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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 396 - (show annotations) (download)
Fri Mar 20 11:23:20 2009 UTC (5 years, 8 months ago) by ph10
File size: 163360 byte(s)
Forgot to commit this one.

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 19. A conditional group that had only one branch was not being correctly
73 recognized as an item that could match an empty string. This meant that an
74 enclosing group might also not be so recognized, causing infinite looping
75 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
76 with the subject "ab", where knowledge that the repeated group can match
77 nothing is needed in order to break the loop.
78
79
80 Version 7.8 05-Sep-08
81 ---------------------
82
83 1. Replaced UCP searching code with optimized version as implemented for Ad
84 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
85 stage table and inline lookup instead of a function, giving speed ups of 2
86 to 5 times on some simple patterns that I tested. Permission was given to
87 distribute the MultiStage2.py script that generates the tables (it's not in
88 the tarball, but is in the Subversion repository).
89
90 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
91 scripts.
92
93 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
94 a group with a zero qualifier. The result of the study could be incorrect,
95 or the function might crash, depending on the pattern.
96
97 4. Caseless matching was not working for non-ASCII characters in back
98 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
99 It now works when Unicode Property Support is available.
100
101 5. In pcretest, an escape such as \x{de} in the data was always generating
102 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
103 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
104 truncation.
105
106 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
107
108 7. Added two (int) casts to pcregrep when printing the difference of two
109 pointers, in case they are 64-bit values.
110
111 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
112 test 2 if it fails.
113
114 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
115 and a #define of that name to empty if it is not externally set. This is to
116 allow users of MSVC to set it if necessary.
117
118 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
119 the convenience functions in the pcre_get.c source file.
120
121 11. An option change at the start of a pattern that had top-level alternatives
122 could cause overwriting and/or a crash. This command provoked a crash in
123 some environments:
124
125 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
126
127 This potential security problem was recorded as CVE-2008-2371.
128
129 12. For a pattern where the match had to start at the beginning or immediately
130 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
131 pcre_dfa_exec() could read past the end of the passed subject if there was
132 no match. To help with detecting such bugs (e.g. with valgrind), I modified
133 pcretest so that it places the subject at the end of its malloc-ed buffer.
134
135 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
136 exec() might read past the end of the data buffer in UTF-8 mode.
137
138 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
139 the data contained the byte 0x85 as part of a UTF-8 character within its
140 first line. This applied both to normal and DFA matching.
141
142 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
143 /^[^d]*?$/8 failed to match "abc".
144
145 16. Added a missing copyright notice to pcrecpp_internal.h.
146
147 17. Make it more clear in the documentation that values returned from
148 pcre_exec() in ovector are byte offsets, not character counts.
149
150 18. Tidied a few places to stop certain compilers from issuing warnings.
151
152 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
153 supplied by Stefan Weber. I made a further small update for 7.8 because
154 there is a change of source arrangements: the pcre_searchfuncs.c module is
155 replaced by pcre_ucd.c.
156
157
158 Version 7.7 07-May-08
159 ---------------------
160
161 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
162 a string to a long long, pretend we don't even have a long long." This is
163 done by checking for the strtoq, strtoll, and _strtoi64 functions.
164
165 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
166 pre-7.6 versions, which defined a global no_arg variable instead of putting
167 it in the RE class. (See also #8 below.)
168
169 3. Remove a line of dead code, identified by coverity and reported by Nuno
170 Lopes.
171
172 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
173
174 (1) The include/exclude patterns were being applied to the whole pathnames
175 of files, instead of just to the final components.
176
177 (2) If there was more than one level of directory, the subdirectories were
178 skipped unless they satisfied the include/exclude conditions. This is
179 inconsistent with GNU grep (and could even be seen as contrary to the
180 pcregrep specification - which I improved to make it absolutely clear).
181 The action now is always to scan all levels of directory, and just
182 apply the include/exclude patterns to regular files.
183
184 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
185 --exclude_dir in the tests to avoid scanning .svn directories.
186
187 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
188 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
189 doesn't support NULs in patterns.
190
191 7. Added some missing "const"s to declarations of static tables in
192 pcre_compile.c and pcre_dfa_exec.c.
193
194 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
195 caused by fix #2 above. (Subsequently also a second patch to fix the
196 first patch. And a third patch - this was a messy problem.)
197
198 9. Applied Craig's patch to remove the use of push_back().
199
200 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
201 matching function regexec().
202
203 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
204 which, however, unlike Perl's \g{...}, are subroutine calls, not back
205 references. PCRE supports relative numbers with this syntax (I don't think
206 Oniguruma does).
207
208 12. Previously, a group with a zero repeat such as (...){0} was completely
209 omitted from the compiled regex. However, this means that if the group
210 was called as a subroutine from elsewhere in the pattern, things went wrong
211 (an internal error was given). Such groups are now left in the compiled
212 pattern, with a new opcode that causes them to be skipped at execution
213 time.
214
215 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
216 to the way PCRE behaves:
217
218 (a) A lone ] character is dis-allowed (Perl treats it as data).
219
220 (b) A back reference to an unmatched subpattern matches an empty string
221 (Perl fails the current match path).
222
223 (c) A data ] in a character class must be notated as \] because if the
224 first data character in a class is ], it defines an empty class. (In
225 Perl it is not possible to have an empty class.) The empty class []
226 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
227 The negative empty class [^] matches any one character, independently
228 of the DOTALL setting.
229
230 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
231 non-existent subpattern following a character class starting with ']' and
232 containing () gave an internal compiling error instead of "reference to
233 non-existent subpattern". Fortunately, when the pattern did exist, the
234 compiled code was correct. (When scanning forwards to check for the
235 existencd of the subpattern, it was treating the data ']' as terminating
236 the class, so got the count wrong. When actually compiling, the reference
237 was subsequently set up correctly.)
238
239 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
240 it was being rejected as not supported by pcre_dfa_exec(), even though
241 other assertions are supported. I have made pcre_dfa_exec() support
242 (*FAIL).
243
244 16. The implementation of 13c above involved the invention of a new opcode,
245 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
246 cannot be changed at match time, I realized I could make a small
247 improvement to matching performance by compiling OP_ALLANY instead of
248 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
249 on the OP_ANY path.
250
251 17. Compiling pcretest on Windows with readline support failed without the
252 following two fixes: (1) Make the unistd.h include conditional on
253 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
254
255 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
256 ncurses library to be included for pcretest when ReadLine support is
257 requested, but also to allow for it to be overridden. This patch came from
258 Daniel Bergström.
259
260 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
261 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
262 any errors with the current Unicode tables. Thanks to Peter Kankowski for
263 spotting this.
264
265
266 Version 7.6 28-Jan-08
267 ---------------------
268
269 1. A character class containing a very large number of characters with
270 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
271 overflow.
272
273 2. Patch to cut out the "long long" test in pcrecpp_unittest when
274 HAVE_LONG_LONG is not defined.
275
276 3. Applied Christian Ehrlicher's patch to update the CMake build files to
277 bring them up to date and include new features. This patch includes:
278
279 - Fixed PH's badly added libz and libbz2 support.
280 - Fixed a problem with static linking.
281 - Added pcredemo. [But later removed - see 7 below.]
282 - Fixed dftables problem and added an option.
283 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
284 HAVE_LONG_LONG.
285 - Added readline support for pcretest.
286 - Added an listing of the option settings after cmake has run.
287
288 4. A user submitted a patch to Makefile that makes it easy to create
289 "pcre.dll" under mingw when using Configure/Make. I added stuff to
290 Makefile.am that cause it to include this special target, without
291 affecting anything else. Note that the same mingw target plus all
292 the other distribution libraries and programs are now supported
293 when configuring with CMake (see 6 below) instead of with
294 Configure/Make.
295
296 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
297 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
298 exported in the Windows port". It has not yet been confirmed that the patch
299 solves the problem, but it does no harm.
300
301 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
302 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
303 with CMake, and also correct the comment about stack recursion.
304
305 7. Remove the automatic building of pcredemo from the ./configure system and
306 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
307 of a program that users should build themselves after PCRE is installed, so
308 building it automatically is not really right. What is more, it gave
309 trouble in some build environments.
310
311 8. Further tidies to CMakeLists.txt from Sheri and Christian.
312
313
314 Version 7.5 10-Jan-08
315 ---------------------
316
317 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
318 values in parens when parsing an RE using the C++ wrapper."
319
320 2. Negative specials like \S did not work in character classes in UTF-8 mode.
321 Characters greater than 255 were excluded from the class instead of being
322 included.
323
324 3. The same bug as (2) above applied to negated POSIX classes such as
325 [:^space:].
326
327 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
328 defined or documented. It seems to have been a typo for PCRE_STATIC, so
329 I have changed it.
330
331 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
332 first named subpattern) and a construct such as (?&a) would reference the
333 first named subpattern whose name started with "a" (in other words, the
334 length check was missing). Both these problems are fixed. "Subpattern name
335 expected" is now given for (?&) (a zero-length name), and this patch also
336 makes it give the same error for \k'' (previously it complained that that
337 was a reference to a non-existent subpattern).
338
339 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
340 this is right because (?- can be followed by option settings as well as by
341 digits. I have, however, made the messages clearer.
342
343 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
344 than the number used in the conditional) now cause a compile-time error.
345 This is actually not compatible with Perl, which accepts such patterns, but
346 treats the conditional as always being FALSE (as PCRE used to), but it
347 seems to me that giving a diagnostic is better.
348
349 8. Change "alphameric" to the more common word "alphanumeric" in comments
350 and messages.
351
352 9. Fix two occurrences of "backslash" in comments that should have been
353 "backspace".
354
355 10. Remove two redundant lines of code that can never be obeyed (their function
356 was moved elsewhere).
357
358 11. The program that makes PCRE's Unicode character property table had a bug
359 which caused it to generate incorrect table entries for sequences of
360 characters that have the same character type, but are in different scripts.
