/[pcre]/code/trunk/ChangeLog
ViewVC logotype

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 221 - (show annotations) (download)
Fri Aug 17 09:25:08 2007 UTC (6 years, 11 months ago) by ph10
File size: 137340 byte(s)
Fix bad fix for repeated \p and \P.

1 ChangeLog for PCRE
2 ------------------
3
4 Version 7.3 17-Aug-07
5 ---------------------
6
7 1. In the rejigging of the build system that eventually resulted in 7.1, the
8 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
9 brackets there is not right, since it causes compilers to look for an
10 installed pcre.h, not the version that is in the source that is being
11 compiled (which of course may be different). I have changed it back to:
12
13 #include "pcre.h"
14
15 I have a vague recollection that the change was concerned with compiling in
16 different directories, but in the new build system, that is taken care of
17 by the VPATH setting the Makefile.
18
19 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
20 when the subject happened to end in the byte 0x85 (e.g. if the last
21 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
22 characters but of course it shouldn't be taken as a newline when it is part
23 of another character. The bug was that, for an unlimited repeat of . in
24 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
25 characters when looking for a newline.
26
27 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
28
29 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
30 in debug output.
31
32 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
33 long printing in the pcrecpp unittest when running under MinGW.
34
35 6. ESC_K was left out of the EBCDIC table.
36
37 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
38 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
39 limit also applies to "virtual nesting" when a pattern is recursive, and in
40 this case 1000 isn't so big. I have been able to remove this limit at the
41 expense of backing off one optimization in certain circumstances. Normally,
42 when pcre_exec() would call its internal match() function recursively and
43 immediately return the result unconditionally, it uses a "tail recursion"
44 feature to save stack. However, when a subpattern that can match an empty
45 string has an unlimited repetition quantifier, it no longer makes this
46 optimization. That gives it a stack frame in which to save the data for
47 checking that an empty string has been matched. Previously this was taken
48 from the 1000-entry workspace that had been reserved. So now there is no
49 explicit limit, but more stack is used.
50
51 8. Applied Daniel's patches to solve problems with the import/export magic
52 syntax that is required for Windows, and which was going wrong for the
53 pcreposix and pcrecpp parts of the library. These were overlooked when this
54 problem was solved for the main library.
55
56 9. There were some crude static tests to avoid integer overflow when computing
57 the size of patterns that contain repeated groups with explicit upper
58 limits. As the maximum quantifier is 65535, the maximum group length was
59 set at 30,000 so that the product of these two numbers did not overflow a
60 32-bit integer. However, it turns out that people want to use groups that
61 are longer than 30,000 bytes (though not repeat them that many times).
62 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
63 made it possible to implement the integer overflow checks in a much more
64 dynamic way, which I have now done. The artificial limitation on group
65 length has been removed - we now have only the limit on the total length of
66 the compiled pattern, which depends on the LINK_SIZE setting.
67
68 10. Fixed a bug in the documentation for get/copy named substring when
69 duplicate names are permitted. If none of the named substrings are set, the
70 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
71 empty string.
72
73 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
74 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
75 because the ] is interpreted as the first data character and the
76 terminating ] is not found. PCRE has been made compatible with Perl in this
77 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
78 cause memory overwriting.
79
80 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
81 string has been matched (to stop an infinite loop). It was not recognizing
82 a conditional subpattern that could match an empty string if that
83 subpattern was within another subpattern. For example, it looped when
84 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
85 condition was not nested. This bug has been fixed.
86
87 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
88 past the start of the subject in the presence of bytes with the top bit
89 set, for example "\x8aBCD".
90
91 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
92 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
93
94 14. Optimized (?!) to (*FAIL).
95
96 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
97 This restricts code points to be within the range 0 to 0x10FFFF, excluding
98 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
99 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
100 does: it's just the validity check that is more restrictive.
101
102 16. Inserted checks for integer overflows during escape sequence (backslash)
103 processing, and also fixed erroneous offset values for syntax errors during
104 backslash processing.
105
106 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
107 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
108
109 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
110 caused an overrun.
111
112 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
113 something other than just ASCII characters) inside a group that had an
114 unlimited repeat caused a loop at compile time (while checking to see
115 whether the group could match an empty string).
116
117 20. Debugging a pattern containing \p or \P could cause a crash. For example,
118 [\P{Any}] did so. (Error in the code for printing property names.)
119
120 21. An orphan \E inside a character class could cause a crash.
121
122 22. A repeated capturing bracket such as (A)? could cause a wild memory
123 reference during compilation.
124
125 23. There are several functions in pcre_compile() that scan along a compiled
126 expression for various reasons (e.g. to see if it's fixed length for look
127 behind). There were bugs in these functions when a repeated \p or \P was
128 present in the pattern. These operators have additional parameters compared
129 with \d, etc, and these were not being taken into account when moving along
130 the compiled data. Specifically:
131
132 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
133 length.
134
135 (b) An item such as \pL+ within a repeated group could cause crashes or
136 loops.
137
138 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
139 "reference to non-existent subpattern" error.
140
141 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
142
143 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
144 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
145
146
147 Version 7.2 19-Jun-07
148 ---------------------
149
150 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
151 which is apparently normally available under Windows.
152
153 2. Re-jig the pcregrep tests with different newline settings in an attempt
154 to make them independent of the local environment's newline setting.
155
156 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
157
158 4. Some of the "internals" tests were previously cut out when the link size
159 was not 2, because the output contained actual offsets. The recent new
160 "Z" feature of pcretest means that these can be cut out, making the tests
161 usable with all link sizes.
162
163 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
164 stack recursion. This gives a massive performance boost under BSD, but just
165 a small improvement under Linux. However, it saves one field in the frame
166 in all cases.
167
168 6. Added more features from the forthcoming Perl 5.10:
169
170 (a) (?-n) (where n is a string of digits) is a relative subroutine or
171 recursion call. It refers to the nth most recently opened parentheses.
172
173 (b) (?+n) is also a relative subroutine call; it refers to the nth next
174 to be opened parentheses.
175
176 (c) Conditions that refer to capturing parentheses can be specified
177 relatively, for example, (?(-2)... or (?(+3)...
178
179 (d) \K resets the start of the current match so that everything before
180 is not part of it.
181
182 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
183
184 (f) \g{name} is another synonym - part of Perl 5.10's unification of
185 reference syntax.
186
187 (g) (?| introduces a group in which the numbering of parentheses in each
188 alternative starts with the same number.
189
190 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
191
192 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
193 PCRE_INFO_JCHANGED.
194
195 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
196 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
197 for detecting groups that can match an empty string.
198
199 9. A pattern with a very large number of alternatives (more than several
200 hundred) was running out of internal workspace during the pre-compile
201 phase, where pcre_compile() figures out how much memory will be needed. A
202 bit of new cunning has reduced the workspace needed for groups with
203 alternatives. The 1000-alternative test pattern now uses 12 bytes of
204 workspace instead of running out of the 4096 that are available.
205
206 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
207
208 11. Applied patch from Google to remove an optimization that didn't quite work.
209 The report of the bug said:
210
211 pcrecpp::RE("a*").FullMatch("aaa") matches, while
212 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
213 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
214
215 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
216 it matched the wrong number of bytes.
217
218
219 Version 7.1 24-Apr-07
220 ---------------------
221
222 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
223 that is more "standard", making use of automake and other Autotools. There
224 is some re-arrangement of the files and adjustment of comments consequent
225 on this.
226
227 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
228 for recursive directory scanning broke on some systems because the files
229 are not scanned in any specific order and on different systems the order
230 was different. A call to "sort" has been inserted into RunGrepTest for the
231 approprate test as a short-term fix. In the longer term there may be an
232 alternative.
233
234 3. I had an email from Eric Raymond about problems translating some of PCRE's
235 man pages to HTML (despite the fact that I distribute HTML pages, some
236 people do their own conversions for various reasons). The problems
237 concerned the use of low-level troff macros .br and .in. I have therefore
238 removed all such uses from the man pages (some were redundant, some could
239 be replaced by .nf/.fi pairs). The 132html script that I use to generate
240 HTML has been updated to handle .nf/.fi and to complain if it encounters
241 .br or .in.
242
243 4. Updated comments in configure.ac that get placed in config.h.in and also
244 arranged for config.h to be included in the distribution, with the name
245 config.h.generic, for the benefit of those who have to compile without
246 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
247
248 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
249 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
250 makevp.bat for the current PCRE, using the additional files
251 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
252
253 6. A Windows user reported a minor discrepancy with test 2, which turned out
254 to be caused by a trailing space on an input line that had got lost in his
255 copy. The trailing space was an accident, so I've just removed it.
256
257 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
258 that is needed.
259
260 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
261 as "const" (a) because they are and (b) because it helps the PHP
262 maintainers who have recently made a script to detect big data structures
263 in the php code that should be moved to the .rodata section. I remembered
264 to update Builducptable as well, so it won't revert if ucptable.h is ever
265 re-created.
266
267 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
268 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
269 order to be able to cut out the UTF-8 tables in the latter when UTF-8
270 support is not required. This saves 1.5-2K of code, which is important in
271 some applications.
272
273 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
274 so as not to refer to the tables, even though these functions will never be
275 called when UTF-8 support is disabled. Otherwise there are problems with a
276 shared library.
277
278 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
279
280 (a) It was defining its arguments as char * instead of void *.
281
282 (b) It was assuming that all moves were upwards in memory; this was true
283 a long time ago when I wrote it, but is no longer the case.
284
285 The emulated memove() is provided for those environments that have neither
286 memmove() nor bcopy(). I didn't think anyone used it these days, but that
287 is clearly not the case, as these two bugs were recently reported.
288
289 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
290 and Detrail to create the HTML documentation, the .txt form of the man
291 pages, and it removes trailing spaces from listed files. It also creates
292 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
293 case, it wraps all the #defines with #ifndefs. This script should be run
294 before "make dist".
295
296 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
297 with Unicode property support.
298
299 (a) For a maximizing quantifier, if the two different cases of the
300 character were of different lengths in their UTF-8 codings (there are
301 some cases like this - I found 11), and the matching function had to
302 back up over a mixture of the two cases, it incorrectly assumed they
303 were both the same length.
304
305 (b) When PCRE was configured to use the heap rather than the stack for
306 recursion during matching, it was not correctly preserving the data for
307 the other case of a UTF-8 character when checking ahead for a match
308 while processing a minimizing repeat. If the check also involved
309 matching a wide character, but failed, corruption could cause an
310 erroneous result when trying to check for a repeat of the original
311 character.
312
313 13. Some tidying changes to the testing mechanism:
314
315 (a) The RunTest script now detects the internal link size and whether there
316 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
317 values substituted by "configure". (The RunGrepTest script already did
318 this for UTF-8.) The configure.ac script no longer substitutes the
319 relevant variables.
320
321 (b) The debugging options /B and /D in pcretest show the compiled bytecode
322 with length and offset values. This means that the output is different
323 for different internal link sizes. Test 2 is skipped for link sizes
324 other than 2 because of this, bypassing the problem. Unfortunately,
325 there was also a test in test 3 (the locale tests) that used /B and
326 failed for link sizes other than 2. Rather than cut the whole test out,
327 I have added a new /Z option to pcretest that replaces the length and
328 offset values with spaces. This is now used to make test 3 independent
329 of link size. (Test 2 will be tidied up later.)
330
331 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
332 segmentation fault instead of returning the appropriate error message.
333
334 15. In multiline mode when the newline sequence was set to "any", the pattern
335 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
336 This doesn't seem right; it now treats the CRLF combination as the line
337 ending, and so does not match in that case. It's only a pattern such as ^$
338 that would hit this one: something like ^ABC$ would have failed after \r
339 and then tried again after \r\n.
340
341 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
342 in an attempt to make files that differ only in their line terminators
343 compare equal. This works on Linux.
344
345 17. Under certain error circumstances pcregrep might try to free random memory
346 as it exited. This is now fixed, thanks to valgrind.
347
348 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
349 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
350 was because its rules for how to advance for /g after matching an empty
351 string at the end of a line did not allow for this case. They now check for
352 it specially.
353
354 20. pcretest is supposed to handle patterns and data of any length, by
355 extending its buffers when necessary. It was getting this wrong when the
356 buffer for a data line had to be extended.
357
358 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
359 CRLF as a newline sequence.
360
361 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
362 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
363 I have nevertheless tidied it up.
364
365 23. Added some casts to kill warnings from HP-UX ia64 compiler.
366
367 24. Added a man page for pcre-config.
368
369
370 Version 7.0 19-Dec-06
371 ---------------------
372
373 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
374 moving to gcc 4.1.1.
375
376 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
377 sys/time.h, which is documented as needed for this function. It doesn't
378 seem to matter on Linux, but it showed up on some releases of OS X.
379
380 3. It seems that there are systems where bytes whose values are greater than
381 127 match isprint() in the "C" locale. The "C" locale should be the
382 default when a C program starts up. In most systems, only ASCII printing
383 characters match isprint(). This difference caused the output from pcretest
384 to vary, making some of the tests fail. I have changed pcretest so that:
385
386 (a) When it is outputting text in the compiled version of a pattern, bytes
387 other than 32-126 are always shown as hex escapes.
