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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 393 - (show annotations) (download)
Tue Mar 17 21:47:54 2009 UTC (5 years, 7 months ago) by ph10
File size: 162918 byte(s)
Force SUPPORT_UTF8 if SUPPORT_UCP is set (for non-"configure" users).

1 ChangeLog for PCRE
2 ------------------
3
4 Version 7.9 xx-xxx-09
5 ---------------------
6
7 1. When building with support for bzlib/zlib (pcregrep) and/or readline
8 (pcretest), all targets were linked against these libraries. This included
9 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
10 libraries. This caused unwanted dependencies to be created. This problem
11 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
12 pcretest is linked with readline.
13
14 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
15 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
16 moved outside it again, because FALSE and TRUE are already defined in AIX,
17 but BOOL is not.
18
19 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
20 PCRE_MATCH_LIMIT_RETURSION values as ints, when they should be long ints.
21
22 4. The pcregrep documentation said spaces were inserted as well as colons (or
23 hyphens) following file names and line numbers when outputting matching
24 lines. This is not true; no spaces are inserted. I have also clarified the
25 wording for the --colour (or --color) option.
26
27 5. In pcregrep, when --colour was used with -o, the list of matching strings
28 was not coloured; this is different to GNU grep, so I have changed it to be
29 the same.
30
31 6. When --colo(u)r was used in pcregrep, only the first matching substring in
32 each matching line was coloured. Now it goes on to look for further matches
33 of any of the test patterns, which is the same behaviour as GNU grep.
34
35 7. A pattern that could match an empty string could cause pcregrep to loop; it
36 doesn't make sense to accept an empty string match in pcregrep, so I have
37 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
38 seems to be how GNU grep behaves.
39
40 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
41 start or after a newline", because the conditional assertion was not being
42 correctly handled. The rule now is that both the assertion and what follows
43 in the first alternative must satisfy the test.
44
45 9. If auto-callout was enabled in a pattern with a conditional group, PCRE
46 could crash during matching.
47
48 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
49 used for matching.
50
51 11. Unicode property support in character classes was not working for
52 characters (bytes) greater than 127 when not in UTF-8 mode.
53
54 12. Added the -M command line option to pcretest.
55
56 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
57
58 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
59
60 16. Added comments and documentation about mis-use of no_arg in the C++
61 wrapper.
62
63 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
64 from Martin Jerabek that uses macro names for all relevant character and
65 string constants.
66
67 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
68 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
69 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
70 these, but not everybody uses configure.
71
72
73 Version 7.8 05-Sep-08
74 ---------------------
75
76 1. Replaced UCP searching code with optimized version as implemented for Ad
77 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
78 stage table and inline lookup instead of a function, giving speed ups of 2
79 to 5 times on some simple patterns that I tested. Permission was given to
80 distribute the MultiStage2.py script that generates the tables (it's not in
81 the tarball, but is in the Subversion repository).
82
83 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
84 scripts.
85
86 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
87 a group with a zero qualifier. The result of the study could be incorrect,
88 or the function might crash, depending on the pattern.
89
90 4. Caseless matching was not working for non-ASCII characters in back
91 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
92 It now works when Unicode Property Support is available.
93
94 5. In pcretest, an escape such as \x{de} in the data was always generating
95 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
96 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
97 truncation.
98
99 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
100
101 7. Added two (int) casts to pcregrep when printing the difference of two
102 pointers, in case they are 64-bit values.
103
104 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
105 test 2 if it fails.
106
107 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
108 and a #define of that name to empty if it is not externally set. This is to
109 allow users of MSVC to set it if necessary.
110
111 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
112 the convenience functions in the pcre_get.c source file.
113
114 11. An option change at the start of a pattern that had top-level alternatives
115 could cause overwriting and/or a crash. This command provoked a crash in
116 some environments:
117
118 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
119
120 This potential security problem was recorded as CVE-2008-2371.
121
122 12. For a pattern where the match had to start at the beginning or immediately
123 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
124 pcre_dfa_exec() could read past the end of the passed subject if there was
125 no match. To help with detecting such bugs (e.g. with valgrind), I modified
126 pcretest so that it places the subject at the end of its malloc-ed buffer.
127
128 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
129 exec() might read past the end of the data buffer in UTF-8 mode.
130
131 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
132 the data contained the byte 0x85 as part of a UTF-8 character within its
133 first line. This applied both to normal and DFA matching.
134
135 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
136 /^[^d]*?$/8 failed to match "abc".
137
138 16. Added a missing copyright notice to pcrecpp_internal.h.
139
140 17. Make it more clear in the documentation that values returned from
141 pcre_exec() in ovector are byte offsets, not character counts.
142
143 18. Tidied a few places to stop certain compilers from issuing warnings.
144
145 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
146 supplied by Stefan Weber. I made a further small update for 7.8 because
147 there is a change of source arrangements: the pcre_searchfuncs.c module is
148 replaced by pcre_ucd.c.
149
150
151 Version 7.7 07-May-08
152 ---------------------
153
154 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
155 a string to a long long, pretend we don't even have a long long." This is
156 done by checking for the strtoq, strtoll, and _strtoi64 functions.
157
158 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
159 pre-7.6 versions, which defined a global no_arg variable instead of putting
160 it in the RE class. (See also #8 below.)
161
162 3. Remove a line of dead code, identified by coverity and reported by Nuno
163 Lopes.
164
165 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
166
167 (1) The include/exclude patterns were being applied to the whole pathnames
168 of files, instead of just to the final components.
169
170 (2) If there was more than one level of directory, the subdirectories were
171 skipped unless they satisfied the include/exclude conditions. This is
172 inconsistent with GNU grep (and could even be seen as contrary to the
173 pcregrep specification - which I improved to make it absolutely clear).
174 The action now is always to scan all levels of directory, and just
175 apply the include/exclude patterns to regular files.
176
177 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
178 --exclude_dir in the tests to avoid scanning .svn directories.
179
180 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
181 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
182 doesn't support NULs in patterns.
183
184 7. Added some missing "const"s to declarations of static tables in
185 pcre_compile.c and pcre_dfa_exec.c.
186
187 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
188 caused by fix #2 above. (Subsequently also a second patch to fix the
189 first patch. And a third patch - this was a messy problem.)
190
191 9. Applied Craig's patch to remove the use of push_back().
192
193 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
194 matching function regexec().
195
196 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
197 which, however, unlike Perl's \g{...}, are subroutine calls, not back
198 references. PCRE supports relative numbers with this syntax (I don't think
199 Oniguruma does).
200
201 12. Previously, a group with a zero repeat such as (...){0} was completely
202 omitted from the compiled regex. However, this means that if the group
203 was called as a subroutine from elsewhere in the pattern, things went wrong
204 (an internal error was given). Such groups are now left in the compiled
205 pattern, with a new opcode that causes them to be skipped at execution
206 time.
207
208 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
209 to the way PCRE behaves:
210
211 (a) A lone ] character is dis-allowed (Perl treats it as data).
212
213 (b) A back reference to an unmatched subpattern matches an empty string
214 (Perl fails the current match path).
215
216 (c) A data ] in a character class must be notated as \] because if the
217 first data character in a class is ], it defines an empty class. (In
218 Perl it is not possible to have an empty class.) The empty class []
219 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
220 The negative empty class [^] matches any one character, independently
221 of the DOTALL setting.
222
223 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
224 non-existent subpattern following a character class starting with ']' and
225 containing () gave an internal compiling error instead of "reference to
226 non-existent subpattern". Fortunately, when the pattern did exist, the
227 compiled code was correct. (When scanning forwards to check for the
228 existencd of the subpattern, it was treating the data ']' as terminating
229 the class, so got the count wrong. When actually compiling, the reference
230 was subsequently set up correctly.)
231
232 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
233 it was being rejected as not supported by pcre_dfa_exec(), even though
234 other assertions are supported. I have made pcre_dfa_exec() support
235 (*FAIL).
236
237 16. The implementation of 13c above involved the invention of a new opcode,
238 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
239 cannot be changed at match time, I realized I could make a small
240 improvement to matching performance by compiling OP_ALLANY instead of
241 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
242 on the OP_ANY path.
243
244 17. Compiling pcretest on Windows with readline support failed without the
245 following two fixes: (1) Make the unistd.h include conditional on
246 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
247
248 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
249 ncurses library to be included for pcretest when ReadLine support is
250 requested, but also to allow for it to be overridden. This patch came from
251 Daniel Bergström.
252
253 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
254 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
255 any errors with the current Unicode tables. Thanks to Peter Kankowski for
256 spotting this.
257
258
259 Version 7.6 28-Jan-08
260 ---------------------
261
262 1. A character class containing a very large number of characters with
263 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
264 overflow.
265
266 2. Patch to cut out the "long long" test in pcrecpp_unittest when
267 HAVE_LONG_LONG is not defined.
268
269 3. Applied Christian Ehrlicher's patch to update the CMake build files to
270 bring them up to date and include new features. This patch includes:
271
272 - Fixed PH's badly added libz and libbz2 support.
273 - Fixed a problem with static linking.
274 - Added pcredemo. [But later removed - see 7 below.]
275 - Fixed dftables problem and added an option.
276 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
277 HAVE_LONG_LONG.
278 - Added readline support for pcretest.
279 - Added an listing of the option settings after cmake has run.
280
281 4. A user submitted a patch to Makefile that makes it easy to create
282 "pcre.dll" under mingw when using Configure/Make. I added stuff to
283 Makefile.am that cause it to include this special target, without
284 affecting anything else. Note that the same mingw target plus all
285 the other distribution libraries and programs are now supported
286 when configuring with CMake (see 6 below) instead of with
287 Configure/Make.
288
289 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
290 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
291 exported in the Windows port". It has not yet been confirmed that the patch
292 solves the problem, but it does no harm.
293
294 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
295 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
296 with CMake, and also correct the comment about stack recursion.
297
298 7. Remove the automatic building of pcredemo from the ./configure system and
299 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
300 of a program that users should build themselves after PCRE is installed, so
301 building it automatically is not really right. What is more, it gave
302 trouble in some build environments.
303
304 8. Further tidies to CMakeLists.txt from Sheri and Christian.
305
306
307 Version 7.5 10-Jan-08
308 ---------------------
309
310 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
311 values in parens when parsing an RE using the C++ wrapper."
312
313 2. Negative specials like \S did not work in character classes in UTF-8 mode.
314 Characters greater than 255 were excluded from the class instead of being
315 included.
316
317 3. The same bug as (2) above applied to negated POSIX classes such as
318 [:^space:].
319
320 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
321 defined or documented. It seems to have been a typo for PCRE_STATIC, so
322 I have changed it.
323
324 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
325 first named subpattern) and a construct such as (?&a) would reference the
326 first named subpattern whose name started with "a" (in other words, the
327 length check was missing). Both these problems are fixed. "Subpattern name
328 expected" is now given for (?&) (a zero-length name), and this patch also
329 makes it give the same error for \k'' (previously it complained that that
330 was a reference to a non-existent subpattern).
331
332 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
333 this is right because (?- can be followed by option settings as well as by
334 digits. I have, however, made the messages clearer.
335
336 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
337 than the number used in the conditional) now cause a compile-time error.
338 This is actually not compatible with Perl, which accepts such patterns, but
339 treats the conditional as always being FALSE (as PCRE used to), but it
340 seems to me that giving a diagnostic is better.
341
342 8. Change "alphameric" to the more common word "alphanumeric" in comments
343 and messages.
344
345 9. Fix two occurrences of "backslash" in comments that should have been
346 "backspace".
347
348 10. Remove two redundant lines of code that can never be obeyed (their function
349 was moved elsewhere).
350
351 11. The program that makes PCRE's Unicode character property table had a bug
352 which caused it to generate incorrect table entries for sequences of
353 characters that have the same character type, but are in different scripts.
354 It amalgamated them into a single range, with the script of the first of
355 them. In other words, some characters were in the wrong script. There were
356 thirteen such cases, affecting characters in the following ranges:
357
358 U+002b0 - U+002c1
359 U+0060c - U+0060d
360 U+0061e - U+00612
361 U+0064b - U+0065e
362 U+0074d - U+0076d
363 U+01800 - U+01805
364 U+01d00 - U+01d77
365 U+01d9b - U+01dbf
366 U+0200b - U+0200f
367 U+030fc - U+030fe
368 U+03260 - U+0327f
369 U+0fb46 - U+0fbb1
370 U+10450 - U+1049d
371
372 12. The -o option (show only the matching part of a line) for pcregrep was not
373 compatible with GNU grep in that, if there was more than one match in a
374 line, it showed only the first of them. It now behaves in the same way as
375 GNU grep.
376
377 13. If the -o and -v options were combined for pcregrep, it printed a blank
378 line for every non-matching line. GNU grep prints nothing, and pcregrep now
379 does the same. The return code can be used to tell if there were any
380 non-matching lines.
381
382 14. Added --file-offsets and --line-offsets to pcregrep.
383
384 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
385 infinitely looping recursion. The bug was that positive lookaheads were not
386 being skipped when checking for a possible empty match (negative lookaheads
387 and both kinds of lookbehind were skipped).
388
389 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
390 inclusion of <windows.h> to before rather than after the definition of
391 INVALID_FILE_ATTRIBUTES (patch from David Byron).
392
393 17. Specifying a possessive quantifier with a specific limit for a Unicode
394 character property caused pcre_compile() to compile bad code, which led at
395 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
396 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
397 caused the error; without that there was no problem.
398
399 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
400
401 19. Added --enable-pcretest-libreadline.
402
403 20. In pcrecpp.cc, the variable 'count' was incremented twice in
404 RE::GlobalReplace(). As a result, the number of replacements returned was
405 double what it should be. I removed one of the increments, but Craig sent a
406 later patch that removed the other one (the right fix) and added unit tests
407 that check the return values (which was not done before).
408
409 21. Several CMake things:
410
411 (1) Arranged that, when cmake is used on Unix, the libraries end up with
412 the names libpcre and libpcreposix, not just pcre and pcreposix.
413
414 (2) The above change means that pcretest and pcregrep are now correctly
415 linked with the newly-built libraries, not previously installed ones.
416
417 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
418
419 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
420 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
421 UTF-8 newline character). The key issue is that the pattern starts .*;
422 this means that the match must be either at the beginning, or after a
423 newline. The bug was in the code for advancing after a failed match and
424 checking that the new position followed a newline. It was not taking
425 account of UTF-8 characters correctly.
426
427 23. PCRE was behaving differently from Perl in the way it recognized POSIX
428 character classes. PCRE was not treating the sequence [:...:] as a
429 character class unless the ... were all letters. Perl, however, seems to
430 allow any characters between [: and :], though of course it rejects as
431 unknown any "names" that contain non-letters, because all the known class
432 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
433 for example, whereas PCRE did not - it did not recognize a POSIX character
434 class. This seemed a bit dangerous, so the code has been changed to be
435 closer to Perl. The behaviour is not identical to Perl, because PCRE will
436 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
437 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
438 Perl does, and where it didn't before.
439
440 24. Rewrite so as to remove the single use of %n from pcregrep because in some
441 Windows environments %n is disabled by default.
442
443
444 Version 7.4 21-Sep-07
445 ---------------------
446
447 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
448 means that a class such as [\s] counted as "explicit reference to CR or
449 LF". That isn't really right - the whole point of the change was to try to
450 help when there was an actual mention of one of the two characters. So now
451 the change happens only if \r or \n (or a literal CR or LF) character is
452 encountered.