361 It amalgamated them into a single range, with the script of the first of
362 them. In other words, some characters were in the wrong script. There were
363 thirteen such cases, affecting characters in the following ranges:
364
365 U+002b0 - U+002c1
366 U+0060c - U+0060d
367 U+0061e - U+00612
368 U+0064b - U+0065e
369 U+0074d - U+0076d
370 U+01800 - U+01805
371 U+01d00 - U+01d77
372 U+01d9b - U+01dbf
373 U+0200b - U+0200f
374 U+030fc - U+030fe
375 U+03260 - U+0327f
376 U+0fb46 - U+0fbb1
377 U+10450 - U+1049d
378
379 12. The -o option (show only the matching part of a line) for pcregrep was not
380 compatible with GNU grep in that, if there was more than one match in a
381 line, it showed only the first of them. It now behaves in the same way as
382 GNU grep.
383
384 13. If the -o and -v options were combined for pcregrep, it printed a blank
385 line for every non-matching line. GNU grep prints nothing, and pcregrep now
386 does the same. The return code can be used to tell if there were any
387 non-matching lines.
388
389 14. Added --file-offsets and --line-offsets to pcregrep.
390
391 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
392 infinitely looping recursion. The bug was that positive lookaheads were not
393 being skipped when checking for a possible empty match (negative lookaheads
394 and both kinds of lookbehind were skipped).
395
396 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
397 inclusion of <windows.h> to before rather than after the definition of
398 INVALID_FILE_ATTRIBUTES (patch from David Byron).
399
400 17. Specifying a possessive quantifier with a specific limit for a Unicode
401 character property caused pcre_compile() to compile bad code, which led at
402 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
403 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
404 caused the error; without that there was no problem.
405
406 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
407
408 19. Added --enable-pcretest-libreadline.
409
410 20. In pcrecpp.cc, the variable 'count' was incremented twice in
411 RE::GlobalReplace(). As a result, the number of replacements returned was
412 double what it should be. I removed one of the increments, but Craig sent a
413 later patch that removed the other one (the right fix) and added unit tests
414 that check the return values (which was not done before).
415
416 21. Several CMake things:
417
418 (1) Arranged that, when cmake is used on Unix, the libraries end up with
419 the names libpcre and libpcreposix, not just pcre and pcreposix.
420
421 (2) The above change means that pcretest and pcregrep are now correctly
422 linked with the newly-built libraries, not previously installed ones.
423
424 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
425
426 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
427 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
428 UTF-8 newline character). The key issue is that the pattern starts .*;
429 this means that the match must be either at the beginning, or after a
430 newline. The bug was in the code for advancing after a failed match and
431 checking that the new position followed a newline. It was not taking
432 account of UTF-8 characters correctly.
433
434 23. PCRE was behaving differently from Perl in the way it recognized POSIX
435 character classes. PCRE was not treating the sequence [:...:] as a
436 character class unless the ... were all letters. Perl, however, seems to
437 allow any characters between [: and :], though of course it rejects as
438 unknown any "names" that contain non-letters, because all the known class
439 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
440 for example, whereas PCRE did not - it did not recognize a POSIX character
441 class. This seemed a bit dangerous, so the code has been changed to be
442 closer to Perl. The behaviour is not identical to Perl, because PCRE will
443 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
444 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
445 Perl does, and where it didn't before.
446
447 24. Rewrite so as to remove the single use of %n from pcregrep because in some
448 Windows environments %n is disabled by default.
449
450
451 Version 7.4 21-Sep-07
452 ---------------------
453
454 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
455 means that a class such as [\s] counted as "explicit reference to CR or
456 LF". That isn't really right - the whole point of the change was to try to
457 help when there was an actual mention of one of the two characters. So now
458 the change happens only if \r or \n (or a literal CR or LF) character is
459 encountered.
460
461 2. The 32-bit options word was also used for 6 internal flags, but the numbers
462 of both had grown to the point where there were only 3 bits left.
463 Fortunately, there was spare space in the data structure, and so I have
464 moved the internal flags into a new 16-bit field to free up more option
465 bits.
466
467 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
468 but did not set the internal JCHANGED flag - either of these is enough to
469 control the way the "get" function works - but the PCRE_INFO_JCHANGED
470 facility is supposed to tell if (?J) was ever used, so now (?J) at the
471 start sets both bits.
472
473 4. Added options (at build time, compile time, exec time) to change \R from
474 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
475
476 5. doc/pcresyntax.html was missing from the distribution.
477
478 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
479 compatibility, even though it is no longer used.
480
481 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
482 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
483 windows.h file is present (where different names are used). [This was
484 reversed later after testing - see 16 below.]
485
486 8. Changed all #include <config.h> to #include "config.h". There were also
487 some further <pcre.h> cases that I changed to "pcre.h".
488
489 9. When pcregrep was used with the --colour option, it missed the line ending
490 sequence off the lines that it output.
491
492 10. It was pointed out to me that arrays of string pointers cause lots of
493 relocations when a shared library is dynamically loaded. A technique of
494 using a single long string with a table of offsets can drastically reduce
495 these. I have refactored PCRE in four places to do this. The result is
496 dramatic:
497
498 Originally: 290
499 After changing UCP table: 187
500 After changing error message table: 43
501 After changing table of "verbs" 36
502 After changing table of Posix names 22
503
504 Thanks to the folks working on Gregex for glib for this insight.
505
506 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
507 unicode-properties was also set.
508
509 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
510
511 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
512 checked only for CRLF.
513
514 14. Added casts to pcretest.c to avoid compiler warnings.
515
516 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
517
518 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
519 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
520 entirely. This removes changes made in 7 above.
521
522 17. The CMake files have been updated, and there is now more information about
523 building with CMake in the NON-UNIX-USE document.
524
525
526 Version 7.3 28-Aug-07
527 ---------------------
528
529 1. In the rejigging of the build system that eventually resulted in 7.1, the
530 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
531 brackets there is not right, since it causes compilers to look for an
532 installed pcre.h, not the version that is in the source that is being
533 compiled (which of course may be different). I have changed it back to:
534
535 #include "pcre.h"
536
537 I have a vague recollection that the change was concerned with compiling in
538 different directories, but in the new build system, that is taken care of
539 by the VPATH setting the Makefile.
540
541 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
542 when the subject happened to end in the byte 0x85 (e.g. if the last
543 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
544 characters but of course it shouldn't be taken as a newline when it is part
545 of another character. The bug was that, for an unlimited repeat of . in
546 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
547 characters when looking for a newline.
548
549 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
550
551 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
552 in debug output.
553
554 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
555 long printing in the pcrecpp unittest when running under MinGW.
556
557 6. ESC_K was left out of the EBCDIC table.
558
559 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
560 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
561 limit also applies to "virtual nesting" when a pattern is recursive, and in
562 this case 1000 isn't so big. I have been able to remove this limit at the
563 expense of backing off one optimization in certain circumstances. Normally,
564 when pcre_exec() would call its internal match() function recursively and
565 immediately return the result unconditionally, it uses a "tail recursion"
566 feature to save stack. However, when a subpattern that can match an empty
567 string has an unlimited repetition quantifier, it no longer makes this
568 optimization. That gives it a stack frame in which to save the data for
569 checking that an empty string has been matched. Previously this was taken
570 from the 1000-entry workspace that had been reserved. So now there is no
571 explicit limit, but more stack is used.
572
573 8. Applied Daniel's patches to solve problems with the import/export magic
574 syntax that is required for Windows, and which was going wrong for the
575 pcreposix and pcrecpp parts of the library. These were overlooked when this
576 problem was solved for the main library.
577
578 9. There were some crude static tests to avoid integer overflow when computing
579 the size of patterns that contain repeated groups with explicit upper
580 limits. As the maximum quantifier is 65535, the maximum group length was
581 set at 30,000 so that the product of these two numbers did not overflow a
582 32-bit integer. However, it turns out that people want to use groups that
583 are longer than 30,000 bytes (though not repeat them that many times).
584 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
585 made it possible to implement the integer overflow checks in a much more
586 dynamic way, which I have now done. The artificial limitation on group
587 length has been removed - we now have only the limit on the total length of
588 the compiled pattern, which depends on the LINK_SIZE setting.
589
590 10. Fixed a bug in the documentation for get/copy named substring when
591 duplicate names are permitted. If none of the named substrings are set, the
592 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
593 empty string.
594
595 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
596 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
597 because the ] is interpreted as the first data character and the
598 terminating ] is not found. PCRE has been made compatible with Perl in this
599 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
600 cause memory overwriting.
601
602 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
603 string has been matched (to stop an infinite loop). It was not recognizing
604 a conditional subpattern that could match an empty string if that
605 subpattern was within another subpattern. For example, it looped when
606 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
607 condition was not nested. This bug has been fixed.
608
609 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
610 past the start of the subject in the presence of bytes with the top bit
611 set, for example "\x8aBCD".
612
613 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
614 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
615
616 14. Optimized (?!) to (*FAIL).
617
618 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
619 This restricts code points to be within the range 0 to 0x10FFFF, excluding
620 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
621 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
622 does: it's just the validity check that is more restrictive.
623
624 16. Inserted checks for integer overflows during escape sequence (backslash)
625 processing, and also fixed erroneous offset values for syntax errors during
626 backslash processing.
627
628 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
629 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
630
631 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
632 caused an overrun.
633
634 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
635 something other than just ASCII characters) inside a group that had an
636 unlimited repeat caused a loop at compile time (while checking to see
637 whether the group could match an empty string).
638
639 20. Debugging a pattern containing \p or \P could cause a crash. For example,
640 [\P{Any}] did so. (Error in the code for printing property names.)
641
642 21. An orphan \E inside a character class could cause a crash.
643
644 22. A repeated capturing bracket such as (A)? could cause a wild memory
645 reference during compilation.
646
647 23. There are several functions in pcre_compile() that scan along a compiled
648 expression for various reasons (e.g. to see if it's fixed length for look
649 behind). There were bugs in these functions when a repeated \p or \P was
650 present in the pattern. These operators have additional parameters compared
651 with \d, etc, and these were not being taken into account when moving along
652 the compiled data. Specifically:
653
654 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
655 length.
656
657 (b) An item such as \pL+ within a repeated group could cause crashes or
658 loops.
659
660 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
661 "reference to non-existent subpattern" error.
662
663 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
664
665 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
666 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
667
668 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
669
670 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
671 character were causing crashes (broken optimization).
672
673 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
674 \p or \P) caused a compile-time loop.