388
389 (b) When it is outputting text that is a matched part of a subject string,
390 it does the same, unless a different locale has been set for the match
391 (using the /L modifier). In this case, it uses isprint() to decide.
392
393 4. Fixed a major bug that caused incorrect computation of the amount of memory
394 required for a compiled pattern when options that changed within the
395 pattern affected the logic of the preliminary scan that determines the
396 length. The relevant options are -x, and -i in UTF-8 mode. The result was
397 that the computed length was too small. The symptoms of this bug were
398 either the PCRE error "internal error: code overflow" from pcre_compile(),
399 or a glibc crash with a message such as "pcretest: free(): invalid next
400 size (fast)". Examples of patterns that provoked this bug (shown in
401 pcretest format) are:
402
403 /(?-x: )/x
404 /(?x)(?-x: \s*#\s*)/
405 /((?i)[\x{c0}])/8
406 /(?i:[\x{c0}])/8
407
408 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
409 is now done differently.
410
411 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
412 wrapper classes; (b) implement a new function in the C++ scanner that is
413 more efficient than the old way of doing things because it avoids levels of
414 recursion in the regex matching; (c) add a paragraph to the documentation
415 for the FullMatch() function.
416
417 6. The escape sequence \n was being treated as whatever was defined as
418 "newline". Not only was this contrary to the documentation, which states
419 that \n is character 10 (hex 0A), but it also went horribly wrong when
420 "newline" was defined as CRLF. This has been fixed.
421
422 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
423 was being set to -1 for the "end of line" case (supposedly a value that no
424 character can have). Though this value is never used (the check for end of
425 line is "zero bytes in current character"), it caused compiler complaints.
426 I've changed it to 0xffffffff.
427
428 8. In pcre_version.c, the version string was being built by a sequence of
429 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
430 string (as it is for production releases) called a macro with an empty
431 argument. The C standard says the result of this is undefined. The gcc
432 compiler treats it as an empty string (which was what was wanted) but it is
433 reported that Visual C gives an error. The source has been hacked around to
434 avoid this problem.
435
436 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
437 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
438 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
439 of them did).
440
441 10. Originally, pcretest opened its input and output without "b"; then I was
442 told that "b" was needed in some environments, so it was added for release
443 5.0 to both the input and output. (It makes no difference on Unix-like
444 systems.) Later I was told that it is wrong for the input on Windows. I've
445 now abstracted the modes into two macros, to make it easier to fiddle with
446 them, and removed "b" from the input mode under Windows.
447
448 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
449
450 12. Added -help and --help to pcretest as an official way of being reminded
451 of the options.
452
453 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
454 and pcrecpp.cc because they annoy compilers at high warning levels.
455
456 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
457
458 15. Fixed an occurrence of == in configure.ac that should have been = (shell
459 scripts are not C programs :-) and which was not noticed because it works
460 on Linux.
461
462 16. pcretest is supposed to handle any length of pattern and data line (as one
463 line or as a continued sequence of lines) by extending its input buffer if
464 necessary. This feature was broken for very long pattern lines, leading to
465 a string of junk being passed to pcre_compile() if the pattern was longer
466 than about 50K.
467
468 17. I have done a major re-factoring of the way pcre_compile() computes the
469 amount of memory needed for a compiled pattern. Previously, there was code
470 that made a preliminary scan of the pattern in order to do this. That was
471 OK when PCRE was new, but as the facilities have expanded, it has become
472 harder and harder to keep it in step with the real compile phase, and there
473 have been a number of bugs (see for example, 4 above). I have now found a
474 cunning way of running the real compile function in a "fake" mode that
475 enables it to compute how much memory it would need, while actually only
476 ever using a few hundred bytes of working memory and without too many
477 tests of the mode. This should make future maintenance and development
478 easier. A side effect of this work is that the limit of 200 on the nesting
479 depth of parentheses has been removed (though this was never a serious
480 limitation, I suspect). However, there is a downside: pcre_compile() now
481 runs more slowly than before (30% or more, depending on the pattern). I
482 hope this isn't a big issue. There is no effect on runtime performance.
483
484 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
485 newline (only possible for the last line of a file) and it was a
486 pattern that set a locale (followed by /Lsomething), pcretest crashed.
487
488 19. Added additional timing features to pcretest. (1) The -tm option now times
489 matching only, not compiling. (2) Both -t and -tm can be followed, as a
490 separate command line item, by a number that specifies the number of
491 repeats to use when timing. The default is 50000; this gives better
492 precision, but takes uncomfortably long for very large patterns.
493
494 20. Extended pcre_study() to be more clever in cases where a branch of a
495 subpattern has no definite first character. For example, (a*|b*)[cd] would
496 previously give no result from pcre_study(). Now it recognizes that the
497 first character must be a, b, c, or d.
498
499 21. There was an incorrect error "recursive call could loop indefinitely" if
500 a subpattern (or the entire pattern) that was being tested for matching an
501 empty string contained only one non-empty item after a nested subpattern.
502 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
503 incorrectly, because the \d was being skipped in the check.
504
505 22. The pcretest program now has a new pattern option /B and a command line
506 option -b, which is equivalent to adding /B to every pattern. This causes
507 it to show the compiled bytecode, without the additional information that
508 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
509 is the same as /B/I).
510
511 23. A new optimization is now able automatically to treat some sequences such
512 as a*b as a*+b. More specifically, if something simple (such as a character
513 or a simple class like \d) has an unlimited quantifier, and is followed by
514 something that cannot possibly match the quantified thing, the quantifier
515 is automatically "possessified".
516
517 24. A recursive reference to a subpattern whose number was greater than 39
518 went wrong under certain circumstances in UTF-8 mode. This bug could also
519 have affected the operation of pcre_study().
520
521 25. Realized that a little bit of performance could be had by replacing
522 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
523
524 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
525
526 27. Possessive quantifiers such as a++ were previously implemented by turning
527 them into atomic groups such as ($>a+). Now they have their own opcodes,
528 which improves performance. This includes the automatically created ones
529 from 23 above.
530
531 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
532 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
533 the first matched character to be a colon. This applied both to named and
534 numbered groups.
535
536 29. The ucpinternal.h header file was missing its idempotency #ifdef.
537
538 30. I was sent a "project" file called libpcre.a.dev which I understand makes
539 building PCRE on Windows easier, so I have included it in the distribution.
540
541 31. There is now a check in pcretest against a ridiculously large number being
542 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
543 loop, the loop is abandoned.
544
545 32. Forward references to subpatterns in conditions such as (?(2)...) where
546 subpattern 2 is defined later cause pcre_compile() to search forwards in
547 the pattern for the relevant set of parentheses. This search went wrong
548 when there were unescaped parentheses in a character class, parentheses
549 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
550
551 33. "Subroutine" calls and backreferences were previously restricted to
552 referencing subpatterns earlier in the regex. This restriction has now
553 been removed.
554
555 34. Added a number of extra features that are going to be in Perl 5.10. On the
556 whole, these are just syntactic alternatives for features that PCRE had
557 previously implemented using the Python syntax or my own invention. The
558 other formats are all retained for compatibility.
559
560 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
561 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
562 also .NET compatible.
563
564 (b) A recursion or subroutine call to a named group can now be defined as
565 (?&name) as well as (?P>name).
566
567 (c) A backreference to a named group can now be defined as \k<name> or
568 \k'name' as well as (?P=name). The new forms, as well as being in Perl
569 5.10, are also .NET compatible.
570
571 (d) A conditional reference to a named group can now use the syntax
572 (?(<name>) or (?('name') as well as (?(name).
573
574 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
575 groups (named and numbered) that are never evaluated inline, but can be
576 called as "subroutines" from elsewhere. In effect, the DEFINE condition
577 is always false. There may be only one alternative in such a group.
578
579 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
580 as the simple (?(R). The condition is true only if the most recent
581 recursion is that of the given number or name. It does not search out
582 through the entire recursion stack.
583
584 (g) The escape \gN or \g{N} has been added, where N is a positive or
585 negative number, specifying an absolute or relative reference.
586
587 35. Tidied to get rid of some further signed/unsigned compiler warnings and
588 some "unreachable code" warnings.
589
590 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
591 things, this adds five new scripts.
592
593 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
594 There were also incompatibilities regarding the handling of \Q..\E inside
595 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
596 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
597
598 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
599 matches an empty string, and forcibly breaks the loop. There were bugs in
600 this code in non-simple cases. For a pattern such as ^(a()*)* matched
601 against aaaa the result was just "a" rather than "aaaa", for example. Two
602 separate and independent bugs (that affected different cases) have been
603 fixed.
604
605 39. Refactored the code to abolish the use of different opcodes for small
606 capturing bracket numbers. This is a tidy that I avoided doing when I
607 removed the limit on the number of capturing brackets for 3.5 back in 2001.
608 The new approach is not only tidier, it makes it possible to reduce the
609 memory needed to fix the previous bug (38).
610
611 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
612 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
613 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
614 mode.
615
616 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
617 report.
618
619 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
620 copy construction and assignment in the C++ wrapper.
621
622 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
623 couple of bugs that could have given wrong results in the "--newline=crlf"
624 case.
625
626 44. Added a number of casts and did some reorganization of signed/unsigned int
627 variables following suggestions from Dair Grant. Also renamed the variable
628 "this" as "item" because it is a C++ keyword.
629
630 45. Arranged for dftables to add
631
632 #include "pcre_internal.h"
633
634 to pcre_chartables.c because without it, gcc 4.x may remove the array
635 definition from the final binary if PCRE is built into a static library and
636 dead code stripping is activated.
637
638 46. For an unanchored pattern, if a match attempt fails at the start of a
639 newline sequence, and the newline setting is CRLF or ANY, and the next two
640 characters are CRLF, advance by two characters instead of one.
641
642
643 Version 6.7 04-Jul-06
644 ---------------------
645
646 1. In order to handle tests when input lines are enormously long, pcretest has
647 been re-factored so that it automatically extends its buffers when
648 necessary. The code is crude, but this _is_ just a test program. The
649 default size has been increased from 32K to 50K.
650
651 2. The code in pcre_study() was using the value of the re argument before
652 testing it for NULL. (Of course, in any sensible call of the function, it
653 won't be NULL.)
654
655 3. The memmove() emulation function in pcre_internal.h, which is used on
656 systems that lack both memmove() and bcopy() - that is, hardly ever -
657 was missing a "static" storage class specifier.
658
659 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
660 containing an extended class (one that cannot be represented by a bitmap
661 because it contains high-valued characters or Unicode property items, e.g.
662 [\pZ]). Almost always one would set UTF-8 mode when processing such a
663 pattern, but PCRE should not loop if you do not (it no longer does).
664 [Detail: two cases were found: (a) a repeated subpattern containing an
665 extended class; (b) a recursive reference to a subpattern that followed a
666 previous extended class. It wasn't skipping over the extended class
667 correctly when UTF-8 mode was not set.]
668
669 5. A negated single-character class was not being recognized as fixed-length
670 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
671 compile error "lookbehind assertion is not fixed length".
672
673 6. The RunPerlTest auxiliary script was showing an unexpected difference
674 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
675 write a Perl script that can interpret lines of an input file either as
676 byte characters or as UTF-8, which is what "perltest" was being required to
677 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
678 can't do is switch easily at run time between having the "use utf8;" pragma
679 or not. In the end, I fudged it by using the RunPerlTest script to insert
680 "use utf8;" explicitly for the UTF-8 tests.
681
682 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
683 the end of the subject string, contrary to the documentation and to what
684 Perl does. This was true of both matching functions. Now it matches only at
685 the start of the subject and immediately after *internal* newlines.
686
687 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
688 a pointer to an int instead of a pointer to an unsigned long int. This
689 caused problems on 64-bit systems.
690
691 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
692 instance of the 'standard' template library not being so standard".
693
694 10. There was no check on the number of named subpatterns nor the maximum
695 length of a subpattern name. The product of these values is used to compute
696 the size of the memory block for a compiled pattern. By supplying a very
697 long subpattern name and a large number of named subpatterns, the size
698 computation could be caused to overflow. This is now prevented by limiting
699 the length of names to 32 characters, and the number of named subpatterns
700 to 10,000.
701
702 11. Subpatterns that are repeated with specific counts have to be replicated in
703 the compiled pattern. The size of memory for this was computed from the
704 length of the subpattern and the repeat count. The latter is limited to
705 65535, but there was no limit on the former, meaning that integer overflow
706 could in principle occur. The compiled length of a repeated subpattern is
707 now limited to 30,000 bytes in order to prevent this.
708
709 12. Added the optional facility to have named substrings with the same name.
710
711 13. Added the ability to use a named substring as a condition, using the
712 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
713 are numbers (not recommended). Forward references are permitted.
714
715 14. Added forward references in named backreferences (if you see what I mean).
716
717 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
718 pattern could run off the end of the subject. For example, the pattern
719 "(?s)(.{1,5})"8 did this with the subject "ab".