453
454 2. The 32-bit options word was also used for 6 internal flags, but the numbers
455 of both had grown to the point where there were only 3 bits left.
456 Fortunately, there was spare space in the data structure, and so I have
457 moved the internal flags into a new 16-bit field to free up more option
458 bits.
459
460 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
461 but did not set the internal JCHANGED flag - either of these is enough to
462 control the way the "get" function works - but the PCRE_INFO_JCHANGED
463 facility is supposed to tell if (?J) was ever used, so now (?J) at the
464 start sets both bits.
465
466 4. Added options (at build time, compile time, exec time) to change \R from
467 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
468
469 5. doc/pcresyntax.html was missing from the distribution.
470
471 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
472 compatibility, even though it is no longer used.
473
474 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
475 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
476 windows.h file is present (where different names are used). [This was
477 reversed later after testing - see 16 below.]
478
479 8. Changed all #include <config.h> to #include "config.h". There were also
480 some further <pcre.h> cases that I changed to "pcre.h".
481
482 9. When pcregrep was used with the --colour option, it missed the line ending
483 sequence off the lines that it output.
484
485 10. It was pointed out to me that arrays of string pointers cause lots of
486 relocations when a shared library is dynamically loaded. A technique of
487 using a single long string with a table of offsets can drastically reduce
488 these. I have refactored PCRE in four places to do this. The result is
489 dramatic:
490
491 Originally: 290
492 After changing UCP table: 187
493 After changing error message table: 43
494 After changing table of "verbs" 36
495 After changing table of Posix names 22
496
497 Thanks to the folks working on Gregex for glib for this insight.
498
499 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
500 unicode-properties was also set.
501
502 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
503
504 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
505 checked only for CRLF.
506
507 14. Added casts to pcretest.c to avoid compiler warnings.
508
509 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
510
511 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
512 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
513 entirely. This removes changes made in 7 above.
514
515 17. The CMake files have been updated, and there is now more information about
516 building with CMake in the NON-UNIX-USE document.
517
518
519 Version 7.3 28-Aug-07
520 ---------------------
521
522 1. In the rejigging of the build system that eventually resulted in 7.1, the
523 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
524 brackets there is not right, since it causes compilers to look for an
525 installed pcre.h, not the version that is in the source that is being
526 compiled (which of course may be different). I have changed it back to:
527
528 #include "pcre.h"
529
530 I have a vague recollection that the change was concerned with compiling in
531 different directories, but in the new build system, that is taken care of
532 by the VPATH setting the Makefile.
533
534 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
535 when the subject happened to end in the byte 0x85 (e.g. if the last
536 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
537 characters but of course it shouldn't be taken as a newline when it is part
538 of another character. The bug was that, for an unlimited repeat of . in
539 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
540 characters when looking for a newline.
541
542 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
543
544 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
545 in debug output.
546
547 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
548 long printing in the pcrecpp unittest when running under MinGW.
549
550 6. ESC_K was left out of the EBCDIC table.
551
552 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
553 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
554 limit also applies to "virtual nesting" when a pattern is recursive, and in
555 this case 1000 isn't so big. I have been able to remove this limit at the
556 expense of backing off one optimization in certain circumstances. Normally,
557 when pcre_exec() would call its internal match() function recursively and
558 immediately return the result unconditionally, it uses a "tail recursion"
559 feature to save stack. However, when a subpattern that can match an empty
560 string has an unlimited repetition quantifier, it no longer makes this
561 optimization. That gives it a stack frame in which to save the data for
562 checking that an empty string has been matched. Previously this was taken
563 from the 1000-entry workspace that had been reserved. So now there is no
564 explicit limit, but more stack is used.
565
566 8. Applied Daniel's patches to solve problems with the import/export magic
567 syntax that is required for Windows, and which was going wrong for the
568 pcreposix and pcrecpp parts of the library. These were overlooked when this
569 problem was solved for the main library.
570
571 9. There were some crude static tests to avoid integer overflow when computing
572 the size of patterns that contain repeated groups with explicit upper
573 limits. As the maximum quantifier is 65535, the maximum group length was
574 set at 30,000 so that the product of these two numbers did not overflow a
575 32-bit integer. However, it turns out that people want to use groups that
576 are longer than 30,000 bytes (though not repeat them that many times).
577 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
578 made it possible to implement the integer overflow checks in a much more
579 dynamic way, which I have now done. The artificial limitation on group
580 length has been removed - we now have only the limit on the total length of
581 the compiled pattern, which depends on the LINK_SIZE setting.
582
583 10. Fixed a bug in the documentation for get/copy named substring when
584 duplicate names are permitted. If none of the named substrings are set, the
585 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
586 empty string.
587
588 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
589 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
590 because the ] is interpreted as the first data character and the
591 terminating ] is not found. PCRE has been made compatible with Perl in this
592 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
593 cause memory overwriting.
594
595 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
596 string has been matched (to stop an infinite loop). It was not recognizing
597 a conditional subpattern that could match an empty string if that
598 subpattern was within another subpattern. For example, it looped when
599 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
600 condition was not nested. This bug has been fixed.
601
602 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
603 past the start of the subject in the presence of bytes with the top bit
604 set, for example "\x8aBCD".
605
606 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
607 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
608
609 14. Optimized (?!) to (*FAIL).
610
611 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
612 This restricts code points to be within the range 0 to 0x10FFFF, excluding
613 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
614 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
615 does: it's just the validity check that is more restrictive.
616
617 16. Inserted checks for integer overflows during escape sequence (backslash)
618 processing, and also fixed erroneous offset values for syntax errors during
619 backslash processing.
620
621 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
622 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
623
624 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
625 caused an overrun.
626
627 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
628 something other than just ASCII characters) inside a group that had an
629 unlimited repeat caused a loop at compile time (while checking to see
630 whether the group could match an empty string).
631
632 20. Debugging a pattern containing \p or \P could cause a crash. For example,
633 [\P{Any}] did so. (Error in the code for printing property names.)
634
635 21. An orphan \E inside a character class could cause a crash.
636
637 22. A repeated capturing bracket such as (A)? could cause a wild memory
638 reference during compilation.
639
640 23. There are several functions in pcre_compile() that scan along a compiled
641 expression for various reasons (e.g. to see if it's fixed length for look
642 behind). There were bugs in these functions when a repeated \p or \P was
643 present in the pattern. These operators have additional parameters compared
644 with \d, etc, and these were not being taken into account when moving along
645 the compiled data. Specifically:
646
647 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
648 length.
649
650 (b) An item such as \pL+ within a repeated group could cause crashes or
651 loops.
652
653 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
654 "reference to non-existent subpattern" error.
655
656 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
657
658 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
659 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
660
661 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
662
663 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
664 character were causing crashes (broken optimization).
665
666 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
667 \p or \P) caused a compile-time loop.
668
669 28. More problems have arisen in unanchored patterns when CRLF is a valid line
670 break. For example, the unstudied pattern [\r\n]A does not match the string
671 "\r\nA" because change 7.0/46 below moves the current point on by two
672 characters after failing to match at the start. However, the pattern \nA
673 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
674 the same is true. There doesn't seem any very clean way out of this, but
675 what I have chosen to do makes the common cases work: PCRE now takes note
676 of whether there can be an explicit match for \r or \n anywhere in the
677 pattern, and if so, 7.0/46 no longer applies. As part of this change,
678 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
679 pattern has explicit CR or LF references.
680
681 29. Added (*CR) etc for changing newline setting at start of pattern.
682
683
684 Version 7.2 19-Jun-07
685 ---------------------
686
687 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
688 which is apparently normally available under Windows.
689
690 2. Re-jig the pcregrep tests with different newline settings in an attempt
691 to make them independent of the local environment's newline setting.
692
693 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
694
695 4. Some of the "internals" tests were previously cut out when the link size
696 was not 2, because the output contained actual offsets. The recent new
697 "Z" feature of pcretest means that these can be cut out, making the tests
698 usable with all link sizes.
699
700 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
701 stack recursion. This gives a massive performance boost under BSD, but just
702 a small improvement under Linux. However, it saves one field in the frame
703 in all cases.
704
705 6. Added more features from the forthcoming Perl 5.10:
706
707 (a) (?-n) (where n is a string of digits) is a relative subroutine or
708 recursion call. It refers to the nth most recently opened parentheses.
709
710 (b) (?+n) is also a relative subroutine call; it refers to the nth next
711 to be opened parentheses.
712
713 (c) Conditions that refer to capturing parentheses can be specified
714 relatively, for example, (?(-2)... or (?(+3)...
715
716 (d) \K resets the start of the current match so that everything before
717 is not part of it.
718
719 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
720
721 (f) \g{name} is another synonym - part of Perl 5.10's unification of
722 reference syntax.
723
724 (g) (?| introduces a group in which the numbering of parentheses in each
725 alternative starts with the same number.
726
727 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
728
729 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
730 PCRE_INFO_JCHANGED.
731
732 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
733 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
734 for detecting groups that can match an empty string.
735
736 9. A pattern with a very large number of alternatives (more than several
737 hundred) was running out of internal workspace during the pre-compile
738 phase, where pcre_compile() figures out how much memory will be needed. A
739 bit of new cunning has reduced the workspace needed for groups with
740 alternatives. The 1000-alternative test pattern now uses 12 bytes of
741 workspace instead of running out of the 4096 that are available.
742
743 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
744
745 11. Applied patch from Google to remove an optimization that didn't quite work.
746 The report of the bug said:
747
748 pcrecpp::RE("a*").FullMatch("aaa") matches, while
749 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
750 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
751
752 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
753 it matched the wrong number of bytes.
754
755
756 Version 7.1 24-Apr-07
757 ---------------------
758
759 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
760 that is more "standard", making use of automake and other Autotools. There
761 is some re-arrangement of the files and adjustment of comments consequent
762 on this.
763
764 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
765 for recursive directory scanning broke on some systems because the files
766 are not scanned in any specific order and on different systems the order
767 was different. A call to "sort" has been inserted into RunGrepTest for the
768 approprate test as a short-term fix. In the longer term there may be an
769 alternative.
770
771 3. I had an email from Eric Raymond about problems translating some of PCRE's
772 man pages to HTML (despite the fact that I distribute HTML pages, some
773 people do their own conversions for various reasons). The problems
774 concerned the use of low-level troff macros .br and .in. I have therefore
775 removed all such uses from the man pages (some were redundant, some could
776 be replaced by .nf/.fi pairs). The 132html script that I use to generate
777 HTML has been updated to handle .nf/.fi and to complain if it encounters
778 .br or .in.
779
780 4. Updated comments in configure.ac that get placed in config.h.in and also
781 arranged for config.h to be included in the distribution, with the name
782 config.h.generic, for the benefit of those who have to compile without
783 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
784
785 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
786 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
787 makevp.bat for the current PCRE, using the additional files
788 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
789
790 6. A Windows user reported a minor discrepancy with test 2, which turned out
791 to be caused by a trailing space on an input line that had got lost in his
792 copy. The trailing space was an accident, so I've just removed it.
793
794 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
795 that is needed.
796
797 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
798 as "const" (a) because they are and (b) because it helps the PHP
799 maintainers who have recently made a script to detect big data structures
800 in the php code that should be moved to the .rodata section. I remembered
801 to update Builducptable as well, so it won't revert if ucptable.h is ever
802 re-created.
803
804 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
805 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
806 order to be able to cut out the UTF-8 tables in the latter when UTF-8
807 support is not required. This saves 1.5-2K of code, which is important in
808 some applications.
809
810 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
811 so as not to refer to the tables, even though these functions will never be
812 called when UTF-8 support is disabled. Otherwise there are problems with a
813 shared library.
814
815 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
816
817 (a) It was defining its arguments as char * instead of void *.
818
819 (b) It was assuming that all moves were upwards in memory; this was true
820 a long time ago when I wrote it, but is no longer the case.
821
822 The emulated memove() is provided for those environments that have neither
823 memmove() nor bcopy(). I didn't think anyone used it these days, but that
824 is clearly not the case, as these two bugs were recently reported.
825
826 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
827 and Detrail to create the HTML documentation, the .txt form of the man
828 pages, and it removes trailing spaces from listed files. It also creates
829 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
830 case, it wraps all the #defines with #ifndefs. This script should be run
831 before "make dist".
832
833 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
834 with Unicode property support.
835
836 (a) For a maximizing quantifier, if the two different cases of the
837 character were of different lengths in their UTF-8 codings (there are
838 some cases like this - I found 11), and the matching function had to
839 back up over a mixture of the two cases, it incorrectly assumed they
840 were both the same length.
841
842 (b) When PCRE was configured to use the heap rather than the stack for
843 recursion during matching, it was not correctly preserving the data for
844 the other case of a UTF-8 character when checking ahead for a match
845 while processing a minimizing repeat. If the check also involved
846 matching a wide character, but failed, corruption could cause an
847 erroneous result when trying to check for a repeat of the original
848 character.
849
850 13. Some tidying changes to the testing mechanism:
851
852 (a) The RunTest script now detects the internal link size and whether there
853 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
854 values substituted by "configure". (The RunGrepTest script already did
855 this for UTF-8.) The configure.ac script no longer substitutes the
856 relevant variables.
857
858 (b) The debugging options /B and /D in pcretest show the compiled bytecode
859 with length and offset values. This means that the output is different
860 for different internal link sizes. Test 2 is skipped for link sizes
861 other than 2 because of this, bypassing the problem. Unfortunately,
862 there was also a test in test 3 (the locale tests) that used /B and
863 failed for link sizes other than 2. Rather than cut the whole test out,
864 I have added a new /Z option to pcretest that replaces the length and
865 offset values with spaces. This is now used to make test 3 independent
866 of link size. (Test 2 will be tidied up later.)
867
868 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
869 segmentation fault instead of returning the appropriate error message.
870
871 15. In multiline mode when the newline sequence was set to "any", the pattern
872 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
873 This doesn't seem right; it now treats the CRLF combination as the line
874 ending, and so does not match in that case. It's only a pattern such as ^$
875 that would hit this one: something like ^ABC$ would have failed after \r
876 and then tried again after \r\n.
877
878 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
879 in an attempt to make files that differ only in their line terminators
880 compare equal. This works on Linux.
881
882 17. Under certain error circumstances pcregrep might try to free random memory
883 as it exited. This is now fixed, thanks to valgrind.
884
885 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
886 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
887 was because its rules for how to advance for /g after matching an empty
888 string at the end of a line did not allow for this case. They now check for
889 it specially.
890
891 20. pcretest is supposed to handle patterns and data of any length, by
892 extending its buffers when necessary. It was getting this wrong when the
893 buffer for a data line had to be extended.
894
895 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
896 CRLF as a newline sequence.
897
898 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
899 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
900 I have nevertheless tidied it up.
901
902 23. Added some casts to kill warnings from HP-UX ia64 compiler.
903
904 24. Added a man page for pcre-config.
905
906
907 Version 7.0 19-Dec-06
908 ---------------------
909
910 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
911 moving to gcc 4.1.1.
912
913 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
914 sys/time.h, which is documented as needed for this function. It doesn't
915 seem to matter on Linux, but it showed up on some releases of OS X.
