are integers, then an integer division is performed and the result is an integer. If either operand is a floating point value, a floating point division occurs, and the result is truncated toward zero to a floating-point integral value. The mod function performs a modulus operation. Firstly, the absolute value of is taken to be a modulus. Then a residue of with respect to is calculated. The residue's sign follows that of the sign of . That is, it is the smallest magnitude (closest to zero) residue of with respect to the absolute value of , having the same sign as . If the operands are integer, the result is an integer. If either operand is of type float, then the result is a float. The modulus operation is then generalized into the floating point domain. For instance the (mod 0.75 0.5) yields a residue of 0.25 because 0.5 "goes into" 0.75 only once, with a "remainder" of 0.25. .SS Function gcd .TP Syntax: (gcd ) .TP Description: The gcd function computes the greatest common divisor: the largest positive integer which divides both arguments. Operands and must be integers, or else an exception is thrown. The value of (gcd 0 x) is 0 for all x, including 0. The value of (gcd x 123) is is (abs x) for all x. Negative operands are permitted; this operation effectivelly ignores sign, so that the value of (gcd x y) is the same as (gcd (abs x) (abs y)) for all x and y. .SS Function abs .TP Syntax: (abs ) .TP Description: The abs function computes the absolute value of . If is positive, it is returned. If the number is negative, its additive inverse is returned: a positive number of the same type with exactly the same magnitude. .SS Functions floor, ceil .TP Syntax: (floor ) (ceil ) .TP Description: The floor function returns the highest integer which does not exceed the value of . The ceiling function returns the lowest integer which does not exceed the value of . If an integer, it is simply returned. If the argument is a float, then the value returned is a float. For instance (floor 1.1) returns 1.0 rather than 1. .SS Functions sin, cos, tan, asin, acos, atan .TP Syntax: (sin ) (cos ) (tan ) (atan ) (asin ) (acos ) .TP Description: These trigonometric functions convert their argument to floating point and return a float result. The sin, cos and tan functions compute the sine and cosine and tangent of the argument which represents an angle expressed in radians. The atan, acos and asin are their respective inverse functions. The argument to asin and acos must be in the range -1.0 to 1.0. .SS Functions log, exp .TP Syntax: (exp ) (log ) .TP Description: The exp function calculates the value of the transcendental number e raised to the specified exponent. The log function calculates the base e logarithm of its argument, which must be a positive value. Integer arguments are converted to floats. .SS Functions expt, sqrt, isqrt .TP Syntax: (expt *) (sqrt ) (isqrt ) .TP Description: The expt function raises to zero or more exponents given by the arguments. (expt x) is equivalent to (expt x 1); and yields x for all x. For three or more arguments, the operation is left associative. That is to say, (expt x y z) is equivalent to (expt (expt x y) z) and so forth. Exponentiation is done pairwise using a binary operation. If both operands to this binary operation are integers, then the result is an integer. If either operand is a float, then the other operand is converted to a float, and a floating point exponentation is performed. Exponentation that would produce a complex number is not supported. The sqrt function produces a floating-point square root. The numeric oeprand is converted from integer to floating-point if necessary. Negative operands are not supported. The isqrt function computes an integer square root: a value which is the greatest integer that is no greater than the true square root of the input value. The input value must be an integer. .SS Function exptmod .TP Syntax: (exptmod ) .TP Description: The exptmod function performs modular exponentiation and accepts only integer arguments. Furthermore, must be a non-negative and must be positive. The return value is raised to , and reduced to the least positive residue modulo . .SS Functions fixnump, bignump, integerp, floatp, numberp .TP Syntax: (fixnump ) (bignump ) (integerp ) (floatp ) (numberp ) .TP Description: These functions test the type of , returning true if it is an object of the implied type. The fixnump, bignump and floatp functions return true if the object is of the basic type fixnum, bignum or float. The function integerp returns true of is either a fixnum or a bignum. The function numberp returns true if the object is either a fixnum, bignum or float. .SS Function zerop .TP Syntax: (zerop ) .TP Description: The zerop function tests for equivalence to zero. The argument must be a number. It returns t for the integer value 0, and for the floating-point value 0.0. For other numbers, it returns nil. .SS Functions evenp, oddp .TP Syntax: (evenp ) (oddp ) .TP Description: The evenp and oddp functions require integer arguments. evenp returns t if is even (divisible by two), otherwise it returns nil. oddp returns t if is not divisible by two (odd), otherwise it returns nil. .SS Functions >, <, >=, <= and = .TP Syntax: (> *) (< *) (>= *) (<= *) (= *) .TP Description: These relational functions compare characters and numbers for numeric equality or inequality. The arguments must be one or more numbers or characters. If just one argument is given, then these functions all return t. If two arguments are given, then, they are compared as follows. First, if the numbers do not have the same type, then the one which has the lower ranking type is converted to the type of the other, according to this ranking: character < integer < float. For instance if a character and integer is compared, the character is converted to integer. Then a straightforward numeric comparison is applied. Three or more arguments may be given, in which case the comparison proceeds pairwise from left to right. For instance in (< a b c), the comparison (< a b) is performed in isolation. If it yields false, then nil is returned, otherwise the comparison (< b c) is performed in isolation, and if that yields false, nil is returned, otherwise t is returned. Note that it is possible for b to undergo two different conversions. For instance in (< ), the character will convert to a floating-point representation of its Unicode, and if that comparison suceeds, then in the second comparison, the character will convert to integer. .SS Function /= .TP Syntax: (/= *) .TP Description: The arguments to /= may be numbers or characters. The /= function returns t if no two of its arguments are numerically equal. That is to say, if there exist some a and b which are distinct arguments such that (= a b) is true, then teh function returns nil. Otherwise it returns t. .SS Functions max and min .TP Syntax: (max *) (min *) .TP Description: The max and min functions determine and return the highest or lowest value from among their arguments. The arguments must be numbers or characters. If only a single argument is given, that value is returned. If two arguments are given, then (max a b) is equivalent to (if (>= a b) a b), and (min a b) is equivalent to (if (<= a b) a b). If the operands do not have the same type, then one of them is converted to the type of the other; however, the original unconverted values are returned. For instance (max 4 3.0) yields the integer 4, not 4.0. If three or