720
721 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
722 PCRE_CASELESS was set when matching characters that were quantified with ?
723 or *.
724
725 17. A character class other than a single negated character that had a minimum
726 but no maximum quantifier - for example [ab]{6,} - was not handled
727 correctly by pce_dfa_exec(). It would match only one character.
728
729 18. A valid (though odd) pattern that looked like a POSIX character
730 class but used an invalid character after [ (for example [[,abc,]]) caused
731 pcre_compile() to give the error "Failed: internal error: code overflow" or
732 in some cases to crash with a glibc free() error. This could even happen if
733 the pattern terminated after [[ but there just happened to be a sequence of
734 letters, a binary zero, and a closing ] in the memory that followed.
735
736 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
737 over the years. Originally (before any Unicode support), just the bottom 8
738 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
739 output from "man perlunicode" includes this:
740
741 The regular expression compiler produces polymorphic opcodes. That
742 is, the pattern adapts to the data and automatically switches to
743 the Unicode character scheme when presented with Unicode data--or
744 instead uses a traditional byte scheme when presented with byte
745 data.
746
747 Sadly, a wide octal escape does not cause a switch, and in a string with
748 no other multibyte characters, these octal escapes are treated as before.
749 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
750 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
751 Unicode string.
752
753 I have not perpetrated such confusion in PCRE. Up till now, it took just
754 the bottom 8 bits, as in old Perl. I have now made octal escapes with
755 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
756 translate to the appropriate multibyte character.
757
758 29. Applied some refactoring to reduce the number of warnings from Microsoft
759 and Borland compilers. This has included removing the fudge introduced
760 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
761 a warning about an unused variable.
762
763 21. PCRE has not included VT (character 0x0b) in the set of whitespace
764 characters since release 4.0, because Perl (from release 5.004) does not.
765 [Or at least, is documented not to: some releases seem to be in conflict
766 with the documentation.] However, when a pattern was studied with
767 pcre_study() and all its branches started with \s, PCRE still included VT
768 as a possible starting character. Of course, this did no harm; it just
769 caused an unnecessary match attempt.
770
771 22. Removed a now-redundant internal flag bit that recorded the fact that case
772 dependency changed within the pattern. This was once needed for "required
773 byte" processing, but is no longer used. This recovers a now-scarce options
774 bit. Also moved the least significant internal flag bit to the most-
775 significant bit of the word, which was not previously used (hangover from
776 the days when it was an int rather than a uint) to free up another bit for
777 the future.
778
779 23. Added support for CRLF line endings as well as CR and LF. As well as the
780 default being selectable at build time, it can now be changed at runtime
781 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
782 specify that it is scanning data with non-default line endings.
783
784 24. Changed the definition of CXXLINK to make it agree with the definition of
785 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
786
787 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
788 recursions. This makes a big different to stack usage for some patterns.
789
790 26. If a subpattern containing a named recursion or subroutine reference such
791 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
792 the space required for the compiled pattern went wrong and gave too small a
793 value. Depending on the environment, this could lead to "Failed: internal
794 error: code overflow at offset 49" or "glibc detected double free or
795 corruption" errors.
796
797 27. Applied patches from Google (a) to support the new newline modes and (b) to
798 advance over multibyte UTF-8 characters in GlobalReplace.
799
800 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
801 difference for some implementation of PCRE in some Windows version.
802
803 29. Added some extra testing facilities to pcretest:
804
805 \q<number> in a data line sets the "match limit" value
806 \Q<number> in a data line sets the "match recursion limt" value
807 -S <number> sets the stack size, where <number> is in megabytes
808
809 The -S option isn't available for Windows.
810
811
812 Version 6.6 06-Feb-06
813 ---------------------
814
815 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
816 in pcreposix.h. I have copied the definition from pcre.h.
817
818 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
819 because pcre.h is no longer a built file.
820
821 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
822 not normally included in the compiled code.
823
824
825 Version 6.5 01-Feb-06
826 ---------------------
827
828 1. When using the partial match feature with pcre_dfa_exec(), it was not
829 anchoring the second and subsequent partial matches at the new starting
830 point. This could lead to incorrect results. For example, with the pattern
831 /1234/, partially matching against "123" and then "a4" gave a match.
832
833 2. Changes to pcregrep:
834
835 (a) All non-match returns from pcre_exec() were being treated as failures
836 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
837 error message is output. Some extra information is given for the
838 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
839 probably the only errors that are likely to be caused by users (by
840 specifying a regex that has nested indefinite repeats, for instance).
841 If there are more than 20 of these errors, pcregrep is abandoned.
842
843 (b) A binary zero was treated as data while matching, but terminated the
844 output line if it was written out. This has been fixed: binary zeroes
845 are now no different to any other data bytes.
846
847 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
848 used to set a locale for matching. The --locale=xxxx long option has
849 been added (no short equivalent) to specify a locale explicitly on the
850 pcregrep command, overriding the environment variables.
851
852 (d) When -B was used with -n, some line numbers in the output were one less
853 than they should have been.
854
855 (e) Added the -o (--only-matching) option.
856
857 (f) If -A or -C was used with -c (count only), some lines of context were
858 accidentally printed for the final match.
859
860 (g) Added the -H (--with-filename) option.
861
862 (h) The combination of options -rh failed to suppress file names for files
863 that were found from directory arguments.
864
865 (i) Added the -D (--devices) and -d (--directories) options.
866
867 (j) Added the -F (--fixed-strings) option.
868
869 (k) Allow "-" to be used as a file name for -f as well as for a data file.
870
871 (l) Added the --colo(u)r option.
872
873 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
874 is not present by default.
875
876 3. A nasty bug was discovered in the handling of recursive patterns, that is,
877 items such as (?R) or (?1), when the recursion could match a number of
878 alternatives. If it matched one of the alternatives, but subsequently,
879 outside the recursion, there was a failure, the code tried to back up into
880 the recursion. However, because of the way PCRE is implemented, this is not
881 possible, and the result was an incorrect result from the match.
882
883 In order to prevent this happening, the specification of recursion has
884 been changed so that all such subpatterns are automatically treated as
885 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
886
887 4. I had overlooked the fact that, in some locales, there are characters for
888 which isalpha() is true but neither isupper() nor islower() are true. In
889 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
890 and ordfeminine) are like this. This affected the treatment of \w and \W
891 when they appeared in character classes, but not when they appeared outside
892 a character class. The bit map for "word" characters is now created
893 separately from the results of isalnum() instead of just taking it from the
894 upper, lower, and digit maps. (Plus the underscore character, of course.)
895
896 5. The above bug also affected the handling of POSIX character classes such as
897 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
898 permanent tables. Instead, the bit maps for such a class were previously
899 created as the appropriate unions of the upper, lower, and digit bitmaps.
900 Now they are created by subtraction from the [[:word:]] class, which has
901 its own bitmap.
902
903 6. The [[:blank:]] character class matches horizontal, but not vertical space.
904 It is created by subtracting the vertical space characters (\x09, \x0a,
905 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
906 subtraction was done in the overall bitmap for a character class, meaning
907 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
908 be recognized. This bug has been fixed.
909
910 7. Patches from the folks at Google:
911
912 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
913 real life, but is still worth protecting against".
914
915 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
916 regular expressions".
917
918 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
919 have it.
920
921 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
922 "configure" and the latter not, in order to fix a problem somebody had
923 with compiling the Arg class on HP-UX.
924
925 (e) Improve the error-handling of the C++ wrapper a little bit.
926
927 (f) New tests for checking recursion limiting.
928
929 8. The pcre_memmove() function, which is used only if the environment does not
930 have a standard memmove() function (and is therefore rarely compiled),
931 contained two bugs: (a) use of int instead of size_t, and (b) it was not
932 returning a result (though PCRE never actually uses the result).
933
934 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
935 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
936 returned instead of calling malloc() with an overflowing number that would
937 most likely cause subsequent chaos.
938
939 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
940
941 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
942 with this option is matched, the nmatch and pmatch options of regexec() are
943 ignored.
944
945 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
946 provided in case anyone wants to the the POSIX interface with UTF-8
947 strings.
948
949 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
950 C++ linking (needed for some HP-UX environments).
951
952 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
953 (unused parameter) and in the pcre_printint() function (omitted "default"
954 switch label when the default is to do nothing).
955
956 15. Added some code to make it possible, when PCRE is compiled as a C++
957 library, to replace subject pointers for pcre_exec() with a smart pointer
958 class, thus making it possible to process discontinuous strings.
959
960 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
961 much the same function. They were added by different people who were trying
962 to make PCRE easy to compile on non-Unix systems. It has been suggested
963 that PCRE_EXPORT be abolished now that there is more automatic apparatus
964 for compiling on Windows systems. I have therefore replaced it with
965 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
966 defaults to "extern" for C or "extern C" for C++, which works fine on
967 Unix-like systems. It is now possible to override the value of PCRE_DATA_
968 SCOPE with something explicit in config.h. In addition:
969
970 (a) pcreposix.h still had just "extern" instead of either of these macros;
971 I have replaced it with PCRE_DATA_SCOPE.
972
973 (b) Functions such as _pcre_xclass(), which are internal to the library,
974 but external in the C sense, all had PCRE_EXPORT in their definitions.
975 This is apparently wrong for the Windows case, so I have removed it.
976 (It makes no difference on Unix-like systems.)
977
978 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
979 of recursive calls to match(). This is different to MATCH_LIMIT because
980 that limits the total number of calls to match(), not all of which increase
981 the depth of recursion. Limiting the recursion depth limits the amount of
982 stack (or heap if NO_RECURSE is set) that is used. The default can be set
983 when PCRE is compiled, and changed at run time. A patch from Google adds
984 this functionality to the C++ interface.
985
986 18. Changes to the handling of Unicode character properties:
987
988 (a) Updated the table to Unicode 4.1.0.
989
990 (b) Recognize characters that are not in the table as "Cn" (undefined).
991
992 (c) I revised the way the table is implemented to a much improved format
993 which includes recognition of ranges. It now supports the ranges that
994 are defined in UnicodeData.txt, and it also amalgamates other
995 characters into ranges. This has reduced the number of entries in the
996 table from around 16,000 to around 3,000, thus reducing its size
997 considerably. I realized I did not need to use a tree structure after
998 all - a binary chop search is just as efficient. Having reduced the
999 number of entries, I extended their size from 6 bytes to 8 bytes to
1000 allow for more data.
1001
1002 (d) Added support for Unicode script names via properties such as \p{Han}.
1003
1004 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
1005 matching that character.
1006
1007 20. When matching a repeated Unicode property with a minimum greater than zero,
1008 (for example \pL{2,}), PCRE could look past the end of the subject if it
1009 reached it while seeking the minimum number of characters. This could
1010 happen only if some of the characters were more than one byte long, because
1011 there is a check for at least the minimum number of bytes.
1012
1013 21. Refactored the implementation of \p and \P so as to be more general, to
1014 allow for more different types of property in future. This has changed the
1015 compiled form incompatibly. Anybody with saved compiled patterns that use
1016 \p or \P will have to recompile them.
1017
1018 22. Added "Any" and "L&" to the supported property types.
1019
1020 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
1021 but give a compile time error if the value is greater than 0xff.
1022
1023 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
1024 accidentally not being installed or uninstalled.
1025
1026 25. The pcre.h file was built from pcre.h.in, but the only changes that were
1027 made were to insert the current release number. This seemed silly, because
1028 it made things harder for people building PCRE on systems that don't run
1029 "configure". I have turned pcre.h into a distributed file, no longer built
1030 by "configure", with the version identification directly included. There is
1031 no longer a pcre.h.in file.
1032
1033 However, this change necessitated a change to the pcre-config script as
1034 well. It is built from pcre-config.in, and one of the substitutions was the
1035 release number. I have updated configure.ac so that ./configure now finds
1036 the release number by grepping pcre.h.
1037
1038 26. Added the ability to run the tests under valgrind.
1039
1040
1041 Version 6.4 05-Sep-05
1042 ---------------------
1043
1044 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
1045 "--" to be printed when multiple files were scanned, even when none of the
1046 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
1047 consider it to be a bug, and have restored the previous behaviour.
1048
1049 2. A couple of code tidies to get rid of compiler warnings.
1050
1051 3. The pcretest program used to cheat by referring to symbols in the library
1052 whose names begin with _pcre_. These are internal symbols that are not
1053 really supposed to be visible externally, and in some environments it is
1054 possible to suppress them. The cheating is now confined to including
1055 certain files from the library's source, which is a bit cleaner.
1056
1057 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
1058 file's purpose clearer.
1059
1060 5. Reorganized pcre_ucp_findchar().
1061
1062
1063 Version 6.3 15-Aug-05
1064 ---------------------
1065
1066 1. The file libpcre.pc.in did not have general read permission in the tarball.
1067
1068 2. There were some problems when building without C++ support:
1069
1070 (a) If C++ support was not built, "make install" and "make test" still
1071 tried to test it.
1072
1073 (b) There were problems when the value of CXX was explicitly set. Some
1074 changes have been made to try to fix these, and ...