916
917 3. It seems that there are systems where bytes whose values are greater than
918 127 match isprint() in the "C" locale. The "C" locale should be the
919 default when a C program starts up. In most systems, only ASCII printing
920 characters match isprint(). This difference caused the output from pcretest
921 to vary, making some of the tests fail. I have changed pcretest so that:
922
923 (a) When it is outputting text in the compiled version of a pattern, bytes
924 other than 32-126 are always shown as hex escapes.
925
926 (b) When it is outputting text that is a matched part of a subject string,
927 it does the same, unless a different locale has been set for the match
928 (using the /L modifier). In this case, it uses isprint() to decide.
929
930 4. Fixed a major bug that caused incorrect computation of the amount of memory
931 required for a compiled pattern when options that changed within the
932 pattern affected the logic of the preliminary scan that determines the
933 length. The relevant options are -x, and -i in UTF-8 mode. The result was
934 that the computed length was too small. The symptoms of this bug were
935 either the PCRE error "internal error: code overflow" from pcre_compile(),
936 or a glibc crash with a message such as "pcretest: free(): invalid next
937 size (fast)". Examples of patterns that provoked this bug (shown in
938 pcretest format) are:
939
940 /(?-x: )/x
941 /(?x)(?-x: \s*#\s*)/
942 /((?i)[\x{c0}])/8
943 /(?i:[\x{c0}])/8
944
945 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
946 is now done differently.
947
948 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
949 wrapper classes; (b) implement a new function in the C++ scanner that is
950 more efficient than the old way of doing things because it avoids levels of
951 recursion in the regex matching; (c) add a paragraph to the documentation
952 for the FullMatch() function.
953
954 6. The escape sequence \n was being treated as whatever was defined as
955 "newline". Not only was this contrary to the documentation, which states
956 that \n is character 10 (hex 0A), but it also went horribly wrong when
957 "newline" was defined as CRLF. This has been fixed.
958
959 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
960 was being set to -1 for the "end of line" case (supposedly a value that no
961 character can have). Though this value is never used (the check for end of
962 line is "zero bytes in current character"), it caused compiler complaints.
963 I've changed it to 0xffffffff.
964
965 8. In pcre_version.c, the version string was being built by a sequence of
966 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
967 string (as it is for production releases) called a macro with an empty
968 argument. The C standard says the result of this is undefined. The gcc
969 compiler treats it as an empty string (which was what was wanted) but it is
970 reported that Visual C gives an error. The source has been hacked around to
971 avoid this problem.
972
973 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
974 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
975 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
976 of them did).
977
978 10. Originally, pcretest opened its input and output without "b"; then I was
979 told that "b" was needed in some environments, so it was added for release
980 5.0 to both the input and output. (It makes no difference on Unix-like
981 systems.) Later I was told that it is wrong for the input on Windows. I've
982 now abstracted the modes into two macros, to make it easier to fiddle with
983 them, and removed "b" from the input mode under Windows.
984
985 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
986
987 12. Added -help and --help to pcretest as an official way of being reminded
988 of the options.
989
990 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
991 and pcrecpp.cc because they annoy compilers at high warning levels.
992
993 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
994
995 15. Fixed an occurrence of == in configure.ac that should have been = (shell
996 scripts are not C programs :-) and which was not noticed because it works
997 on Linux.
998
999 16. pcretest is supposed to handle any length of pattern and data line (as one
1000 line or as a continued sequence of lines) by extending its input buffer if
1001 necessary. This feature was broken for very long pattern lines, leading to
1002 a string of junk being passed to pcre_compile() if the pattern was longer
1003 than about 50K.
1004
1005 17. I have done a major re-factoring of the way pcre_compile() computes the
1006 amount of memory needed for a compiled pattern. Previously, there was code
1007 that made a preliminary scan of the pattern in order to do this. That was
1008 OK when PCRE was new, but as the facilities have expanded, it has become
1009 harder and harder to keep it in step with the real compile phase, and there
1010 have been a number of bugs (see for example, 4 above). I have now found a
1011 cunning way of running the real compile function in a "fake" mode that
1012 enables it to compute how much memory it would need, while actually only
1013 ever using a few hundred bytes of working memory and without too many
1014 tests of the mode. This should make future maintenance and development
1015 easier. A side effect of this work is that the limit of 200 on the nesting
1016 depth of parentheses has been removed (though this was never a serious
1017 limitation, I suspect). However, there is a downside: pcre_compile() now
1018 runs more slowly than before (30% or more, depending on the pattern). I
1019 hope this isn't a big issue. There is no effect on runtime performance.
1020
1021 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
1022 newline (only possible for the last line of a file) and it was a
1023 pattern that set a locale (followed by /Lsomething), pcretest crashed.
1024
1025 19. Added additional timing features to pcretest. (1) The -tm option now times
1026 matching only, not compiling. (2) Both -t and -tm can be followed, as a
1027 separate command line item, by a number that specifies the number of
1028 repeats to use when timing. The default is 50000; this gives better
1029 precision, but takes uncomfortably long for very large patterns.
1030
1031 20. Extended pcre_study() to be more clever in cases where a branch of a
1032 subpattern has no definite first character. For example, (a*|b*)[cd] would
1033 previously give no result from pcre_study(). Now it recognizes that the
1034 first character must be a, b, c, or d.
1035
1036 21. There was an incorrect error "recursive call could loop indefinitely" if
1037 a subpattern (or the entire pattern) that was being tested for matching an
1038 empty string contained only one non-empty item after a nested subpattern.
1039 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
1040 incorrectly, because the \d was being skipped in the check.
1041
1042 22. The pcretest program now has a new pattern option /B and a command line
1043 option -b, which is equivalent to adding /B to every pattern. This causes
1044 it to show the compiled bytecode, without the additional information that
1045 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
1046 is the same as /B/I).
1047
1048 23. A new optimization is now able automatically to treat some sequences such
1049 as a*b as a*+b. More specifically, if something simple (such as a character
1050 or a simple class like \d) has an unlimited quantifier, and is followed by
1051 something that cannot possibly match the quantified thing, the quantifier
1052 is automatically "possessified".
1053
1054 24. A recursive reference to a subpattern whose number was greater than 39
1055 went wrong under certain circumstances in UTF-8 mode. This bug could also
1056 have affected the operation of pcre_study().
1057
1058 25. Realized that a little bit of performance could be had by replacing
1059 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
1060
1061 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
1062
1063 27. Possessive quantifiers such as a++ were previously implemented by turning
1064 them into atomic groups such as ($>a+). Now they have their own opcodes,
1065 which improves performance. This includes the automatically created ones
1066 from 23 above.
1067
1068 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
1069 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
1070 the first matched character to be a colon. This applied both to named and
1071 numbered groups.
1072
1073 29. The ucpinternal.h header file was missing its idempotency #ifdef.
1074
1075 30. I was sent a "project" file called libpcre.a.dev which I understand makes
1076 building PCRE on Windows easier, so I have included it in the distribution.
1077
1078 31. There is now a check in pcretest against a ridiculously large number being
1079 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
1080 loop, the loop is abandoned.
1081
1082 32. Forward references to subpatterns in conditions such as (?(2)...) where
1083 subpattern 2 is defined later cause pcre_compile() to search forwards in
1084 the pattern for the relevant set of parentheses. This search went wrong
1085 when there were unescaped parentheses in a character class, parentheses
1086 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
1087
1088 33. "Subroutine" calls and backreferences were previously restricted to
1089 referencing subpatterns earlier in the regex. This restriction has now
1090 been removed.
1091
1092 34. Added a number of extra features that are going to be in Perl 5.10. On the
1093 whole, these are just syntactic alternatives for features that PCRE had
1094 previously implemented using the Python syntax or my own invention. The
1095 other formats are all retained for compatibility.
1096
1097 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
1098 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
1099 also .NET compatible.
1100
1101 (b) A recursion or subroutine call to a named group can now be defined as
1102 (?&name) as well as (?P>name).
1103
1104 (c) A backreference to a named group can now be defined as \k<name> or
1105 \k'name' as well as (?P=name). The new forms, as well as being in Perl
1106 5.10, are also .NET compatible.
1107
1108 (d) A conditional reference to a named group can now use the syntax
1109 (?(<name>) or (?('name') as well as (?(name).
1110
1111 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
1112 groups (named and numbered) that are never evaluated inline, but can be
1113 called as "subroutines" from elsewhere. In effect, the DEFINE condition
1114 is always false. There may be only one alternative in such a group.
1115
1116 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
1117 as the simple (?(R). The condition is true only if the most recent
1118 recursion is that of the given number or name. It does not search out
1119 through the entire recursion stack.
1120
1121 (g) The escape \gN or \g{N} has been added, where N is a positive or
1122 negative number, specifying an absolute or relative reference.
1123
1124 35. Tidied to get rid of some further signed/unsigned compiler warnings and
1125 some "unreachable code" warnings.
1126
1127 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
1128 things, this adds five new scripts.
1129
1130 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
1131 There were also incompatibilities regarding the handling of \Q..\E inside
1132 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
1133 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
1134
1135 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
1136 matches an empty string, and forcibly breaks the loop. There were bugs in
1137 this code in non-simple cases. For a pattern such as ^(a()*)* matched
1138 against aaaa the result was just "a" rather than "aaaa", for example. Two
1139 separate and independent bugs (that affected different cases) have been
1140 fixed.
1141
1142 39. Refactored the code to abolish the use of different opcodes for small
1143 capturing bracket numbers. This is a tidy that I avoided doing when I
1144 removed the limit on the number of capturing brackets for 3.5 back in 2001.
1145 The new approach is not only tidier, it makes it possible to reduce the
1146 memory needed to fix the previous bug (38).
1147
1148 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
1149 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
1150 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
1151 mode.
1152
1153 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
1154 report.
1155
1156 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
1157 copy construction and assignment in the C++ wrapper.
1158
1159 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
1160 couple of bugs that could have given wrong results in the "--newline=crlf"
1161 case.
1162
1163 44. Added a number of casts and did some reorganization of signed/unsigned int
1164 variables following suggestions from Dair Grant. Also renamed the variable
1165 "this" as "item" because it is a C++ keyword.
1166
1167 45. Arranged for dftables to add
1168
1169 #include "pcre_internal.h"
1170
1171 to pcre_chartables.c because without it, gcc 4.x may remove the array
1172 definition from the final binary if PCRE is built into a static library and
1173 dead code stripping is activated.
1174
1175 46. For an unanchored pattern, if a match attempt fails at the start of a
1176 newline sequence, and the newline setting is CRLF or ANY, and the next two
1177 characters are CRLF, advance by two characters instead of one.
1178
1179
1180 Version 6.7 04-Jul-06
1181 ---------------------
1182
1183 1. In order to handle tests when input lines are enormously long, pcretest has
1184 been re-factored so that it automatically extends its buffers when
1185 necessary. The code is crude, but this _is_ just a test program. The
1186 default size has been increased from 32K to 50K.
1187
1188 2. The code in pcre_study() was using the value of the re argument before
1189 testing it for NULL. (Of course, in any sensible call of the function, it
1190 won't be NULL.)
1191
1192 3. The memmove() emulation function in pcre_internal.h, which is used on
1193 systems that lack both memmove() and bcopy() - that is, hardly ever -
1194 was missing a "static" storage class specifier.
1195
1196 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
1197 containing an extended class (one that cannot be represented by a bitmap
1198 because it contains high-valued characters or Unicode property items, e.g.
1199 [\pZ]). Almost always one would set UTF-8 mode when processing such a
1200 pattern, but PCRE should not loop if you do not (it no longer does).
1201 [Detail: two cases were found: (a) a repeated subpattern containing an
1202 extended class; (b) a recursive reference to a subpattern that followed a
1203 previous extended class. It wasn't skipping over the extended class
1204 correctly when UTF-8 mode was not set.]
1205
1206 5. A negated single-character class was not being recognized as fixed-length
1207 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
1208 compile error "lookbehind assertion is not fixed length".
1209
1210 6. The RunPerlTest auxiliary script was showing an unexpected difference
1211 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
1212 write a Perl script that can interpret lines of an input file either as
1213 byte characters or as UTF-8, which is what "perltest" was being required to
1214 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
1215 can't do is switch easily at run time between having the "use utf8;" pragma
1216 or not. In the end, I fudged it by using the RunPerlTest script to insert
1217 "use utf8;" explicitly for the UTF-8 tests.
1218
1219 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
1220 the end of the subject string, contrary to the documentation and to what
1221 Perl does. This was true of both matching functions. Now it matches only at
1222 the start of the subject and immediately after *internal* newlines.
1223
1224 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
1225 a pointer to an int instead of a pointer to an unsigned long int. This
1226 caused problems on 64-bit systems.
1227
1228 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
1229 instance of the 'standard' template library not being so standard".
1230
1231 10. There was no check on the number of named subpatterns nor the maximum
1232 length of a subpattern name. The product of these values is used to compute
1233 the size of the memory block for a compiled pattern. By supplying a very
1234 long subpattern name and a large number of named subpatterns, the size
1235 computation could be caused to overflow. This is now prevented by limiting
1236 the length of names to 32 characters, and the number of named subpatterns
1237 to 10,000.
1238
1239 11. Subpatterns that are repeated with specific counts have to be replicated in
1240 the compiled pattern. The size of memory for this was computed from the
1241 length of the subpattern and the repeat count. The latter is limited to
1242 65535, but there was no limit on the former, meaning that integer overflow
1243 could in principle occur. The compiled length of a repeated subpattern is
1244 now limited to 30,000 bytes in order to prevent this.
1245
1246 12. Added the optional facility to have named substrings with the same name.
1247
1248 13. Added the ability to use a named substring as a condition, using the
1249 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
1250 are numbers (not recommended). Forward references are permitted.
1251
1252 14. Added forward references in named backreferences (if you see what I mean).
1253
1254 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
1255 pattern could run off the end of the subject. For example, the pattern
1256 "(?s)(.{1,5})"8 did this with the subject "ab".
1257
1258 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
1259 PCRE_CASELESS was set when matching characters that were quantified with ?
1260 or *.
1261
1262 17. A character class other than a single negated character that had a minimum
1263 but no maximum quantifier - for example [ab]{6,} - was not handled
1264 correctly by pce_dfa_exec(). It would match only one character.
1265
1266 18. A valid (though odd) pattern that looked like a POSIX character
1267 class but used an invalid character after [ (for example [[,abc,]]) caused
1268 pcre_compile() to give the error "Failed: internal error: code overflow" or
1269 in some cases to crash with a glibc free() error. This could even happen if
1270 the pattern terminated after [[ but there just happened to be a sequence of
1271 letters, a binary zero, and a closing ] in the memory that followed.
1272
1273 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
1274 over the years. Originally (before any Unicode support), just the bottom 8
1275 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
1276 output from "man perlunicode" includes this:
1277
1278 The regular expression compiler produces polymorphic opcodes. That
1279 is, the pattern adapts to the data and automatically switches to
1280 the Unicode character scheme when presented with Unicode data--or
1281 instead uses a traditional byte scheme when presented with byte
1282 data.
1283
1284 Sadly, a wide octal escape does not cause a switch, and in a string with
1285 no other multibyte characters, these octal escapes are treated as before.
1286 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
1287 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
1288 Unicode string.
1289
1290 I have not perpetrated such confusion in PCRE. Up till now, it took just
1291 the bottom 8 bits, as in old Perl. I have now made octal escapes with
1292 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
1293 translate to the appropriate multibyte character.
1294
1295 29. Applied some refactoring to reduce the number of warnings from Microsoft
1296 and Borland compilers. This has included removing the fudge introduced
1297 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
1298 a warning about an unused variable.
