| Commit message (Collapse) | Author | Age | Files | Lines |
|---|
| ... | |
| |
|
|
|
|
|
|
|
|
|
| |
* hash.c (hash_join): New function.
(hash_init): hash-join intrinsic registered.
* hash.h (hash_join): Declared.
* tests/010/hash.tl: New tests.
* txr.1: Documented.
|
| |
|
|
|
|
|
| |
* tests/010/hash.tl: Add test cases for the hash set operations.
* txr.1: Clarify that in hash-uni, the mapping functions are
used on all items, not just ones subject to joinfun.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
In March 2012, b7f1f4c5bbea86e288b6a4d68595c1d2d07217bd
introduced the feature that the @nil variable matches and
discards. This was incompletely implemented. Some cases
of a nil variable with modifiers fail to match.
* match.c (dest_bind): This function must correctly handle
the case when pattern is nil: it should just return bindings
without extending them. If the pattern is any nonbindable
symbol, it should indicate a failed match using t.
The logic has not been touched since 2009, at which time
an additional bogosity was introduced of calling
funcall(testfun, pattern, value) when pattern is a
non-bindable symbol. If value is a string, that could
never work. Possibly the idea is that the value could come
from a symbol-valued expression, such as one producing
a keyword symbol. We are not going to support that, unless
someone complains.
* tests/000/nilvar.txr, tests/000/nilvar.expected: New files,
providing a test case that fails without this commit.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
When TXR executes a top-level program, such that it will
exit when the last form in that program terminates,
it simulates a load. There is a block named load visible,
and the program can evaluate a (return-from load <expr>).
The value of that <expr> is thrown away, and the
termination status is always unsuccessful.
In this patch, (return-from load <expr>) is made to work
such that the value of <expr> will determine the exit
status, according to the same interpretation that
(exit <expr>) would give to the value.
* sysif.[ch] (exit_wrap): Static function becomes external.
* txr.c (txr_main): In the cases where we execute a file
and return from main, we now call exit_wrap instead.
The termination status is not simply based on whether
the file was successfully read, but takes into account the
load block.
* tests/019/load-ret/{script.tl,bad.tl}: New files.
* tests/019/load-ret/load-ret.tl: New tests.
* txr.1: Documented.
|
| |
|
|
|
|
|
|
|
|
| |
* lib.c (obj_print_impl): Do not print (rcons X Y)
as X..Y if X looks like (rcons ...). This
causes the problem that (rcons (rcons 1 2) 3)
prints as 1..2..3, a notation which unambiguously
means (rcons 1 (rcons 2 3)).
* tests/012/syntax.tl: New test cases.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Not all special characters can just be backslash escaped.
Spaces and newlines must be quoted.
* stream.c (sh_esc_common): New function. Handles both
sh-esc and sh-esc-all logic, distinguished by a flag.
Quoting is used, rather than backslash escaping.
If the string contains no special characters, it is just
erturned. If it can be double quoted, it is double quoted.
Otherwise it is single quoted and any contained single
quotes are replaced by '\''.
(sh_esc, sh_esc_all): Now just wrap sh_esc_common.
(sh_esc_dq): Remove the newline from the set of escaped
characters. Escaping a newline generates a continuation
sequence which eats the newline.
* tests/018/sh-esc.tl: Most test cases deleted; many new test
cases added.
* txr.1: Documentation revised.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
* glob.c (glob_wrap): When converting the glob array to the
returned list, suppress consecutive duplicates. This has to be
done separately for each call to glob or super_glob, so we now
interleave the production of the output list with the glob
calls. It has to be done separately because there can be
duplicates between different patterns. E.g. if (glob "?")
matches one path then (glob '("?" "?")) must return two copies
of it. Furthermore, the brace expansion implementation in
glob* produces multiple glob calls and appends their results.