1075
1076 (c) --disable-cpp can now be used to explicitly disable C++ support.
1077
1078 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
1079 backslash in a target when C++ was disabled. This confuses some
1080 versions of "make", apparently. Using an intermediate variable solves
1081 this. (Same for CPP_LOBJ.)
1082
1083 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
1084 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
1085 necessary on certain architectures.
1086
1087 4. Added a setting of -export-symbols-regex to the link command to remove
1088 those symbols that are exported in the C sense, but actually are local
1089 within the library, and not documented. Their names all begin with
1090 "_pcre_". This is not a perfect job, because (a) we have to except some
1091 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
1092 available (and never for static libraries). I have made a note to try to
1093 find a way round (a) in the future.
1094
1095
1096 Version 6.2 01-Aug-05
1097 ---------------------
1098
1099 1. There was no test for integer overflow of quantifier values. A construction
1100 such as {1111111111111111} would give undefined results. What is worse, if
1101 a minimum quantifier for a parenthesized subpattern overflowed and became
1102 negative, the calculation of the memory size went wrong. This could have
1103 led to memory overwriting.
1104
1105 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
1106
1107 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
1108 operating environments where this matters.
1109
1110 4. Applied Giuseppe Maxia's patch to add additional features for controlling
1111 PCRE options from within the C++ wrapper.
1112
1113 5. Named capturing subpatterns were not being correctly counted when a pattern
1114 was compiled. This caused two problems: (a) If there were more than 100
1115 such subpatterns, the calculation of the memory needed for the whole
1116 compiled pattern went wrong, leading to an overflow error. (b) Numerical
1117 back references of the form \12, where the number was greater than 9, were
1118 not recognized as back references, even though there were sufficient
1119 previous subpatterns.
1120
1121 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
1122 versions of gcc, e.g. 2.95.4.
1123
1124
1125 Version 6.1 21-Jun-05
1126 ---------------------
1127
1128 1. There was one reference to the variable "posix" in pcretest.c that was not
1129 surrounded by "#if !defined NOPOSIX".
1130
1131 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
1132 the cross-check on the old pcre_info() function, for the benefit of the
1133 cut-down version of PCRE that is currently imported into Exim.
1134
1135 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
1136 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
1137 patterns that start (?i) but I don't think that matters. The use of (?i) is
1138 just an example; this all applies to the other options as well.
1139
1140 4. Since libtool seems to echo the compile commands it is issuing, the output
1141 from "make" can be reduced a bit by putting "@" in front of each libtool
1142 compile command.
1143
1144 5. Patch from the folks at Google for configure.in to be a bit more thorough
1145 in checking for a suitable C++ installation before trying to compile the
1146 C++ stuff. This should fix a reported problem when a compiler was present,
1147 but no suitable headers.
1148
1149 6. The man pages all had just "PCRE" as their title. I have changed them to
1150 be the relevant file name. I have also arranged that these names are
1151 retained in the file doc/pcre.txt, which is a concatenation in text format
1152 of all the man pages except the little individual ones for each function.
1153
1154 7. The NON-UNIX-USE file had not been updated for the different set of source
1155 files that come with release 6. I also added a few comments about the C++
1156 wrapper.
1157
1158
1159 Version 6.0 07-Jun-05
1160 ---------------------
1161
1162 1. Some minor internal re-organization to help with my DFA experiments.
1163
1164 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
1165 didn't matter for the library itself when fully configured, but did matter
1166 when compiling without UCP support, or within Exim, where the ucp files are
1167 not imported.
1168
1169 3. Refactoring of the library code to split up the various functions into
1170 different source modules. The addition of the new DFA matching code (see
1171 below) to a single monolithic source would have made it really too
1172 unwieldy, quite apart from causing all the code to be include in a
1173 statically linked application, when only some functions are used. This is
1174 relevant even without the DFA addition now that patterns can be compiled in
1175 one application and matched in another.
1176
1177 The downside of splitting up is that there have to be some external
1178 functions and data tables that are used internally in different modules of
1179 the library but which are not part of the API. These have all had their
1180 names changed to start with "_pcre_" so that they are unlikely to clash
1181 with other external names.
1182
1183 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
1184 a different (DFA) algorithm. Although it is slower than the original
1185 function, it does have some advantages for certain types of matching
1186 problem.
1187
1188 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
1189 including restarting after a partial match.
1190
1191 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
1192 defined when compiling for Windows was sent to me. I have put it into the
1193 code, though I have no means of testing or verifying it.
1194
1195 7. Added the pcre_refcount() auxiliary function.
1196
1197 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
1198 match before or at the first newline in the subject string. In pcretest,
1199 the /f option on a pattern can be used to set this.
1200
1201 9. A repeated \w when used in UTF-8 mode with characters greater than 256
1202 would behave wrongly. This has been present in PCRE since release 4.0.
1203
1204 10. A number of changes to the pcregrep command:
1205
1206 (a) Refactored how -x works; insert ^(...)$ instead of setting
1207 PCRE_ANCHORED and checking the length, in preparation for adding
1208 something similar for -w.
1209
1210 (b) Added the -w (match as a word) option.
1211
1212 (c) Refactored the way lines are read and buffered so as to have more
1213 than one at a time available.
1214
1215 (d) Implemented a pcregrep test script.
1216
1217 (e) Added the -M (multiline match) option. This allows patterns to match
1218 over several lines of the subject. The buffering ensures that at least
1219 8K, or the rest of the document (whichever is the shorter) is available
1220 for matching (and similarly the previous 8K for lookbehind assertions).
1221
1222 (f) Changed the --help output so that it now says
1223
1224 -w, --word-regex(p)
1225
1226 instead of two lines, one with "regex" and the other with "regexp"
1227 because that confused at least one person since the short forms are the
1228 same. (This required a bit of code, as the output is generated
1229 automatically from a table. It wasn't just a text change.)
1230
1231 (g) -- can be used to terminate pcregrep options if the next thing isn't an
1232 option but starts with a hyphen. Could be a pattern or a path name
1233 starting with a hyphen, for instance.
1234
1235 (h) "-" can be given as a file name to represent stdin.
1236
1237 (i) When file names are being printed, "(standard input)" is used for
1238 the standard input, for compatibility with GNU grep. Previously
1239 "<stdin>" was used.
1240
1241 (j) The option --label=xxx can be used to supply a name to be used for
1242 stdin when file names are being printed. There is no short form.
1243
1244 (k) Re-factored the options decoding logic because we are going to add
1245 two more options that take data. Such options can now be given in four
1246 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
1247
1248 (l) Added the -A, -B, and -C options for requesting that lines of context
1249 around matches be printed.
1250
1251 (m) Added the -L option to print the names of files that do not contain
1252 any matching lines, that is, the complement of -l.
1253
1254 (n) The return code is 2 if any file cannot be opened, but pcregrep does
1255 continue to scan other files.
1256
1257 (o) The -s option was incorrectly implemented. For compatibility with other
1258 greps, it now suppresses the error message for a non-existent or non-
1259 accessible file (but not the return code). There is a new option called
1260 -q that suppresses the output of matching lines, which was what -s was
1261 previously doing.
1262
1263 (p) Added --include and --exclude options to specify files for inclusion
1264 and exclusion when recursing.
1265
1266 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
1267 Hopefully, it now does.
1268
1269 12. Missing cast in pcre_study().
1270
1271 13. Added an "uninstall" target to the makefile.
1272
1273 14. Replaced "extern" in the function prototypes in Makefile.in with
1274 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
1275 world, but is set differently for Windows.
1276
1277 15. Added a second compiling function called pcre_compile2(). The only
1278 difference is that it has an extra argument, which is a pointer to an
1279 integer error code. When there is a compile-time failure, this is set
1280 non-zero, in addition to the error test pointer being set to point to an
1281 error message. The new argument may be NULL if no error number is required
1282 (but then you may as well call pcre_compile(), which is now just a
1283 wrapper). This facility is provided because some applications need a
1284 numeric error indication, but it has also enabled me to tidy up the way
1285 compile-time errors are handled in the POSIX wrapper.
1286
1287 16. Added VPATH=.libs to the makefile; this should help when building with one
1288 prefix path and installing with another. (Or so I'm told by someone who
1289 knows more about this stuff than I do.)
1290
1291 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
1292 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
1293 match everything, including newlines. This is not POSIX-compatible, but
1294 somebody wanted the feature. From pcretest it can be activated by using
1295 both the P and the s flags.
1296
1297 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
1298
1299 19. libpcre.pc was being incorrectly installed as executable.
1300
1301 20. A couple of places in pcretest check for end-of-line by looking for '\n';
1302 it now also looks for '\r' so that it will work unmodified on Windows.
1303
1304 21. Added Google's contributed C++ wrapper to the distribution.
1305
1306 22. Added some untidy missing memory free() calls in pcretest, to keep
1307 Electric Fence happy when testing.
1308
1309
1310
1311 Version 5.0 13-Sep-04
1312 ---------------------
1313
1314 1. Internal change: literal characters are no longer packed up into items
1315 containing multiple characters in a single byte-string. Each character
1316 is now matched using a separate opcode. However, there may be more than one
1317 byte in the character in UTF-8 mode.
1318
1319 2. The pcre_callout_block structure has two new fields: pattern_position and
1320 next_item_length. These contain the offset in the pattern to the next match
1321 item, and its length, respectively.
1322
1323 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
1324 insertion of callouts before each pattern item. Added the /C option to
1325 pcretest to make use of this.
1326
1327 4. On the advice of a Windows user, the lines
1328
1329 #if defined(_WIN32) || defined(WIN32)
1330 _setmode( _fileno( stdout ), 0x8000 );
1331 #endif /* defined(_WIN32) || defined(WIN32) */
1332
1333 have been added to the source of pcretest. This apparently does useful
1334 magic in relation to line terminators.
1335
1336 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
1337 for the benefit of those environments where the "b" makes a difference.
1338
1339 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
1340 to know about it. I have put a hack into configure.in that adds in code
1341 to set GCC=yes if CC=icc. This seems to end up at a point in the
1342 generated configure script that is early enough to affect the setting of
1343 compiler options, which is what is needed, but I have no means of testing
1344 whether it really works. (The user who reported this had patched the
1345 generated configure script, which of course I cannot do.)
1346
1347 LATER: After change 22 below (new libtool files), the configure script
1348 seems to know about icc (and also ecc). Therefore, I have commented out
1349 this hack in configure.in.
1350
1351 7. Added support for pkg-config (2 patches were sent in).
1352
1353 8. Negated POSIX character classes that used a combination of internal tables
1354 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
1355 [[:^ascii]]. Typically, they would match almost any characters. The other
1356 POSIX classes were not broken in this way.
1357
1358 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
1359 to find the match, as PCRE was deluded into thinking that the match had to
1360 start at the start point or following a newline. The same bug applied to
1361 patterns with negative forward assertions or any backward assertions
1362 preceding ".*" at the start, unless the pattern required a fixed first
1363 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
1364
1365 10. In UTF-8 mode, when moving forwards in the subject after a failed match
1366 starting at the last subject character, bytes beyond the end of the subject
1367 string were read.
1368
1369 11. Renamed the variable "class" as "classbits" to make life easier for C++
1370 users. (Previously there was a macro definition, but it apparently wasn't
1371 enough.)
1372
1373 12. Added the new field "tables" to the extra data so that tables can be passed
1374 in at exec time, or the internal tables can be re-selected. This allows
1375 a compiled regex to be saved and re-used at a later time by a different
1376 program that might have everything at different addresses.
1377
1378 13. Modified the pcre-config script so that, when run on Solaris, it shows a
1379 -R library as well as a -L library.
1380
1381 14. The debugging options of pcretest (-d on the command line or D on a
1382 pattern) showed incorrect output for anything following an extended class
1383 that contained multibyte characters and which was followed by a quantifier.
1384
1385 15. Added optional support for general category Unicode character properties
1386 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
1387 support. It adds about 90K to the size of the library. The meanings of the
1388 inbuilt class escapes such as \d and \s have NOT been changed.
1389
1390 16. Updated pcredemo.c to include calls to free() to release the memory for the
1391 compiled pattern.
1392
1393 17. The generated file chartables.c was being created in the source directory
1394 instead of in the building directory. This caused the build to fail if the
1395 source directory was different from the building directory, and was
1396 read-only.
1397
1398 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
1399 file. No doubt somebody will tell me if they don't make sense... Also added
1400 Dan Mooney's comments about building on OpenVMS.
1401
1402 19. Added support for partial matching via the PCRE_PARTIAL option for
1403 pcre_exec() and the \P data escape in pcretest.
1404
1405 20. Extended pcretest with 3 new pattern features:
1406
1407 (i) A pattern option of the form ">rest-of-line" causes pcretest to
1408 write the compiled pattern to the file whose name is "rest-of-line".
1409 This is a straight binary dump of the data, with the saved pointer to
1410 the character tables forced to be NULL. The study data, if any, is
1411 written too. After writing, pcretest reads a new pattern.