1299
1300 21. PCRE has not included VT (character 0x0b) in the set of whitespace
1301 characters since release 4.0, because Perl (from release 5.004) does not.
1302 [Or at least, is documented not to: some releases seem to be in conflict
1303 with the documentation.] However, when a pattern was studied with
1304 pcre_study() and all its branches started with \s, PCRE still included VT
1305 as a possible starting character. Of course, this did no harm; it just
1306 caused an unnecessary match attempt.
1307
1308 22. Removed a now-redundant internal flag bit that recorded the fact that case
1309 dependency changed within the pattern. This was once needed for "required
1310 byte" processing, but is no longer used. This recovers a now-scarce options
1311 bit. Also moved the least significant internal flag bit to the most-
1312 significant bit of the word, which was not previously used (hangover from
1313 the days when it was an int rather than a uint) to free up another bit for
1314 the future.
1315
1316 23. Added support for CRLF line endings as well as CR and LF. As well as the
1317 default being selectable at build time, it can now be changed at runtime
1318 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
1319 specify that it is scanning data with non-default line endings.
1320
1321 24. Changed the definition of CXXLINK to make it agree with the definition of
1322 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
1323
1324 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
1325 recursions. This makes a big different to stack usage for some patterns.
1326
1327 26. If a subpattern containing a named recursion or subroutine reference such
1328 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
1329 the space required for the compiled pattern went wrong and gave too small a
1330 value. Depending on the environment, this could lead to "Failed: internal
1331 error: code overflow at offset 49" or "glibc detected double free or
1332 corruption" errors.
1333
1334 27. Applied patches from Google (a) to support the new newline modes and (b) to
1335 advance over multibyte UTF-8 characters in GlobalReplace.
1336
1337 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
1338 difference for some implementation of PCRE in some Windows version.
1339
1340 29. Added some extra testing facilities to pcretest:
1341
1342 \q<number> in a data line sets the "match limit" value
1343 \Q<number> in a data line sets the "match recursion limt" value
1344 -S <number> sets the stack size, where <number> is in megabytes
1345
1346 The -S option isn't available for Windows.
1347
1348
1349 Version 6.6 06-Feb-06
1350 ---------------------
1351
1352 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
1353 in pcreposix.h. I have copied the definition from pcre.h.
1354
1355 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
1356 because pcre.h is no longer a built file.
1357
1358 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
1359 not normally included in the compiled code.
1360
1361
1362 Version 6.5 01-Feb-06
1363 ---------------------
1364
1365 1. When using the partial match feature with pcre_dfa_exec(), it was not
1366 anchoring the second and subsequent partial matches at the new starting
1367 point. This could lead to incorrect results. For example, with the pattern
1368 /1234/, partially matching against "123" and then "a4" gave a match.
1369
1370 2. Changes to pcregrep:
1371
1372 (a) All non-match returns from pcre_exec() were being treated as failures
1373 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
1374 error message is output. Some extra information is given for the
1375 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
1376 probably the only errors that are likely to be caused by users (by
1377 specifying a regex that has nested indefinite repeats, for instance).
1378 If there are more than 20 of these errors, pcregrep is abandoned.
1379
1380 (b) A binary zero was treated as data while matching, but terminated the
1381 output line if it was written out. This has been fixed: binary zeroes
1382 are now no different to any other data bytes.
1383
1384 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
1385 used to set a locale for matching. The --locale=xxxx long option has
1386 been added (no short equivalent) to specify a locale explicitly on the
1387 pcregrep command, overriding the environment variables.
1388
1389 (d) When -B was used with -n, some line numbers in the output were one less
1390 than they should have been.
1391
1392 (e) Added the -o (--only-matching) option.
1393
1394 (f) If -A or -C was used with -c (count only), some lines of context were
1395 accidentally printed for the final match.
1396
1397 (g) Added the -H (--with-filename) option.
1398
1399 (h) The combination of options -rh failed to suppress file names for files
1400 that were found from directory arguments.
1401
1402 (i) Added the -D (--devices) and -d (--directories) options.
1403
1404 (j) Added the -F (--fixed-strings) option.
1405
1406 (k) Allow "-" to be used as a file name for -f as well as for a data file.
1407
1408 (l) Added the --colo(u)r option.
1409
1410 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
1411 is not present by default.
1412
1413 3. A nasty bug was discovered in the handling of recursive patterns, that is,
1414 items such as (?R) or (?1), when the recursion could match a number of
1415 alternatives. If it matched one of the alternatives, but subsequently,
1416 outside the recursion, there was a failure, the code tried to back up into
1417 the recursion. However, because of the way PCRE is implemented, this is not
1418 possible, and the result was an incorrect result from the match.
1419
1420 In order to prevent this happening, the specification of recursion has
1421 been changed so that all such subpatterns are automatically treated as
1422 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
1423
1424 4. I had overlooked the fact that, in some locales, there are characters for
1425 which isalpha() is true but neither isupper() nor islower() are true. In
1426 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
1427 and ordfeminine) are like this. This affected the treatment of \w and \W
1428 when they appeared in character classes, but not when they appeared outside
1429 a character class. The bit map for "word" characters is now created
1430 separately from the results of isalnum() instead of just taking it from the
1431 upper, lower, and digit maps. (Plus the underscore character, of course.)
1432
1433 5. The above bug also affected the handling of POSIX character classes such as
1434 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
1435 permanent tables. Instead, the bit maps for such a class were previously
1436 created as the appropriate unions of the upper, lower, and digit bitmaps.
1437 Now they are created by subtraction from the [[:word:]] class, which has
1438 its own bitmap.
1439
1440 6. The [[:blank:]] character class matches horizontal, but not vertical space.
1441 It is created by subtracting the vertical space characters (\x09, \x0a,
1442 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
1443 subtraction was done in the overall bitmap for a character class, meaning
1444 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
1445 be recognized. This bug has been fixed.
1446
1447 7. Patches from the folks at Google:
1448
1449 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
1450 real life, but is still worth protecting against".
1451
1452 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
1453 regular expressions".
1454
1455 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
1456 have it.
1457
1458 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
1459 "configure" and the latter not, in order to fix a problem somebody had
1460 with compiling the Arg class on HP-UX.
1461
1462 (e) Improve the error-handling of the C++ wrapper a little bit.
1463
1464 (f) New tests for checking recursion limiting.
1465
1466 8. The pcre_memmove() function, which is used only if the environment does not
1467 have a standard memmove() function (and is therefore rarely compiled),
1468 contained two bugs: (a) use of int instead of size_t, and (b) it was not
1469 returning a result (though PCRE never actually uses the result).
1470
1471 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
1472 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
1473 returned instead of calling malloc() with an overflowing number that would
1474 most likely cause subsequent chaos.
1475
1476 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
1477
1478 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
1479 with this option is matched, the nmatch and pmatch options of regexec() are
1480 ignored.
1481
1482 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
1483 provided in case anyone wants to the the POSIX interface with UTF-8
1484 strings.
1485
1486 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
1487 C++ linking (needed for some HP-UX environments).
1488
1489 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
1490 (unused parameter) and in the pcre_printint() function (omitted "default"
1491 switch label when the default is to do nothing).
1492
1493 15. Added some code to make it possible, when PCRE is compiled as a C++
1494 library, to replace subject pointers for pcre_exec() with a smart pointer
1495 class, thus making it possible to process discontinuous strings.
1496
1497 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
1498 much the same function. They were added by different people who were trying
1499 to make PCRE easy to compile on non-Unix systems. It has been suggested
1500 that PCRE_EXPORT be abolished now that there is more automatic apparatus
1501 for compiling on Windows systems. I have therefore replaced it with
1502 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
1503 defaults to "extern" for C or "extern C" for C++, which works fine on
1504 Unix-like systems. It is now possible to override the value of PCRE_DATA_
1505 SCOPE with something explicit in config.h. In addition:
1506
1507 (a) pcreposix.h still had just "extern" instead of either of these macros;
1508 I have replaced it with PCRE_DATA_SCOPE.
1509
1510 (b) Functions such as _pcre_xclass(), which are internal to the library,
1511 but external in the C sense, all had PCRE_EXPORT in their definitions.
1512 This is apparently wrong for the Windows case, so I have removed it.
1513 (It makes no difference on Unix-like systems.)
1514
1515 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
1516 of recursive calls to match(). This is different to MATCH_LIMIT because
1517 that limits the total number of calls to match(), not all of which increase
1518 the depth of recursion. Limiting the recursion depth limits the amount of
1519 stack (or heap if NO_RECURSE is set) that is used. The default can be set
1520 when PCRE is compiled, and changed at run time. A patch from Google adds
1521 this functionality to the C++ interface.
1522
1523 18. Changes to the handling of Unicode character properties:
1524
1525 (a) Updated the table to Unicode 4.1.0.
1526
1527 (b) Recognize characters that are not in the table as "Cn" (undefined).
1528
1529 (c) I revised the way the table is implemented to a much improved format
1530 which includes recognition of ranges. It now supports the ranges that
1531 are defined in UnicodeData.txt, and it also amalgamates other
1532 characters into ranges. This has reduced the number of entries in the
1533 table from around 16,000 to around 3,000, thus reducing its size
1534 considerably. I realized I did not need to use a tree structure after
1535 all - a binary chop search is just as efficient. Having reduced the
1536 number of entries, I extended their size from 6 bytes to 8 bytes to
1537 allow for more data.
1538
1539 (d) Added support for Unicode script names via properties such as \p{Han}.
1540
1541 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
1542 matching that character.
1543
1544 20. When matching a repeated Unicode property with a minimum greater than zero,
1545 (for example \pL{2,}), PCRE could look past the end of the subject if it
1546 reached it while seeking the minimum number of characters. This could
1547 happen only if some of the characters were more than one byte long, because
1548 there is a check for at least the minimum number of bytes.
1549
1550 21. Refactored the implementation of \p and \P so as to be more general, to
1551 allow for more different types of property in future. This has changed the
1552 compiled form incompatibly. Anybody with saved compiled patterns that use
1553 \p or \P will have to recompile them.
1554
1555 22. Added "Any" and "L&" to the supported property types.
1556
1557 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
1558 but give a compile time error if the value is greater than 0xff.
1559
1560 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
1561 accidentally not being installed or uninstalled.
1562
1563 25. The pcre.h file was built from pcre.h.in, but the only changes that were
1564 made were to insert the current release number. This seemed silly, because
1565 it made things harder for people building PCRE on systems that don't run
1566 "configure". I have turned pcre.h into a distributed file, no longer built
1567 by "configure", with the version identification directly included. There is
1568 no longer a pcre.h.in file.
1569
1570 However, this change necessitated a change to the pcre-config script as
1571 well. It is built from pcre-config.in, and one of the substitutions was the
1572 release number. I have updated configure.ac so that ./configure now finds
1573 the release number by grepping pcre.h.
1574
1575 26. Added the ability to run the tests under valgrind.
1576
1577
1578 Version 6.4 05-Sep-05
1579 ---------------------
1580
1581 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
1582 "--" to be printed when multiple files were scanned, even when none of the
1583 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
1584 consider it to be a bug, and have restored the previous behaviour.
1585
1586 2. A couple of code tidies to get rid of compiler warnings.
1587
1588 3. The pcretest program used to cheat by referring to symbols in the library
1589 whose names begin with _pcre_. These are internal symbols that are not
1590 really supposed to be visible externally, and in some environments it is
1591 possible to suppress them. The cheating is now confined to including
1592 certain files from the library's source, which is a bit cleaner.
1593
1594 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
1595 file's purpose clearer.
1596
1597 5. Reorganized pcre_ucp_findchar().
1598
1599
1600 Version 6.3 15-Aug-05
1601 ---------------------
1602
1603 1. The file libpcre.pc.in did not have general read permission in the tarball.
1604
1605 2. There were some problems when building without C++ support:
1606
1607 (a) If C++ support was not built, "make install" and "make test" still
1608 tried to test it.
1609
1610 (b) There were problems when the value of CXX was explicitly set. Some
1611 changes have been made to try to fix these, and ...
1612
1613 (c) --disable-cpp can now be used to explicitly disable C++ support.
1614
1615 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
1616 backslash in a target when C++ was disabled. This confuses some
1617 versions of "make", apparently. Using an intermediate variable solves
1618 this. (Same for CPP_LOBJ.)
1619
1620 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
1621 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
1622 necessary on certain architectures.
1623
1624 4. Added a setting of -export-symbols-regex to the link command to remove
1625 those symbols that are exported in the C sense, but actually are local
1626 within the library, and not documented. Their names all begin with
1627 "_pcre_". This is not a perfect job, because (a) we have to except some
1628 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
1629 available (and never for static libraries). I have made a note to try to
1630 find a way round (a) in the future.
1631
1632
1633 Version 6.2 01-Aug-05
1634 ---------------------
1635
1636 1. There was no test for integer overflow of quantifier values. A construction
1637 such as {1111111111111111} would give undefined results. What is worse, if
1638 a minimum quantifier for a parenthesized subpattern overflowed and became
1639 negative, the calculation of the memory size went wrong. This could have
1640 led to memory overwriting.
1641
1642 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
1643
1644 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
1645 operating environments where this matters.
1646
1647 4. Applied Giuseppe Maxia's patch to add additional features for controlling
1648 PCRE options from within the C++ wrapper.
1649
1650 5. Named capturing subpatterns were not being correctly counted when a pattern
1651 was compiled. This caused two problems: (a) If there were more than 100
1652 such subpatterns, the calculation of the memory needed for the whole
1653 compiled pattern went wrong, leading to an overflow error. (b) Numerical
1654 back references of the form \12, where the number was greater than 9, were
1655 not recognized as back references, even though there were sufficient
1656 previous subpatterns.
1657
1658 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
1659 versions of gcc, e.g. 2.95.4.
1660
1661
1662 Version 6.1 21-Jun-05
1663 ---------------------
1664
1665 1. There was one reference to the variable "posix" in pcretest.c that was not
1666 surrounded by "#if !defined NOPOSIX".
1667
1668 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
1669 the cross-check on the old pcre_info() function, for the benefit of the
1670 cut-down version of PCRE that is currently imported into Exim.
1671
1672 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
1673 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
1674 patterns that start (?i) but I don't think that matters. The use of (?i) is
1675 just an example; this all applies to the other options as well.
1676
1677 4. Since libtool seems to echo the compile commands it is issuing, the output
1678 from "make" can be reduced a bit by putting "@" in front of each libtool
1679 compile command.
1680
1681 5. Patch from the folks at Google for configure.in to be a bit more thorough
1682 in checking for a suitable C++ installation before trying to compile the
1683 C++ stuff. This should fix a reported problem when a compiler was present,
1684 but no suitable headers.
1685
1686 6. The man pages all had just "PCRE" as their title. I have changed them to
1687 be the relevant file name. I have also arranged that these names are
1688 retained in the file doc/pcre.txt, which is a concatenation in text format
1689 of all the man pages except the little individual ones for each function.
1690
1691 7. The NON-UNIX-USE file had not been updated for the different set of source
1692 files that come with release 6. I also added a few comments about the C++
1693 wrapper.
1694
1695
1696 Version 6.0 07-Jun-05
1697 ---------------------
1698
1699 1. Some minor internal re-organization to help with my DFA experiments.