675
676 28. More problems have arisen in unanchored patterns when CRLF is a valid line
677 break. For example, the unstudied pattern [\r\n]A does not match the string
678 "\r\nA" because change 7.0/46 below moves the current point on by two
679 characters after failing to match at the start. However, the pattern \nA
680 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
681 the same is true. There doesn't seem any very clean way out of this, but
682 what I have chosen to do makes the common cases work: PCRE now takes note
683 of whether there can be an explicit match for \r or \n anywhere in the
684 pattern, and if so, 7.0/46 no longer applies. As part of this change,
685 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
686 pattern has explicit CR or LF references.
687
688 29. Added (*CR) etc for changing newline setting at start of pattern.
689
690
691 Version 7.2 19-Jun-07
692 ---------------------
693
694 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
695 which is apparently normally available under Windows.
696
697 2. Re-jig the pcregrep tests with different newline settings in an attempt
698 to make them independent of the local environment's newline setting.
699
700 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
701
702 4. Some of the "internals" tests were previously cut out when the link size
703 was not 2, because the output contained actual offsets. The recent new
704 "Z" feature of pcretest means that these can be cut out, making the tests
705 usable with all link sizes.
706
707 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
708 stack recursion. This gives a massive performance boost under BSD, but just
709 a small improvement under Linux. However, it saves one field in the frame
710 in all cases.
711
712 6. Added more features from the forthcoming Perl 5.10:
713
714 (a) (?-n) (where n is a string of digits) is a relative subroutine or
715 recursion call. It refers to the nth most recently opened parentheses.
716
717 (b) (?+n) is also a relative subroutine call; it refers to the nth next
718 to be opened parentheses.
719
720 (c) Conditions that refer to capturing parentheses can be specified
721 relatively, for example, (?(-2)... or (?(+3)...
722
723 (d) \K resets the start of the current match so that everything before
724 is not part of it.
725
726 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
727
728 (f) \g{name} is another synonym - part of Perl 5.10's unification of
729 reference syntax.
730
731 (g) (?| introduces a group in which the numbering of parentheses in each
732 alternative starts with the same number.
733
734 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
735
736 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
737 PCRE_INFO_JCHANGED.
738
739 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
740 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
741 for detecting groups that can match an empty string.
742
743 9. A pattern with a very large number of alternatives (more than several
744 hundred) was running out of internal workspace during the pre-compile
745 phase, where pcre_compile() figures out how much memory will be needed. A
746 bit of new cunning has reduced the workspace needed for groups with
747 alternatives. The 1000-alternative test pattern now uses 12 bytes of
748 workspace instead of running out of the 4096 that are available.
749
750 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
751
752 11. Applied patch from Google to remove an optimization that didn't quite work.
753 The report of the bug said:
754
755 pcrecpp::RE("a*").FullMatch("aaa") matches, while
756 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
757 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
758
759 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
760 it matched the wrong number of bytes.
761
762
763 Version 7.1 24-Apr-07
764 ---------------------
765
766 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
767 that is more "standard", making use of automake and other Autotools. There
768 is some re-arrangement of the files and adjustment of comments consequent
769 on this.
770
771 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
772 for recursive directory scanning broke on some systems because the files
773 are not scanned in any specific order and on different systems the order
774 was different. A call to "sort" has been inserted into RunGrepTest for the
775 approprate test as a short-term fix. In the longer term there may be an
776 alternative.
777
778 3. I had an email from Eric Raymond about problems translating some of PCRE's
779 man pages to HTML (despite the fact that I distribute HTML pages, some
780 people do their own conversions for various reasons). The problems
781 concerned the use of low-level troff macros .br and .in. I have therefore
782 removed all such uses from the man pages (some were redundant, some could
783 be replaced by .nf/.fi pairs). The 132html script that I use to generate
784 HTML has been updated to handle .nf/.fi and to complain if it encounters
785 .br or .in.
786
787 4. Updated comments in configure.ac that get placed in config.h.in and also
788 arranged for config.h to be included in the distribution, with the name
789 config.h.generic, for the benefit of those who have to compile without
790 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
791
792 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
793 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
794 makevp.bat for the current PCRE, using the additional files
795 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
796
797 6. A Windows user reported a minor discrepancy with test 2, which turned out
798 to be caused by a trailing space on an input line that had got lost in his
799 copy. The trailing space was an accident, so I've just removed it.
800
801 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
802 that is needed.
803
804 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
805 as "const" (a) because they are and (b) because it helps the PHP
806 maintainers who have recently made a script to detect big data structures
807 in the php code that should be moved to the .rodata section. I remembered
808 to update Builducptable as well, so it won't revert if ucptable.h is ever
809 re-created.
810
811 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
812 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
813 order to be able to cut out the UTF-8 tables in the latter when UTF-8
814 support is not required. This saves 1.5-2K of code, which is important in
815 some applications.
816
817 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
818 so as not to refer to the tables, even though these functions will never be
819 called when UTF-8 support is disabled. Otherwise there are problems with a
820 shared library.
821
822 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
823
824 (a) It was defining its arguments as char * instead of void *.
825
826 (b) It was assuming that all moves were upwards in memory; this was true
827 a long time ago when I wrote it, but is no longer the case.
828
829 The emulated memove() is provided for those environments that have neither
830 memmove() nor bcopy(). I didn't think anyone used it these days, but that
831 is clearly not the case, as these two bugs were recently reported.
832
833 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
834 and Detrail to create the HTML documentation, the .txt form of the man
835 pages, and it removes trailing spaces from listed files. It also creates
836 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
837 case, it wraps all the #defines with #ifndefs. This script should be run
838 before "make dist".
839
840 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
841 with Unicode property support.
842
843 (a) For a maximizing quantifier, if the two different cases of the
844 character were of different lengths in their UTF-8 codings (there are
845 some cases like this - I found 11), and the matching function had to
846 back up over a mixture of the two cases, it incorrectly assumed they
847 were both the same length.
848
849 (b) When PCRE was configured to use the heap rather than the stack for
850 recursion during matching, it was not correctly preserving the data for
851 the other case of a UTF-8 character when checking ahead for a match
852 while processing a minimizing repeat. If the check also involved
853 matching a wide character, but failed, corruption could cause an
854 erroneous result when trying to check for a repeat of the original
855 character.
856
857 13. Some tidying changes to the testing mechanism:
858
859 (a) The RunTest script now detects the internal link size and whether there
860 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
861 values substituted by "configure". (The RunGrepTest script already did
862 this for UTF-8.) The configure.ac script no longer substitutes the
863 relevant variables.
864
865 (b) The debugging options /B and /D in pcretest show the compiled bytecode
866 with length and offset values. This means that the output is different
867 for different internal link sizes. Test 2 is skipped for link sizes
868 other than 2 because of this, bypassing the problem. Unfortunately,
869 there was also a test in test 3 (the locale tests) that used /B and
870 failed for link sizes other than 2. Rather than cut the whole test out,
871 I have added a new /Z option to pcretest that replaces the length and
872 offset values with spaces. This is now used to make test 3 independent
873 of link size. (Test 2 will be tidied up later.)
874
875 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
876 segmentation fault instead of returning the appropriate error message.
877
878 15. In multiline mode when the newline sequence was set to "any", the pattern
879 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
880 This doesn't seem right; it now treats the CRLF combination as the line
881 ending, and so does not match in that case. It's only a pattern such as ^$
882 that would hit this one: something like ^ABC$ would have failed after \r
883 and then tried again after \r\n.
884
885 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
886 in an attempt to make files that differ only in their line terminators
887 compare equal. This works on Linux.
888
889 17. Under certain error circumstances pcregrep might try to free random memory
890 as it exited. This is now fixed, thanks to valgrind.
891
892 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
893 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
894 was because its rules for how to advance for /g after matching an empty
895 string at the end of a line did not allow for this case. They now check for
896 it specially.
897
898 20. pcretest is supposed to handle patterns and data of any length, by
899 extending its buffers when necessary. It was getting this wrong when the
900 buffer for a data line had to be extended.
901
902 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
903 CRLF as a newline sequence.
904
905 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
906 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
907 I have nevertheless tidied it up.
908
909 23. Added some casts to kill warnings from HP-UX ia64 compiler.
910
911 24. Added a man page for pcre-config.
912
913
914 Version 7.0 19-Dec-06
915 ---------------------
916
917 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
918 moving to gcc 4.1.1.
919
920 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
921 sys/time.h, which is documented as needed for this function. It doesn't
922 seem to matter on Linux, but it showed up on some releases of OS X.
923
924 3. It seems that there are systems where bytes whose values are greater than
925 127 match isprint() in the "C" locale. The "C" locale should be the
926 default when a C program starts up. In most systems, only ASCII printing
927 characters match isprint(). This difference caused the output from pcretest
928 to vary, making some of the tests fail. I have changed pcretest so that:
929
930 (a) When it is outputting text in the compiled version of a pattern, bytes
931 other than 32-126 are always shown as hex escapes.
932
933 (b) When it is outputting text that is a matched part of a subject string,
934 it does the same, unless a different locale has been set for the match
935 (using the /L modifier). In this case, it uses isprint() to decide.
936
937 4. Fixed a major bug that caused incorrect computation of the amount of memory
938 required for a compiled pattern when options that changed within the
939 pattern affected the logic of the preliminary scan that determines the
940 length. The relevant options are -x, and -i in UTF-8 mode. The result was
941 that the computed length was too small. The symptoms of this bug were
942 either the PCRE error "internal error: code overflow" from pcre_compile(),
943 or a glibc crash with a message such as "pcretest: free(): invalid next
944 size (fast)". Examples of patterns that provoked this bug (shown in
945 pcretest format) are:
946
947 /(?-x: )/x
948 /(?x)(?-x: \s*#\s*)/
949 /((?i)[\x{c0}])/8
950 /(?i:[\x{c0}])/8
951
952 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
953 is now done differently.
954
955 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
956 wrapper classes; (b) implement a new function in the C++ scanner that is
957 more efficient than the old way of doing things because it avoids levels of
958 recursion in the regex matching; (c) add a paragraph to the documentation
959 for the FullMatch() function.
960
961 6. The escape sequence \n was being treated as whatever was defined as
962 "newline". Not only was this contrary to the documentation, which states
963 that \n is character 10 (hex 0A), but it also went horribly wrong when
964 "newline" was defined as CRLF. This has been fixed.