more arguments are given, max and min are left-associative. Thus (max a b c) is (max (max a b) c). .SS Functions int-str, flo-str and num-str .TP Syntax: (int-str ) (flo-str ) (num-str ) .TP Description: These functions extract numbers . Leading whitespace, if any, is skipped. If no digits can be successfully extracted, then nil is returned. Trailing material which does not contribute to the number is ignored. The int-str function converts a string of digits in the specified radix to an integer value. For radices above 10, letters of the alphabet are used for digits: A represent a digit whose value is 10, B represents 11 and so forth until Z. For values of radix above 36, the returned value is unspecified. Upper and lower case letters are recognized. Any character which is not a digit of the specified radix is regarded as the start of trailing junk at which the extraction of the digits stops. The flo-str function converts a floating-point decimal notation to a nearby floating point value. The material which contributes to the value is the longest match for optional leading space, followed by a mantissa which consists of an optional sign followed by a mixture of at least one digit, and at most one decimal point, optionally followed by an exponent part denoted by the letter E or e, an optional sign and one or more optional exponent digits. The num-str function converts a decimal notation to either an integer as if by a radix 10 application of int-str, or to a floating point value as if by flo-str. The floating point interpretation is chosen if the possibly empty initial sequence of digits (following any whitespace and optional sign) is followed by a period, e or E. .SS Functions int-flo and flo-int .TP Syntax: (int-flo ) (flo-int ) .TP Description: These functions perform numeric conversion between integer and floating point type. The int-flo function returns an integer by truncating toward zero. The flo-int function returns an exact floating point value corresponding to the integer argument, if possible, otherwise an approximation using a nearby floating point value. .SH EXCEPTIONS .SS Functions throw, throwf and error .SS Operator catch .TP Syntax: (catch {( (*) *)}*) .TP Description: The catch operator establishes an exception catching block around the . The is followed by zero or more catch clauses. Each catch clause consists of a symbol which denotes an exception type, an argument list, and zero or more body forms. If terminates normally, then the catch clauses are ignored. The catch itself terminates, and its return value is that of the . If throws an exception which is a subtype of one or more of the type symbols given in the exception clauses, then the first (leftmost) such clause becomes the exit point where the exception is handled. The exception is converted into arguments for the clause, and the clause body is executed. When the clause body terminates, the catch terminates, and the return value of the catch is that of the clause body. If throws an exception which is not a subtype of any of the symbols given in the clauses, then the search for an exit point for the exception continues through the enclosing forms. The catch clauses are not involved in the handling of that exception. When a clause catches an exception, the number of arguments in the catch must match the number of elements in the exception. A catch argument list resembles a function or lambda argument list, and may be dotted. For instance the clause (foo (a . b)) catches an exception subtyped from foo, with one or more elements. The first element binds to parameter a, and the rest, if any, bind to parameter b. If there is only one element, b takes on the value nil. Also see: the unwind-protect operator, and the functions throw, throwf and error. .SH REGULAR EXPRESSION LIBRARY .SS Functions search-regex and match-regex .TP Syntax: (search-regex : ) (match-regex : ) .TP Description These functions perform regular-expression-based searching and matching. The search-regex function searches through starting at position for a match for . If is omitted, the search starts at position 0. If is specified, the search proceeds in reverse, from the last position in the string, toward . This function returns nil if no match is found, otherwise it returns a cons pair, whose car indicates the position of the match, and whose cdr indicates the length of the match. The match-regex function tests whether matches at in . If is not specified, it is taken to be zero. If the regex matches, then the length of the match is returned. If it does not match, then nil is returned. . .SS Function regsub .TP Syntax: (regsub ) .TP Description: The regsub function searches for multiple occurences of non-overlapping matches for . A new string is constructed similar to but in which each matching region is replaced with using as follows. The object may be a character or a string, in which case it is simply taken to be the replacement for each match of the regular expression. The object may be a function of one argument, in which case for every match which is found, this function is invoked, with the matching piece of text as an argument. The function's return value is then taken to be the replacement text. .TP Examples: ;; match every lower case e or o, and replace by filtering ;; through the upcase-str function: [regsub #/[eo]/ upcase-str "Hello world!"] -> "HEllO wOrld!" ;; Replace Hello with Goodbye: (regsub #/Hello/ "Goodbye" "Hello world!") -> "Goodbye world!" .SS Function regexp .TP Syntax: (regexp ) .TP Description: The regexp function returns t if is a compiled regular expression object. For any other object type, it returns nil. .SS Function regex-compile .TP Syntax: (regex-compile ) .TP Description: The regex compile function takes the source code of a regular expression, expressed as a Lisp data structure, and compiles it to a regular expression object. .TP Examples: ;; the equivalent of #/[a-zA-Z0-9_/ (regex-compile '(set (#\ea . #\ez) (#\eA . #\eZ) (#\e0 . #\e9) #\e_)) ;; the equivalent of #/.*/ and #/.+/ (regex-compile '(0+ wild)) (regex-compile '(1+ wild)) ;; #/a|b|c/ (regex-compile '(or (or #\ea #\eb) #\ec)) .SH HASHING LIBRARY .SS Functions make-hash, hash .TP Syntax: (make-hash ) (hash [ :weak-keys ] [ :weak-vals ] [ :equal-based ]) .TP Description: These functions construct a new hash table, using different syntax. A hash table is an object which retains an association between pairs of objects. Each pair consists of a key and value. Given an object which is similar to a key in the hash table, it is possible to retrieve the corresponding value. Entries in a hash table are not ordered in any way, and lookup is facilitated by hashing: quickly mapping a key object to a numeric value which is then used to index into one of many buckets where the matching key will be found (if such a key is present in the hash table). make-hash takes three mandatory boolean arguments. The argument specifies, whether the hash table shall have weak keys. The argument specifies whether it shall have weak values, and specifies whether it is equal-based. The hash function defaults all three of these properties to false, and allows them to be overridden to true by the presence of keyword arguments. The hash function provides an alternative interface. It accepts optional arguments which are keyword symbols. Any combination of the three symbols :weak-keys, :weak-vals and :equal-based can be specified in any order to turn on the corresponding properties