1700
1701 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
1702 didn't matter for the library itself when fully configured, but did matter
1703 when compiling without UCP support, or within Exim, where the ucp files are
1704 not imported.
1705
1706 3. Refactoring of the library code to split up the various functions into
1707 different source modules. The addition of the new DFA matching code (see
1708 below) to a single monolithic source would have made it really too
1709 unwieldy, quite apart from causing all the code to be include in a
1710 statically linked application, when only some functions are used. This is
1711 relevant even without the DFA addition now that patterns can be compiled in
1712 one application and matched in another.
1713
1714 The downside of splitting up is that there have to be some external
1715 functions and data tables that are used internally in different modules of
1716 the library but which are not part of the API. These have all had their
1717 names changed to start with "_pcre_" so that they are unlikely to clash
1718 with other external names.
1719
1720 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
1721 a different (DFA) algorithm. Although it is slower than the original
1722 function, it does have some advantages for certain types of matching
1723 problem.
1724
1725 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
1726 including restarting after a partial match.
1727
1728 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
1729 defined when compiling for Windows was sent to me. I have put it into the
1730 code, though I have no means of testing or verifying it.
1731
1732 7. Added the pcre_refcount() auxiliary function.
1733
1734 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
1735 match before or at the first newline in the subject string. In pcretest,
1736 the /f option on a pattern can be used to set this.
1737
1738 9. A repeated \w when used in UTF-8 mode with characters greater than 256
1739 would behave wrongly. This has been present in PCRE since release 4.0.
1740
1741 10. A number of changes to the pcregrep command:
1742
1743 (a) Refactored how -x works; insert ^(...)$ instead of setting
1744 PCRE_ANCHORED and checking the length, in preparation for adding
1745 something similar for -w.
1746
1747 (b) Added the -w (match as a word) option.
1748
1749 (c) Refactored the way lines are read and buffered so as to have more
1750 than one at a time available.
1751
1752 (d) Implemented a pcregrep test script.
1753
1754 (e) Added the -M (multiline match) option. This allows patterns to match
1755 over several lines of the subject. The buffering ensures that at least
1756 8K, or the rest of the document (whichever is the shorter) is available
1757 for matching (and similarly the previous 8K for lookbehind assertions).
1758
1759 (f) Changed the --help output so that it now says
1760
1761 -w, --word-regex(p)
1762
1763 instead of two lines, one with "regex" and the other with "regexp"
1764 because that confused at least one person since the short forms are the
1765 same. (This required a bit of code, as the output is generated
1766 automatically from a table. It wasn't just a text change.)
1767
1768 (g) -- can be used to terminate pcregrep options if the next thing isn't an
1769 option but starts with a hyphen. Could be a pattern or a path name
1770 starting with a hyphen, for instance.
1771
1772 (h) "-" can be given as a file name to represent stdin.
1773
1774 (i) When file names are being printed, "(standard input)" is used for
1775 the standard input, for compatibility with GNU grep. Previously
1776 "<stdin>" was used.
1777
1778 (j) The option --label=xxx can be used to supply a name to be used for
1779 stdin when file names are being printed. There is no short form.
1780
1781 (k) Re-factored the options decoding logic because we are going to add
1782 two more options that take data. Such options can now be given in four
1783 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
1784
1785 (l) Added the -A, -B, and -C options for requesting that lines of context
1786 around matches be printed.
1787
1788 (m) Added the -L option to print the names of files that do not contain
1789 any matching lines, that is, the complement of -l.
1790
1791 (n) The return code is 2 if any file cannot be opened, but pcregrep does
1792 continue to scan other files.
1793
1794 (o) The -s option was incorrectly implemented. For compatibility with other
1795 greps, it now suppresses the error message for a non-existent or non-
1796 accessible file (but not the return code). There is a new option called
1797 -q that suppresses the output of matching lines, which was what -s was
1798 previously doing.
1799
1800 (p) Added --include and --exclude options to specify files for inclusion
1801 and exclusion when recursing.
1802
1803 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
1804 Hopefully, it now does.
1805
1806 12. Missing cast in pcre_study().
1807
1808 13. Added an "uninstall" target to the makefile.
1809
1810 14. Replaced "extern" in the function prototypes in Makefile.in with
1811 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
1812 world, but is set differently for Windows.
1813
1814 15. Added a second compiling function called pcre_compile2(). The only
1815 difference is that it has an extra argument, which is a pointer to an
1816 integer error code. When there is a compile-time failure, this is set
1817 non-zero, in addition to the error test pointer being set to point to an
1818 error message. The new argument may be NULL if no error number is required
1819 (but then you may as well call pcre_compile(), which is now just a
1820 wrapper). This facility is provided because some applications need a
1821 numeric error indication, but it has also enabled me to tidy up the way
1822 compile-time errors are handled in the POSIX wrapper.
1823
1824 16. Added VPATH=.libs to the makefile; this should help when building with one
1825 prefix path and installing with another. (Or so I'm told by someone who
1826 knows more about this stuff than I do.)
1827
1828 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
1829 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
1830 match everything, including newlines. This is not POSIX-compatible, but
1831 somebody wanted the feature. From pcretest it can be activated by using
1832 both the P and the s flags.
1833
1834 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
1835
1836 19. libpcre.pc was being incorrectly installed as executable.
1837
1838 20. A couple of places in pcretest check for end-of-line by looking for '\n';
1839 it now also looks for '\r' so that it will work unmodified on Windows.
1840
1841 21. Added Google's contributed C++ wrapper to the distribution.
1842
1843 22. Added some untidy missing memory free() calls in pcretest, to keep
1844 Electric Fence happy when testing.
1845
1846
1847
1848 Version 5.0 13-Sep-04
1849 ---------------------
1850
1851 1. Internal change: literal characters are no longer packed up into items
1852 containing multiple characters in a single byte-string. Each character
1853 is now matched using a separate opcode. However, there may be more than one
1854 byte in the character in UTF-8 mode.
1855
1856 2. The pcre_callout_block structure has two new fields: pattern_position and
1857 next_item_length. These contain the offset in the pattern to the next match
1858 item, and its length, respectively.
1859
1860 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
1861 insertion of callouts before each pattern item. Added the /C option to
1862 pcretest to make use of this.
1863
1864 4. On the advice of a Windows user, the lines
1865
1866 #if defined(_WIN32) || defined(WIN32)
1867 _setmode( _fileno( stdout ), 0x8000 );
1868 #endif /* defined(_WIN32) || defined(WIN32) */
1869
1870 have been added to the source of pcretest. This apparently does useful
1871 magic in relation to line terminators.
1872
1873 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
1874 for the benefit of those environments where the "b" makes a difference.
1875
1876 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
1877 to know about it. I have put a hack into configure.in that adds in code
1878 to set GCC=yes if CC=icc. This seems to end up at a point in the
1879 generated configure script that is early enough to affect the setting of
1880 compiler options, which is what is needed, but I have no means of testing
1881 whether it really works. (The user who reported this had patched the
1882 generated configure script, which of course I cannot do.)
1883
1884 LATER: After change 22 below (new libtool files), the configure script
1885 seems to know about icc (and also ecc). Therefore, I have commented out
1886 this hack in configure.in.
1887
1888 7. Added support for pkg-config (2 patches were sent in).
1889
1890 8. Negated POSIX character classes that used a combination of internal tables
1891 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
1892 [[:^ascii]]. Typically, they would match almost any characters. The other
1893 POSIX classes were not broken in this way.
1894
1895 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
1896 to find the match, as PCRE was deluded into thinking that the match had to
1897 start at the start point or following a newline. The same bug applied to
1898 patterns with negative forward assertions or any backward assertions
1899 preceding ".*" at the start, unless the pattern required a fixed first
1900 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
1901
1902 10. In UTF-8 mode, when moving forwards in the subject after a failed match
1903 starting at the last subject character, bytes beyond the end of the subject
1904 string were read.
1905
1906 11. Renamed the variable "class" as "classbits" to make life easier for C++
1907 users. (Previously there was a macro definition, but it apparently wasn't
1908 enough.)
1909
1910 12. Added the new field "tables" to the extra data so that tables can be passed
1911 in at exec time, or the internal tables can be re-selected. This allows
1912 a compiled regex to be saved and re-used at a later time by a different
1913 program that might have everything at different addresses.
1914
1915 13. Modified the pcre-config script so that, when run on Solaris, it shows a
1916 -R library as well as a -L library.
1917
1918 14. The debugging options of pcretest (-d on the command line or D on a
1919 pattern) showed incorrect output for anything following an extended class
1920 that contained multibyte characters and which was followed by a quantifier.
1921
1922 15. Added optional support for general category Unicode character properties
1923 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
1924 support. It adds about 90K to the size of the library. The meanings of the
1925 inbuilt class escapes such as \d and \s have NOT been changed.
1926
1927 16. Updated pcredemo.c to include calls to free() to release the memory for the
1928 compiled pattern.
1929
1930 17. The generated file chartables.c was being created in the source directory
1931 instead of in the building directory. This caused the build to fail if the
1932 source directory was different from the building directory, and was
1933 read-only.
1934
1935 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
1936 file. No doubt somebody will tell me if they don't make sense... Also added
1937 Dan Mooney's comments about building on OpenVMS.
1938
1939 19. Added support for partial matching via the PCRE_PARTIAL option for
1940 pcre_exec() and the \P data escape in pcretest.
1941
1942 20. Extended pcretest with 3 new pattern features:
1943
1944 (i) A pattern option of the form ">rest-of-line" causes pcretest to
1945 write the compiled pattern to the file whose name is "rest-of-line".
1946 This is a straight binary dump of the data, with the saved pointer to
1947 the character tables forced to be NULL. The study data, if any, is
1948 written too. After writing, pcretest reads a new pattern.
1949
1950 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
1951 compiled pattern from the given file. There must not be any
1952 occurrences of "<" in the file name (pretty unlikely); if there are,
1953 pcretest will instead treat the initial "<" as a pattern delimiter.
1954 After reading in the pattern, pcretest goes on to read data lines as
1955 usual.
1956
1957 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
1958 and 16-bit fields in a compiled pattern, to simulate a pattern that
1959 was compiled on a host of opposite endianness.
1960
1961 21. The pcre-exec() function can now cope with patterns that were compiled on
1962 hosts of opposite endianness, with this restriction:
1963
1964 As for any compiled expression that is saved and used later, the tables
1965 pointer field cannot be preserved; the extra_data field in the arguments
1966 to pcre_exec() should be used to pass in a tables address if a value
1967 other than the default internal tables were used at compile time.
1968
1969 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
1970 now diagnosed as an error. Previously, most of the time, a negative number
1971 would have been treated as zero, but if in addition "ovector" was passed as
1972 NULL, a crash could occur.
1973
1974 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
1975 new versions from the libtool 1.5 distribution (the last one is a copy of
1976 a file called libtool.m4). This seems to have fixed the need to patch
1977 "configure" to support Darwin 1.3 (which I used to do). However, I still
1978 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
1979 workstation).
1980
1981 24. Changed the PCRE licence to be the more standard "BSD" licence.
1982
1983
1984 Version 4.5 01-Dec-03
1985 ---------------------
1986
1987 1. There has been some re-arrangement of the code for the match() function so
1988 that it can be compiled in a version that does not call itself recursively.
1989 Instead, it keeps those local variables that need separate instances for
1990 each "recursion" in a frame on the heap, and gets/frees frames whenever it
1991 needs to "recurse". Keeping track of where control must go is done by means
1992 of setjmp/longjmp. The whole thing is implemented by a set of macros that
1993 hide most of the details from the main code, and operates only if
1994 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
1995 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
1996 operating.
1997
1998 To make it easier for callers to provide specially tailored get/free
1999 functions for this usage, two new functions, pcre_stack_malloc, and
2000 pcre_stack_free, are used. They are always called in strict stacking order,
2001 and the size of block requested is always the same.
2002
2003 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
2004 PCRE has been compiled to use the stack or the heap for recursion. The
2005 -C option of pcretest uses this to show which version is compiled.
2006
2007 A new data escape \S, is added to pcretest; it causes the amounts of store
2008 obtained and freed by both kinds of malloc/free at match time to be added
2009 to the output.
2010
2011 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
2012 what's available on my current Linux desktop machine.
2013
2014 3. When matching a UTF-8 string, the test for a valid string at the start has
2015 been extended. If start_offset is not zero, PCRE now checks that it points
2016 to a byte that is the start of a UTF-8 character. If not, it returns
2017 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
2018 this is necessary because there may be backward assertions in the pattern.
2019 When matching the same subject several times, it may save resources to use
2020 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
2021
2022 4. The code for checking the validity of UTF-8 strings has been tightened so
2023 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
2024 containing "overlong sequences".
2025
2026 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
2027 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
2028 should have been "&", but it just so happened that all the cases this let
2029 through by mistake were picked up later in the function.
2030
2031 6. I had used a variable called "isblank" - this is a C99 function, causing
2032 some compilers to warn. To avoid this, I renamed it (as "blankclass").
2033
2034 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
2035 prompting; (b) run "./pcretest /dev/null" at the start of the test script
2036 so the version is shown; (c) stop "make test" echoing "./RunTest".
2037
2038 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
2039
2040 9. The prototype for memmove() for systems that don't have it was using
2041 size_t, but the inclusion of the header that defines size_t was later. I've
2042 moved the #includes for the C headers earlier to avoid this.
2043
2044 10. Added some adjustments to the code to make it easier to compiler on certain
2045 special systems:
2046
2047 (a) Some "const" qualifiers were missing.
2048 (b) Added the macro EXPORT before all exported functions; by default this
2049 is defined to be empty.
2050 (c) Changed the dftables auxiliary program (that builds chartables.c) so
2051 that it reads its output file name as an argument instead of writing
2052 to the standard output and assuming this can be redirected.
2053
2054 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
2055 class containing characters with values greater than 255, PCRE compilation
2056 went into a loop.
2057
2058 12. A recursive reference to a subpattern that was within another subpattern
2059 that had a minimum quantifier of zero caused PCRE to crash. For example,
2060 (x(y(?2))z)? provoked this bug with a subject that got as far as the
2061 recursion. If the recursively-called subpattern itself had a zero repeat,
2062 that was OK.
2063
2064 13. In pcretest, the buffer for reading a data line was set at 30K, but the
2065 buffer into which it was copied (for escape processing) was still set at
2066 1024, so long lines caused crashes.
2067
2068 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
2069 "internal error: code overflow...". This applied to any character class
2070 that was followed by a possessive quantifier.
2071
2072 15. Modified the Makefile to add libpcre.la as a prerequisite for
2073 libpcreposix.la because I was told this is needed for a parallel build to
2074 work.
2075
2076 16. If a pattern that contained .* following optional items at the start was
2077 studied, the wrong optimizing data was generated, leading to matching
2078 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
2079 matching string must start with a or b or c. The correct conclusion for
2080 this pattern is that a match can start with any character.
2081
2082
2083 Version 4.4 13-Aug-03
2084 ---------------------
2085
2086 1. In UTF-8 mode, a character class containing characters with values between
2087 127 and 255 was not handled correctly if the compiled pattern was studied.
2088 In fixing this, I have also improved the studying algorithm for such
2089 classes (slightly).
2090
2091 2. Three internal functions had redundant arguments passed to them. Removal
2092 might give a very teeny performance improvement.
2093
2094 3. Documentation bug: the value of the capture_top field in a callout is *one
2095 more than* the number of the hightest numbered captured substring.
