Record-delimiting stream adapter.

* regex.c (read_until_match): New function.
(regex_init): Registered read-until-match intrinsic.

* regex.h (read_until_match): Declared.

* stream.c (struct delegate_base): New struct type.
(delegate_base_mark, delegate_put_string, delegate_put_char,
delegate_put_byte, delegate_get_char, delegate_get_byte,
delegate_unget_char, delegate_unget_byte, delegate_close,
delegate_flush, delegate_seek, delegate_truncate,
delegate_get_prop, delegate_set_prop, delegate_get_error,
delegate_get_error_str, delegate_clear_error,
make_delegate_stream): New static functions.
(struct record_adapter_base): New struct type.
(record_adapter_base_mark, record_adapter_mark_op,
record_adapter_get_line): New static functions.
(record_adapter_ops): New static structure.
(record_adapter): New function.
(stream_init): Registered record-adapter intrinsic.

* stream.h (record_adapter): Declared.

* txr.1: Documented read-until-match and record-adapter.

1 files changed, 89 insertions, 0 deletions

diff --git a/txr.1 b/txr.1
index 6b870a89..7db1b98b 100644
--- a/txr.1
+++ b/txr.1
@@ -29825,6 +29825,49 @@ The double backslash in the string literal produces a single backslash
 in the resulting string object that is processed by
 .codn regex-parse .
 
+.coNP Function @ read-until-match
+.synb
+.mets (read-until-match < regex <> [ stream ])
+.syne
+.desc
+The
+.code read-until-match
+function reads characters from
+.metn stream ,
+accumulating them into a string, which is returned.
+
+If an argument is not specified for
+.metn stream ,
+then the
+.code *std-input*
+stream is used.
+
+The accumulation of characters is terminated by a match on
+.metn regex ,
+the end of the stream, or an error.
+
+This means that characters are read from the stream and accumulated while the
+stream has more characters available, and while its prefix does not match
+.metn regex .
+
+If
+.meta regex
+matches the stream before any characters are accumulated,
+then an empty string is returned.
+
+If the stream ends or an error occurs before any characters
+are accumulated, the function returns
+.codn nil .
+
+When the accumulation of characters terminates by a match on
+.metn regex ,
+reading characters from the stream continues. The longest
+possible prefix of the stream which matches
+.meta regex
+is read, and discarded. If an error is encountered, the
+matching procedure terminates and returns the previously
+accumulated string.
+
 .SS* Hashing Library
 .coNP Functions @, make-hash and @ hash
 .synb
@@ -33746,6 +33789,52 @@ possibility that the expression will be further extended by means of the dot or
 dotdot operators. An explicit end-of-input signal must be given from the
 terminal to terminate the expression.
 
+.coNP Function @ record-adapter
+.synb
+.mets (record-adapter < regex <> [ stream ])
+.syne
+.desc
+The
+.code record-adapter
+function returns a new stream object which acts as an
+.I adapter
+to the existing
+.metn stream .
+
+If an argument is not specified for
+.metn stream ,
+then the
+.code *std-input*
+stream is used.
+
+With the exception of
+.metn get-line ,
+all operations on the returned adapter transparently delegate to the original
+.meta stream
+object.
+
+When the
+.code get-line
+function is used on the adapter, it behaves differently. A string is
+extracted from
+.metn stream ,
+and returned. However, the string isn't a line delimited by a newline
+character, but a record delimited by
+.metn regex
+as if using the
+.code read-until-match
+function.
+
+All behavior which is built on the
+.code get-lines
+functions is affected by the record-delimiting semantics of a record adapter's
+.code get-line
+implementation. Notably, the
+.code get-lines
+and
+.code lazy-stream-cons
+functions return a lazy list of delimited records rather than of lines.
+
 .SS* Stream Output Indentation
 \*(TL streams provide support for establishing hanging indentations
 in text output. Each stream which supports output has a built-in state variable