Duplicates inside a single super_glob call result when there
are multiple ** (double star) patterns present, which are
matched by the same path in more than one way. If the results
are sorted, then the duplicates appear consecutively and we
will squash them. Also, a memory leak is fixed here: we
must free(pat_u8) unconditionally, before testing for the
early exit situation.
|
| |
|
|
|
|
|
| |
* tests/017/setjmp.tl: Solaris has libpng.so, but but some
version without png_set_longjmp_fn. We add a test for the
presence of this function as a precondition for running
the real test.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
There used to be a hack in the Makefile whereby the
compilation of stdlib/error.tl was forced to occur earlier.
I got rid of it. Now, the issue that was solving reproduced.
A situation can occur whereby loading error.tl triggers
loading some other files, which end up performing an expansion
that needs sys:bind-mac-check: but that function has not yet
been defined because error.tl has not yet loaded that far.
The issue occurs when stdlib/place.tl is compiled before
stdlib/error.tl. The compiled place.tl has a run-time
dependency on functions in error.tl, because the compiled
version of mac-param-bind and other forms relies on a run-time
support function sys:bind-mac-check defined in stdlib/error.tl.
* stdlib/error.tl (sys:dig): This function triggers the
problem, but it's not the only cause. Here, the problem is
because the (set ...) macro is used which triggers loading the
stdlib/place module. That brings in the need for
bind-mac-params. So here we use sys:setq instead. That is not
a complete solution. The changes in eval.c are also required,
because built-in macros like whilet expand to code that uses
the (set ...) macro. Note how sys:dig uses whilet.
(sys:bind-mac-check, sys:bind-mac-error): We move these
functions above compile-warning. This addresses remaining
circularity problem. The compile-warning function uses the
catch macro which brings in stdlib/except.tl, which pulls in
stdlib/op.tl due to its use of (do ...), which pulls in
stdlib/place.tl. So if we already define sys:bind-mac-check
at that point, we are good.
* eval.c: Sweep the file for almost all places where macros
generate code that invokes (set <symbol> <value>) and replace
that with (sys:setq <symbol> <value>) to eliminate the
dependency on loading the stdlib/place.tl module.
(me_def_variable, me_gun, me_while_until_star, me_case,
me_whilet, me_mlet, me_load_for, me_pop_after_load):
In all these macro expanders, use sys:setq rather than set
in the generated code.
* tests/019/load-hook.tl: Some test cases here look for a
macro expansion containing (set ...), needing to be fixed
to look for (sys:setq ...) due to the change in eval.c.
|
| |
|
|
|
|
|
|
|
|
| |
* lib.c (dwim_del): Remove check against structures from
OBJ case; we just let this pass through to the logic that
invokes replace.
* tests/012/aseq.tl: New test cases.
* txr.1: Document how del works on a [obj index] place.
|
| |
|
|
|
|
|
|
|
|
|
|
|
| |
* struct.c (invalidate_special_slots): New static function.
(invalidate_special_slot_nonexistence): Move static function
up in file, to be next to invalidate_special_slots.
(make_struct_type, static_slot_ens_rec): Call the new
invalidate_special_slots function in addition to calling
static_slot_home_fixup whenever the stslots array is resized.
The spslot array contains pointers to the elements of stslots,
which become invalid when that is resized.
* tests/012/oop-seq.tl: Repro test case added.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
| |
* eval.c (eval_init): Register mref intrinsic.
* lib.[ch] (mref): New function.
* stdlib/place.tl (sys:mref1): New place.
(mref): New place macro, defined in terms
of sys:merf1, ref place and mref function.
* tests/012/seq.tl: New tests.
* txr.1: Documented.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The ref function is not defined in the documentation
as an accessor, but there is a ref place. Unfortunately,
deletion is broken: (del (ref x y)) does not store the
new sequence back into place x, and so it does not work
correctly for lists; if x is a list, it doesn't change.
Various accessors are defined in terms of ref, as place
macros, such as the first, second, third, ... accessors.
This fixes the bug for them also; (del (second list))
must update list.