2096
2097 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
2098 in incorrectly linking with a previously installed version. They now link
2099 explicitly with libpcre.la.
2100
2101 5. configure.in no longer needs to recognize Cygwin specially.
2102
2103 6. A problem in pcre.in for Windows platforms is fixed.
2104
2105 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
2106 pcretest, it used to include the size of the study block as part of its
2107 output. Unfortunately, the structure contains a field that has a different
2108 size on different hardware architectures. This meant that the tests that
2109 showed this size failed. As the block is currently always of a fixed size,
2110 this information isn't actually particularly useful in pcretest output, so
2111 I have just removed it.
2112
2113 8. Three pre-processor statements accidentally did not start in column 1.
2114 Sadly, there are *still* compilers around that complain, even though
2115 standard C has not required this for well over a decade. Sigh.
2116
2117 9. In pcretest, the code for checking callouts passed small integers in the
2118 callout_data field, which is a void * field. However, some picky compilers
2119 complained about the casts involved for this on 64-bit systems. Now
2120 pcretest passes the address of the small integer instead, which should get
2121 rid of the warnings.
2122
2123 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
2124 both compile and run time, and gives an error if an invalid UTF-8 sequence
2125 is found. There is a option for disabling this check in cases where the
2126 string is known to be correct and/or the maximum performance is wanted.
2127
2128 11. In response to a bug report, I changed one line in Makefile.in from
2129
2130 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
2131 to
2132 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
2133
2134 to look similar to other lines, but I have no way of telling whether this
2135 is the right thing to do, as I do not use Windows. No doubt I'll get told
2136 if it's wrong...
2137
2138
2139 Version 4.3 21-May-03
2140 ---------------------
2141
2142 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
2143 Makefile.
2144
2145 2. Some refactoring to improve the quality of the code:
2146
2147 (i) The utf8_table... variables are now declared "const".
2148
2149 (ii) The code for \cx, which used the "case flipping" table to upper case
2150 lower case letters, now just substracts 32. This is ASCII-specific,
2151 but the whole concept of \cx is ASCII-specific, so it seems
2152 reasonable.
2153
2154 (iii) PCRE was using its character types table to recognize decimal and
2155 hexadecimal digits in the pattern. This is silly, because it handles
2156 only 0-9, a-f, and A-F, but the character types table is locale-
2157 specific, which means strange things might happen. A private
2158 table is now used for this - though it costs 256 bytes, a table is
2159 much faster than multiple explicit tests. Of course, the standard
2160 character types table is still used for matching digits in subject
2161 strings against \d.
2162
2163 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
2164 ending in _t are). So I've renamed it as ESC_tee.
2165
2166 3. The first argument for regexec() in the POSIX wrapper should have been
2167 defined as "const".
2168
2169 4. Changed pcretest to use malloc() for its buffers so that they can be
2170 Electric Fenced for debugging.
2171
2172 5. There were several places in the code where, in UTF-8 mode, PCRE would try
2173 to read one or more bytes before the start of the subject string. Often this
2174 had no effect on PCRE's behaviour, but in some circumstances it could
2175 provoke a segmentation fault.
2176
2177 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
2178 to try to read one or more bytes before the start of the subject string.
2179
2180 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
2181 UTF-8 support could misbehave in various ways if the subject string
2182 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
2183 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
2184 back over UTF-8 characters.)
2185
2186
2187 Version 4.2 14-Apr-03
2188 ---------------------
2189
2190 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
2191
2192 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
2193 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
2194 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
2195 [WIN_PREFIX]: new variable, "cyg" for Cygwin
2196 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
2197 and BUILD_EXEEXT
2198 Note: automatic setting of the BUILD variables is not yet working
2199 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
2200 compile-time but not at link-time
2201 [LINK]: use for linking executables only
2202 make different versions for Windows and non-Windows
2203 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
2204 libraries
2205 [LINK_FOR_BUILD]: new variable
2206 [OBJEXT]: use throughout
2207 [EXEEXT]: use throughout
2208 <winshared>: new target
2209 <wininstall>: new target
2210 <dftables.o>: use native compiler
2211 <dftables>: use native linker
2212 <install>: handle Windows platform correctly
2213 <clean>: ditto
2214 <check>: ditto
2215 copy DLL to top builddir before testing
2216
2217 As part of these changes, -no-undefined was removed again. This was reported
2218 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
2219 in any case.
2220
2221 3. Some tidies to get rid of compiler warnings:
2222
2223 . In the match_data structure, match_limit was an unsigned long int, whereas
2224 match_call_count was an int. I've made them both unsigned long ints.
2225
2226 . In pcretest the fact that a const uschar * doesn't automatically cast to
2227 a void * provoked a warning.
2228
2229 . Turning on some more compiler warnings threw up some "shadow" variables
2230 and a few more missing casts.
2231
2232 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
2233 option, a class that contained a single character with a value between 128
2234 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
2235
2236 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
2237 option, a class that contained several characters, but with at least one
2238 whose value was between 128 and 255 caused PCRE to crash.
2239
2240
2241 Version 4.1 12-Mar-03
2242 ---------------------
2243
2244 1. Compiling with gcc -pedantic found a couple of places where casts were
2245 needed, and a string in dftables.c that was longer than standard compilers are
2246 required to support.
2247
2248 2. Compiling with Sun's compiler found a few more places where the code could
2249 be tidied up in order to avoid warnings.
2250
2251 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
2252 first of these names is deprecated in the latest Autoconf in favour of the name
2253 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
2254 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
2255 analogy I have changed it to CFLAGS_FOR_BUILD.
2256
2257 4. Added -no-undefined to the linking command in the Makefile, because this is
2258 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
2259 linking step for the pcreposix library.
2260
2261 5. PCRE was failing to diagnose the case of two named groups with the same
2262 name.
2263
2264 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
2265 literal character that is needed in the subject for a match, and scans along to
2266 ensure that it is present before embarking on the full matching process. This
2267 saves time in cases of nested unlimited repeats that are never going to match.
2268 Problem: the scan can take a lot of time if the subject is very long (e.g.
2269 megabytes), thus penalizing straightforward matches. It is now done only if the
2270 amount of subject to be scanned is less than 1000 bytes.
2271
2272 7. A lesser problem with the same optimization is that it was recording the
2273 first character of an anchored pattern as "needed", thus provoking a search
2274 right along the subject, even when the first match of the pattern was going to
2275 fail. The "needed" character is now not set for anchored patterns, unless it
2276 follows something in the pattern that is of non-fixed length. Thus, it still
2277 fulfils its original purpose of finding quick non-matches in cases of nested
2278 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
2279
2280
2281 Version 4.0 17-Feb-03
2282 ---------------------
2283
2284 1. If a comment in an extended regex that started immediately after a meta-item
2285 extended to the end of string, PCRE compiled incorrect data. This could lead to
2286 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
2287
2288 2. Moved to autoconf 2.53 and libtool 1.4.2.
2289
2290 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
2291 the special perltest8 script is no longer needed - all the tests can be run
2292 from a single perltest script.
2293
2294 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
2295 by \s. It has now been removed in PCRE. This means it isn't recognized as
2296 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
2297 class [:space:] *does* include VT, thereby creating a mess.
2298
2299 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
2300 space and tab.
2301
2302 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
2303 its new features into the main test script, reducing the number of scripts.
2304
2305 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
2306 were backward compatible, and made the (?i) apply to the whole pattern, as if
2307 /i were given. Now it behaves more logically, and applies the option setting
2308 only to what follows. PCRE has been changed to follow suit. However, if it
2309 finds options settings right at the start of the pattern, it extracts them into
2310 the global options, as before. Thus, they show up in the info data.
2311
2312 8. Added support for the \Q...\E escape sequence. Characters in between are
2313 treated as literals. This is slightly different from Perl in that $ and @ are
2314 also handled as literals inside the quotes. In Perl, they will cause variable
2315 interpolation. Note the following examples:
2316
2317 Pattern PCRE matches Perl matches
2318
2319 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
2320 \Qabc\$xyz\E abc\$xyz abc\$xyz
2321 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
2322
2323 For compatibility with Perl, \Q...\E sequences are recognized inside character
2324 classes as well as outside them.
2325
2326 9. Re-organized 3 code statements in pcretest to avoid "overflow in
2327 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
2328 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
2329 signed/unsigned warnings.
2330
2331 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
2332 option for pcretest, so I've replaced it by a simple function that does just
2333 that job.
2334
2335 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
2336 "pcregrep -".
2337
2338 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
2339 Java package. This provides some syntactic sugar for simple cases of what my
2340 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
2341 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
2342 item, you can use this simplified notation. Note that only makes sense with
2343 greedy quantifiers. Consequently, the use of the possessive quantifier forces
2344 greediness, whatever the setting of the PCRE_UNGREEDY option.
2345
2346 13. A change of greediness default within a pattern was not taking effect at
2347 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
2348 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
2349 was abstracted outside.
2350
2351 14. PCRE now supports the \G assertion. It is true when the current matching
2352 position is at the start point of the match. This differs from \A when the
2353 starting offset is non-zero. Used with the /g option of pcretest (or similar
2354 code), it works in the same way as it does for Perl's /g option. If all
2355 alternatives of a regex begin with \G, the expression is anchored to the start
2356 match position, and the "anchored" flag is set in the compiled expression.
2357
2358 15. Some bugs concerning the handling of certain option changes within patterns
2359 have been fixed. These applied to options other than (?ims). For example,
2360 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
2361 been the other way round. Some of this was related to change 7 above.
2362
2363 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
2364 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
2365 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
2366 POSIX classes only within a class (e.g. /[[:alpha:]]/).
2367
2368 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
2369 mode. Unlike ".", it always matches newline, whatever the setting of
2370 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
2371 assertions. Perl allows it, but it doesn't (in general) work because it can't
2372 calculate the length of the lookbehind. At least, that's the case for Perl
2373 5.8.0 - I've been told they are going to document that it doesn't work in
2374 future.
2375
2376 18. Added an error diagnosis for escapes that PCRE does not support: these are
2377 \L, \l, \N, \P, \p, \U, \u, and \X.
2378
2379 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
2380 reading past the end of the pattern in cases such as /[abcd/.
2381
2382 20. PCRE was getting more memory than necessary for patterns with classes that
2383 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
2384
2385 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
2386 compiling PCRE for use with Virtual Pascal.
2387
2388 22. Small fix to the Makefile to make it work properly if the build is done
2389 outside the source tree.
2390
2391 23. Added a new extension: a condition to go with recursion. If a conditional
2392 subpattern starts with (?(R) the "true" branch is used if recursion has
2393 happened, whereas the "false" branch is used only at the top level.
2394
2395 24. When there was a very long string of literal characters (over 255 bytes
2396 without UTF support, over 250 bytes with UTF support), the computation of how
2397 much memory was required could be incorrect, leading to segfaults or other
2398 strange effects.
2399
2400 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
2401 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
2402 there was a subsequent back reference to those brackets. This meant that, for
2403 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
2404 possible to check for precisely this case. All we can do is abandon the
2405 optimization if .* occurs inside capturing brackets when there are any back
2406 references whatsoever. (See below for a better fix that came later.)
2407
2408 26. The handling of the optimization for finding the first character of a
2409 non-anchored pattern, and for finding a character that is required later in the
2410 match were failing in some cases. This didn't break the matching; it just
2411 failed to optimize when it could. The way this is done has been re-implemented.
2412
2413 27. Fixed typo in error message for invalid (?R item (it said "(?p").
2414
2415 28. Added a new feature that provides some of the functionality that Perl
2416 provides with (?{...}). The facility is termed a "callout". The way it is done
2417 in PCRE is for the caller to provide an optional function, by setting
2418 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
2419 global variable. By default it is unset, which disables all calling out. To get
2420 the function called, the regex must include (?C) at appropriate points. This
2421 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
2422 This provides a means of identifying different callout points. When PCRE
2423 reaches such a point in the regex, if pcre_callout has been set, the external
2424 function is called. It is provided with data in a structure called
2425 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
2426 matching continues; if it returns a non-zero value, the match at the current
2427 point fails. However, backtracking will occur if possible. [This was changed
2428 later and other features added - see item 49 below.]
2429
2430 29. pcretest is upgraded to test the callout functionality. It provides a
2431 callout function that displays information. By default, it shows the start of
2432 the match and the current position in the text. There are some new data escapes
2433 to vary what happens:
2434
2435 \C+ in addition, show current contents of captured substrings
2436 \C- do not supply a callout function
2437 \C!n return 1 when callout number n is reached
2438 \C!n!m return 1 when callout number n is reached for the mth time
2439
2440 30. If pcregrep was called with the -l option and just a single file name, it
2441 output "<stdin>" if a match was found, instead of the file name.
2442
2443 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
2444 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
2445 pcre_exec(). This saves a malloc/free per call. The default value of
2446 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
2447 when configuring.
2448
2449 32. The default maximum size of a compiled pattern is 64K. There have been a
2450 few cases of people hitting this limit. The code now uses macros to handle the
2451 storing of links as offsets within the compiled pattern. It defaults to 2-byte
2452 links, but this can be changed to 3 or 4 bytes by --with-link-size when
2453 configuring. Tests 2 and 5 work only with 2-byte links because they output
2454 debugging information about compiled patterns.
2455
2456 33. Internal code re-arrangements:
2457
2458 (a) Moved the debugging function for printing out a compiled regex into
2459 its own source file (printint.c) and used #include to pull it into
2460 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
2461 separate copies.
2462
2463 (b) Defined the list of op-code names for debugging as a macro in
2464 internal.h so that it is next to the definition of the opcodes.
2465
2466 (c) Defined a table of op-code lengths for simpler skipping along compiled
2467 code. This is again a macro in internal.h so that it is next to the
2468 definition of the opcodes.
2469
2470 34. Added support for recursive calls to individual subpatterns, along the
2471 lines of Robin Houston's patch (but implemented somewhat differently).
2472
2473 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
2474 allow it to read and process whole directories in Win32. This code was
2475 contributed by Lionel Fourquaux; it has not been tested by me.
2476
2477 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
2478 used to name a group. Names consist of alphanumerics and underscores, and must
2479 be unique. Back references use the syntax (?P=name) and recursive calls use
2480 (?P>name) which is a PCRE extension to the Python extension. Groups still have
2481 numbers. The function pcre_fullinfo() can be used after compilation to extract
2482 a name/number map. There are three relevant calls:
2483
2484 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
2485 PCRE_INFO_NAMECOUNT yields the number of entries
2486 PCRE_INFO_NAMETABLE yields a pointer to the map.
2487
2488 The map is a vector of fixed-size entries. The size of each entry depends on
2489 the length of the longest name used. The first two bytes of each entry are the
2490 group number, most significant byte first. There follows the corresponding
2491 name, zero terminated. The names are in alphabetical order.
2492
2493 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
2494 case instead of 255. Making it the same both with and without UTF-8 support
2495 means that the same test output works with both.
2496
2497 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
2498 calling malloc() with a zero argument.
2499
2500 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
2501 optimization. I've improved things by keeping a bitmap of backreferences with
2502 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
2503 fact referenced, the optimization can be applied. It is unlikely that a
2504 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
2505 the match to follow \n) will appear inside brackets with a number greater than
2506 31, but if it does, any back reference > 31 suppresses the optimization.
2507
2508 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
2509 of disabling numbered capturing parentheses. Any opening parenthesis that is
2510 not followed by ? behaves as if it were followed by ?: but named parentheses
2511 can still be used for capturing (and they will acquire numbers in the usual
2512 way).