in the newly constructed hash table. If any of the keywords is not specified, the corresponding property defaults to nil. If a hash table has weak keys, this means that from the point of view of garbage collection, that table holds only weak references to the keys stored in it. Similarly, if a hash table has weak values, it means that it holds a weak reference to each value stored. A weak reference is one which does not prevent the reclamation of an object by the garbage collector. That is to say, when the garbage collector discovers that the only references to some object are weak references, then that object is considered garbage, just as if it had no references to it. The object is reclaimed, and the weak references "lapse" in some way, which depends on what kind they are. Hash table weak references lapse by entry removal: if either a key or a value object is reclaimed, then the corresponding key-value entry is erased from the hash table. Important to the operation of a hash table is the criterion by which keys are considered same. By default, this similarity follows the eql function. A hash table will search for a stored key which is eql to the given search key. A hash table constructed with the equal-based property compares keys using the equal function instead. In addition to storing key-value pairs, a hash table can have a piece of information associated with it, called the user data. .SS Functions make-similar-hash and copy-hash .TP Syntax: (make-similar-hash ) (copy-hash ) .TP Description: The make-similar-hash and copy-hash functions create a new hash object based on the existing object. The make-similar-hash produces an empty hash table which inherits all of the attributes of . It uses the same kind of key equality, the same configuration of weak keys and values, and has the same user data (see the set-hash-userdata function). The copy-hash function is like make-similar-hash, except that instead of producing an empty hash table, it produces one which has all the same elements as : it contains the same key and value objects. .SS Function sethash .TP Syntax: (sethash ) .TP Description: The sethash function places a value into table under the given . If a similar key already exists in the hash table, then that key's value is replaced by . Otherwise, the and pair is newly inserted into . .SS Function pushhash .TP Syntax: (pushhash ) .TP Description: The pushhash function is useful when the values stored in a hash table are lists. If the given does not already exist in , then a list of length one is made which contains , and stored in table under . If the already exists in the hash table, then the corresponding value must be a list. The value is added to the front of that list, and the extended list then becomes the new value under . .SS Function remhash .TP Syntax: (remhash ) .TP Description: The remhash function searches for a key similar to the . If that key is found, then that key and its corresponding value are removed from the hash table. .SS Function hash-count .TP Syntax: (hash-count ) .TP Description: The hash-count function returns an integer representing the number of key-value pairs stored in . .SS Function get-hash-userdata .TP Syntax: (get-hash-userdata ) .TP Description: This function retrieves the user data object associated with . The user data object of a newly created hash table is initialized to nil. .SS Function set-hash-userdata .TP Syntax: (set-hash-userdata ) .TP Description: The set-hash-userdata replaces, with the , the user data object associated with . .SS Function hashp .TP Syntax: (hashp ) .TP Description: The hashp function returns t if the is a hash table, otherwise it returns nil. .SS Function maphash .TP Syntax: (maphash ) .TP Description: The maphash function successively invokes binary-function for each key-value pair present in . The key and value are passed as arguments to . .SS Functions hash-eql and hash-equal .TP Syntax: (hash-eql ) (hash-equal ) .TP Description: These functions each compute an integer hash value from the internal representation of , which satisifes the following properties. If two objects A and B are the same under the eql function, then (hash-eql A) and (hash-eql B) produce the same integer hash value. Similarly, if two objects A and B are the same under the equal function, then (hash-equal A) and (hash-equal B) each produce the same integer hash value. In all other circumstances, the hash values of two distinct objects are unrelated, and may or may not be equal. .SS Functions hash_keys, hash_values, hash_pairs, and hash_alist .TP Syntax: (hash-keys ) (hash-values ) (hash-pairs ) (hash-alist ) .TP Description: These functions retrieve the bulk key-value data of hash table in various ways. hash-keys retrieves a list of the keys. hash-values retrieves a list of the values. hash-pairs retrieves a list of pairs, which are two-element lists consisting of the key, followed by the value. Finally, hash-pairs retrieves the key-value pairs as a Lisp association list: a list of cons cells whose car fields are keys, and whose cdr fields are the values. These functions all retrieve the keys and values in the same order. For example, if the keys are retrieved with hash-keys, and the values with hash-values, then the corresponding entries from each list correspond to the pairs in the hash table. .SS Operator dohash .TP Syntax: (dohash ( []) *) .TP Description: The dohash operator iterates over a hash table. The expression must evaluate to an object of hash table type. The and arguents must be symbols suitable for use as variable names. Bindings are established for these variables over the scope of the -s and the optional result form. For each element in the hash table, the and variables are set to the key and value of that entry, respectively, and each , if there are any, is evaluated. When all of the entries of the table are thus processed, the is evaluated, and its return value becomes the return value of the dohash form. If there is no , the return value is nil. The and -s are in the scope of an implicit anonymous block, which means that it is possible to terminate the execution of dohash early using (return) or (return ). .SS Functions hash-uni, hash-diff and hash-isec .TP Syntax: (hash-uni ) (hash-diff ) (hash-isec ) .TP Description: These functions perform basic set operations on hash tables in a nondestructive way, returning a new hash table without altering the inputs. The arguments and must be compatible hash tables. This means that their keys must use the same kind of equality. The resulting hash table inherits attributes from , as if by the make-similar-hash operation. If has userdata, the resulting hash table has the same userdata. If has weak keys, the resulting table has weak keys, and so forth. The hash-uni function performs a set union. The resulting hash contains all of the keys from and all of the keys from , and their corresponding values. If a key occurs both in and , then it occurs only once in the resulting hash. The value for this common key is the one from . The hash-diff function performs a set difference. First, a copy of is made as if by the copy-has function. Then from this copy, all keys which occur in are deleted. The hash-isec function performs a set intersection. The resulting hash contains only those keys which occur both in and . The values selected for these common keys are those from . .SH PARTIAL EVALUATION AND COMBINATORS .SS Operator op .TP Syntax: (op {}+) .TP Description: Like