965
966 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
967 was being set to -1 for the "end of line" case (supposedly a value that no
968 character can have). Though this value is never used (the check for end of
969 line is "zero bytes in current character"), it caused compiler complaints.
970 I've changed it to 0xffffffff.
971
972 8. In pcre_version.c, the version string was being built by a sequence of
973 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
974 string (as it is for production releases) called a macro with an empty
975 argument. The C standard says the result of this is undefined. The gcc
976 compiler treats it as an empty string (which was what was wanted) but it is
977 reported that Visual C gives an error. The source has been hacked around to
978 avoid this problem.
979
980 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
981 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
982 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
983 of them did).
984
985 10. Originally, pcretest opened its input and output without "b"; then I was
986 told that "b" was needed in some environments, so it was added for release
987 5.0 to both the input and output. (It makes no difference on Unix-like
988 systems.) Later I was told that it is wrong for the input on Windows. I've
989 now abstracted the modes into two macros, to make it easier to fiddle with
990 them, and removed "b" from the input mode under Windows.
991
992 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
993
994 12. Added -help and --help to pcretest as an official way of being reminded
995 of the options.
996
997 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
998 and pcrecpp.cc because they annoy compilers at high warning levels.
999
1000 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
1001
1002 15. Fixed an occurrence of == in configure.ac that should have been = (shell
1003 scripts are not C programs :-) and which was not noticed because it works
1004 on Linux.
1005
1006 16. pcretest is supposed to handle any length of pattern and data line (as one
1007 line or as a continued sequence of lines) by extending its input buffer if
1008 necessary. This feature was broken for very long pattern lines, leading to
1009 a string of junk being passed to pcre_compile() if the pattern was longer
1010 than about 50K.
1011
1012 17. I have done a major re-factoring of the way pcre_compile() computes the
1013 amount of memory needed for a compiled pattern. Previously, there was code
1014 that made a preliminary scan of the pattern in order to do this. That was
1015 OK when PCRE was new, but as the facilities have expanded, it has become
1016 harder and harder to keep it in step with the real compile phase, and there
1017 have been a number of bugs (see for example, 4 above). I have now found a
1018 cunning way of running the real compile function in a "fake" mode that
1019 enables it to compute how much memory it would need, while actually only
1020 ever using a few hundred bytes of working memory and without too many
1021 tests of the mode. This should make future maintenance and development
1022 easier. A side effect of this work is that the limit of 200 on the nesting
1023 depth of parentheses has been removed (though this was never a serious
1024 limitation, I suspect). However, there is a downside: pcre_compile() now
1025 runs more slowly than before (30% or more, depending on the pattern). I
1026 hope this isn't a big issue. There is no effect on runtime performance.
1027
1028 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
1029 newline (only possible for the last line of a file) and it was a
1030 pattern that set a locale (followed by /Lsomething), pcretest crashed.
1031
1032 19. Added additional timing features to pcretest. (1) The -tm option now times
1033 matching only, not compiling. (2) Both -t and -tm can be followed, as a
1034 separate command line item, by a number that specifies the number of
1035 repeats to use when timing. The default is 50000; this gives better
1036 precision, but takes uncomfortably long for very large patterns.
1037
1038 20. Extended pcre_study() to be more clever in cases where a branch of a
1039 subpattern has no definite first character. For example, (a*|b*)[cd] would
1040 previously give no result from pcre_study(). Now it recognizes that the
1041 first character must be a, b, c, or d.
1042
1043 21. There was an incorrect error "recursive call could loop indefinitely" if
1044 a subpattern (or the entire pattern) that was being tested for matching an
1045 empty string contained only one non-empty item after a nested subpattern.
1046 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
1047 incorrectly, because the \d was being skipped in the check.
1048
1049 22. The pcretest program now has a new pattern option /B and a command line
1050 option -b, which is equivalent to adding /B to every pattern. This causes
1051 it to show the compiled bytecode, without the additional information that
1052 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
1053 is the same as /B/I).
1054
1055 23. A new optimization is now able automatically to treat some sequences such
1056 as a*b as a*+b. More specifically, if something simple (such as a character
1057 or a simple class like \d) has an unlimited quantifier, and is followed by
1058 something that cannot possibly match the quantified thing, the quantifier
1059 is automatically "possessified".
1060
1061 24. A recursive reference to a subpattern whose number was greater than 39
1062 went wrong under certain circumstances in UTF-8 mode. This bug could also
1063 have affected the operation of pcre_study().
1064
1065 25. Realized that a little bit of performance could be had by replacing
1066 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
1067
1068 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
1069
1070 27. Possessive quantifiers such as a++ were previously implemented by turning
1071 them into atomic groups such as ($>a+). Now they have their own opcodes,
1072 which improves performance. This includes the automatically created ones
1073 from 23 above.
1074
1075 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
1076 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
1077 the first matched character to be a colon. This applied both to named and
1078 numbered groups.
1079
1080 29. The ucpinternal.h header file was missing its idempotency #ifdef.
1081
1082 30. I was sent a "project" file called libpcre.a.dev which I understand makes
1083 building PCRE on Windows easier, so I have included it in the distribution.
1084
1085 31. There is now a check in pcretest against a ridiculously large number being
1086 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
1087 loop, the loop is abandoned.
1088
1089 32. Forward references to subpatterns in conditions such as (?(2)...) where
1090 subpattern 2 is defined later cause pcre_compile() to search forwards in
1091 the pattern for the relevant set of parentheses. This search went wrong
1092 when there were unescaped parentheses in a character class, parentheses
1093 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
1094
1095 33. "Subroutine" calls and backreferences were previously restricted to
1096 referencing subpatterns earlier in the regex. This restriction has now
1097 been removed.
1098
1099 34. Added a number of extra features that are going to be in Perl 5.10. On the
1100 whole, these are just syntactic alternatives for features that PCRE had
1101 previously implemented using the Python syntax or my own invention. The
1102 other formats are all retained for compatibility.
1103
1104 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
1105 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
1106 also .NET compatible.
1107
1108 (b) A recursion or subroutine call to a named group can now be defined as
1109 (?&name) as well as (?P>name).
1110
1111 (c) A backreference to a named group can now be defined as \k<name> or
1112 \k'name' as well as (?P=name). The new forms, as well as being in Perl
1113 5.10, are also .NET compatible.
1114
1115 (d) A conditional reference to a named group can now use the syntax
1116 (?(<name>) or (?('name') as well as (?(name).
1117
1118 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
1119 groups (named and numbered) that are never evaluated inline, but can be
1120 called as "subroutines" from elsewhere. In effect, the DEFINE condition
1121 is always false. There may be only one alternative in such a group.
1122
1123 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
1124 as the simple (?(R). The condition is true only if the most recent
1125 recursion is that of the given number or name. It does not search out
1126 through the entire recursion stack.
1127
1128 (g) The escape \gN or \g{N} has been added, where N is a positive or
1129 negative number, specifying an absolute or relative reference.
1130
1131 35. Tidied to get rid of some further signed/unsigned compiler warnings and
1132 some "unreachable code" warnings.
1133
1134 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
1135 things, this adds five new scripts.
1136
1137 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
1138 There were also incompatibilities regarding the handling of \Q..\E inside
1139 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
1140 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
1141
1142 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
1143 matches an empty string, and forcibly breaks the loop. There were bugs in
1144 this code in non-simple cases. For a pattern such as ^(a()*)* matched
1145 against aaaa the result was just "a" rather than "aaaa", for example. Two
1146 separate and independent bugs (that affected different cases) have been
1147 fixed.
1148
1149 39. Refactored the code to abolish the use of different opcodes for small
1150 capturing bracket numbers. This is a tidy that I avoided doing when I
1151 removed the limit on the number of capturing brackets for 3.5 back in 2001.
1152 The new approach is not only tidier, it makes it possible to reduce the
1153 memory needed to fix the previous bug (38).
1154
1155 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
1156 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
1157 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
1158 mode.
1159
1160 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
1161 report.
1162
1163 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
1164 copy construction and assignment in the C++ wrapper.
1165
1166 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
1167 couple of bugs that could have given wrong results in the "--newline=crlf"
1168 case.
1169
1170 44. Added a number of casts and did some reorganization of signed/unsigned int
1171 variables following suggestions from Dair Grant. Also renamed the variable
1172 "this" as "item" because it is a C++ keyword.
1173
1174 45. Arranged for dftables to add
1175
1176 #include "pcre_internal.h"
1177
1178 to pcre_chartables.c because without it, gcc 4.x may remove the array
1179 definition from the final binary if PCRE is built into a static library and
1180 dead code stripping is activated.
1181
1182 46. For an unanchored pattern, if a match attempt fails at the start of a
1183 newline sequence, and the newline setting is CRLF or ANY, and the next two
1184 characters are CRLF, advance by two characters instead of one.
1185
1186
1187 Version 6.7 04-Jul-06
1188 ---------------------
1189
1190 1. In order to handle tests when input lines are enormously long, pcretest has
1191 been re-factored so that it automatically extends its buffers when
1192 necessary. The code is crude, but this _is_ just a test program. The
1193 default size has been increased from 32K to 50K.
1194
1195 2. The code in pcre_study() was using the value of the re argument before
1196 testing it for NULL. (Of course, in any sensible call of the function, it
1197 won't be NULL.)
1198
1199 3. The memmove() emulation function in pcre_internal.h, which is used on
1200 systems that lack both memmove() and bcopy() - that is, hardly ever -
1201 was missing a "static" storage class specifier.
1202
1203 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
1204 containing an extended class (one that cannot be represented by a bitmap
1205 because it contains high-valued characters or Unicode property items, e.g.
1206 [\pZ]). Almost always one would set UTF-8 mode when processing such a
1207 pattern, but PCRE should not loop if you do not (it no longer does).
1208 [Detail: two cases were found: (a) a repeated subpattern containing an
1209 extended class; (b) a recursive reference to a subpattern that followed a
1210 previous extended class. It wasn't skipping over the extended class
1211 correctly when UTF-8 mode was not set.]
1212
1213 5. A negated single-character class was not being recognized as fixed-length
1214 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
1215 compile error "lookbehind assertion is not fixed length".