* stdlib/place.tl (ref): Fix the delete-expander to
fetch the clobber expander of the sequence place,
and use the simple setter to put the edited sequence into
that place.
* tests/012/seq.tl: Test case, which breaks without
this fix. Test the (second ...) place also, which is defined
in terms of ref.
* txr.1: Split documentation for ref and refset, mainly
because one is an Accessor and one is a Function. Removing
some discussions about the equivalences between DWIM brackets
and ref; there are subtleties there not worth going into.
Description of refset is simplified. We mention the possibility
of del over a ref place; only in that case is the sequence
itself required to be a place.
|
| |
|
|
|
|
|
|
|
|
|
| |
* autoload.c (op_set_entries): Add tap symbol as autoload
trigger for op module.
* stdlib/op.tl (tap): New macro.
* tests/012/op.tl: New test.
* txr.1: Documented.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
These are functions for testing whether a list or
sequence is shorter than a given integer. This is cheaper
than calculating the length of lists, which is in
some cases impossible if they are infinite.
A length-str-< function already exists, useful
with lazy strings.
length-< uses length-list-< or length-str-<
as appropriate
* lib.[ch] (length_list_lt, length_lt): New functions.
* eval.c (eval_init): length-list-< and length-<
intrinsics registered.
* tests/012/seq.tl: New tests.
* txr.1: Documented.
|
| |
|
|
| |
* tests/012/seq.tl: New tests.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
* lib.c (lazy_flatten_scan): Fix a problem which results
in cases like (()), ((())) ... to incorrectly flatten
to (nil). The do loop in this function which iteratively
descends into a nested left-nesting of a list does not handle
all cases, and therefore the function may not return at that
point. Removing the return fixes the problem, but so does
removing the loop so that in that case we just descend one
level into the nested list, and continue in the main loop.
What is incorrect is that when the consp(a) test fails and the
do loop terminates, we need to distinguish the cases off
a being an atom versus nil. Continuing in the loop does that.
This bug was spotted by a reviewer in the comp.lang.c
Usenet newsgroup.
(lazy_flatten): We neglect to handle the case here that
the input is an empty list, resulting in (flatten* nil)
returning (nil) rather than nil. The flatten function
is correct.
* tests/012/seq.tl: New tests.
* txr.1: Documentation improved. In particular, these
functions don't handle improper lists. Also, it needs
to be documented that the argument may be an atom.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Defining libpng bindings, with longjmp catching, is
now possible.
* autoload.c (ffi_set_entries): Add setjmp symbol, which is
a new macro in stdlib/ffi.tl.
* ffi.c (jmp_buf_s): New symbol variable.
(mk_jmp_buf, rt_setjmp, longjmp_wrap): New functions.
(ffi_init): Initialize jmp_buf_s. Register
sys:rt-setjmp and longjmp intrinsics.
* ffi.h (jmp_buf_s): Declared.
* stdlib/ffi.h (setjmp): New macro. Rather than introducing
a new special operator, we use a run-time support function
called sys:rt-setjmp, which takes functional arguments.
* unwind.[ch] (uw_snapshot, uw_restore): New functions.
The rt_setjmp function needs these to restore our unwind
frame stack into a sane state after catching a longjmp,
which bails without unwinding it, leaving the pointers
referring to frames that no longer exist.
* tests/017/setjmp.tl,
* tests/017/setjmp.expected: New files.
* txr.1: Documented.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
* eval.c (eval_init): Register hist-sort intrinsic.
* lib.c (gt_f): New global variable.
(hist_succ_f): New static variable.
(hist_succ): New static function.
(hist_sort): New function.
* lib.h (gt_f, hist_sort): Declared.
* tests/012/sort.tl: New tests.
* txr.1: Documented.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
* eval.c (eval_init): Register nestd-vec-of and nested-vec
intrinsics.
* lib.[ch] (vec_allocate, vec_own, vec_init): New static functions.
(vector, copy_vec): Expressed in terms of new functions.
(nested_vec_of_v, nested_vec_v): New functions.