the lambda operator, the op operator creates an anonymous function. The difference is that the arguments of the function are implicit, or optionally specified within the function body. Also, the arguments of op are implicitly turned into a DWIM expression, which means that argument evaluation follows Lisp-1 rules. (See the dwim operator below). The argument forms are arbitrary expressions, within which a special convention is permitted: .IP @ A number preceded by a @ is a metanumber. This is a special syntax which denotes an argument. For instance @2 means that the second argument of the anonymous function is to be substituted in place of the @2. If @2 is used it means that @1 also has to appear somewhere, otherwise the op construct is erroneous. .IP @rest The meta-symbol @rest indicates that any trailing arguments to the function are to be inserted. If the @ syntax is not used anywhere, it means that the function only has trailing arguments. If @1 is used, it means that the second and subsequent arguments are trailing arguments. If @rest is not used anywhere, then the rest arguments are automatically applied to the op form. If @rest appears, then this is suppressed. The actions of form may be understood by these examples, which show how op is rewritten to lambda. However, note that the real translator uses generated symbols for the arguments, which are not equal to any symbols in the program. (op) -> invalid (op +) -> (lambda rest [+ . rest]) (op @1 @2) -> (lambda (arg1 arg2 . rest) [arg1 arg2 . rest]) (op foo @1 (@2) (bar @3)) -> (lambda (arg1 arg2 arg3 . rest) [foo arg1 (arg2) (bar arg3) . rest]) (op foo @rest @1) -> (lambda (arg1 . rest) [foo rest arg1]) .TP Examples: ;; Take a list of pairs and produce a list in which those pairs ;; are reversed. (mapcar (op list @2 @1) '((1 2) (a b))) -> ((2 1) (b a)) .SS Function chain .TP Syntax: (chain {}*) .TP Description: The chain function accepts zero or more functions as arguments, and returns a single function, called the chained function, which represents the chained application of those functions, in left to right order. If chain is given no arguments, then it returns a variadic function which ignores all of its arguments and returns nil. Otherwise, the first function may accept any number of arguments. The second and subsequent functions, if any, must accept one argument. The chained function can be called with an argument list which is acceptable to the first function. Those arguments are in fact passed to the first function. The return value of that call is then passed to the second function, and the return value of that call is passed to the third function and so on. The final return value is returned to the caller. .TP Example: (call [chain + (op * 2)] 3 4) -> 14 In this example, a two-element chain is formed from the + function and the function produced by (op * 2) which is a one-argument function that returns the value of its argument multiplied by two. (See the definition of the op operator). The chained function is invoked using the call function, with the arguments 3 and 4. The chained evaluation begins by passing 3 and 4 to the + function, which yields 7. This 7 is then passed to the (op * 2) doubling function, resulting in 14. A way to write the above example without the use of the DWIM brackets and the op operator is this: (call (chain (fun +) (lambda (x) (* 2 x))) 3 4) .SS Functions andf and orf .TP Syntax: (andf {}*) (orf {}*) .TP Description: The andf and orf functions are the functional equivalent of the and and or operators. These functions accept multiple functions and return a new function which represents the logical combination of those functions. The input functions should have the same arity. Or, rather, there should exist some common argument arity with which they can all be invoked. The resulting combined function is then callable with that many arguments. The andf operator returns a function which combines the input functions with short-circuiting logical conjunction. The resulting function passes its arguments to the functions successively, in left to right order. As soon as any of the functions returns nil, then nil is returned immediately, and the remaining functions are not called. Otherwise, if none of the functions return nil, then the value returned by the last function is returned. If the list of functions is empty, then t is returned. That is, (andf) returns a function which accepts any arguments, and returns t. The orf operator combines the input functions with a short-circuiting logical disjunction. The function produced by orf passes its arguments down to the functions successively, in left to right order. As soon as any function returns a non-nil value, that value is returned and the remaining functions are not called. If all functions return nil, then nil is returned. The expression (orf) returns a function which accepts any arguments and returns nil. .SS Function iff .TP Syntax: (iff []) .TP Description: The iff function is the functional equivalent of the if operator. It accepts functional arguments and returns a function. The resulting function takes its arguments and applies them to . If yields true, then the arguments are passed to and the resulting value is returned. Otherwise if yields a false result, and there is no then nil is returned. If yields false, and an exists, then the original arguments are passed to and the resulting value is returned. .SH INPUT AND OUTPUT TXR Lisp supports input and output streams of various kinds, with generic operations that work across the stream types. In general, I/O errors are usually turned into exceptions. When the description of error reporting is omitted from the description of a function, it can be assumed that it throws an error. .SS Variables *stdout*, *stddebug*, *stdin* and *stderr* These variables hold predefined stream objects. The *stdin*, *stdout* and *stderr* streams closely correspond to the underlying operating system streams. Various I/O functions require stream objects as arguments. The *stddebug* stream goes to the same destination as *stdout*, but is a separate object which can be redirected independently, allowing debugging output to be separated from normal output. .SS Function format .TP Syntax: (format {}*) .TP Description: The format function performs output to a stream given by , by interpreting the actions implicit in a , incorporating material pulled from additional arguments given by . Though the function is simple to invoke, there is complexity in format string language, which is documented below. The argument can be a stream object, or one of the values t and nil. The value t serves as a shorthand for *stdout*. The value nil means that the function will send output into a newly instantiated string output stream, and then return the resulting string. .TP Format string syntax: Within , most characters represent themselves. Those characters are simply output. The character ~ (tilde) introduces formatting directives, which are denoted by a letter. The special sequence ~~ (tilde-tilde) codes a single tilde. Nothing is permitted between the two tildes. The syntax of a directive is generally as follows: ~[[,]] The is a single alphabetic character which determines the general action of the directive. The optional width and precision can be numeric digits, or special codes documented below. The width specifier gives the width of the character field in which the object's printed representation is placed. If the printed