1216
1217 6. The RunPerlTest auxiliary script was showing an unexpected difference
1218 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
1219 write a Perl script that can interpret lines of an input file either as
1220 byte characters or as UTF-8, which is what "perltest" was being required to
1221 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
1222 can't do is switch easily at run time between having the "use utf8;" pragma
1223 or not. In the end, I fudged it by using the RunPerlTest script to insert
1224 "use utf8;" explicitly for the UTF-8 tests.
1225
1226 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
1227 the end of the subject string, contrary to the documentation and to what
1228 Perl does. This was true of both matching functions. Now it matches only at
1229 the start of the subject and immediately after *internal* newlines.
1230
1231 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
1232 a pointer to an int instead of a pointer to an unsigned long int. This
1233 caused problems on 64-bit systems.
1234
1235 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
1236 instance of the 'standard' template library not being so standard".
1237
1238 10. There was no check on the number of named subpatterns nor the maximum
1239 length of a subpattern name. The product of these values is used to compute
1240 the size of the memory block for a compiled pattern. By supplying a very
1241 long subpattern name and a large number of named subpatterns, the size
1242 computation could be caused to overflow. This is now prevented by limiting
1243 the length of names to 32 characters, and the number of named subpatterns
1244 to 10,000.
1245
1246 11. Subpatterns that are repeated with specific counts have to be replicated in
1247 the compiled pattern. The size of memory for this was computed from the
1248 length of the subpattern and the repeat count. The latter is limited to
1249 65535, but there was no limit on the former, meaning that integer overflow
1250 could in principle occur. The compiled length of a repeated subpattern is
1251 now limited to 30,000 bytes in order to prevent this.
1252
1253 12. Added the optional facility to have named substrings with the same name.
1254
1255 13. Added the ability to use a named substring as a condition, using the
1256 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
1257 are numbers (not recommended). Forward references are permitted.
1258
1259 14. Added forward references in named backreferences (if you see what I mean).
1260
1261 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
1262 pattern could run off the end of the subject. For example, the pattern
1263 "(?s)(.{1,5})"8 did this with the subject "ab".
1264
1265 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
1266 PCRE_CASELESS was set when matching characters that were quantified with ?
1267 or *.
1268
1269 17. A character class other than a single negated character that had a minimum
1270 but no maximum quantifier - for example [ab]{6,} - was not handled
1271 correctly by pce_dfa_exec(). It would match only one character.
1272
1273 18. A valid (though odd) pattern that looked like a POSIX character
1274 class but used an invalid character after [ (for example [[,abc,]]) caused
1275 pcre_compile() to give the error "Failed: internal error: code overflow" or
1276 in some cases to crash with a glibc free() error. This could even happen if
1277 the pattern terminated after [[ but there just happened to be a sequence of
1278 letters, a binary zero, and a closing ] in the memory that followed.
1279
1280 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
1281 over the years. Originally (before any Unicode support), just the bottom 8
1282 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
1283 output from "man perlunicode" includes this:
1284
1285 The regular expression compiler produces polymorphic opcodes. That
1286 is, the pattern adapts to the data and automatically switches to
1287 the Unicode character scheme when presented with Unicode data--or
1288 instead uses a traditional byte scheme when presented with byte
1289 data.
1290
1291 Sadly, a wide octal escape does not cause a switch, and in a string with
1292 no other multibyte characters, these octal escapes are treated as before.
1293 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
1294 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
1295 Unicode string.
1296
1297 I have not perpetrated such confusion in PCRE. Up till now, it took just
1298 the bottom 8 bits, as in old Perl. I have now made octal escapes with
1299 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
1300 translate to the appropriate multibyte character.
1301
1302 29. Applied some refactoring to reduce the number of warnings from Microsoft
1303 and Borland compilers. This has included removing the fudge introduced
1304 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
1305 a warning about an unused variable.
1306
1307 21. PCRE has not included VT (character 0x0b) in the set of whitespace
1308 characters since release 4.0, because Perl (from release 5.004) does not.
1309 [Or at least, is documented not to: some releases seem to be in conflict
1310 with the documentation.] However, when a pattern was studied with
1311 pcre_study() and all its branches started with \s, PCRE still included VT
1312 as a possible starting character. Of course, this did no harm; it just
1313 caused an unnecessary match attempt.
1314
1315 22. Removed a now-redundant internal flag bit that recorded the fact that case
1316 dependency changed within the pattern. This was once needed for "required
1317 byte" processing, but is no longer used. This recovers a now-scarce options
1318 bit. Also moved the least significant internal flag bit to the most-
1319 significant bit of the word, which was not previously used (hangover from
1320 the days when it was an int rather than a uint) to free up another bit for
1321 the future.
1322
1323 23. Added support for CRLF line endings as well as CR and LF. As well as the
1324 default being selectable at build time, it can now be changed at runtime
1325 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
1326 specify that it is scanning data with non-default line endings.
1327
1328 24. Changed the definition of CXXLINK to make it agree with the definition of
1329 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
1330
1331 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
1332 recursions. This makes a big different to stack usage for some patterns.
1333
1334 26. If a subpattern containing a named recursion or subroutine reference such
1335 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
1336 the space required for the compiled pattern went wrong and gave too small a
1337 value. Depending on the environment, this could lead to "Failed: internal
1338 error: code overflow at offset 49" or "glibc detected double free or
1339 corruption" errors.
1340
1341 27. Applied patches from Google (a) to support the new newline modes and (b) to
1342 advance over multibyte UTF-8 characters in GlobalReplace.
1343
1344 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
1345 difference for some implementation of PCRE in some Windows version.
1346
1347 29. Added some extra testing facilities to pcretest:
1348
1349 \q<number> in a data line sets the "match limit" value
1350 \Q<number> in a data line sets the "match recursion limt" value
1351 -S <number> sets the stack size, where <number> is in megabytes
1352
1353 The -S option isn't available for Windows.
1354
1355
1356 Version 6.6 06-Feb-06
1357 ---------------------
1358
1359 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
1360 in pcreposix.h. I have copied the definition from pcre.h.
1361
1362 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
1363 because pcre.h is no longer a built file.
1364
1365 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
1366 not normally included in the compiled code.
1367
1368
1369 Version 6.5 01-Feb-06
1370 ---------------------
1371
1372 1. When using the partial match feature with pcre_dfa_exec(), it was not
1373 anchoring the second and subsequent partial matches at the new starting
1374 point. This could lead to incorrect results. For example, with the pattern
1375 /1234/, partially matching against "123" and then "a4" gave a match.
1376
1377 2. Changes to pcregrep:
1378
1379 (a) All non-match returns from pcre_exec() were being treated as failures
1380 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
1381 error message is output. Some extra information is given for the
1382 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
1383 probably the only errors that are likely to be caused by users (by
1384 specifying a regex that has nested indefinite repeats, for instance).
1385 If there are more than 20 of these errors, pcregrep is abandoned.
1386
1387 (b) A binary zero was treated as data while matching, but terminated the
1388 output line if it was written out. This has been fixed: binary zeroes
1389 are now no different to any other data bytes.
1390
1391 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
1392 used to set a locale for matching. The --locale=xxxx long option has
1393 been added (no short equivalent) to specify a locale explicitly on the
1394 pcregrep command, overriding the environment variables.
1395
1396 (d) When -B was used with -n, some line numbers in the output were one less
1397 than they should have been.
1398
1399 (e) Added the -o (--only-matching) option.
1400
1401 (f) If -A or -C was used with -c (count only), some lines of context were
1402 accidentally printed for the final match.
1403
1404 (g) Added the -H (--with-filename) option.
1405
1406 (h) The combination of options -rh failed to suppress file names for files
1407 that were found from directory arguments.
1408
1409 (i) Added the -D (--devices) and -d (--directories) options.
1410
1411 (j) Added the -F (--fixed-strings) option.
1412
1413 (k) Allow "-" to be used as a file name for -f as well as for a data file.
1414
1415 (l) Added the --colo(u)r option.
1416
1417 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
1418 is not present by default.
1419
1420 3. A nasty bug was discovered in the handling of recursive patterns, that is,
1421 items such as (?R) or (?1), when the recursion could match a number of
1422 alternatives. If it matched one of the alternatives, but subsequently,
1423 outside the recursion, there was a failure, the code tried to back up into
1424 the recursion. However, because of the way PCRE is implemented, this is not
1425 possible, and the result was an incorrect result from the match.
1426
1427 In order to prevent this happening, the specification of recursion has
1428 been changed so that all such subpatterns are automatically treated as
1429 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
1430
1431 4. I had overlooked the fact that, in some locales, there are characters for
1432 which isalpha() is true but neither isupper() nor islower() are true. In
1433 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
1434 and ordfeminine) are like this. This affected the treatment of \w and \W
1435 when they appeared in character classes, but not when they appeared outside
1436 a character class. The bit map for "word" characters is now created
1437 separately from the results of isalnum() instead of just taking it from the
1438 upper, lower, and digit maps. (Plus the underscore character, of course.)
1439
1440 5. The above bug also affected the handling of POSIX character classes such as
1441 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
1442 permanent tables. Instead, the bit maps for such a class were previously
1443 created as the appropriate unions of the upper, lower, and digit bitmaps.
1444 Now they are created by subtraction from the [[:word:]] class, which has
1445 its own bitmap.
1446
1447 6. The [[:blank:]] character class matches horizontal, but not vertical space.
1448 It is created by subtracting the vertical space characters (\x09, \x0a,
1449 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
1450 subtraction was done in the overall bitmap for a character class, meaning
1451 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
1452 be recognized. This bug has been fixed.
1453
1454 7. Patches from the folks at Google:
1455
1456 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
1457 real life, but is still worth protecting against".
1458
1459 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
1460 regular expressions".
1461
1462 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
1463 have it.
1464
1465 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
1466 "configure" and the latter not, in order to fix a problem somebody had
1467 with compiling the Arg class on HP-UX.
1468
1469 (e) Improve the error-handling of the C++ wrapper a little bit.
1470
1471 (f) New tests for checking recursion limiting.