1412
1413 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
1414 compiled pattern from the given file. There must not be any
1415 occurrences of "<" in the file name (pretty unlikely); if there are,
1416 pcretest will instead treat the initial "<" as a pattern delimiter.
1417 After reading in the pattern, pcretest goes on to read data lines as
1418 usual.
1419
1420 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
1421 and 16-bit fields in a compiled pattern, to simulate a pattern that
1422 was compiled on a host of opposite endianness.
1423
1424 21. The pcre-exec() function can now cope with patterns that were compiled on
1425 hosts of opposite endianness, with this restriction:
1426
1427 As for any compiled expression that is saved and used later, the tables
1428 pointer field cannot be preserved; the extra_data field in the arguments
1429 to pcre_exec() should be used to pass in a tables address if a value
1430 other than the default internal tables were used at compile time.
1431
1432 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
1433 now diagnosed as an error. Previously, most of the time, a negative number
1434 would have been treated as zero, but if in addition "ovector" was passed as
1435 NULL, a crash could occur.
1436
1437 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
1438 new versions from the libtool 1.5 distribution (the last one is a copy of
1439 a file called libtool.m4). This seems to have fixed the need to patch
1440 "configure" to support Darwin 1.3 (which I used to do). However, I still
1441 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
1442 workstation).
1443
1444 24. Changed the PCRE licence to be the more standard "BSD" licence.
1445
1446
1447 Version 4.5 01-Dec-03
1448 ---------------------
1449
1450 1. There has been some re-arrangement of the code for the match() function so
1451 that it can be compiled in a version that does not call itself recursively.
1452 Instead, it keeps those local variables that need separate instances for
1453 each "recursion" in a frame on the heap, and gets/frees frames whenever it
1454 needs to "recurse". Keeping track of where control must go is done by means
1455 of setjmp/longjmp. The whole thing is implemented by a set of macros that
1456 hide most of the details from the main code, and operates only if
1457 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
1458 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
1459 operating.
1460
1461 To make it easier for callers to provide specially tailored get/free
1462 functions for this usage, two new functions, pcre_stack_malloc, and
1463 pcre_stack_free, are used. They are always called in strict stacking order,
1464 and the size of block requested is always the same.
1465
1466 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
1467 PCRE has been compiled to use the stack or the heap for recursion. The
1468 -C option of pcretest uses this to show which version is compiled.
1469
1470 A new data escape \S, is added to pcretest; it causes the amounts of store
1471 obtained and freed by both kinds of malloc/free at match time to be added
1472 to the output.
1473
1474 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
1475 what's available on my current Linux desktop machine.
1476
1477 3. When matching a UTF-8 string, the test for a valid string at the start has
1478 been extended. If start_offset is not zero, PCRE now checks that it points
1479 to a byte that is the start of a UTF-8 character. If not, it returns
1480 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
1481 this is necessary because there may be backward assertions in the pattern.
1482 When matching the same subject several times, it may save resources to use
1483 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
1484
1485 4. The code for checking the validity of UTF-8 strings has been tightened so
1486 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
1487 containing "overlong sequences".
1488
1489 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
1490 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
1491 should have been "&", but it just so happened that all the cases this let
1492 through by mistake were picked up later in the function.
1493
1494 6. I had used a variable called "isblank" - this is a C99 function, causing
1495 some compilers to warn. To avoid this, I renamed it (as "blankclass").
1496
1497 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
1498 prompting; (b) run "./pcretest /dev/null" at the start of the test script
1499 so the version is shown; (c) stop "make test" echoing "./RunTest".
1500
1501 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
1502
1503 9. The prototype for memmove() for systems that don't have it was using
1504 size_t, but the inclusion of the header that defines size_t was later. I've
1505 moved the #includes for the C headers earlier to avoid this.
1506
1507 10. Added some adjustments to the code to make it easier to compiler on certain
1508 special systems:
1509
1510 (a) Some "const" qualifiers were missing.
1511 (b) Added the macro EXPORT before all exported functions; by default this
1512 is defined to be empty.
1513 (c) Changed the dftables auxiliary program (that builds chartables.c) so
1514 that it reads its output file name as an argument instead of writing
1515 to the standard output and assuming this can be redirected.
1516
1517 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
1518 class containing characters with values greater than 255, PCRE compilation
1519 went into a loop.
1520
1521 12. A recursive reference to a subpattern that was within another subpattern
1522 that had a minimum quantifier of zero caused PCRE to crash. For example,
1523 (x(y(?2))z)? provoked this bug with a subject that got as far as the
1524 recursion. If the recursively-called subpattern itself had a zero repeat,
1525 that was OK.
1526
1527 13. In pcretest, the buffer for reading a data line was set at 30K, but the
1528 buffer into which it was copied (for escape processing) was still set at
1529 1024, so long lines caused crashes.
1530
1531 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
1532 "internal error: code overflow...". This applied to any character class
1533 that was followed by a possessive quantifier.
1534
1535 15. Modified the Makefile to add libpcre.la as a prerequisite for
1536 libpcreposix.la because I was told this is needed for a parallel build to
1537 work.
1538
1539 16. If a pattern that contained .* following optional items at the start was
1540 studied, the wrong optimizing data was generated, leading to matching
1541 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
1542 matching string must start with a or b or c. The correct conclusion for
1543 this pattern is that a match can start with any character.
1544
1545
1546 Version 4.4 13-Aug-03
1547 ---------------------
1548
1549 1. In UTF-8 mode, a character class containing characters with values between
1550 127 and 255 was not handled correctly if the compiled pattern was studied.
1551 In fixing this, I have also improved the studying algorithm for such
1552 classes (slightly).
1553
1554 2. Three internal functions had redundant arguments passed to them. Removal
1555 might give a very teeny performance improvement.
1556
1557 3. Documentation bug: the value of the capture_top field in a callout is *one
1558 more than* the number of the hightest numbered captured substring.
1559
1560 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
1561 in incorrectly linking with a previously installed version. They now link
1562 explicitly with libpcre.la.
1563
1564 5. configure.in no longer needs to recognize Cygwin specially.
1565
1566 6. A problem in pcre.in for Windows platforms is fixed.
1567
1568 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
1569 pcretest, it used to include the size of the study block as part of its
1570 output. Unfortunately, the structure contains a field that has a different
1571 size on different hardware architectures. This meant that the tests that
1572 showed this size failed. As the block is currently always of a fixed size,
1573 this information isn't actually particularly useful in pcretest output, so
1574 I have just removed it.
1575
1576 8. Three pre-processor statements accidentally did not start in column 1.
1577 Sadly, there are *still* compilers around that complain, even though
1578 standard C has not required this for well over a decade. Sigh.
1579
1580 9. In pcretest, the code for checking callouts passed small integers in the
1581 callout_data field, which is a void * field. However, some picky compilers
1582 complained about the casts involved for this on 64-bit systems. Now
1583 pcretest passes the address of the small integer instead, which should get
1584 rid of the warnings.
1585
1586 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
1587 both compile and run time, and gives an error if an invalid UTF-8 sequence
1588 is found. There is a option for disabling this check in cases where the
1589 string is known to be correct and/or the maximum performance is wanted.
1590
1591 11. In response to a bug report, I changed one line in Makefile.in from
1592
1593 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
1594 to
1595 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
1596
1597 to look similar to other lines, but I have no way of telling whether this
1598 is the right thing to do, as I do not use Windows. No doubt I'll get told
1599 if it's wrong...
1600
1601
1602 Version 4.3 21-May-03
1603 ---------------------
1604
1605 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
1606 Makefile.
1607
1608 2. Some refactoring to improve the quality of the code:
1609
1610 (i) The utf8_table... variables are now declared "const".
1611
1612 (ii) The code for \cx, which used the "case flipping" table to upper case
1613 lower case letters, now just substracts 32. This is ASCII-specific,
1614 but the whole concept of \cx is ASCII-specific, so it seems
1615 reasonable.
1616
1617 (iii) PCRE was using its character types table to recognize decimal and
1618 hexadecimal digits in the pattern. This is silly, because it handles
1619 only 0-9, a-f, and A-F, but the character types table is locale-
1620 specific, which means strange things might happen. A private
1621 table is now used for this - though it costs 256 bytes, a table is
1622 much faster than multiple explicit tests. Of course, the standard
1623 character types table is still used for matching digits in subject
1624 strings against \d.
1625
1626 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
1627 ending in _t are). So I've renamed it as ESC_tee.
1628
1629 3. The first argument for regexec() in the POSIX wrapper should have been
1630 defined as "const".
1631
1632 4. Changed pcretest to use malloc() for its buffers so that they can be
1633 Electric Fenced for debugging.
1634
1635 5. There were several places in the code where, in UTF-8 mode, PCRE would try
1636 to read one or more bytes before the start of the subject string. Often this
1637 had no effect on PCRE's behaviour, but in some circumstances it could
1638 provoke a segmentation fault.
1639
1640 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
1641 to try to read one or more bytes before the start of the subject string.
1642
1643 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
1644 UTF-8 support could misbehave in various ways if the subject string
1645 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
1646 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
1647 back over UTF-8 characters.)
1648
1649
1650 Version 4.2 14-Apr-03
1651 ---------------------
1652
1653 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
1654
1655 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
1656 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
1657 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
1658 [WIN_PREFIX]: new variable, "cyg" for Cygwin
1659 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
1660 and BUILD_EXEEXT
1661 Note: automatic setting of the BUILD variables is not yet working
1662 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
1663 compile-time but not at link-time
1664 [LINK]: use for linking executables only
1665 make different versions for Windows and non-Windows
1666 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
1667 libraries
1668 [LINK_FOR_BUILD]: new variable
1669 [OBJEXT]: use throughout
1670 [EXEEXT]: use throughout
1671 <winshared>: new target
1672 <wininstall>: new target
1673 <dftables.o>: use native compiler
1674 <dftables>: use native linker
1675 <install>: handle Windows platform correctly
1676 <clean>: ditto
1677 <check>: ditto
1678 copy DLL to top builddir before testing
1679
1680 As part of these changes, -no-undefined was removed again. This was reported
1681 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
1682 in any case.
1683
1684 3. Some tidies to get rid of compiler warnings:
1685
1686 . In the match_data structure, match_limit was an unsigned long int, whereas
1687 match_call_count was an int. I've made them both unsigned long ints.
1688
1689 . In pcretest the fact that a const uschar * doesn't automatically cast to
1690 a void * provoked a warning.
1691
1692 . Turning on some more compiler warnings threw up some "shadow" variables
1693 and a few more missing casts.
1694
1695 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
1696 option, a class that contained a single character with a value between 128
1697 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
1698
1699 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
1700 option, a class that contained several characters, but with at least one
1701 whose value was between 128 and 255 caused PCRE to crash.
1702
1703
1704 Version 4.1 12-Mar-03
1705 ---------------------
1706
1707 1. Compiling with gcc -pedantic found a couple of places where casts were
1708 needed, and a string in dftables.c that was longer than standard compilers are
1709 required to support.
1710
1711 2. Compiling with Sun's compiler found a few more places where the code could
1712 be tidied up in order to avoid warnings.
1713
1714 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
1715 first of these names is deprecated in the latest Autoconf in favour of the name
1716 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
1717 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
1718 analogy I have changed it to CFLAGS_FOR_BUILD.
1719
1720 4. Added -no-undefined to the linking command in the Makefile, because this is
1721 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
1722 linking step for the pcreposix library.
1723
1724 5. PCRE was failing to diagnose the case of two named groups with the same
1725 name.
1726
1727 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
1728 literal character that is needed in the subject for a match, and scans along to
1729 ensure that it is present before embarking on the full matching process. This
1730 saves time in cases of nested unlimited repeats that are never going to match.
1731 Problem: the scan can take a lot of time if the subject is very long (e.g.
1732 megabytes), thus penalizing straightforward matches. It is now done only if the
1733 amount of subject to be scanned is less than 1000 bytes.
1734
1735 7. A lesser problem with the same optimization is that it was recording the
1736 first character of an anchored pattern as "needed", thus provoking a search
1737 right along the subject, even when the first match of the pattern was going to
1738 fail. The "needed" character is now not set for anchored patterns, unless it
1739 follows something in the pattern that is of non-fixed length. Thus, it still
1740 fulfils its original purpose of finding quick non-matches in cases of nested
1741 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
1742
1743
1744 Version 4.0 17-Feb-03
1745 ---------------------
1746
1747 1. If a comment in an extended regex that started immediately after a meta-item
1748 extended to the end of string, PCRE compiled incorrect data. This could lead to
1749 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
1750
1751 2. Moved to autoconf 2.53 and libtool 1.4.2.
1752
1753 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
1754 the special perltest8 script is no longer needed - all the tests can be run
1755 from a single perltest script.
1756
1757 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
1758 by \s. It has now been removed in PCRE. This means it isn't recognized as
1759 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
1760 class [:space:] *does* include VT, thereby creating a mess.
1761
1762 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
1763 space and tab.
1764
1765 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
1766 its new features into the main test script, reducing the number of scripts.