2513
2514 41. Redesigned the return codes from the match() function into yes/no/error so
2515 that errors can be passed back from deep inside the nested calls. A malloc
2516 failure while inside a recursive subpattern call now causes the
2517 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
2518
2519 42. It is now possible to set a limit on the number of times the match()
2520 function is called in a call to pcre_exec(). This facility makes it possible to
2521 limit the amount of recursion and backtracking, though not in a directly
2522 obvious way, because the match() function is used in a number of different
2523 circumstances. The count starts from zero for each position in the subject
2524 string (for non-anchored patterns). The default limit is, for compatibility, a
2525 large number, namely 10 000 000. You can change this in two ways:
2526
2527 (a) When configuring PCRE before making, you can use --with-match-limit=n
2528 to set a default value for the compiled library.
2529
2530 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
2531 a different value is set. See 45 below.
2532
2533 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
2534
2535 43. Added a new function pcre_config(int, void *) to enable run-time extraction
2536 of things that can be changed at compile time. The first argument specifies
2537 what is wanted and the second points to where the information is to be placed.
2538 The current list of available information is:
2539
2540 PCRE_CONFIG_UTF8
2541
2542 The output is an integer that is set to one if UTF-8 support is available;
2543 otherwise it is set to zero.
2544
2545 PCRE_CONFIG_NEWLINE
2546
2547 The output is an integer that it set to the value of the code that is used for
2548 newline. It is either LF (10) or CR (13).
2549
2550 PCRE_CONFIG_LINK_SIZE
2551
2552 The output is an integer that contains the number of bytes used for internal
2553 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
2554
2555 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
2556
2557 The output is an integer that contains the threshold above which the POSIX
2558 interface uses malloc() for output vectors. See item 31 above.
2559
2560 PCRE_CONFIG_MATCH_LIMIT
2561
2562 The output is an unsigned integer that contains the default limit of the number
2563 of match() calls in a pcre_exec() execution. See 42 above.
2564
2565 44. pcretest has been upgraded by the addition of the -C option. This causes it
2566 to extract all the available output from the new pcre_config() function, and to
2567 output it. The program then exits immediately.
2568
2569 45. A need has arisen to pass over additional data with calls to pcre_exec() in
2570 order to support additional features. One way would have been to define
2571 pcre_exec2() (for example) with extra arguments, but this would not have been
2572 extensible, and would also have required all calls to the original function to
2573 be mapped to the new one. Instead, I have chosen to extend the mechanism that
2574 is used for passing in "extra" data from pcre_study().
2575
2576 The pcre_extra structure is now exposed and defined in pcre.h. It currently
2577 contains the following fields:
2578
2579 flags a bitmap indicating which of the following fields are set
2580 study_data opaque data from pcre_study()
2581 match_limit a way of specifying a limit on match() calls for a specific
2582 call to pcre_exec()
2583 callout_data data for callouts (see 49 below)
2584
2585 The flag bits are also defined in pcre.h, and are
2586
2587 PCRE_EXTRA_STUDY_DATA
2588 PCRE_EXTRA_MATCH_LIMIT
2589 PCRE_EXTRA_CALLOUT_DATA
2590
2591 The pcre_study() function now returns one of these new pcre_extra blocks, with
2592 the actual study data pointed to by the study_data field, and the
2593 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
2594 before. That is, this change is entirely upwards-compatible and requires no
2595 change to existing code.
2596
2597 If you want to pass in additional data to pcre_exec(), you can either place it
2598 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
2599 block.
2600
2601 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
2602 data string contains the escape sequence \M, pcretest calls pcre_exec() several
2603 times with different match limits, until it finds the minimum value needed for
2604 pcre_exec() to complete. The value is then output. This can be instructive; for
2605 most simple matches the number is quite small, but for pathological cases it
2606 gets very large very quickly.
2607
2608 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
2609 returns the size of the data block pointed to by the study_data field in a
2610 pcre_extra block, that is, the value that was passed as the argument to
2611 pcre_malloc() when PCRE was getting memory in which to place the information
2612 created by pcre_study(). The fourth argument should point to a size_t variable.
2613 pcretest has been extended so that this information is shown after a successful
2614 pcre_study() call when information about the compiled regex is being displayed.
2615
2616 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
2617 because what follows is always an absolute path. (Later: it turns out that this
2618 is more than cosmetic for MinGW, because it doesn't like empty path
2619 components.)
2620
2621 49. Some changes have been made to the callout feature (see 28 above):
2622
2623 (i) A callout function now has three choices for what it returns:
2624
2625 0 => success, carry on matching
2626 > 0 => failure at this point, but backtrack if possible
2627 < 0 => serious error, return this value from pcre_exec()
2628
2629 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
2630 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
2631 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
2632 use by callout functions. It will never be used by PCRE itself.
2633
2634 (ii) The pcre_extra structure (see 45 above) has a void * field called
2635 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
2636 pcre_callout_block structure has a field of the same name. The contents of
2637 the field passed in the pcre_extra structure are passed to the callout
2638 function in the corresponding field in the callout block. This makes it
2639 easier to use the same callout-containing regex from multiple threads. For
2640 testing, the pcretest program has a new data escape
2641
2642 \C*n pass the number n (may be negative) as callout_data
2643
2644 If the callout function in pcretest receives a non-zero value as
2645 callout_data, it returns that value.
2646
2647 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
2648 there were some redundant $(CFLAGS) in commands that are now specified as
2649 $(LINK), which already includes $(CFLAGS).
2650
2651 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
2652 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
2653 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
2654 one-byte characters throughout. Note that case-insensitive matching applies
2655 only to characters whose values are less than 256. PCRE doesn't support the
2656 notion of cases for higher-valued characters.
2657
2658 (i) A character class whose characters are all within 0-255 is handled as
2659 a bit map, and the map is inverted for negative classes. Previously, a
2660 character > 255 always failed to match such a class; however it should
2661 match if the class was a negative one (e.g. [^ab]). This has been fixed.
2662
2663 (ii) A negated character class with a single character < 255 is coded as
2664 "not this character" (OP_NOT). This wasn't working properly when the test
2665 character was multibyte, either singly or repeated.
2666
2667 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
2668 mode, for example: \x{100}{2,3}.
2669
2670 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
2671 singly or repeated) now correctly test multibyte characters. However,
2672 PCRE doesn't recognize any characters with values greater than 255 as
2673 digits, spaces, or word characters. Such characters always match \D, \S,
2674 and \W, and never match \d, \s, or \w.
2675
2676 (v) Classes may now contain characters and character ranges with values
2677 greater than 255. For example: [ab\x{100}-\x{400}].
2678
2679 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
2680 PCRE in UTF-8 mode.
2681
2682 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
2683 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
2684 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
2685 value.)
2686
2687 53. The single man page has become too large. I have therefore split it up into
2688 a number of separate man pages. These also give rise to individual HTML pages;
2689 these are now put in a separate directory, and there is an index.html page that
2690 lists them all. Some hyperlinking between the pages has been installed.
2691
2692 54. Added convenience functions for handling named capturing parentheses.
2693
2694 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
2695 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
2696 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
2697 are faulted.
2698
2699 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
2700 calling configure. These values are used when compiling the dftables.c program
2701 which is run to generate the source of the default character tables. They
2702 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
2703 you will need to set these values.
2704
2705 57. Updated the building process for Windows DLL, as provided by Fred Cox.
2706
2707
2708 Version 3.9 02-Jan-02
2709 ---------------------
2710
2711 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
2712
2713 2. If --disable-static was given, the building process failed when trying to
2714 build pcretest and pcregrep. (For some reason it was using libtool to compile
2715 them, which is not right, as they aren't part of the library.)
2716
2717
2718 Version 3.8 18-Dec-01
2719 ---------------------
2720
2721 1. The experimental UTF-8 code was completely screwed up. It was packing the
2722 bytes in the wrong order. How dumb can you get?
2723
2724
2725 Version 3.7 29-Oct-01
2726 ---------------------
2727
2728 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
2729 This caused pcretest, when used on the test data, to segfault. Unfortunately,
2730 this didn't happen under Solaris 8, where I normally test things.
2731
2732 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
2733 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
2734 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
2735 this fix an hour or so after the initial 3.7 release.)
2736
2737
2738 Version 3.6 23-Oct-01
2739 ---------------------
2740
2741 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
2742 offsets passed as NULL with zero offset count.
2743
2744 2. The config.guess and config.sub files had not been updated when I moved to
2745 the latest autoconf.
2746
2747
2748 Version 3.5 15-Aug-01
2749 ---------------------
2750
2751 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
2752 had been forgotten.
2753
2754 2. By using declared but undefined structures, we can avoid using "void"
2755 definitions in pcre.h while keeping the internal definitions of the structures
2756 private.
2757
2758 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
2759 user point of view, this means that both static and shared libraries are built
2760 by default, but this can be individually controlled. More of the work of
2761 handling this static/shared cases is now inside libtool instead of PCRE's make
2762 file.
2763
2764 4. The pcretest utility is now installed along with pcregrep because it is
2765 useful for users (to test regexs) and by doing this, it automatically gets
2766 relinked by libtool. The documentation has been turned into a man page, so
2767 there are now .1, .txt, and .html versions in /doc.
2768
2769 5. Upgrades to pcregrep:
2770 (i) Added long-form option names like gnu grep.
2771 (ii) Added --help to list all options with an explanatory phrase.
2772 (iii) Added -r, --recursive to recurse into sub-directories.
2773 (iv) Added -f, --file to read patterns from a file.
2774
2775 6. pcre_exec() was referring to its "code" argument before testing that
2776 argument for NULL (and giving an error if it was NULL).
2777
2778 7. Upgraded Makefile.in to allow for compiling in a different directory from
2779 the source directory.
2780
2781 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
2782 options bits, the pointer it was passed was to an int instead of to an unsigned
2783 long int. This mattered only on 64-bit systems.
2784
2785 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
2786 generated) instead of pcre.in, which it its source. Also made the same change
2787 in several of the .c files.
2788
2789 10. A new release of gcc defines printf() as a macro, which broke pcretest
2790 because it had an ifdef in the middle of a string argument for printf(). Fixed
2791 by using separate calls to printf().
2792
2793 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
2794 script, to force use of CR or LF instead of \n in the source. On non-Unix
2795 systems, the value can be set in config.h.
2796
2797 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
2798 absolute limit. Changed the text of the error message to make this clear, and
2799 likewise updated the man page.
2800
2801 13. The limit of 99 on the number of capturing subpatterns has been removed.
2802 The new limit is 65535, which I hope will not be a "real" limit.
2803
2804
2805 Version 3.4 22-Aug-00
2806 ---------------------
2807
2808 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
2809
2810 2. Diagnose condition (?(0) as an error instead of crashing on matching.
2811
2812
2813 Version 3.3 01-Aug-00
2814 ---------------------
2815
2816 1. If an octal character was given, but the value was greater than \377, it
2817 was not getting masked to the least significant bits, as documented. This could
2818 lead to crashes in some systems.
2819
2820 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
2821 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
2822
2823 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
2824 These just pass their arguments on to (pcre_free)(), but they are provided
2825 because some uses of PCRE bind it to non-C systems that can call its functions,
2826 but cannot call free() or pcre_free() directly.
2827
2828 4. Add "make test" as a synonym for "make check". Corrected some comments in
2829 the Makefile.
2830
2831 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
2832 Makefile.
2833
2834 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
2835 command called pgrep for grepping around the active processes.
2836
2837 7. Added the beginnings of support for UTF-8 character strings.
2838
2839 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
2840 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
2841 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
2842 out for the ar command.)
2843
2844
2845 Version 3.2 12-May-00
2846 ---------------------
2847
2848 This is purely a bug fixing release.
2849
2850 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
2851 of ZA. This was just one example of several cases that could provoke this bug,
2852 which was introduced by change 9 of version 2.00. The code for breaking
2853 infinite loops after an iteration that matches an empty string was't working
2854 correctly.
2855
2856 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
2857 when matched against abbab (for example). After matching an empty string, it
2858 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
2859 caused it to match further down the string than it should.
2860
2861 3. The code contained an inclusion of sys/types.h. It isn't clear why this
2862 was there because it doesn't seem to be needed, and it causes trouble on some
2863 systems, as it is not a Standard C header. It has been removed.
2864
2865 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
2866 were reported on the Macintosh. The changes were from
2867
2868 while ((c = *(++ptr)) != 0 && c != '\n');
2869 to
2870 while ((c = *(++ptr)) != 0 && c != '\n') ;
2871
2872 Totally extraordinary, but if that's what it takes...
2873
2874 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
2875 available. Added HAVE_BCOPY and an autoconf test for it; if neither
2876 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
2877 assumes the way PCRE uses memmove() (always moving upwards).
2878
2879 6. PCRE is being used in one environment where strchr() is not available. There
2880 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
2881 faster code anyway.
2882
2883
2884 Version 3.1 09-Feb-00
2885 ---------------------
2886
2887 The only change in this release is the fixing of some bugs in Makefile.in for
2888 the "install" target:
2889
2890 (1) It was failing to install pcreposix.h.
2891
2892 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
2893
2894
2895 Version 3.0 01-Feb-00
2896 ---------------------
2897
2898 1. Add support for the /+ modifier to perltest (to output $` like it does in
2899 pcretest).
2900
2901 2. Add support for the /g modifier to perltest.
2902
2903 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
2904 matches null strings.
2905
2906 4. Fix perltest so that it doesn't do unwanted things when fed an empty
2907 pattern. Perl treats empty patterns specially - it reuses the most recent
2908 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
2909 effect.
2910
2911 5. The POSIX interface was broken in that it was just handing over the POSIX
2912 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
2913 required a bigger vector, with some working space on the end. This means that
2914 the POSIX wrapper now has to get and free some memory, and copy the results.
2915
2916 6. Added some simple autoconf support, placing the test data and the
2917 documentation in separate directories, re-organizing some of the
2918 information files, and making it build pcre-config (a GNU standard). Also added
2919 libtool support for building PCRE as a shared library, which is now the
2920 default.
2921
2922 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
2923 09 are not valid octal constants. Single digits will be used for minor values
2924 less than 10.
2925
2926 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
2927 existing programs that set these in the POSIX interface can use PCRE without
2928 modification.
2929
2930 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
2931 return all that pcre_info() returns, plus additional data. The pcre_info()
2932 function is retained for compatibility, but is considered to be obsolete.
2933
2934 10. Added experimental recursion feature (?R) to handle one common case that
2935 Perl 5.6 will be able to do with (?p{...}).
2936
2937 11. Added support for POSIX character classes like [:alpha:], which Perl is
2938 adopting.
2939
2940
2941 Version 2.08 31-Aug-99
2942 ----------------------
2943
2944 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
2945 trying to match at the startoffset position, but instead was moving forward to
2946 the next newline as if a previous match had failed.
2947
2948 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
2949 and could get into a loop if a null string was matched other than at the start
2950 of the subject.
2951
2952 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
2953 be distinguished at compile time, and for completeness also added PCRE_DATE.
2954
2955 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
2956 in GnuWin32 environments.
2957
2958
2959 Version 2.07 29-Jul-99
2960 ----------------------
2961
2962 1. The documentation is now supplied in plain text form and HTML as well as in
2963 the form of man page sources.