1472
1473 8. The pcre_memmove() function, which is used only if the environment does not
1474 have a standard memmove() function (and is therefore rarely compiled),
1475 contained two bugs: (a) use of int instead of size_t, and (b) it was not
1476 returning a result (though PCRE never actually uses the result).
1477
1478 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
1479 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
1480 returned instead of calling malloc() with an overflowing number that would
1481 most likely cause subsequent chaos.
1482
1483 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
1484
1485 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
1486 with this option is matched, the nmatch and pmatch options of regexec() are
1487 ignored.
1488
1489 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
1490 provided in case anyone wants to the the POSIX interface with UTF-8
1491 strings.
1492
1493 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
1494 C++ linking (needed for some HP-UX environments).
1495
1496 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
1497 (unused parameter) and in the pcre_printint() function (omitted "default"
1498 switch label when the default is to do nothing).
1499
1500 15. Added some code to make it possible, when PCRE is compiled as a C++
1501 library, to replace subject pointers for pcre_exec() with a smart pointer
1502 class, thus making it possible to process discontinuous strings.
1503
1504 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
1505 much the same function. They were added by different people who were trying
1506 to make PCRE easy to compile on non-Unix systems. It has been suggested
1507 that PCRE_EXPORT be abolished now that there is more automatic apparatus
1508 for compiling on Windows systems. I have therefore replaced it with
1509 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
1510 defaults to "extern" for C or "extern C" for C++, which works fine on
1511 Unix-like systems. It is now possible to override the value of PCRE_DATA_
1512 SCOPE with something explicit in config.h. In addition:
1513
1514 (a) pcreposix.h still had just "extern" instead of either of these macros;
1515 I have replaced it with PCRE_DATA_SCOPE.
1516
1517 (b) Functions such as _pcre_xclass(), which are internal to the library,
1518 but external in the C sense, all had PCRE_EXPORT in their definitions.
1519 This is apparently wrong for the Windows case, so I have removed it.
1520 (It makes no difference on Unix-like systems.)
1521
1522 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
1523 of recursive calls to match(). This is different to MATCH_LIMIT because
1524 that limits the total number of calls to match(), not all of which increase
1525 the depth of recursion. Limiting the recursion depth limits the amount of
1526 stack (or heap if NO_RECURSE is set) that is used. The default can be set
1527 when PCRE is compiled, and changed at run time. A patch from Google adds
1528 this functionality to the C++ interface.
1529
1530 18. Changes to the handling of Unicode character properties:
1531
1532 (a) Updated the table to Unicode 4.1.0.
1533
1534 (b) Recognize characters that are not in the table as "Cn" (undefined).
1535
1536 (c) I revised the way the table is implemented to a much improved format
1537 which includes recognition of ranges. It now supports the ranges that
1538 are defined in UnicodeData.txt, and it also amalgamates other
1539 characters into ranges. This has reduced the number of entries in the
1540 table from around 16,000 to around 3,000, thus reducing its size
1541 considerably. I realized I did not need to use a tree structure after
1542 all - a binary chop search is just as efficient. Having reduced the
1543 number of entries, I extended their size from 6 bytes to 8 bytes to
1544 allow for more data.
1545
1546 (d) Added support for Unicode script names via properties such as \p{Han}.
1547
1548 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
1549 matching that character.
1550
1551 20. When matching a repeated Unicode property with a minimum greater than zero,
1552 (for example \pL{2,}), PCRE could look past the end of the subject if it
1553 reached it while seeking the minimum number of characters. This could
1554 happen only if some of the characters were more than one byte long, because
1555 there is a check for at least the minimum number of bytes.
1556
1557 21. Refactored the implementation of \p and \P so as to be more general, to
1558 allow for more different types of property in future. This has changed the
1559 compiled form incompatibly. Anybody with saved compiled patterns that use
1560 \p or \P will have to recompile them.
1561
1562 22. Added "Any" and "L&" to the supported property types.
1563
1564 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
1565 but give a compile time error if the value is greater than 0xff.
1566
1567 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
1568 accidentally not being installed or uninstalled.
1569
1570 25. The pcre.h file was built from pcre.h.in, but the only changes that were
1571 made were to insert the current release number. This seemed silly, because
1572 it made things harder for people building PCRE on systems that don't run
1573 "configure". I have turned pcre.h into a distributed file, no longer built
1574 by "configure", with the version identification directly included. There is
1575 no longer a pcre.h.in file.
1576
1577 However, this change necessitated a change to the pcre-config script as
1578 well. It is built from pcre-config.in, and one of the substitutions was the
1579 release number. I have updated configure.ac so that ./configure now finds
1580 the release number by grepping pcre.h.
1581
1582 26. Added the ability to run the tests under valgrind.
1583
1584
1585 Version 6.4 05-Sep-05
1586 ---------------------
1587
1588 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
1589 "--" to be printed when multiple files were scanned, even when none of the
1590 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
1591 consider it to be a bug, and have restored the previous behaviour.
1592
1593 2. A couple of code tidies to get rid of compiler warnings.
1594
1595 3. The pcretest program used to cheat by referring to symbols in the library
1596 whose names begin with _pcre_. These are internal symbols that are not
1597 really supposed to be visible externally, and in some environments it is
1598 possible to suppress them. The cheating is now confined to including
1599 certain files from the library's source, which is a bit cleaner.
1600
1601 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
1602 file's purpose clearer.
1603
1604 5. Reorganized pcre_ucp_findchar().
1605
1606
1607 Version 6.3 15-Aug-05
1608 ---------------------
1609
1610 1. The file libpcre.pc.in did not have general read permission in the tarball.
1611
1612 2. There were some problems when building without C++ support:
1613
1614 (a) If C++ support was not built, "make install" and "make test" still
1615 tried to test it.
1616
1617 (b) There were problems when the value of CXX was explicitly set. Some
1618 changes have been made to try to fix these, and ...
1619
1620 (c) --disable-cpp can now be used to explicitly disable C++ support.
1621
1622 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
1623 backslash in a target when C++ was disabled. This confuses some
1624 versions of "make", apparently. Using an intermediate variable solves
1625 this. (Same for CPP_LOBJ.)
1626
1627 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
1628 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
1629 necessary on certain architectures.
1630
1631 4. Added a setting of -export-symbols-regex to the link command to remove
1632 those symbols that are exported in the C sense, but actually are local
1633 within the library, and not documented. Their names all begin with
1634 "_pcre_". This is not a perfect job, because (a) we have to except some
1635 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
1636 available (and never for static libraries). I have made a note to try to
1637 find a way round (a) in the future.
1638
1639
1640 Version 6.2 01-Aug-05
1641 ---------------------
1642
1643 1. There was no test for integer overflow of quantifier values. A construction
1644 such as {1111111111111111} would give undefined results. What is worse, if
1645 a minimum quantifier for a parenthesized subpattern overflowed and became
1646 negative, the calculation of the memory size went wrong. This could have
1647 led to memory overwriting.
1648
1649 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
1650
1651 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
1652 operating environments where this matters.
1653
1654 4. Applied Giuseppe Maxia's patch to add additional features for controlling
1655 PCRE options from within the C++ wrapper.
1656
1657 5. Named capturing subpatterns were not being correctly counted when a pattern
1658 was compiled. This caused two problems: (a) If there were more than 100
1659 such subpatterns, the calculation of the memory needed for the whole
1660 compiled pattern went wrong, leading to an overflow error. (b) Numerical
1661 back references of the form \12, where the number was greater than 9, were
1662 not recognized as back references, even though there were sufficient
1663 previous subpatterns.
1664
1665 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
1666 versions of gcc, e.g. 2.95.4.
1667
1668
1669 Version 6.1 21-Jun-05
1670 ---------------------
1671
1672 1. There was one reference to the variable "posix" in pcretest.c that was not
1673 surrounded by "#if !defined NOPOSIX".
1674
1675 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
1676 the cross-check on the old pcre_info() function, for the benefit of the
1677 cut-down version of PCRE that is currently imported into Exim.
1678
1679 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
1680 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
1681 patterns that start (?i) but I don't think that matters. The use of (?i) is
1682 just an example; this all applies to the other options as well.
1683
1684 4. Since libtool seems to echo the compile commands it is issuing, the output
1685 from "make" can be reduced a bit by putting "@" in front of each libtool
1686 compile command.
1687
1688 5. Patch from the folks at Google for configure.in to be a bit more thorough
1689 in checking for a suitable C++ installation before trying to compile the
1690 C++ stuff. This should fix a reported problem when a compiler was present,
1691 but no suitable headers.
1692
1693 6. The man pages all had just "PCRE" as their title. I have changed them to
1694 be the relevant file name. I have also arranged that these names are
1695 retained in the file doc/pcre.txt, which is a concatenation in text format
1696 of all the man pages except the little individual ones for each function.
1697
1698 7. The NON-UNIX-USE file had not been updated for the different set of source
1699 files that come with release 6. I also added a few comments about the C++
1700 wrapper.
1701
1702
1703 Version 6.0 07-Jun-05
1704 ---------------------
1705
1706 1. Some minor internal re-organization to help with my DFA experiments.
1707
1708 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
1709 didn't matter for the library itself when fully configured, but did matter
1710 when compiling without UCP support, or within Exim, where the ucp files are
1711 not imported.
1712
1713 3. Refactoring of the library code to split up the various functions into
1714 different source modules. The addition of the new DFA matching code (see
1715 below) to a single monolithic source would have made it really too
1716 unwieldy, quite apart from causing all the code to be include in a
1717 statically linked application, when only some functions are used. This is
1718 relevant even without the DFA addition now that patterns can be compiled in
1719 one application and matched in another.
1720
1721 The downside of splitting up is that there have to be some external
1722 functions and data tables that are used internally in different modules of
1723 the library but which are not part of the API. These have all had their
1724 names changed to start with "_pcre_" so that they are unlikely to clash
1725 with other external names.
1726
1727 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
1728 a different (DFA) algorithm. Although it is slower than the original
1729 function, it does have some advantages for certain types of matching
1730 problem.
1731
1732 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
1733 including restarting after a partial match.
1734
1735 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
1736 defined when compiling for Windows was sent to me. I have put it into the
1737 code, though I have no means of testing or verifying it.
1738
1739 7. Added the pcre_refcount() auxiliary function.