1767
1768 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
1769 were backward compatible, and made the (?i) apply to the whole pattern, as if
1770 /i were given. Now it behaves more logically, and applies the option setting
1771 only to what follows. PCRE has been changed to follow suit. However, if it
1772 finds options settings right at the start of the pattern, it extracts them into
1773 the global options, as before. Thus, they show up in the info data.
1774
1775 8. Added support for the \Q...\E escape sequence. Characters in between are
1776 treated as literals. This is slightly different from Perl in that $ and @ are
1777 also handled as literals inside the quotes. In Perl, they will cause variable
1778 interpolation. Note the following examples:
1779
1780 Pattern PCRE matches Perl matches
1781
1782 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
1783 \Qabc\$xyz\E abc\$xyz abc\$xyz
1784 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
1785
1786 For compatibility with Perl, \Q...\E sequences are recognized inside character
1787 classes as well as outside them.
1788
1789 9. Re-organized 3 code statements in pcretest to avoid "overflow in
1790 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
1791 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
1792 signed/unsigned warnings.
1793
1794 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
1795 option for pcretest, so I've replaced it by a simple function that does just
1796 that job.
1797
1798 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
1799 "pcregrep -".
1800
1801 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
1802 Java package. This provides some syntactic sugar for simple cases of what my
1803 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
1804 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
1805 item, you can use this simplified notation. Note that only makes sense with
1806 greedy quantifiers. Consequently, the use of the possessive quantifier forces
1807 greediness, whatever the setting of the PCRE_UNGREEDY option.
1808
1809 13. A change of greediness default within a pattern was not taking effect at
1810 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
1811 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
1812 was abstracted outside.
1813
1814 14. PCRE now supports the \G assertion. It is true when the current matching
1815 position is at the start point of the match. This differs from \A when the
1816 starting offset is non-zero. Used with the /g option of pcretest (or similar
1817 code), it works in the same way as it does for Perl's /g option. If all
1818 alternatives of a regex begin with \G, the expression is anchored to the start
1819 match position, and the "anchored" flag is set in the compiled expression.
1820
1821 15. Some bugs concerning the handling of certain option changes within patterns
1822 have been fixed. These applied to options other than (?ims). For example,
1823 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
1824 been the other way round. Some of this was related to change 7 above.
1825
1826 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
1827 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
1828 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
1829 POSIX classes only within a class (e.g. /[[:alpha:]]/).
1830
1831 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
1832 mode. Unlike ".", it always matches newline, whatever the setting of
1833 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
1834 assertions. Perl allows it, but it doesn't (in general) work because it can't
1835 calculate the length of the lookbehind. At least, that's the case for Perl
1836 5.8.0 - I've been told they are going to document that it doesn't work in
1837 future.
1838
1839 18. Added an error diagnosis for escapes that PCRE does not support: these are
1840 \L, \l, \N, \P, \p, \U, \u, and \X.
1841
1842 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
1843 reading past the end of the pattern in cases such as /[abcd/.
1844
1845 20. PCRE was getting more memory than necessary for patterns with classes that
1846 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
1847
1848 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
1849 compiling PCRE for use with Virtual Pascal.
1850
1851 22. Small fix to the Makefile to make it work properly if the build is done
1852 outside the source tree.
1853
1854 23. Added a new extension: a condition to go with recursion. If a conditional
1855 subpattern starts with (?(R) the "true" branch is used if recursion has
1856 happened, whereas the "false" branch is used only at the top level.
1857
1858 24. When there was a very long string of literal characters (over 255 bytes
1859 without UTF support, over 250 bytes with UTF support), the computation of how
1860 much memory was required could be incorrect, leading to segfaults or other
1861 strange effects.
1862
1863 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
1864 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
1865 there was a subsequent back reference to those brackets. This meant that, for
1866 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
1867 possible to check for precisely this case. All we can do is abandon the
1868 optimization if .* occurs inside capturing brackets when there are any back
1869 references whatsoever. (See below for a better fix that came later.)
1870
1871 26. The handling of the optimization for finding the first character of a
1872 non-anchored pattern, and for finding a character that is required later in the
1873 match were failing in some cases. This didn't break the matching; it just
1874 failed to optimize when it could. The way this is done has been re-implemented.
1875
1876 27. Fixed typo in error message for invalid (?R item (it said "(?p").
1877
1878 28. Added a new feature that provides some of the functionality that Perl
1879 provides with (?{...}). The facility is termed a "callout". The way it is done
1880 in PCRE is for the caller to provide an optional function, by setting
1881 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
1882 global variable. By default it is unset, which disables all calling out. To get
1883 the function called, the regex must include (?C) at appropriate points. This
1884 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
1885 This provides a means of identifying different callout points. When PCRE
1886 reaches such a point in the regex, if pcre_callout has been set, the external
1887 function is called. It is provided with data in a structure called
1888 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
1889 matching continues; if it returns a non-zero value, the match at the current
1890 point fails. However, backtracking will occur if possible. [This was changed
1891 later and other features added - see item 49 below.]
1892
1893 29. pcretest is upgraded to test the callout functionality. It provides a
1894 callout function that displays information. By default, it shows the start of
1895 the match and the current position in the text. There are some new data escapes
1896 to vary what happens:
1897
1898 \C+ in addition, show current contents of captured substrings
1899 \C- do not supply a callout function
1900 \C!n return 1 when callout number n is reached
1901 \C!n!m return 1 when callout number n is reached for the mth time
1902
1903 30. If pcregrep was called with the -l option and just a single file name, it
1904 output "<stdin>" if a match was found, instead of the file name.
1905
1906 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
1907 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
1908 pcre_exec(). This saves a malloc/free per call. The default value of
1909 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
1910 when configuring.
1911
1912 32. The default maximum size of a compiled pattern is 64K. There have been a
1913 few cases of people hitting this limit. The code now uses macros to handle the
1914 storing of links as offsets within the compiled pattern. It defaults to 2-byte
1915 links, but this can be changed to 3 or 4 bytes by --with-link-size when
1916 configuring. Tests 2 and 5 work only with 2-byte links because they output
1917 debugging information about compiled patterns.
1918
1919 33. Internal code re-arrangements:
1920
1921 (a) Moved the debugging function for printing out a compiled regex into
1922 its own source file (printint.c) and used #include to pull it into
1923 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
1924 separate copies.
1925
1926 (b) Defined the list of op-code names for debugging as a macro in
1927 internal.h so that it is next to the definition of the opcodes.
1928
1929 (c) Defined a table of op-code lengths for simpler skipping along compiled
1930 code. This is again a macro in internal.h so that it is next to the
1931 definition of the opcodes.
1932
1933 34. Added support for recursive calls to individual subpatterns, along the
1934 lines of Robin Houston's patch (but implemented somewhat differently).
1935
1936 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
1937 allow it to read and process whole directories in Win32. This code was
1938 contributed by Lionel Fourquaux; it has not been tested by me.
1939
1940 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
1941 used to name a group. Names consist of alphanumerics and underscores, and must
1942 be unique. Back references use the syntax (?P=name) and recursive calls use
1943 (?P>name) which is a PCRE extension to the Python extension. Groups still have
1944 numbers. The function pcre_fullinfo() can be used after compilation to extract
1945 a name/number map. There are three relevant calls:
1946
1947 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
1948 PCRE_INFO_NAMECOUNT yields the number of entries
1949 PCRE_INFO_NAMETABLE yields a pointer to the map.
1950
1951 The map is a vector of fixed-size entries. The size of each entry depends on
1952 the length of the longest name used. The first two bytes of each entry are the
1953 group number, most significant byte first. There follows the corresponding
1954 name, zero terminated. The names are in alphabetical order.
1955
1956 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
1957 case instead of 255. Making it the same both with and without UTF-8 support
1958 means that the same test output works with both.
1959
1960 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
1961 calling malloc() with a zero argument.
1962
1963 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
1964 optimization. I've improved things by keeping a bitmap of backreferences with
1965 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
1966 fact referenced, the optimization can be applied. It is unlikely that a
1967 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
1968 the match to follow \n) will appear inside brackets with a number greater than
1969 31, but if it does, any back reference > 31 suppresses the optimization.
1970
1971 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
1972 of disabling numbered capturing parentheses. Any opening parenthesis that is
1973 not followed by ? behaves as if it were followed by ?: but named parentheses
1974 can still be used for capturing (and they will acquire numbers in the usual
1975 way).
1976
1977 41. Redesigned the return codes from the match() function into yes/no/error so
1978 that errors can be passed back from deep inside the nested calls. A malloc
1979 failure while inside a recursive subpattern call now causes the
1980 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
1981
1982 42. It is now possible to set a limit on the number of times the match()
1983 function is called in a call to pcre_exec(). This facility makes it possible to
1984 limit the amount of recursion and backtracking, though not in a directly
1985 obvious way, because the match() function is used in a number of different
1986 circumstances. The count starts from zero for each position in the subject
1987 string (for non-anchored patterns). The default limit is, for compatibility, a
1988 large number, namely 10 000 000. You can change this in two ways:
1989
1990 (a) When configuring PCRE before making, you can use --with-match-limit=n
1991 to set a default value for the compiled library.
1992
1993 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
1994 a different value is set. See 45 below.
1995
1996 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
1997
1998 43. Added a new function pcre_config(int, void *) to enable run-time extraction
1999 of things that can be changed at compile time. The first argument specifies
2000 what is wanted and the second points to where the information is to be placed.
2001 The current list of available information is:
2002
2003 PCRE_CONFIG_UTF8
2004
2005 The output is an integer that is set to one if UTF-8 support is available;
2006 otherwise it is set to zero.
2007
2008 PCRE_CONFIG_NEWLINE
2009
2010 The output is an integer that it set to the value of the code that is used for
2011 newline. It is either LF (10) or CR (13).
2012
2013 PCRE_CONFIG_LINK_SIZE
2014
2015 The output is an integer that contains the number of bytes used for internal
2016 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
2017
2018 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
2019
2020 The output is an integer that contains the threshold above which the POSIX
2021 interface uses malloc() for output vectors. See item 31 above.
2022
2023 PCRE_CONFIG_MATCH_LIMIT
2024
2025 The output is an unsigned integer that contains the default limit of the number
2026 of match() calls in a pcre_exec() execution. See 42 above.
2027
2028 44. pcretest has been upgraded by the addition of the -C option. This causes it
2029 to extract all the available output from the new pcre_config() function, and to
2030 output it. The program then exits immediately.
2031
2032 45. A need has arisen to pass over additional data with calls to pcre_exec() in
2033 order to support additional features. One way would have been to define
2034 pcre_exec2() (for example) with extra arguments, but this would not have been
2035 extensible, and would also have required all calls to the original function to
2036 be mapped to the new one. Instead, I have chosen to extend the mechanism that
2037 is used for passing in "extra" data from pcre_study().
2038
2039 The pcre_extra structure is now exposed and defined in pcre.h. It currently
2040 contains the following fields:
2041
2042 flags a bitmap indicating which of the following fields are set
2043 study_data opaque data from pcre_study()
2044 match_limit a way of specifying a limit on match() calls for a specific
2045 call to pcre_exec()
2046 callout_data data for callouts (see 49 below)
2047
2048 The flag bits are also defined in pcre.h, and are
2049
2050 PCRE_EXTRA_STUDY_DATA
2051 PCRE_EXTRA_MATCH_LIMIT
2052 PCRE_EXTRA_CALLOUT_DATA
2053
2054 The pcre_study() function now returns one of these new pcre_extra blocks, with
2055 the actual study data pointed to by the study_data field, and the
2056 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
2057 before. That is, this change is entirely upwards-compatible and requires no
2058 change to existing code.
2059
2060 If you want to pass in additional data to pcre_exec(), you can either place it
2061 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
2062 block.
2063
2064 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
2065 data string contains the escape sequence \M, pcretest calls pcre_exec() several
2066 times with different match limits, until it finds the minimum value needed for
2067 pcre_exec() to complete. The value is then output. This can be instructive; for
2068 most simple matches the number is quite small, but for pathological cases it
2069 gets very large very quickly.
2070
2071 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
2072 returns the size of the data block pointed to by the study_data field in a
2073 pcre_extra block, that is, the value that was passed as the argument to
2074 pcre_malloc() when PCRE was getting memory in which to place the information
2075 created by pcre_study(). The fourth argument should point to a size_t variable.
2076 pcretest has been extended so that this information is shown after a successful
2077 pcre_study() call when information about the compiled regex is being displayed.
2078
2079 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
2080 because what follows is always an absolute path. (Later: it turns out that this
2081 is more than cosmetic for MinGW, because it doesn't like empty path
2082 components.)
2083
2084 49. Some changes have been made to the callout feature (see 28 above):
2085
2086 (i) A callout function now has three choices for what it returns:
2087
2088 0 => success, carry on matching
2089 > 0 => failure at this point, but backtrack if possible
2090 < 0 => serious error, return this value from pcre_exec()
2091
2092 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
2093 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
2094 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
2095 use by callout functions. It will never be used by PCRE itself.
