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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 215 - (show annotations) (download)
Wed Aug 15 14:20:05 2007 UTC (7 years ago) by ph10
File size: 135799 byte(s)
Fixed overrun for missing ] with a forward reference, e.g. /(?1)\c[/.

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

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