2964
2965 2. C++ compilers don't like assigning (void *) values to other pointer types.
2966 In particular this affects malloc(). Although there is no problem in Standard
2967 C, I've put in casts to keep C++ compilers happy.
2968
2969 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
2970 should be (const char *).
2971
2972 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
2973 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
2974 However, I haven't made this a standard facility. The documentation doesn't
2975 mention it, and the Makefile doesn't support it.
2976
2977 5. The Makefile now contains an "install" target, with editable destinations at
2978 the top of the file. The pcretest program is not installed.
2979
2980 6. pgrep -V now gives the PCRE version number and date.
2981
2982 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
2983 causing the entire string to be ignored, instead of just the last character.
2984
2985 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
2986 non-matching string, it can take a very, very long time, even for strings of
2987 quite modest length, because of the nested recursion. PCRE now does better in
2988 some of these cases. It does this by remembering the last required literal
2989 character in the pattern, and pre-searching the subject to ensure it is present
2990 before running the real match. In other words, it applies a heuristic to detect
2991 some types of certain failure quickly, and in the above example, if presented
2992 with a string that has no trailing " it gives "no match" very quickly.
2993
2994 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
2995 other alternatives are tried instead.
2996
2997
2998 Version 2.06 09-Jun-99
2999 ----------------------
3000
3001 1. Change pcretest's output for amount of store used to show just the code
3002 space, because the remainder (the data block) varies in size between 32-bit and
3003 64-bit systems.
3004
3005 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
3006 start matching at. This allows lookbehinds to work when searching for multiple
3007 occurrences in a string.
3008
3009 3. Added additional options to pcretest for testing multiple occurrences:
3010
3011 /+ outputs the rest of the string that follows a match
3012 /g loops for multiple occurrences, using the new startoffset argument
3013 /G loops for multiple occurrences by passing an incremented pointer
3014
3015 4. PCRE wasn't doing the "first character" optimization for patterns starting
3016 with \b or \B, though it was doing it for other lookbehind assertions. That is,
3017 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
3018 the letter 'x'. On long subject strings, this gives a significant speed-up.
3019
3020
3021 Version 2.05 21-Apr-99
3022 ----------------------
3023
3024 1. Changed the type of magic_number from int to long int so that it works
3025 properly on 16-bit systems.
3026
3027 2. Fixed a bug which caused patterns starting with .* not to work correctly
3028 when the subject string contained newline characters. PCRE was assuming
3029 anchoring for such patterns in all cases, which is not correct because .* will
3030 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
3031 DOTALL is set at top level; otherwise it knows that patterns starting with .*
3032 must be retried after every newline in the subject.
3033
3034
3035 Version 2.04 18-Feb-99
3036 ----------------------
3037
3038 1. For parenthesized subpatterns with repeats whose minimum was zero, the
3039 computation of the store needed to hold the pattern was incorrect (too large).
3040 If such patterns were nested a few deep, this could multiply and become a real
3041 problem.
3042
3043 2. Added /M option to pcretest to show the memory requirement of a specific
3044 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
3045
3046 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
3047 compiled in such a way that the backtracking after subsequent failure was
3048 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
3049 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
3050
3051
3052 Version 2.03 02-Feb-99
3053 ----------------------
3054
3055 1. Fixed typo and small mistake in man page.
3056
3057 2. Added 4th condition (GPL supersedes if conflict) and created separate
3058 LICENCE file containing the conditions.
3059
3060 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
3061 Perl, that is the internal \ allows the delimiter to be included in the
3062 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
3063 the final delimiter, add \ to the end of the pattern (to test the error).
3064
3065 4. Added the convenience functions for extracting substrings after a successful
3066 match. Updated pcretest to make it able to test these functions.
3067
3068
3069 Version 2.02 14-Jan-99
3070 ----------------------
3071
3072 1. Initialized the working variables associated with each extraction so that
3073 their saving and restoring doesn't refer to uninitialized store.
3074
3075 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
3076 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
3077 fix the problem.
3078
3079 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
3080 calls, and wasn't printing the correct value for compiling calls. Increased the
3081 default value of LOOPREPEAT, and the number of significant figures in the
3082 times.
3083
3084 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
3085
3086 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
3087 a building problem on Windows NT with a FAT file system.
3088
3089
3090 Version 2.01 21-Oct-98
3091 ----------------------
3092
3093 1. Changed the API for pcre_compile() to allow for the provision of a pointer
3094 to character tables built by pcre_maketables() in the current locale. If NULL
3095 is passed, the default tables are used.
3096
3097
3098 Version 2.00 24-Sep-98
3099 ----------------------
3100
3101 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
3102 it any more.
3103
3104 2. Allow quantification of (?>) groups, and make it work correctly.
3105
3106 3. The first character computation wasn't working for (?>) groups.
3107
3108 4. Correct the implementation of \Z (it is permitted to match on the \n at the
3109 end of the subject) and add 5.005's \z, which really does match only at the
3110 very end of the subject.
3111
3112 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
3113
3114 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
3115 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
3116 localized options. All options to pcre_study() were also removed.
3117
3118 7. Add other new features from 5.005:
3119
3120 $(?<= positive lookbehind
3121 $(?<! negative lookbehind
3122 (?imsx-imsx) added the unsetting capability
3123 such a setting is global if at outer level; local otherwise
3124 (?imsx-imsx:) non-capturing groups with option setting
3125 (?(cond)re|re) conditional pattern matching
3126
3127 A backreference to itself in a repeated group matches the previous
3128 captured string.
3129
3130 8. General tidying up of studying (both automatic and via "study")
3131 consequential on the addition of new assertions.
3132
3133 9. As in 5.005, unlimited repeated groups that could match an empty substring
3134 are no longer faulted at compile time. Instead, the loop is forcibly broken at
3135 runtime if any iteration does actually match an empty substring.
3136
3137 10. Include the RunTest script in the distribution.
3138
3139 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
3140 discrepancies, some of which were old and were also with respect to 5.004. They
3141 have now been fixed.
3142
3143
3144 Version 1.09 28-Apr-98
3145 ----------------------
3146
3147 1. A negated single character class followed by a quantifier with a minimum
3148 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
3149 program crashes, or just wrong answers. This did not apply to negated classes
3150 containing more than one character, or to minima other than one.
3151
3152
3153 Version 1.08 27-Mar-98
3154 ----------------------
3155
3156 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
3157
3158 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
3159 latter must appear before anything that relies on it in the pattern.
3160
3161
3162 Version 1.07 16-Feb-98
3163 ----------------------
3164
3165 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
3166 repeat of a potentially empty string).
3167
3168
3169 Version 1.06 23-Jan-98
3170 ----------------------
3171
3172 1. Added Markus Oberhumer's little patches for C++.
3173
3174 2. Literal strings longer than 255 characters were broken.
3175
3176
3177 Version 1.05 23-Dec-97
3178 ----------------------
3179
3180 1. Negated character classes containing more than one character were failing if
3181 PCRE_CASELESS was set at run time.
3182
3183
3184 Version 1.04 19-Dec-97
3185 ----------------------
3186
3187 1. Corrected the man page, where some "const" qualifiers had been omitted.
3188
3189 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
3190 input syntax.
3191
3192 3. Fixed memory leak which occurred when a regex with back references was
3193 matched with an offsets vector that wasn't big enough. The temporary memory
3194 that is used in this case wasn't being freed if the match failed.
3195
3196 4. Tidied pcretest to ensure it frees memory that it gets.
3197
3198 5. Temporary memory was being obtained in the case where the passed offsets
3199 vector was exactly big enough.
3200
3201 6. Corrected definition of offsetof() from change 5 below.
3202
3203 7. I had screwed up change 6 below and broken the rules for the use of
3204 setjmp(). Now fixed.
3205
3206
3207 Version 1.03 18-Dec-97
3208 ----------------------
3209
3210 1. A erroneous regex with a missing opening parenthesis was correctly
3211 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
3212 on some systems.
3213
3214 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
3215 it was reported that one broken compiler failed on the former because "code" is
3216 also an independent variable.
3217
3218 3. The erroneous regex a[]b caused an array overrun reference.
3219
3220 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
3221 fail on data ending with that character. (It was going on too far, and checking
3222 the next character, typically a binary zero.) This was specific to the
3223 optimized code for single-character negative classes.
3224
3225 5. Added a contributed patch from the TIN world which does the following:
3226
3227 + Add an undef for memmove, in case the the system defines a macro for it.
3228
3229 + Add a definition of offsetof(), in case there isn't one. (I don't know
3230 the reason behind this - offsetof() is part of the ANSI standard - but
3231 it does no harm).
3232
3233 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
3234 most of the places where whitespace preceded '#'. I have given up and
3235 allowed the remaining 2 cases to be at the margin.
3236
3237 + Rename some variables in pcre to eliminate shadowing. This seems very
3238 pedantic, but does no harm, of course.
3239
3240 6. Moved the call to setjmp() into its own function, to get rid of warnings
3241 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
3242
3243 7. Constructs such as \d{8,} were compiling into the equivalent of
3244 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
3245 outcome, but in this particular case used more store than had been allocated,
3246 which caused the bug to be discovered because it threw up an internal error.
3247
3248 8. The debugging code in both pcre and pcretest for outputting the compiled
3249 form of a regex was going wrong in the case of back references followed by
3250 curly-bracketed repeats.
3251
3252
3253 Version 1.02 12-Dec-97
3254 ----------------------
3255
3256 1. Typos in pcre.3 and comments in the source fixed.
3257
3258 2. Applied a contributed patch to get rid of places where it used to remove
3259 'const' from variables, and fixed some signed/unsigned and uninitialized
3260 variable warnings.
3261
3262 3. Added the "runtest" target to Makefile.
3263
3264 4. Set default compiler flag to -O2 rather than just -O.
3265
3266
3267 Version 1.01 19-Nov-97
3268 ----------------------
3269
3270 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
3271 like /([ab]*)*/, that is, for classes with more than one character in them.
3272
3273 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
3274 as /((?>a*))*/ (a PCRE_EXTRA facility).
3275
3276
3277 Version 1.00 18-Nov-97
3278 ----------------------
3279
3280 1. Added compile-time macros to support systems such as SunOS4 which don't have
3281 memmove() or strerror() but have other things that can be used instead.
3282
3283 2. Arranged that "make clean" removes the executables.
3284
3285
3286 Version 0.99 27-Oct-97
3287 ----------------------
3288
3289 1. Fixed bug in code for optimizing classes with only one character. It was
3290 initializing a 32-byte map regardless, which could cause it to run off the end
3291 of the memory it had got.
3292
3293 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
3294
3295
3296 Version 0.98 22-Oct-97
3297 ----------------------
3298
3299 1. Fixed bug in code for handling temporary memory usage when there are more
3300 back references than supplied space in the ovector. This could cause segfaults.
3301
3302
3303 Version 0.97 21-Oct-97
3304 ----------------------
3305
3306 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
3307
3308 2. Optimized negated single characters not to use a bit map.
3309
3310 3. Brought error texts together as macro definitions; clarified some of them;
3311 fixed one that was wrong - it said "range out of order" when it meant "invalid
3312 escape sequence".
3313
3314 4. Changed some char * arguments to const char *.
3315
3316 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
3317
3318 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
3319 pcretest.
3320
3321
3322 Version 0.96 16-Oct-97
3323 ----------------------
3324
3325 1. Added a simple "pgrep" utility to the distribution.
3326
3327 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
3328 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
3329 where "ddd" means "one or more decimal digits".
3330
3331 3. Fixed serious bug. If a pattern had a back reference, but the call to
3332 pcre_exec() didn't supply a large enough ovector to record the related
3333 identifying subpattern, the match always failed. PCRE now remembers the number
3334 of the largest back reference, and gets some temporary memory in which to save
3335 the offsets during matching if necessary, in order to ensure that
3336 backreferences always work.
3337
3338 4. Increased the compatibility with Perl in a number of ways:
3339
3340 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
3341 to request this handling. The option can be set at compile or exec time.
3342
3343 (b) $ matches before a terminating newline by default; an option
3344 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
3345 mode). The option can be set at compile or exec time.
3346
3347 (c) The handling of \ followed by a digit other than 0 is now supposed to be
3348 the same as Perl's. If the decimal number it represents is less than 10
3349 or there aren't that many previous left capturing parentheses, an octal
3350 escape is read. Inside a character class, it's always an octal escape,
3351 even if it is a single digit.
3352
3353 (d) An escaped but undefined alphabetic character is taken as a literal,
3354 unless PCRE_EXTRA is set. Currently this just reserves the remaining
3355 escapes.
3356
3357 (e) {0} is now permitted. (The previous item is removed from the compiled
3358 pattern).
3359
3360 5. Changed all the names of code files so that the basic parts are no longer
3361 than 10 characters, and abolished the teeny "globals.c" file.
3362
3363 6. Changed the handling of character classes; they are now done with a 32-byte
3364 bit map always.
3365
3366 7. Added the -d and /D options to pcretest to make it possible to look at the
3367 internals of compilation without having to recompile pcre.
3368
3369
3370 Version 0.95 23-Sep-97
3371 ----------------------
3372
3373 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
3374 \x20 at the start of a run of normal characters. These were being treated as
3375 real characters, instead of the source characters being re-checked.
3376
3377
3378 Version 0.94 18-Sep-97
3379 ----------------------
3380
3381 1. The functions are now thread-safe, with the caveat that the global variables
3382 containing pointers to malloc() and free() or alternative functions are the
3383 same for all threads.
3384
3385 2. Get pcre_study() to generate a bitmap of initial characters for non-
3386 anchored patterns when this is possible, and use it if passed to pcre_exec().
3387
3388
3389 Version 0.93 15-Sep-97
3390 ----------------------
3391
3392 1. /(b)|(:+)/ was computing an incorrect first character.
3393
3394 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
3395 but not actually doing anything yet.
3396
3397 3. Treat "-" characters in classes that cannot be part of ranges as literals,
3398 as Perl does (e.g. [-az] or [az-]).
3399
3400 4. Set the anchored flag if a branch starts with .* or .*? because that tests
3401 all possible positions.
3402
3403 5. Split up into different modules to avoid including unneeded functions in a
3404 compiled binary. However, compile and exec are still in one module. The "study"
3405 function is split off.
3406
3407 6. The character tables are now in a separate module whose source is generated
3408 by an auxiliary program - but can then be edited by hand if required. There are
3409 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
3410 toupper() in the code.
3411
3412 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
3413 make them global. Abolish the function for setting them, as the caller can now
3414 set them directly.
3415
3416
3417 Version 0.92 11-Sep-97
3418 ----------------------
3419
3420 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
3421 (e.g. /a{1,3}/) was broken (I mis-optimized it).
3422
3423 2. Caseless matching was not working in character classes if the characters in
3424 the pattern were in upper case.
3425
3426 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
3427
3428 4. Make PCRE_ANCHORED public and accept as a compile option.
3429
3430 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
3431 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
3432 pass them.
3433
3434 6. Give an error if bad option bits passed at compile or run time.
3435
3436 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
3437 pcretest to cause it to pass that flag.
3438
3439 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
3440 options, and the first character, if set.
3441
3442 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
3443
3444
3445 Version 0.91 10-Sep-97
3446 ----------------------
3447
3448 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
3449 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
3450
3451 2. PCRE was looping on encountering an indefinitely repeated back reference to
3452 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
3453 Perl does - treats the match as successful.
3454
3455 ****

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