2096
2097 (ii) The pcre_extra structure (see 45 above) has a void * field called
2098 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
2099 pcre_callout_block structure has a field of the same name. The contents of
2100 the field passed in the pcre_extra structure are passed to the callout
2101 function in the corresponding field in the callout block. This makes it
2102 easier to use the same callout-containing regex from multiple threads. For
2103 testing, the pcretest program has a new data escape
2104
2105 \C*n pass the number n (may be negative) as callout_data
2106
2107 If the callout function in pcretest receives a non-zero value as
2108 callout_data, it returns that value.
2109
2110 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
2111 there were some redundant $(CFLAGS) in commands that are now specified as
2112 $(LINK), which already includes $(CFLAGS).
2113
2114 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
2115 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
2116 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
2117 one-byte characters throughout. Note that case-insensitive matching applies
2118 only to characters whose values are less than 256. PCRE doesn't support the
2119 notion of cases for higher-valued characters.
2120
2121 (i) A character class whose characters are all within 0-255 is handled as
2122 a bit map, and the map is inverted for negative classes. Previously, a
2123 character > 255 always failed to match such a class; however it should
2124 match if the class was a negative one (e.g. [^ab]). This has been fixed.
2125
2126 (ii) A negated character class with a single character < 255 is coded as
2127 "not this character" (OP_NOT). This wasn't working properly when the test
2128 character was multibyte, either singly or repeated.
2129
2130 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
2131 mode, for example: \x{100}{2,3}.
2132
2133 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
2134 singly or repeated) now correctly test multibyte characters. However,
2135 PCRE doesn't recognize any characters with values greater than 255 as
2136 digits, spaces, or word characters. Such characters always match \D, \S,
2137 and \W, and never match \d, \s, or \w.
2138
2139 (v) Classes may now contain characters and character ranges with values
2140 greater than 255. For example: [ab\x{100}-\x{400}].
2141
2142 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
2143 PCRE in UTF-8 mode.
2144
2145 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
2146 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
2147 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
2148 value.)
2149
2150 53. The single man page has become too large. I have therefore split it up into
2151 a number of separate man pages. These also give rise to individual HTML pages;
2152 these are now put in a separate directory, and there is an index.html page that
2153 lists them all. Some hyperlinking between the pages has been installed.
2154
2155 54. Added convenience functions for handling named capturing parentheses.
2156
2157 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
2158 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
2159 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
2160 are faulted.
2161
2162 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
2163 calling configure. These values are used when compiling the dftables.c program
2164 which is run to generate the source of the default character tables. They
2165 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
2166 you will need to set these values.
2167
2168 57. Updated the building process for Windows DLL, as provided by Fred Cox.
2169
2170
2171 Version 3.9 02-Jan-02
2172 ---------------------
2173
2174 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
2175
2176 2. If --disable-static was given, the building process failed when trying to
2177 build pcretest and pcregrep. (For some reason it was using libtool to compile
2178 them, which is not right, as they aren't part of the library.)
2179
2180
2181 Version 3.8 18-Dec-01
2182 ---------------------
2183
2184 1. The experimental UTF-8 code was completely screwed up. It was packing the
2185 bytes in the wrong order. How dumb can you get?
2186
2187
2188 Version 3.7 29-Oct-01
2189 ---------------------
2190
2191 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
2192 This caused pcretest, when used on the test data, to segfault. Unfortunately,
2193 this didn't happen under Solaris 8, where I normally test things.
2194
2195 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
2196 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
2197 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
2198 this fix an hour or so after the initial 3.7 release.)
2199
2200
2201 Version 3.6 23-Oct-01
2202 ---------------------
2203
2204 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
2205 offsets passed as NULL with zero offset count.
2206
2207 2. The config.guess and config.sub files had not been updated when I moved to
2208 the latest autoconf.
2209
2210
2211 Version 3.5 15-Aug-01
2212 ---------------------
2213
2214 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
2215 had been forgotten.
2216
2217 2. By using declared but undefined structures, we can avoid using "void"
2218 definitions in pcre.h while keeping the internal definitions of the structures
2219 private.
2220
2221 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
2222 user point of view, this means that both static and shared libraries are built
2223 by default, but this can be individually controlled. More of the work of
2224 handling this static/shared cases is now inside libtool instead of PCRE's make
2225 file.
2226
2227 4. The pcretest utility is now installed along with pcregrep because it is
2228 useful for users (to test regexs) and by doing this, it automatically gets
2229 relinked by libtool. The documentation has been turned into a man page, so
2230 there are now .1, .txt, and .html versions in /doc.
2231
2232 5. Upgrades to pcregrep:
2233 (i) Added long-form option names like gnu grep.
2234 (ii) Added --help to list all options with an explanatory phrase.
2235 (iii) Added -r, --recursive to recurse into sub-directories.
2236 (iv) Added -f, --file to read patterns from a file.
2237
2238 6. pcre_exec() was referring to its "code" argument before testing that
2239 argument for NULL (and giving an error if it was NULL).
2240
2241 7. Upgraded Makefile.in to allow for compiling in a different directory from
2242 the source directory.
2243
2244 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
2245 options bits, the pointer it was passed was to an int instead of to an unsigned
2246 long int. This mattered only on 64-bit systems.
2247
2248 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
2249 generated) instead of pcre.in, which it its source. Also made the same change
2250 in several of the .c files.
2251
2252 10. A new release of gcc defines printf() as a macro, which broke pcretest
2253 because it had an ifdef in the middle of a string argument for printf(). Fixed
2254 by using separate calls to printf().
2255
2256 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
2257 script, to force use of CR or LF instead of \n in the source. On non-Unix
2258 systems, the value can be set in config.h.
2259
2260 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
2261 absolute limit. Changed the text of the error message to make this clear, and
2262 likewise updated the man page.
2263
2264 13. The limit of 99 on the number of capturing subpatterns has been removed.
2265 The new limit is 65535, which I hope will not be a "real" limit.
2266
2267
2268 Version 3.4 22-Aug-00
2269 ---------------------
2270
2271 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
2272
2273 2. Diagnose condition (?(0) as an error instead of crashing on matching.
2274
2275
2276 Version 3.3 01-Aug-00
2277 ---------------------
2278
2279 1. If an octal character was given, but the value was greater than \377, it
2280 was not getting masked to the least significant bits, as documented. This could
2281 lead to crashes in some systems.
2282
2283 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
2284 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
2285
2286 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
2287 These just pass their arguments on to (pcre_free)(), but they are provided
2288 because some uses of PCRE bind it to non-C systems that can call its functions,
2289 but cannot call free() or pcre_free() directly.
2290
2291 4. Add "make test" as a synonym for "make check". Corrected some comments in
2292 the Makefile.
2293
2294 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
2295 Makefile.
2296
2297 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
2298 command called pgrep for grepping around the active processes.
2299
2300 7. Added the beginnings of support for UTF-8 character strings.
2301
2302 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
2303 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
2304 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
2305 out for the ar command.)
2306
2307
2308 Version 3.2 12-May-00
2309 ---------------------
2310
2311 This is purely a bug fixing release.
2312
2313 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
2314 of ZA. This was just one example of several cases that could provoke this bug,
2315 which was introduced by change 9 of version 2.00. The code for breaking
2316 infinite loops after an iteration that matches an empty string was't working
2317 correctly.
2318
2319 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
2320 when matched against abbab (for example). After matching an empty string, it
2321 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
2322 caused it to match further down the string than it should.
2323
2324 3. The code contained an inclusion of sys/types.h. It isn't clear why this
2325 was there because it doesn't seem to be needed, and it causes trouble on some
2326 systems, as it is not a Standard C header. It has been removed.
2327
2328 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
2329 were reported on the Macintosh. The changes were from
2330
2331 while ((c = *(++ptr)) != 0 && c != '\n');
2332 to
2333 while ((c = *(++ptr)) != 0 && c != '\n') ;
2334
2335 Totally extraordinary, but if that's what it takes...
2336
2337 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
2338 available. Added HAVE_BCOPY and an autoconf test for it; if neither
2339 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
2340 assumes the way PCRE uses memmove() (always moving upwards).
2341
2342 6. PCRE is being used in one environment where strchr() is not available. There
2343 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
2344 faster code anyway.
2345
2346
2347 Version 3.1 09-Feb-00
2348 ---------------------
2349
2350 The only change in this release is the fixing of some bugs in Makefile.in for
2351 the "install" target:
2352
2353 (1) It was failing to install pcreposix.h.
2354
2355 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
2356
2357
2358 Version 3.0 01-Feb-00
2359 ---------------------
2360
2361 1. Add support for the /+ modifier to perltest (to output $` like it does in
2362 pcretest).
2363
2364 2. Add support for the /g modifier to perltest.
2365
2366 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
2367 matches null strings.
2368
2369 4. Fix perltest so that it doesn't do unwanted things when fed an empty
2370 pattern. Perl treats empty patterns specially - it reuses the most recent
2371 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
2372 effect.
2373
2374 5. The POSIX interface was broken in that it was just handing over the POSIX
2375 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
2376 required a bigger vector, with some working space on the end. This means that
2377 the POSIX wrapper now has to get and free some memory, and copy the results.
2378
2379 6. Added some simple autoconf support, placing the test data and the
2380 documentation in separate directories, re-organizing some of the
2381 information files, and making it build pcre-config (a GNU standard). Also added
2382 libtool support for building PCRE as a shared library, which is now the
2383 default.
2384
2385 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
2386 09 are not valid octal constants. Single digits will be used for minor values
2387 less than 10.
2388
2389 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
2390 existing programs that set these in the POSIX interface can use PCRE without
2391 modification.
2392
2393 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
2394 return all that pcre_info() returns, plus additional data. The pcre_info()
2395 function is retained for compatibility, but is considered to be obsolete.
2396
2397 10. Added experimental recursion feature (?R) to handle one common case that
2398 Perl 5.6 will be able to do with (?p{...}).
2399
2400 11. Added support for POSIX character classes like [:alpha:], which Perl is
2401 adopting.
2402
2403
2404 Version 2.08 31-Aug-99
2405 ----------------------
2406
2407 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
2408 trying to match at the startoffset position, but instead was moving forward to
2409 the next newline as if a previous match had failed.
2410
2411 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
2412 and could get into a loop if a null string was matched other than at the start
2413 of the subject.
2414
2415 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
2416 be distinguished at compile time, and for completeness also added PCRE_DATE.
2417
2418 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
2419 in GnuWin32 environments.
2420
2421
2422 Version 2.07 29-Jul-99
2423 ----------------------
2424
2425 1. The documentation is now supplied in plain text form and HTML as well as in
2426 the form of man page sources.
2427
2428 2. C++ compilers don't like assigning (void *) values to other pointer types.
2429 In particular this affects malloc(). Although there is no problem in Standard
2430 C, I've put in casts to keep C++ compilers happy.
2431
2432 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
2433 should be (const char *).
2434
2435 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
2436 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
2437 However, I haven't made this a standard facility. The documentation doesn't
2438 mention it, and the Makefile doesn't support it.
2439
2440 5. The Makefile now contains an "install" target, with editable destinations at
2441 the top of the file. The pcretest program is not installed.
2442
2443 6. pgrep -V now gives the PCRE version number and date.
2444
2445 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
2446 causing the entire string to be ignored, instead of just the last character.
2447
2448 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
2449 non-matching string, it can take a very, very long time, even for strings of
2450 quite modest length, because of the nested recursion. PCRE now does better in
2451 some of these cases. It does this by remembering the last required literal
2452 character in the pattern, and pre-searching the subject to ensure it is present
2453 before running the real match. In other words, it applies a heuristic to detect
2454 some types of certain failure quickly, and in the above example, if presented
2455 with a string that has no trailing " it gives "no match" very quickly.
2456
2457 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
2458 other alternatives are tried instead.
2459
2460
2461 Version 2.06 09-Jun-99
2462 ----------------------
2463
2464 1. Change pcretest's output for amount of store used to show just the code
2465 space, because the remainder (the data block) varies in size between 32-bit and
2466 64-bit systems.
2467
2468 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
2469 start matching at. This allows lookbehinds to work when searching for multiple
2470 occurrences in a string.
2471
2472 3. Added additional options to pcretest for testing multiple occurrences:
2473
2474 /+ outputs the rest of the string that follows a match
2475 /g loops for multiple occurrences, using the new startoffset argument
2476 /G loops for multiple occurrences by passing an incremented pointer
2477
2478 4. PCRE wasn't doing the "first character" optimization for patterns starting
2479 with \b or \B, though it was doing it for other lookbehind assertions. That is,
2480 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
2481 the letter 'x'. On long subject strings, this gives a significant speed-up.
2482
2483
2484 Version 2.05 21-Apr-99
2485 ----------------------
2486
2487 1. Changed the type of magic_number from int to long int so that it works
2488 properly on 16-bit systems.
2489
2490 2. Fixed a bug which caused patterns starting with .* not to work correctly
2491 when the subject string contained newline characters. PCRE was assuming
2492 anchoring for such patterns in all cases, which is not correct because .* will
2493 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
2494 DOTALL is set at top level; otherwise it knows that patterns starting with .*
2495 must be retried after every newline in the subject.