representation overflows the field width, then the field width is ignored. The default field width is zero. If the width specifier begins with the < (left angle bracket) character, then the printing will be left-adjusted within this field. Otherwise it will be right-adjusted by default. The width can be specified as decimal digits, or as the character *. The * notation means that instead of digits, the value of the next argument is consumed, and expected to be an integer which specifies the width. If that integer value is negative, then the field will be left-adjusted. (If that value is positive, but the < notation is present prior to the * notation, the field will also be left-adjusted). The meaning of the precision specifier depends on the format directive. Also, when the precision is not specified, the default behavior may be different from any specific precision value. If the precision has a leading zero, this has special semantics for numeric conversions. The precision may also have a leading sign character, which may be a plus or space. .TP Format directives: Format directives are case sensitive, so that for example ~x and ~X have a different effect, and ~A doesn't exist wheras ~a does. They are: .IP a Prints any object in an aesthetic way, as if by the pprint function. The aesthetic notation violates read-print consistency: this notation is not necessarily readable if it is implanted in TXR source code. The field width specifier is honored, including the left-right adjustment semantics. The precision field has a meaning as follows. For integer objects, the precision specifies the minimum number of digits to print. (A leading sign does not count as a digit). If the precision field has a leading zero, then the number is padded with zeros to the required number of digits, otherwise the number is padded with spaces instead of zeros. If zero or space padding is present, and a leading positive or negative sign must be printed, then it is placed before leading zeros, or after leading spaces, as the case may be. If the precision specifier has a leading + sign, then a + sign is printed for positive numbers. If the precision specifier has a leading space instead of a + sign, then the + sign is rendered as a space for positive numbers. If there is no leading space or +, then a sign character is omitted for positive numbers. For floating point values, the precision specifies the maximum number of total significant figures, which do not include any digits in the exponent, if one is printed. Numbers are printed in exponential notation if their magnitude is small, or else if their exponent exceeds their precision. (If the precision is not specified, then it defaults to the system-dependent number of digits in a floating point value, derived from the C language DBL_DIG constant.) Floating point values which are integers are printed without a trailing .0 (point zero). For all other objects, the precision specifies the maximum number of characters to print. The object's printed representation is crudely truncated at that number of characters. .IP s Prints any object in a standard way, as if by the print function. Objects for which read-print consistency is possible are printed in a way such that if their notation is implanted in TXR source, they are readable. The field width specifier is honored, including the left-right adjustment semantics. The precision field is treated very similarly to the ~a format directive, except that non-exponentiated floating point numbers that would be mistaken for integers include a trailing ".0" for the sake read-print consistency. Objects truncated by precision may not have read-print consitency. For instance, if a string object is truncated, it loses its trailing closing quote, so that the resulting representation is no longer a properly formed string object. .IP x Requires an argument of character or integer type. The integer value or character code is printed in hexadecimal, using lower-case letters for the digits "a" through "f". Width and precision semantics are as described for the "a" format directive, for integers. .IP X Like the "x" directive, but the digits "a" through "f" are rendered in upper case. .IP o Like the "x" directive, but octal is used instead of hexadecimal. .IP f The "f" directive prints numbers in a fixed point decimal notation, with a fixed number of digits after the decimal point. It requires a numeric argument. (Unlike "x", "X" and "o", characters are not permitted). The precision specifier gives the number of digits past the decimal point. The number is rounded off to the specified precision, if necessary. Furthermore, that many digits are always printed, regardless of the actual precision of the number or its type. If it is omitted, then the default value is three: three digits past the decimal point. A precision of zero means no digits pas the decimal point, and in this case the decimal point is suppressed (regardless of whether the numeric argument is floating-point or integer). .IP e The "e" directive prints numbers in exponential notation. It requires a numeric argument. (Unlike "x", "X" and "o", characters are not permitted). The precision specifier gives the number of digits past the decimal point printed in the exponential notation, not counting the digits in the exponent. Exactly that many digits are printed, regardless of the precision of the number. If the precision is omitted, then the number of digits after the decimal point is three. If the precision is zero, then a decimal portion is truncated off entirely, including the decimal point. .SS Functions print, pprint, tostring, tostringp .TP Syntax: (print []) (pprint []) (tostring ) (tostringp ) .TP Description: The print and pprint functions render a printed character representation of the argument into . If a stream argument is not supplied, then the destination is the stream currently stored in the *stdout* variable. The print function renders in a way which strives for read-print consistency: an object is printed in a notation which is recognized as a similar object of the same kind when it appears in TXR source code. The pprint function ("pretty print") does not strive for read-print consistency. For instance it prints a string object simply by dumping its characters, rather than by adding the surrounding quotes and rendering escape syntax for special characters. The tostring and tostringp functions are like print and pprint, but they do not accept a stream argument, instead printing to a freshly instantiated string stream, and returning the resulting string. The following equivalences hold between calls to the format function and calls to these functions: (format stream "~s" obj) <--> (print obj stream) (format t "~s" obj) <--> (print obj) (format nil "~s" obj) <--> (tostring obj) For pprint and tostringp, the equivalence is produced by using "~a" in format rather than "~s". .SS Function make-string-input-stream .TP Syntax: (make-string-input-stream ) .TP Description: This function produces an input stream object. Character read operations on the stream object read successive characters from . Output operations and byte operations are not supported. .SS Function make-string-byte-input-stream .TP Syntax: (make-string-byte-input-stream ) .TP Description: This function produces an input stream object. Byte read operations on this stream object read successive byte values obtained by encoding into UTF-8. Character read operations are not supported, and neither are