1740
1741 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
1742 match before or at the first newline in the subject string. In pcretest,
1743 the /f option on a pattern can be used to set this.
1744
1745 9. A repeated \w when used in UTF-8 mode with characters greater than 256
1746 would behave wrongly. This has been present in PCRE since release 4.0.
1747
1748 10. A number of changes to the pcregrep command:
1749
1750 (a) Refactored how -x works; insert ^(...)$ instead of setting
1751 PCRE_ANCHORED and checking the length, in preparation for adding
1752 something similar for -w.
1753
1754 (b) Added the -w (match as a word) option.
1755
1756 (c) Refactored the way lines are read and buffered so as to have more
1757 than one at a time available.
1758
1759 (d) Implemented a pcregrep test script.
1760
1761 (e) Added the -M (multiline match) option. This allows patterns to match
1762 over several lines of the subject. The buffering ensures that at least
1763 8K, or the rest of the document (whichever is the shorter) is available
1764 for matching (and similarly the previous 8K for lookbehind assertions).
1765
1766 (f) Changed the --help output so that it now says
1767
1768 -w, --word-regex(p)
1769
1770 instead of two lines, one with "regex" and the other with "regexp"
1771 because that confused at least one person since the short forms are the
1772 same. (This required a bit of code, as the output is generated
1773 automatically from a table. It wasn't just a text change.)
1774
1775 (g) -- can be used to terminate pcregrep options if the next thing isn't an
1776 option but starts with a hyphen. Could be a pattern or a path name
1777 starting with a hyphen, for instance.
1778
1779 (h) "-" can be given as a file name to represent stdin.
1780
1781 (i) When file names are being printed, "(standard input)" is used for
1782 the standard input, for compatibility with GNU grep. Previously
1783 "<stdin>" was used.
1784
1785 (j) The option --label=xxx can be used to supply a name to be used for
1786 stdin when file names are being printed. There is no short form.
1787
1788 (k) Re-factored the options decoding logic because we are going to add
1789 two more options that take data. Such options can now be given in four
1790 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
1791
1792 (l) Added the -A, -B, and -C options for requesting that lines of context
1793 around matches be printed.
1794
1795 (m) Added the -L option to print the names of files that do not contain
1796 any matching lines, that is, the complement of -l.
1797
1798 (n) The return code is 2 if any file cannot be opened, but pcregrep does
1799 continue to scan other files.
1800
1801 (o) The -s option was incorrectly implemented. For compatibility with other
1802 greps, it now suppresses the error message for a non-existent or non-
1803 accessible file (but not the return code). There is a new option called
1804 -q that suppresses the output of matching lines, which was what -s was
1805 previously doing.
1806
1807 (p) Added --include and --exclude options to specify files for inclusion
1808 and exclusion when recursing.
1809
1810 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
1811 Hopefully, it now does.
1812
1813 12. Missing cast in pcre_study().
1814
1815 13. Added an "uninstall" target to the makefile.
1816
1817 14. Replaced "extern" in the function prototypes in Makefile.in with
1818 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
1819 world, but is set differently for Windows.
1820
1821 15. Added a second compiling function called pcre_compile2(). The only
1822 difference is that it has an extra argument, which is a pointer to an
1823 integer error code. When there is a compile-time failure, this is set
1824 non-zero, in addition to the error test pointer being set to point to an
1825 error message. The new argument may be NULL if no error number is required
1826 (but then you may as well call pcre_compile(), which is now just a
1827 wrapper). This facility is provided because some applications need a
1828 numeric error indication, but it has also enabled me to tidy up the way
1829 compile-time errors are handled in the POSIX wrapper.
1830
1831 16. Added VPATH=.libs to the makefile; this should help when building with one
1832 prefix path and installing with another. (Or so I'm told by someone who
1833 knows more about this stuff than I do.)
1834
1835 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
1836 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
1837 match everything, including newlines. This is not POSIX-compatible, but
1838 somebody wanted the feature. From pcretest it can be activated by using
1839 both the P and the s flags.
1840
1841 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
1842
1843 19. libpcre.pc was being incorrectly installed as executable.
1844
1845 20. A couple of places in pcretest check for end-of-line by looking for '\n';
1846 it now also looks for '\r' so that it will work unmodified on Windows.
1847
1848 21. Added Google's contributed C++ wrapper to the distribution.
1849
1850 22. Added some untidy missing memory free() calls in pcretest, to keep
1851 Electric Fence happy when testing.
1852
1853
1854
1855 Version 5.0 13-Sep-04
1856 ---------------------
1857
1858 1. Internal change: literal characters are no longer packed up into items
1859 containing multiple characters in a single byte-string. Each character
1860 is now matched using a separate opcode. However, there may be more than one
1861 byte in the character in UTF-8 mode.
1862
1863 2. The pcre_callout_block structure has two new fields: pattern_position and
1864 next_item_length. These contain the offset in the pattern to the next match
1865 item, and its length, respectively.
1866
1867 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
1868 insertion of callouts before each pattern item. Added the /C option to
1869 pcretest to make use of this.
1870
1871 4. On the advice of a Windows user, the lines
1872
1873 #if defined(_WIN32) || defined(WIN32)
1874 _setmode( _fileno( stdout ), 0x8000 );
1875 #endif /* defined(_WIN32) || defined(WIN32) */
1876
1877 have been added to the source of pcretest. This apparently does useful
1878 magic in relation to line terminators.
1879
1880 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
1881 for the benefit of those environments where the "b" makes a difference.
1882
1883 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
1884 to know about it. I have put a hack into configure.in that adds in code
1885 to set GCC=yes if CC=icc. This seems to end up at a point in the
1886 generated configure script that is early enough to affect the setting of
1887 compiler options, which is what is needed, but I have no means of testing
1888 whether it really works. (The user who reported this had patched the
1889 generated configure script, which of course I cannot do.)
1890
1891 LATER: After change 22 below (new libtool files), the configure script
1892 seems to know about icc (and also ecc). Therefore, I have commented out
1893 this hack in configure.in.
1894
1895 7. Added support for pkg-config (2 patches were sent in).
1896
1897 8. Negated POSIX character classes that used a combination of internal tables
1898 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
1899 [[:^ascii]]. Typically, they would match almost any characters. The other
1900 POSIX classes were not broken in this way.
1901
1902 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
1903 to find the match, as PCRE was deluded into thinking that the match had to
1904 start at the start point or following a newline. The same bug applied to
1905 patterns with negative forward assertions or any backward assertions
1906 preceding ".*" at the start, unless the pattern required a fixed first
1907 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
1908
1909 10. In UTF-8 mode, when moving forwards in the subject after a failed match
1910 starting at the last subject character, bytes beyond the end of the subject
1911 string were read.
1912
1913 11. Renamed the variable "class" as "classbits" to make life easier for C++
1914 users. (Previously there was a macro definition, but it apparently wasn't
1915 enough.)
1916
1917 12. Added the new field "tables" to the extra data so that tables can be passed
1918 in at exec time, or the internal tables can be re-selected. This allows
1919 a compiled regex to be saved and re-used at a later time by a different
1920 program that might have everything at different addresses.
1921
1922 13. Modified the pcre-config script so that, when run on Solaris, it shows a
1923 -R library as well as a -L library.
1924
1925 14. The debugging options of pcretest (-d on the command line or D on a
1926 pattern) showed incorrect output for anything following an extended class
1927 that contained multibyte characters and which was followed by a quantifier.
1928
1929 15. Added optional support for general category Unicode character properties
1930 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
1931 support. It adds about 90K to the size of the library. The meanings of the
1932 inbuilt class escapes such as \d and \s have NOT been changed.
1933
1934 16. Updated pcredemo.c to include calls to free() to release the memory for the
1935 compiled pattern.
1936
1937 17. The generated file chartables.c was being created in the source directory
1938 instead of in the building directory. This caused the build to fail if the
1939 source directory was different from the building directory, and was
1940 read-only.
1941
1942 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
1943 file. No doubt somebody will tell me if they don't make sense... Also added
1944 Dan Mooney's comments about building on OpenVMS.
1945
1946 19. Added support for partial matching via the PCRE_PARTIAL option for
1947 pcre_exec() and the \P data escape in pcretest.
1948
1949 20. Extended pcretest with 3 new pattern features:
1950
1951 (i) A pattern option of the form ">rest-of-line" causes pcretest to
1952 write the compiled pattern to the file whose name is "rest-of-line".
1953 This is a straight binary dump of the data, with the saved pointer to
1954 the character tables forced to be NULL. The study data, if any, is
1955 written too. After writing, pcretest reads a new pattern.
1956
1957 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
1958 compiled pattern from the given file. There must not be any
1959 occurrences of "<" in the file name (pretty unlikely); if there are,
1960 pcretest will instead treat the initial "<" as a pattern delimiter.
1961 After reading in the pattern, pcretest goes on to read data lines as
1962 usual.
1963
1964 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
1965 and 16-bit fields in a compiled pattern, to simulate a pattern that
1966 was compiled on a host of opposite endianness.
1967
1968 21. The pcre-exec() function can now cope with patterns that were compiled on
1969 hosts of opposite endianness, with this restriction:
1970
1971 As for any compiled expression that is saved and used later, the tables
1972 pointer field cannot be preserved; the extra_data field in the arguments
1973 to pcre_exec() should be used to pass in a tables address if a value
1974 other than the default internal tables were used at compile time.
1975
1976 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
1977 now diagnosed as an error. Previously, most of the time, a negative number
1978 would have been treated as zero, but if in addition "ovector" was passed as
1979 NULL, a crash could occur.
1980
1981 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
1982 new versions from the libtool 1.5 distribution (the last one is a copy of
1983 a file called libtool.m4). This seems to have fixed the need to patch
1984 "configure" to support Darwin 1.3 (which I used to do). However, I still
1985 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
1986 workstation).
1987
1988 24. Changed the PCRE licence to be the more standard "BSD" licence.
1989
1990
1991 Version 4.5 01-Dec-03
1992 ---------------------
1993
1994 1. There has been some re-arrangement of the code for the match() function so
1995 that it can be compiled in a version that does not call itself recursively.