* args.[ch] (args_cat_from): New function.
* tests/010/vec.tl: New tests.
* txr.1: Documented.
|
| |
|
|
|
|
|
|
|
|
|
|
|
| |
* glob.c (glob_wrap): #ifdef GLOB_BRACE around code that removes
the flag.
(super_glob_find_inner): Initialize pst. The older compiler I'm
using on Solaris 10. isn't smart enough to figure out that it
is not used uninitialized.
* tests/018/glob.t: Skip the ...\/** test on Solaris. It
takes a long time, and produces nil in the end. We don't
care how it behaves, only that we pass through that
pattern to glob without interpreting it as a double star.
|
| |
|
|
| |
* tests/018/glob.tl: exit successfully on Cygwin.
|
| |
|
|
|
|
|
|
| |
* glob.c (glob_path_cmp): Compare bytes as unsigned.
After the loop, don't test whether the pointer are null;
they never are. Test whether they point to null.
* tests/018/glob.tl: Expected data replaced.
|
| |
|
|
|
|
|
| |
* glob.c (super_glob_rec): Do not recognize a trailing /**
if it is preceded by a backslash.
* tests/018/glob.tl: Test case added.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The glob* function supports brace expansion, the **
pattern for matching zero or more path components,
as well as a sane sort for path names.
glob* relies on brace expansion written in Lisp;
the ** processing and sorting is done by a glob-compatible
C function called super_glob that uses glob.
* autoload.c (glob_set_entries, glob_instantiate): New static
functions.
(autoload_init): Register autoload of stdlib/glob module.
* glob.c (GLOB_XNOBRACE, GLOB_XSTAR): New macros.
(glob_wrap): Call super_glob instead of glob if GLOB_XSTAR
is present in flags. Avoid passing extension flags to glob.
(super_glob_find_inner, super_glob_rec, glob_path_cmp,
glob_str_cmp, super_glob): New static functions.
(glob_init): Register sys:glob-xstar, and glob-xnobrace.
sys:glob-xstar is used by glob* to request support for
the ** pattern from glob.
* stdlib/glob.tl: New file.
* tests/018/glob.tl: New file.
* txr.1: Documented.
|
| |
|
|
|
|
|
|
| |
* tests/018/crypt.tl: replace (crypt "a" "*$") test with
(crypt "a" "::"). Musl's crypt treats all unrecognized hashes
through DES, and the DES module accepts almost anything
as salt characters, except '\0', '\n' and ':', since
those characters would wreck the password file.
|
| |
|
|
|
|
|
|
|
|
|
|
| |
* lib.c (out_json_rec): Handle NUM and BGNUM cases same as FLNUM
so integers get printed. The restriction against integers has
been largely unhelpful and bothersome. Handle LCONS together with
CONS. Lists that are not special notation fall through to the VEC
case, which now uses seq_iter_t iteration to handle vectors and lists.
* tests/010/json.tl: New tests.
* txr.1: Documented support for printing integers and lists.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
It has been reported by user cielesti that some of our
crypt tests fail on the Musl library.
Musl has some additional agorithms so it yields
a meaningful hash for a "$0$" salt, as well as for "$9$".
Musl uses "*" and "x" as error tokens rather than "*0"
and "*1". We need to change how we detect error tokens.
* sysif.c (crypt_wrap): Detect error tokens only by their
length: if a string emerges from crypt or crypt_r, whose
length is less than 13, it's an error token.
* tests/018/crypt.tl: Drop the tests that require :error
for salts "$0$" and "$9$", replacing them with a test
for a salt that is almost certainly invalid in all C libraries
on Linux.
* txr.1: Document that crypt throws an error exception and
under what circumstances (when the C library function does
what).
|
| |
|
|
|
|
|
|
|
|
|
| |
* stream.c (sh_esc, sh_esc_all, sh_esc_dq, sh_esc_sq): New static
functions.
(stream_init): sh-esc, sh-esc-all, sh-esc-dq, sh-esc-sq: Intrinsics
registered.