output operations. .SS Function make-string-output-stream .TP Syntax: (make-string-output-stream) .TP Description: This function, which takes no arguments, creates a string output stream. Data sent to this stream is accumulated into a string object. String output streams supports both character and byte output operations. Bytes are assumed to represent a UTF-8 encoding, and are decoded in order to form characters which are stored into the string. If an incomplete UTF-8 code is output, and a character output operation then takes place, that code is assumed to be terminated and is decoded as invalid bytes. The UTF-8 decoding machine is reset and ready for the start of a new code. The get-string-from-stream function is used to retrieve the accumulated string. If the null character is written to a string output stream, either via a character output operation or as a byte operation, the resulting string will appear to be prematurely terminated. TXR strings cannot contain null bytes. .SS Function get-string-from-stream .TP Syntax: (get-string-from-stream ) .TP Description: The argument must be a string output stream. This function finalizes the data sent to the stream and retrieves the accumulated character string. If a partial UTF-8 code has been written to , and then this function is called, the byte stream is considered complete and the partial code is decoded as invalid bytes. After this function is called, further output on the stream is not possible. .SS Function make-strlist-output-stream .TP Syntax: (make-strlist-output-stream) .TP Description: This function is closely analogous to make-string-output-stream. However, instead of producing a string, it produces a list of strings. The data is broken into multiple strings by newline characters written to the stream. Newline characters do not appear in the string list. Also, byte output operations are not supported. .SS Function get-list-from-stream .TP Syntax: (get-list-from-stream ) .TP Description: This function returns the string list which has accumulated inside a string output stream given by . The string output stream is finalized, so that further output is no longer possible. .SS Function close-stream .TP Syntax: (close-stream ) .TP Description: The close-stream function performs a close operation on , whose meaning is depends on the type of the stream. For some types of streams, such as string streams, it does nothing. For streams which are connected to operating system files or devices, will perform a close of the underlying file descriptor, and dissociate that descriptor from the stream. Any buffered data is flushed first. .SS Functions get-line, get-char and get-byte .TP Syntax: (get-line []) (get-char []) (get-byte []) .TP Description: These fundamental stream functions perform input. The argument is optional. If it is specified, it should be an input stream which supports the given operation. If it is not specified, then the *stdin* stream is used. The get-char function pulls a character from a stream which supports character input. Streams which support character input also support the get-line function which extracts a line of text delimited by the end of the stream or a newline character and returns it as a string. (The newline character does not appear in the string which is returned). Character input from streams based on bytes requires UTF-8 decoding, so that get-char actually may read several bytes from the underlying low level operating system stream. The get-byte function bypasses UTF-8 decoding and reads raw bytes from any stream which supports byte input. Bytes are represented as integer values in the range 0 to 255. Note that if a stream supports both byte input and character input, then mixing the two operations will interfere with the UTF-8 decoding. .SS Functions put-string, put-line, put-char and put-byte .TP Syntax: (put-line []) (put-string []) (put-char []) (put-byte []) .TP Description: These functions perform output on an output stream. The argument must be an output stream which supports the given operation. If it is omitted, then *stdout* is used. The put-char function writes a character given by to a stream. If the stream is based on bytes, then the character is encoded into UTF-8 and multiple bytes are written. Streams which support put-char also support put-line, and put-string. The put-string function writes the characters of a string out to the stream as if by multiple calls to put-char. The put-line function is like put-string, but also writes an additional newline character. The put-byte function writes a raw byte given by the argument to , if supports a byte write operation. The byte value is specified as an integer value in the range 0 to 255. .SS Function flush-stream .TP Syntax: (flush-stream ) .TP Description: This function is meaningful for output streams which accumulate data which is passed on to the operating system in larger transfer units. Calling this function causes all accumulated data inside to be passed to the operating system. If called on streams for which this function is not meaningful, it does nothing. .SH FILESYSTEM ACCESS .SS Function stat .SS Function open-directory .SS Functions open-file .SH COPROCESSES .SS Functions open-command, open-process .SH SYMBOLS AND PACKAGES .SS Variables *user-package*, *keyword-package*, *system-package* .SS Function make-sym .SS Function make-package .SS Function find-package .SS Function intern .SS Function symbolp .SS Function symbol-name .SS Function symbol-package .SS Function keywordp .SH PSEUDO-RANDOM NUMBERS .SS Variable *random-state* .SS Function make-random-state .SS Function random-state-p .SS Functions random-fixnum and random .SS Functions range and range* .SH TIME .SS Functions time and time-usec .SH WEB PROGRAMMING SUPPORT .SS Functions url-encode and url-decode .SH ACCESS TO TXR PATTERN LANGUAGE FROM LISP .SS Function match-fun .SH DEBUGGING FUNCTIONS .SS Functions source-loc and source-loc-str .SH MODULARIZATION .SS Variable *self-path* .SH DEBUGGER .B TXR has a simple, crude, built-in debugger. The debugger is invoked by adding the the -d command line option to an invocation of txr. In this debugger it is possible to step through code, set breakpoints, and examine the variable binding environment. Prior to executing any code, the debugger waits at the txr> prompt, allowing for the opportunity to set breakpoints. Help can be obtained with the h or ? command. Whenever the program stosp at the debugger, it prints the Lisp-ified piece of syntax tree that is about to be interpreted. It also shows the context of the input being matched. The s command can be used to step into a form; n to step over. Sometimes the behavior seems counter-intuitive. For instance stepping over a @(next) directive actually means skipping everything which follows it. This is because the query material after a @(next) is actually child nodes in the abstract syntax tree node of the next directive. .SS Sample Session Here is an example of the debugger beign applied to a web scraping program which connects to a US NAVY clock server to retrieve a dynamically-generated web page, from which the current time is extracted, in various time zones. The handling of the web request is done by the wget command; the txr query opens a wget command as and scans the body of the HTTP response containing HTML. This is the code, saved in a file called navytime.txr: @(next `!wget -c http://tycho.usno.navy.mil/cgi-bin/timer.pl -O - 2> /dev/null`) What time is it?