2496
2497
2498 Version 2.04 18-Feb-99
2499 ----------------------
2500
2501 1. For parenthesized subpatterns with repeats whose minimum was zero, the
2502 computation of the store needed to hold the pattern was incorrect (too large).
2503 If such patterns were nested a few deep, this could multiply and become a real
2504 problem.
2505
2506 2. Added /M option to pcretest to show the memory requirement of a specific
2507 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
2508
2509 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
2510 compiled in such a way that the backtracking after subsequent failure was
2511 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
2512 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
2513
2514
2515 Version 2.03 02-Feb-99
2516 ----------------------
2517
2518 1. Fixed typo and small mistake in man page.
2519
2520 2. Added 4th condition (GPL supersedes if conflict) and created separate
2521 LICENCE file containing the conditions.
2522
2523 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
2524 Perl, that is the internal \ allows the delimiter to be included in the
2525 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
2526 the final delimiter, add \ to the end of the pattern (to test the error).
2527
2528 4. Added the convenience functions for extracting substrings after a successful
2529 match. Updated pcretest to make it able to test these functions.
2530
2531
2532 Version 2.02 14-Jan-99
2533 ----------------------
2534
2535 1. Initialized the working variables associated with each extraction so that
2536 their saving and restoring doesn't refer to uninitialized store.
2537
2538 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
2539 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
2540 fix the problem.
2541
2542 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
2543 calls, and wasn't printing the correct value for compiling calls. Increased the
2544 default value of LOOPREPEAT, and the number of significant figures in the
2545 times.
2546
2547 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
2548
2549 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
2550 a building problem on Windows NT with a FAT file system.
2551
2552
2553 Version 2.01 21-Oct-98
2554 ----------------------
2555
2556 1. Changed the API for pcre_compile() to allow for the provision of a pointer
2557 to character tables built by pcre_maketables() in the current locale. If NULL
2558 is passed, the default tables are used.
2559
2560
2561 Version 2.00 24-Sep-98
2562 ----------------------
2563
2564 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
2565 it any more.
2566
2567 2. Allow quantification of (?>) groups, and make it work correctly.
2568
2569 3. The first character computation wasn't working for (?>) groups.
2570
2571 4. Correct the implementation of \Z (it is permitted to match on the \n at the
2572 end of the subject) and add 5.005's \z, which really does match only at the
2573 very end of the subject.
2574
2575 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
2576
2577 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
2578 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
2579 localized options. All options to pcre_study() were also removed.
2580
2581 7. Add other new features from 5.005:
2582
2583 $(?<= positive lookbehind
2584 $(?<! negative lookbehind
2585 (?imsx-imsx) added the unsetting capability
2586 such a setting is global if at outer level; local otherwise
2587 (?imsx-imsx:) non-capturing groups with option setting
2588 (?(cond)re|re) conditional pattern matching
2589
2590 A backreference to itself in a repeated group matches the previous
2591 captured string.
2592
2593 8. General tidying up of studying (both automatic and via "study")
2594 consequential on the addition of new assertions.
2595
2596 9. As in 5.005, unlimited repeated groups that could match an empty substring
2597 are no longer faulted at compile time. Instead, the loop is forcibly broken at
2598 runtime if any iteration does actually match an empty substring.
2599
2600 10. Include the RunTest script in the distribution.
2601
2602 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
2603 discrepancies, some of which were old and were also with respect to 5.004. They
2604 have now been fixed.
2605
2606
2607 Version 1.09 28-Apr-98
2608 ----------------------
2609
2610 1. A negated single character class followed by a quantifier with a minimum
2611 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
2612 program crashes, or just wrong answers. This did not apply to negated classes
2613 containing more than one character, or to minima other than one.
2614
2615
2616 Version 1.08 27-Mar-98
2617 ----------------------
2618
2619 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
2620
2621 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
2622 latter must appear before anything that relies on it in the pattern.
2623
2624
2625 Version 1.07 16-Feb-98
2626 ----------------------
2627
2628 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
2629 repeat of a potentially empty string).
2630
2631
2632 Version 1.06 23-Jan-98
2633 ----------------------
2634
2635 1. Added Markus Oberhumer's little patches for C++.
2636
2637 2. Literal strings longer than 255 characters were broken.
2638
2639
2640 Version 1.05 23-Dec-97
2641 ----------------------
2642
2643 1. Negated character classes containing more than one character were failing if
2644 PCRE_CASELESS was set at run time.
2645
2646
2647 Version 1.04 19-Dec-97
2648 ----------------------
2649
2650 1. Corrected the man page, where some "const" qualifiers had been omitted.
2651
2652 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
2653 input syntax.
2654
2655 3. Fixed memory leak which occurred when a regex with back references was
2656 matched with an offsets vector that wasn't big enough. The temporary memory
2657 that is used in this case wasn't being freed if the match failed.
2658
2659 4. Tidied pcretest to ensure it frees memory that it gets.
2660
2661 5. Temporary memory was being obtained in the case where the passed offsets
2662 vector was exactly big enough.
2663
2664 6. Corrected definition of offsetof() from change 5 below.
2665
2666 7. I had screwed up change 6 below and broken the rules for the use of
2667 setjmp(). Now fixed.
2668
2669
2670 Version 1.03 18-Dec-97
2671 ----------------------
2672
2673 1. A erroneous regex with a missing opening parenthesis was correctly
2674 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
2675 on some systems.
2676
2677 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
2678 it was reported that one broken compiler failed on the former because "code" is
2679 also an independent variable.
2680
2681 3. The erroneous regex a[]b caused an array overrun reference.
2682
2683 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
2684 fail on data ending with that character. (It was going on too far, and checking
2685 the next character, typically a binary zero.) This was specific to the
2686 optimized code for single-character negative classes.
2687
2688 5. Added a contributed patch from the TIN world which does the following:
2689
2690 + Add an undef for memmove, in case the the system defines a macro for it.
2691
2692 + Add a definition of offsetof(), in case there isn't one. (I don't know
2693 the reason behind this - offsetof() is part of the ANSI standard - but
2694 it does no harm).
2695
2696 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
2697 most of the places where whitespace preceded '#'. I have given up and
2698 allowed the remaining 2 cases to be at the margin.
2699
2700 + Rename some variables in pcre to eliminate shadowing. This seems very
2701 pedantic, but does no harm, of course.
2702
2703 6. Moved the call to setjmp() into its own function, to get rid of warnings
2704 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
2705
2706 7. Constructs such as \d{8,} were compiling into the equivalent of
2707 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
2708 outcome, but in this particular case used more store than had been allocated,
2709 which caused the bug to be discovered because it threw up an internal error.
2710
2711 8. The debugging code in both pcre and pcretest for outputting the compiled
2712 form of a regex was going wrong in the case of back references followed by
2713 curly-bracketed repeats.
2714
2715
2716 Version 1.02 12-Dec-97
2717 ----------------------
2718
2719 1. Typos in pcre.3 and comments in the source fixed.
2720
2721 2. Applied a contributed patch to get rid of places where it used to remove
2722 'const' from variables, and fixed some signed/unsigned and uninitialized
2723 variable warnings.
2724
2725 3. Added the "runtest" target to Makefile.
2726
2727 4. Set default compiler flag to -O2 rather than just -O.
2728
2729
2730 Version 1.01 19-Nov-97
2731 ----------------------
2732
2733 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
2734 like /([ab]*)*/, that is, for classes with more than one character in them.
2735
2736 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
2737 as /((?>a*))*/ (a PCRE_EXTRA facility).
2738
2739
2740 Version 1.00 18-Nov-97
2741 ----------------------
2742
2743 1. Added compile-time macros to support systems such as SunOS4 which don't have
2744 memmove() or strerror() but have other things that can be used instead.
2745
2746 2. Arranged that "make clean" removes the executables.
2747
2748
2749 Version 0.99 27-Oct-97
2750 ----------------------
2751
2752 1. Fixed bug in code for optimizing classes with only one character. It was
2753 initializing a 32-byte map regardless, which could cause it to run off the end
2754 of the memory it had got.
2755
2756 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
2757
2758
2759 Version 0.98 22-Oct-97
2760 ----------------------
2761
2762 1. Fixed bug in code for handling temporary memory usage when there are more
2763 back references than supplied space in the ovector. This could cause segfaults.
2764
2765
2766 Version 0.97 21-Oct-97
2767 ----------------------
2768
2769 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
2770
2771 2. Optimized negated single characters not to use a bit map.
2772
2773 3. Brought error texts together as macro definitions; clarified some of them;
2774 fixed one that was wrong - it said "range out of order" when it meant "invalid
2775 escape sequence".
2776
2777 4. Changed some char * arguments to const char *.
2778
2779 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
2780
2781 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
2782 pcretest.
2783
2784
2785 Version 0.96 16-Oct-97
2786 ----------------------
2787
2788 1. Added a simple "pgrep" utility to the distribution.
2789
2790 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
2791 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
2792 where "ddd" means "one or more decimal digits".
2793
2794 3. Fixed serious bug. If a pattern had a back reference, but the call to
2795 pcre_exec() didn't supply a large enough ovector to record the related
2796 identifying subpattern, the match always failed. PCRE now remembers the number
2797 of the largest back reference, and gets some temporary memory in which to save
2798 the offsets during matching if necessary, in order to ensure that
2799 backreferences always work.
2800
2801 4. Increased the compatibility with Perl in a number of ways:
2802
2803 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
2804 to request this handling. The option can be set at compile or exec time.
2805
2806 (b) $ matches before a terminating newline by default; an option
2807 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
2808 mode). The option can be set at compile or exec time.
2809
2810 (c) The handling of \ followed by a digit other than 0 is now supposed to be
2811 the same as Perl's. If the decimal number it represents is less than 10
2812 or there aren't that many previous left capturing parentheses, an octal
2813 escape is read. Inside a character class, it's always an octal escape,
2814 even if it is a single digit.
2815
2816 (d) An escaped but undefined alphabetic character is taken as a literal,
2817 unless PCRE_EXTRA is set. Currently this just reserves the remaining
2818 escapes.
2819
2820 (e) {0} is now permitted. (The previous item is removed from the compiled
2821 pattern).
2822
2823 5. Changed all the names of code files so that the basic parts are no longer
2824 than 10 characters, and abolished the teeny "globals.c" file.
2825
2826 6. Changed the handling of character classes; they are now done with a 32-byte
2827 bit map always.
2828
2829 7. Added the -d and /D options to pcretest to make it possible to look at the
2830 internals of compilation without having to recompile pcre.
2831
2832
2833 Version 0.95 23-Sep-97
2834 ----------------------
2835
2836 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
2837 \x20 at the start of a run of normal characters. These were being treated as
2838 real characters, instead of the source characters being re-checked.
2839
2840
2841 Version 0.94 18-Sep-97
2842 ----------------------
2843
2844 1. The functions are now thread-safe, with the caveat that the global variables
2845 containing pointers to malloc() and free() or alternative functions are the
2846 same for all threads.
2847
2848 2. Get pcre_study() to generate a bitmap of initial characters for non-
2849 anchored patterns when this is possible, and use it if passed to pcre_exec().
2850
2851
2852 Version 0.93 15-Sep-97
2853 ----------------------
2854
2855 1. /(b)|(:+)/ was computing an incorrect first character.
2856
2857 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
2858 but not actually doing anything yet.
2859
2860 3. Treat "-" characters in classes that cannot be part of ranges as literals,
2861 as Perl does (e.g. [-az] or [az-]).
2862
2863 4. Set the anchored flag if a branch starts with .* or .*? because that tests
2864 all possible positions.
2865
2866 5. Split up into different modules to avoid including unneeded functions in a
2867 compiled binary. However, compile and exec are still in one module. The "study"
2868 function is split off.
2869
2870 6. The character tables are now in a separate module whose source is generated
2871 by an auxiliary program - but can then be edited by hand if required. There are
2872 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
2873 toupper() in the code.
2874
2875 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
2876 make them global. Abolish the function for setting them, as the caller can now
2877 set them directly.
2878
2879
2880 Version 0.92 11-Sep-97
2881 ----------------------
2882
2883 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
2884 (e.g. /a{1,3}/) was broken (I mis-optimized it).
2885
2886 2. Caseless matching was not working in character classes if the characters in
2887 the pattern were in upper case.
2888
2889 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
2890
2891 4. Make PCRE_ANCHORED public and accept as a compile option.
2892
2893 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
2894 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
2895 pass them.
2896
2897 6. Give an error if bad option bits passed at compile or run time.
2898
2899 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
2900 pcretest to cause it to pass that flag.
2901
2902 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
2903 options, and the first character, if set.
2904
2905 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
2906
2907
2908 Version 0.91 10-Sep-97
2909 ----------------------
2910
2911 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
2912 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
2913
2914 2. PCRE was looping on encountering an indefinitely repeated back reference to
2915 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
2916 Perl does - treats the match as successful.
2917
2918 ****

Properties

Name Value
svn:eol-style native
svn:keywords "Author Date Id Revision Url"

webmaster@exim.org
ViewVC Help
Powered by ViewVC 1.1.12