* tests/018/sh-esc.tl: New file.
* txr.1: Documented.
|
| |
|
|
|
|
|
|
|
|
| |
* lib.[ch] (str_esc): New function.
* eval.c (eval_init): str-esc intrinsic registered.
* tests/015/esc.tl: New file.
* txr.1: Documented.
|
| |
|
|
|
|
|
|
|
| |
* stdlib/awk.tl (awk-state prn): Return nil in the no-argument
case instead of returning whatever put-string returns.
* tests/015/awk-misc.tl: New file.
* txr.1: Documented.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
| |
These remove repetitive (op ...) syntax from
the arguments of functional combinators.
* stdlib/opt.tl (opf, lopf): New macros.
* autoload.c (op_set_entries): Register opf and
lopf as autoload triggers.
* tests/012/op.tl: New tests.
* txr.1: Documented.
|
| |
|
|
|
|
|
|
|
|
|
|
|
| |
* autoload.c (csort_set_entries): Register csort-group
as autoload trigger for stdlib/csort.tl.
* stdlib/csort.tl (csort-group): New function.
* tests/012/sort.tl: Tests for sort-group and csort-group.
* txr.1: Documented.
* stdlib/doc-syms.tl: Updated.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
* arith.c (tofloat_s, toint_s): New symbol variables.
(tofloat, toint): If the argument is a COBJ, handle
it via do_unary_method.
(arith_init): Initialize new symbol variables.
The functions tofloat, toint, tofloatz and tointz.
are now registered here, rather than eval_init.
* eval.c (eval_init): Remove registrations of tofloat,
toint, tofloatz and tointz.
* tests/016/ud-arith.tl: New tests.
* txr.1: Documented.
* stdlib/doc-syms.tl: Updated.
|
| |
|
|
|
|
| |
* tests/016/ud-arith.tl (numbase): Add methods for
the newer functions: cbrt, erf, ...
Add tests covering these.
|
| |
|
|
|
|
|
|
|
|
|
| |
* tree.c (tr_delete_specific): We cant' juse use key(node)
as the search key; we must apply the tree's key function
to the node key field to retrieve the correct search key.
* tests/010/tree.tl: New test case which fails without
this bugfix: a node which is the left subtree of the root
node doesn't get deleted since the search is led astray
by the wrong key object.
|
| |
|
|
|
|
|
|
|
|
|
|
|
| |
* lib.c (unuse_sym): A used symbol may now appear in a package
under a different name. So if we don't find a symbol under
the symbol's name, or find a different symbol, we must try
a reverse hash search before giving up.
* txr.1: Add notes to use-sym-as that unuse-sym must be
used to undo its effect. Add notes to unuse-sym discussing
similarities and differences versus unintern.
* tests/012/use-as.tl: New test cases.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The new function use-sym-as can bring a foreign
symbol into a package under a different name,
which is not that symbol's name. This is also
featured in a new defpackage clause, :use-syms-as.
With this simple relaxation in the package system,
we don't require package local nicknames, which is
more complicated to implement and less ergonomic,
because it doesn't actually vanquish the use of
ugly package prefixes on clashing symbols.
* eval.c (eval_init): Register use-syms-as.
* lib.c (use_sym_as): New function, made out of
use_sym.
(use_sym): Now a wrapper for use_sym_as.
* lib.h (use_sym_as): Declared.
* stdlib/package.tl (defpackage): Implement :use-syms-as
clause.
* tests/012/use-as.tl: New file.
* txr.1: Documented,
* stdlib/doc-syms.tl: Updated.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
* stdlib/op.tl (opip-expand): Take arguments which specify
the op and do operators to be inserted. Pass these
through the recursive calls.
(opip, oand): Pass op and do for the new arguments.
(lopip, loand): New macros like opip and oand, but
passing lop and ldo to the expander.
(lflow): New macro.
* autoload.c (op_set_entries): Add autoload entries
for lopip, loand and lflow.