US Naval Observatory Master Clock Time

  @(collect :vars (MO DD HH MM SS (PM "  ") TZ TZNAME))
  
@MO. @DD, @HH:@MM:@SS @(maybe)@{PM /PM/} @(end)@TZ@/\t+/@TZNAME
  @  (until)
@/.*/ @(end)

US Naval Observatory @(output) @ (repeat) @MO-@DD @HH:@MM:@SS @PM @TZ @ (end) @(end) This is the debug session: $ txr -d navytime.txr stopped at line 1 of navytime.txr form: (next (sys:quasi "!wget -c http://tycho.usno.navy.mil/cgi-bin/timer.pl -O - 2> /dev/null")) depth: 1 data (nil): nil The user types s to step into the (next ...) form. txr> s stopped at line 2 of navytime.txr form: (sys:text " 1+ #\space) "HTML" (# 1+ #\space) "PUBLIC" (# 1+ #\space) "\"-//W3C//DTD" (# 1+ #\space) "HTML" (# 1+ #\space) "3.2" (# 1+ #\space) "Final\"//EN>") depth: 2 data (1): "" txr> s The current form now is a syt:text form which is an internal representation of a block of horizontal material. The pattern matching is in vertical mode at this point, and so the line of data is printed without an indication of character position. stopped at line 2 of navytime.txr form: (sys:text " 1+ #\space) "HTML" (# 1+ #\space) "PUBLIC" (# 1+ #\space) "\"-//W3C//DTD" (# 1+ #\space) "HTML" (# 1+ #\space) "3.2" (# 1+ #\space) "Final\"//EN>") depth: 3 data (1:0): "" . "" The user types s to step in. txr> s stopped at line 2 of navytime.txr form: "" Now, the form about to be processed is the first item of the (sys:text ...), a the string " s stopped at line 2 of navytime.txr form: (# 1+ #\espace) depth: 4 data (1:9): "" Control has now passed to the second element of the (sys:text ...), a regular expression which matches one or more spaces, generated by a single space in the source code according to the language rules. The input context shows taht " s stopped at line 2 of navytime.txr form: "HTML" depth: 4 data (1:10): "" Now, the regular expression has matched and moved the psoition past the space; the facing input is now "HTML ...". The programmer then repeats the s command by hitting Enter. txr> stopped at line 2 of navytime.txr form: (# 1+ #\espace) depth: 4 data (1:14): "" txr> stopped at line 2 of navytime.txr form: "PUBLIC" depth: 4 data (1:15): "" txr> stopped at line 2 of navytime.txr form: (# 1+ #\espace) depth: 4 data (1:21): "" txr> stopped at line 2 of navytime.txr form: "\e"-//W3C//DTD" depth: 4 data (1:22): "" txr> stopped at line 2 of navytime.txr form: (# 1+ #\espace) depth: 4 data (1:34): "" It is not evident from the session transcript, but during interactive use, the input context appears to be animated. Whenever the programmer hits Enter, the new context is printed and the dot appears to advance. Eventually the programmer becomes bored and place a breakpoint on line 15, where the @(output) block begins, and invokes the c command to continue the execution: txr> b 15 txr> c stopped at line 15 of navytime.txr form: (output (((repeat nil (((sys:var MO nil nil) "-" (sys:var DD nil nil) " " (sys:var HH nil nil) ":" (sys:var MM nil nil) ":" (sys:var SS nil nil) " " (sys:var PM nil nil) " " (sys:var TZ nil nil))) nil nil nil nil nil nil)))) depth: 2 data (16): "" The programmer issues a v command to take a look at the variable bindings, which indicate that the @(collect) has produced some lists. txr> v bindings: 0: ((PM " " "PM" "PM" "PM" "PM" "PM" "PM") (TZNAME "Universal Time" "Eastern Time" "Central Time" "Mountain Time" "Pacific Time" "Alaska Time" "Hawaii-Aleutian Time") (TZ "UTC" "EDT" "CDT" "MDT" "PDT" "AKDT" "HAST") (SS "35" "35" "35" "35" "35" "35" "35") (MM "32" "32" "32" "32" "32" "32" "32") (HH "23" "07" "06" "05" "04" "03" "01") (DD "30" "30" "30" "30" "30" "30" "30") (MO "Mar" "Mar" "Mar" "Mar" "Mar" "Mar" "Mar")) Then a continue command, which finishes the program, whose output appears: txr> c Mar-30 23:22:52 UTC Mar-30 07:22:52 PM EDT Mar-30 06:22:52 PM CDT Mar-30 05:22:52 PM MDT Mar-30 04:22:52 PM PDT Mar-30 03:22:52 PM AKDT Mar-30 01:22:52 PM HAST .SH APPENDIX A: NOTES ON EXOTIC REGULAR EXPRESSIONS Users familiar with regular expressions may not be familiar with the complement and intersection operators, which are often absent from text processing tools that support regular expressions. The following remarks are offered in hope that they are of some use. .SS Equivalence to Sets Regexp intersection is not essential; it may be obtained from complement and union as follows, since De Morgan's law applies to regular expression algebra: (R1)&(R2) = ~(~(R1)|~(R2)). (The complement of the union of the complements of R1 and R2 constitutes the intersection.) This law works because the regular expression operators denote set operations in a straightforward way. A regular expression denotes a set of strings (a potentially infinite one) in a condensed way. The union of two regular expressions R1 | R2 denotes the union of the set of texts denoted by R1 and that denoted by R2. Similarly R1 & R2 denotes a set intersection, and ~R denotes a set complement. Thus algebraic laws that apply to set operations apply to regular expressions. It's useful to keep in mind this relationship between regular expressions and sets in understanding intersection and complement. Given a finite set of strings, like the set { "abc", "def" }, which corresponds to the regular expression (abc|def), the complement is the set which contains an infinite number of strings: it consists of all possible strings except "abc" and "def". It includes the empty string, all strings of length 1, all strings of length 2, all strings of length 3 other than "abc" and "def", all strings of length 4, etc. This means that a "harmless looking" expression like ~(abc|def) can actually match arbitrarily long inputs. .SS Set Difference How about matching only three-character-long strings other than "abc" or "def"? To express this, regex intersection can be used: these strings are the intersection of the set of all three-character strings, and the set of all strings which are not "abc" or "def". The straightforward set-based reasoning leads us to this: ...&~(abc|def). This A&~B idiom is also called set difference, sometimes notated with a minus sign: A-B (which is not supported in .B TXR regular expression syntax). Elements which are in the set A, but not B, are those elements which are in the intersection of A with the complement of B. This is similar to the arithmetic rule A - B = A + -B: subtraction is equivalent to addition of the additive inverse. Set difference is a useful tool: it enables us to write a positive match which captures a more general set than what is intended (but one whose regular expression is far simpler than a positive match for the exact set we want), then we can intersect this over-generalized set with the complemented set of another regular expression which matches the particulars that we wish excluded. .SS Expressivity versus Power It turns out that regular expressions which do not make use of the complement or intersection operators are just as powerful as expressions that do. That is to say, with or without these operators, regular expressions can match the same sets of strings (all regular languages). This means that for a given regular expression which uses intersection and complement, it is possible to find a regular expression which doesn't use these operators, yet matches the same set of strings. But, though they exist, such equivalent regular expressions are often much more complicated, which makes them difficult to design. Such expressions do not necessarily . B express what it is they match; they merely capture the equivalent set. They perform a job, without making it obvious what it is they do. The use of complement and intersection leads to natural ways of expressing many kinds of matching sets, which not only demonstrate the power to carry out an operation, but also easily express the concept. .SS Example: Matching C Language Comments For instance, using complement, we can write a straightforward regular expression which matches C language comments. A C language comment is the digraph /*, followed by any string which does not contain the closing sequence */, followed by that closing sequence. Examples of valid comments are /**/, /* abc */ or /***/. But C comments do not nest (cannot contain comments), so that /* /* nested */ */ actually consists of the comment /* /* nested */, which is followed by the trailing junk */. Our simple characterization of interior part of a C comment as a string which does not contain the terminating digraph makes use of the complement, and can be expressed using the complemented regular expression like this: (~.