* tests/012/op.tl: A few new tests.
* txr.1: Documented.
* stdlib/doc-syms.tl: Regenerated.
|
| |
|
|
|
|
|
|
|
|
|
|
|
| |
* stream.c (close_stream): If the underlying method
returns the colon symbol :, then keep the cached
close_result as nil, so that the method can be called
again, but return t to the caller to indicate success.
* tests/018/close-delegate.tl: Test case added.
* tests/018/close-delegate.expected: Updated.
* txr.1: Documented.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This is motivated by trying to implement a struct delegate
stream which performs reference counting in close, in
order to close the real stream when the count hits zero.
The caching behavior of close-stream is a problem.
* stream.c (strm_base_init): Initialize close_result to nil,
rather than nao.
(strm_base_mark): Don't check close_result for nao.
(close_stream): Suppress the call to op->close if close_result
has a non-nil value, rather than a value other than nao.
* tests/018/close-delegate.tl,
* tests/018/close-delegate.expected: New files.
* txr.1: Document that only a non-nil return is cached
by close-stream.
|
| |
|
|
|
|
|
|
|
|
|
|
|
| |
* stdlib/op.tl (sys:opip-single-let-p,
sys:opip-let-p): New functions.
(sys:opip-expand): Restructure from collect loop
to car/cdr recursive form, because the new let operators
in opip need access to the rest of the pipeline.
Implement let operators.
* tests/012/op.tl: New tests.
* txr.1: Documented.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This is a regression due to a March 2016 commit which
introduced the ability for :vars in an output-side @(repeat)
block to have initial values.
The bug has the effect that all arguments in @(repeat)
which are conses/lists get duplicated, which messes up
the property list structure.
* parser.y (expand_repeat_rep_args): Do not unconditionally
add reg to the output at the bottom of the loop. A few
cases above in the consp(arg) case handle that themselves, and
do not continue the loop, so control ends up at the bottom,
adding a spurious item. By removing this list_collect,
we have to introduce it to just one case which relies on it.
* tests/008/repeat.txr,
* tests/008/repeat.expected: New files.
* y.tab.c.shipped: Updated.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
When a pattern variable match like @foo references a global
symbol macro, that's treated as an existing expression to
match, and not a new binding. However, local symbol macros
are not treated this way; they are invisible to variable
patterns. That is an unintended inconsistency.
* stdlib/match.tl (var-list exists): Use lexical-binding-kind
rather than lexical-var-p. This returns true for lexical
symbol macros also.
* tests/011/patmatch.tl: New test cases.
* txr.1: Documentation revised to clarify that both global
and local symbol macros are considered to be existing variable
bindings by pattern matching.
|
| |
|
|
|
|
| |
* tests/011/patmatch.tl: Move the form which compiles the
entire file to the end of the file, so that all the
interpreted test cases complete before we compile.
|
| |
|
|
|
|
|
|
|
|
|
|
|
| |
* stdlib/path-test.tl (path-volume): Don't return
:abs for a path whose empty first component
isn't followed by any more items. Otherwise
we return :abs for a path formed by splitting
the empty string, and then calls like (rel-path "" "a")
complain about a mixture of absolute and relative.
With this change, empty paths given to rel-path
behave as if they were ".".
* tests/018/rel-path.tl: New test cases.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This commit does two things. The replace function, implemented
under the hood by four specializations: replace-list, replace-vec,
replace-str and replace-buf, will handle the index-list case
a little differently. This is needed to fix the ability of the
del macro work on place designated by an index list, such as:
(del [sequence '(1 3 5 6)]
which now deletes elements 1, 3, 5 and 6 from the sequence,
and returns a sequence of those items. The underlying
implementation uses replace with an index-list, which is now
capable of deleting items. Previously, replace would stop
processing the index list when the replacement-sequence
corresponding to the index list ran out of items. Now,
when the replacement-sequence runs out of items, the
remaining index-list sequence elements specify items to
be deleted. For instance if str holds "abcdefg" then:
(set [str '(1 3 5)] "xy")
will change str to "axcyeg". Elements 1 and 3 are replaced
by x and y, respectively. Element 5, the letter f, is
deleted, because the replacement "xy" has no element
corresponding to 5.