*[*][/].*). That is to say, strings which contain */ are matched by the expression .*[*][/].*: zero or more arbitrary characters, followed by */, followed by zero or more arbitrary characters. Therefore, the complement of this expression matches all other strings: those which do not contain */. These strings up the inside of a C comment between the /* and */. The equivalent simple regex is quite a bit more complicated. Without complement, we must somehow write a positive match for all strings such that we avoid matching */. Obviously, sequences of characters other than * are included: [^*]*. Occurrences of * are also allowed, but only if followed by something other than a slash, so let's include this via union: ([^*]|[*][^/])*. Alas, already, we have a bug in this expression. The subexpression [*][^/] can match "**", since a * is not a /. If the next character in the input is /, we missed a comment close. To fix the problem we revise to this: ([^*]|[*][^*/])* (The interior of a C language comment is a any mixture of zero or more non-asterisks, or digraphs consisting of an asterisk followed by something other than a slash or another asterisk). Oops, now we have a problem again. What if two asterisks occur in a comment? They are not matched by [^*], and they are not matched by [*][^*/]. Actually, our regex must not simply match asterisk-non-asterisk digraphs, but rather sequences of one or more asterisks followed by a non-asterisk: ([^*]|[*]*[^*/])* This is still not right, because, for instance, it fails to match the interior of a comment which is terminated by asterisks, including the simple test cases where the comment interior is nothing but asterisks. We have no provision in our expression for this case; the expression requires all runs of asterisks to be followed by something which is not a slash or asterisk. The way to fix this is to add on a subexpression which optionally matches a run of zero or more interior asterisks before the comment close: ([^*]|[*]*[^*/])*[*]* Thus our the semi-final regular expression is [/][*]([^*]|[*]*[^*/])*[*]*[*][/] (A C comment is an interior string enclosed in /* */, where this interior part consists of a mixture of non-asterisk characters, as well as runs of asterisk characters which are terminated by a character other than a slash, except for possibly one rightmost run of asterisks which extends to the end of the interior, touching the comment close. Phew!) One final simplification is possible: the tail part [*]*[*][/] can be reduced to [*]+[/] such that the final run of asterisks is regarded as part of an extended comment terminator which consists of one or more asterisks followed by a slash. The regular expression works, but it's cryptic; to someone who has not developed it, it isn't obvious what it is intended to match. Working out complemented matching without complement support from the language is not impossible, but it may be difficult and error-prone, possibly requiring multiple iterations of trial-and-error development involving numerous test cases, resulting in an expression that doesn't have a straightforward relationship to the original idea. .SS The Non-Greedy Operator The non-greedy operator % is actually defined in terms of a set difference, which is in turn based on intersection and complement. The uninteresting case (R%) where the right operand is empty reduces to (R*): if there is no trailing context, the non-greedy operator matches R as far as possible, possibly to the end of the input, exactly like the greedy Kleene. The interesting case (R%T) is defined as a "syntactic sugar" for the equivalent expression ((R*)&(~.*(T&.+).*))T which means: match the longest string which is matched by R*, but which does not contain a non-empty match for T; then, match T. This is a useful and expressive notation. With it, we can write the regular expression for matching C language comments simply like this: [/][*].%[*][/] (match the opening sequence /*, then match a sequence of zero or more characters non-greedily, and then the closing sequence */. With the non-greedy operator, we don't have to think about the interior of the comment as set of strings which excludes */. Though the non-greedy operator appears expressive, its apparent simplicity may be deceptive. It looks as if it works "magically" by itself; "somehow" this .% "knows" only to consume enough characters so that it doesn't swallow an occurrence of the trailing context. Care must be taken that the trailing context passed to the operator really is the correct text that should be excluded by the non-greedy match. For instance, take the expression .%abc. If you intend the trailing context to be merely a, you must be careful to write (.%a)bc. Otherwise the trailing context is abc, and this means that the .% match will consume the longest string that does not contain "abc", when in fact what was intended was to consume the longest string that does not contain a. The change in behavior of the % operator upon modifying the trailing context is not as intuitive as that of the * operator, because the trailing context is deeply involved in its logic. For single-character trailing contexts, it may be a good idea to use a complemented character class instead. That is to say, rather than (.%a)bc, consider [^a]*bc. The set of strings which don't contain the character a is adequately expressed by [^a]*. .SH APPENDIX B: NOTES ON FALSE The reason for printing the word .IR false on standard output when a query doesn't match, in addition to returning a failed termination status, is that the output of .B TXR may be collected by a shell script, by the application of eval to command substitution syntax. Printing .IR false will cause eval to evaluate the .IR false command, and thus failed status will propagate from the eval itself. The eval command conceals the termination status of a program run via command substitution. That is to say, if a program fails, without producing output, its output is substituted into the eval command which then succeeds, masking the failure of the program. For example: eval "$(false)" appears successful: the false utility indicates a failed status, but produces no output. Eval evaluates an empty script and reports success; the failed status of the false program is forgotten. Note the difference between the above and this: eval "$(echo false)" This command has a failed status. The echo prints the word false and succeeds; this false word is then evaluated as a script, and thus interpreted as the false command which fails. This failure .B is propagated as the result of the eval command.