* lib.c (replace_list, replace_str, replace_vec): Implement
new deleteion semantics for the case when the replacement
sequence runs out of items.
* buf.c (replace_buf): Likewise.
* tests/010/seq.txr: Some new test cases here for
deletion.
* tests/010/seq.expected: Updated.
* txr.1: Documented new semantics of replace, including
a new restriction that if elements are being deleted,
the indices should be monotonically increasing regardless
of the type of the sequence (not only list).
A value of 289 for the -C option documented, which restores
the previous behavior of replace (breaking deletion by
index-list, unfortunately: you don't always get to
simulate an old version of TXR while using new features.)
|
| |
|
|
|
|
|
|
| |
* eval.c (op_defsymacro, rt_defsymacro): We must call
vm_invalidate_binding so the VM forgets a cached binding
for this variable.
* tests/019/redef.tl: Test added.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
There was a bug in rt_defun in that it was not calling
vm_invalidate_binding. This mean that compiled functions
were not picking up redefinitions. This bug is fixed now
because rt_defun now calls sethash on the top_fb directly,
which modifies the existing binding cell; it is not
allocating a new cell.
We put in new test cases to confirm the proper redefinition
behaviors.
The proper redefinition behavior exposes an issue in
pattern matching.
* tests/019/redef.tl: New file.
* stdlib/match.tl (transform-quote): This function's compiled
image, when deposited into a .tlo file, becomes incorrect
because (sys:hash-lit) turns into #H() syntax, which reads
back as something else. In other words (sys:hash-lit)
deosn't have print-read consistency and so doesn't
externalize. To fix this right we would need a print mode
which ensures machine readability rather than human
readability, like in Common Lisp. For now, we just break up
the pattern so that it's not a literal match. This bug was
hidden due to theredefinition issue. When match.tl is
being compiled, it defines non-triv-pat-p twice. Due to
redefinitions not kicking in properly, the first definition
of non-triv-pat-p remains in effect for some functions.
When transform-qquote is being expanded, the (sys:hash-lit)
pattern is treated as non-trivial, even though it is
is trivial, and so it is turned into pattern matching code.
The code doesn't contain a (sys:hash-lit) literal and so
the issue doesn't occur.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The compiler handles trivial constant folding over the
source code, as a source to source transformation.
However, there are more opportunities for constant folding
after data flow optimizations of the VM code.
Early constant folding will not fold, for instance,
(let ((a 2) (b 3)) (* a b))
but we can reduce this to an end instruction that returns
the value of a D register that holds 6. Data flow optimizations
will propagate the D registers for 2 and 3 into the gcall
instruction. We can then recognize that we have a gcall with
nothing but D register operands, calling a constant-foldable
function. We can allocate a new D register to hold the result
of that calculation and just move that D register's value
into the target register of the original gcall.
* stdlib/compiler.tl (compiler get-dreg): When allocating
a new D reg, we must invalidate the datavec slot which is
calculated from the data hash. This didn't matter before,
because until now, get-datavec was called after compilation,
at which point no new D regs will exist. That is changing;
the optimizer can allocate D regs.
(compiler null-dregs, compiler null-stab): New methods.
(compiler optimize): Pass self to constructor for basic-blocks.
basic-blocks now references back to the compiler.
At optimization level 5 or higher, constant folding can
now happen, so we call the new method in the optimizer to
null the unused data. This overwrites unused D registers
and unused parts of the symbol vector with nil.
* stdlib/optimize (basic-blocks): Boa constructor now takes
a new leftmost param, the compiler.
(basic-blocks do-peephole-block): New optimization case:
gcall instruction invoking const-foldable function, with
all arguments being dregs.
(basic-blocks null-unused-data): New method.
|