1996 Instead, it keeps those local variables that need separate instances for
1997 each "recursion" in a frame on the heap, and gets/frees frames whenever it
1998 needs to "recurse". Keeping track of where control must go is done by means
1999 of setjmp/longjmp. The whole thing is implemented by a set of macros that
2000 hide most of the details from the main code, and operates only if
2001 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
2002 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
2003 operating.
2004
2005 To make it easier for callers to provide specially tailored get/free
2006 functions for this usage, two new functions, pcre_stack_malloc, and
2007 pcre_stack_free, are used. They are always called in strict stacking order,
2008 and the size of block requested is always the same.
2009
2010 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
2011 PCRE has been compiled to use the stack or the heap for recursion. The
2012 -C option of pcretest uses this to show which version is compiled.
2013
2014 A new data escape \S, is added to pcretest; it causes the amounts of store
2015 obtained and freed by both kinds of malloc/free at match time to be added
2016 to the output.
2017
2018 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
2019 what's available on my current Linux desktop machine.
2020
2021 3. When matching a UTF-8 string, the test for a valid string at the start has
2022 been extended. If start_offset is not zero, PCRE now checks that it points
2023 to a byte that is the start of a UTF-8 character. If not, it returns
2024 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
2025 this is necessary because there may be backward assertions in the pattern.
2026 When matching the same subject several times, it may save resources to use
2027 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
2028
2029 4. The code for checking the validity of UTF-8 strings has been tightened so
2030 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
2031 containing "overlong sequences".
2032
2033 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
2034 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
2035 should have been "&", but it just so happened that all the cases this let
2036 through by mistake were picked up later in the function.
2037
2038 6. I had used a variable called "isblank" - this is a C99 function, causing
2039 some compilers to warn. To avoid this, I renamed it (as "blankclass").
2040
2041 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
2042 prompting; (b) run "./pcretest /dev/null" at the start of the test script
2043 so the version is shown; (c) stop "make test" echoing "./RunTest".
2044
2045 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
2046
2047 9. The prototype for memmove() for systems that don't have it was using
2048 size_t, but the inclusion of the header that defines size_t was later. I've
2049 moved the #includes for the C headers earlier to avoid this.
2050
2051 10. Added some adjustments to the code to make it easier to compiler on certain
2052 special systems:
2053
2054 (a) Some "const" qualifiers were missing.
2055 (b) Added the macro EXPORT before all exported functions; by default this
2056 is defined to be empty.
2057 (c) Changed the dftables auxiliary program (that builds chartables.c) so
2058 that it reads its output file name as an argument instead of writing
2059 to the standard output and assuming this can be redirected.
2060
2061 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
2062 class containing characters with values greater than 255, PCRE compilation
2063 went into a loop.
2064
2065 12. A recursive reference to a subpattern that was within another subpattern
2066 that had a minimum quantifier of zero caused PCRE to crash. For example,
2067 (x(y(?2))z)? provoked this bug with a subject that got as far as the
2068 recursion. If the recursively-called subpattern itself had a zero repeat,
2069 that was OK.
2070
2071 13. In pcretest, the buffer for reading a data line was set at 30K, but the
2072 buffer into which it was copied (for escape processing) was still set at
2073 1024, so long lines caused crashes.
2074
2075 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
2076 "internal error: code overflow...". This applied to any character class
2077 that was followed by a possessive quantifier.
2078
2079 15. Modified the Makefile to add libpcre.la as a prerequisite for
2080 libpcreposix.la because I was told this is needed for a parallel build to
2081 work.
2082
2083 16. If a pattern that contained .* following optional items at the start was
2084 studied, the wrong optimizing data was generated, leading to matching
2085 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
2086 matching string must start with a or b or c. The correct conclusion for
2087 this pattern is that a match can start with any character.
2088
2089
2090 Version 4.4 13-Aug-03
2091 ---------------------
2092
2093 1. In UTF-8 mode, a character class containing characters with values between
2094 127 and 255 was not handled correctly if the compiled pattern was studied.
2095 In fixing this, I have also improved the studying algorithm for such
2096 classes (slightly).
2097
2098 2. Three internal functions had redundant arguments passed to them. Removal
2099 might give a very teeny performance improvement.
2100
2101 3. Documentation bug: the value of the capture_top field in a callout is *one
2102 more than* the number of the hightest numbered captured substring.
2103
2104 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
2105 in incorrectly linking with a previously installed version. They now link
2106 explicitly with libpcre.la.
2107
2108 5. configure.in no longer needs to recognize Cygwin specially.
2109
2110 6. A problem in pcre.in for Windows platforms is fixed.
2111
2112 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
2113 pcretest, it used to include the size of the study block as part of its
2114 output. Unfortunately, the structure contains a field that has a different
2115 size on different hardware architectures. This meant that the tests that
2116 showed this size failed. As the block is currently always of a fixed size,
2117 this information isn't actually particularly useful in pcretest output, so
2118 I have just removed it.
2119
2120 8. Three pre-processor statements accidentally did not start in column 1.
2121 Sadly, there are *still* compilers around that complain, even though
2122 standard C has not required this for well over a decade. Sigh.
2123
2124 9. In pcretest, the code for checking callouts passed small integers in the
2125 callout_data field, which is a void * field. However, some picky compilers
2126 complained about the casts involved for this on 64-bit systems. Now
2127 pcretest passes the address of the small integer instead, which should get
2128 rid of the warnings.
2129
2130 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
2131 both compile and run time, and gives an error if an invalid UTF-8 sequence
2132 is found. There is a option for disabling this check in cases where the
2133 string is known to be correct and/or the maximum performance is wanted.
2134
2135 11. In response to a bug report, I changed one line in Makefile.in from
2136
2137 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
2138 to
2139 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
2140
2141 to look similar to other lines, but I have no way of telling whether this
2142 is the right thing to do, as I do not use Windows. No doubt I'll get told
2143 if it's wrong...
2144
2145
2146 Version 4.3 21-May-03
2147 ---------------------
2148
2149 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
2150 Makefile.
2151
2152 2. Some refactoring to improve the quality of the code:
2153
2154 (i) The utf8_table... variables are now declared "const".
2155
2156 (ii) The code for \cx, which used the "case flipping" table to upper case
2157 lower case letters, now just substracts 32. This is ASCII-specific,
2158 but the whole concept of \cx is ASCII-specific, so it seems
2159 reasonable.
2160
2161 (iii) PCRE was using its character types table to recognize decimal and
2162 hexadecimal digits in the pattern. This is silly, because it handles
2163 only 0-9, a-f, and A-F, but the character types table is locale-
2164 specific, which means strange things might happen. A private
2165 table is now used for this - though it costs 256 bytes, a table is
2166 much faster than multiple explicit tests. Of course, the standard
2167 character types table is still used for matching digits in subject
2168 strings against \d.
2169
2170 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
2171 ending in _t are). So I've renamed it as ESC_tee.
2172
2173 3. The first argument for regexec() in the POSIX wrapper should have been
2174 defined as "const".
2175
2176 4. Changed pcretest to use malloc() for its buffers so that they can be
2177 Electric Fenced for debugging.
2178
2179 5. There were several places in the code where, in UTF-8 mode, PCRE would try
2180 to read one or more bytes before the start of the subject string. Often this
2181 had no effect on PCRE's behaviour, but in some circumstances it could
2182 provoke a segmentation fault.
2183
2184 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
2185 to try to read one or more bytes before the start of the subject string.
2186
2187 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
2188 UTF-8 support could misbehave in various ways if the subject string
2189 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
2190 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
2191 back over UTF-8 characters.)
2192
2193
2194 Version 4.2 14-Apr-03
2195 ---------------------
2196
2197 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
2198
2199 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
2200 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
2201 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
2202 [WIN_PREFIX]: new variable, "cyg" for Cygwin
2203 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
2204 and BUILD_EXEEXT
2205 Note: automatic setting of the BUILD variables is not yet working
2206 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
2207 compile-time but not at link-time
2208 [LINK]: use for linking executables only
2209 make different versions for Windows and non-Windows
2210 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
2211 libraries
2212 [LINK_FOR_BUILD]: new variable
2213 [OBJEXT]: use throughout
2214 [EXEEXT]: use throughout
2215 <winshared>: new target
2216 <wininstall>: new target
2217 <dftables.o>: use native compiler
2218 <dftables>: use native linker
2219 <install>: handle Windows platform correctly
2220 <clean>: ditto
2221 <check>: ditto
2222 copy DLL to top builddir before testing
2223
2224 As part of these changes, -no-undefined was removed again. This was reported
2225 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
2226 in any case.
2227
2228 3. Some tidies to get rid of compiler warnings:
2229
2230 . In the match_data structure, match_limit was an unsigned long int, whereas
2231 match_call_count was an int. I've made them both unsigned long ints.
2232
2233 . In pcretest the fact that a const uschar * doesn't automatically cast to
2234 a void * provoked a warning.
2235
2236 . Turning on some more compiler warnings threw up some "shadow" variables
2237 and a few more missing casts.
2238
2239 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
2240 option, a class that contained a single character with a value between 128
2241 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
2242
2243 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
2244 option, a class that contained several characters, but with at least one
2245 whose value was between 128 and 255 caused PCRE to crash.
2246
2247
2248 Version 4.1 12-Mar-03
2249 ---------------------
2250
2251 1. Compiling with gcc -pedantic found a couple of places where casts were
2252 needed, and a string in dftables.c that was longer than standard compilers are
2253 required to support.
2254
2255 2. Compiling with Sun's compiler found a few more places where the code could
2256 be tidied up in order to avoid warnings.
2257
2258 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
2259 first of these names is deprecated in the latest Autoconf in favour of the name
2260 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
2261 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
2262 analogy I have changed it to CFLAGS_FOR_BUILD.
2263
2264 4. Added -no-undefined to the linking command in the Makefile, because this is
2265 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
2266 linking step for the pcreposix library.
2267
2268 5. PCRE was failing to diagnose the case of two named groups with the same
2269 name.
2270
2271 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
2272 literal character that is needed in the subject for a match, and scans along to
2273 ensure that it is present before embarking on the full matching process. This
2274 saves time in cases of nested unlimited repeats that are never going to match.
2275 Problem: the scan can take a lot of time if the subject is very long (e.g.
2276 megabytes), thus penalizing straightforward matches. It is now done only if the
2277 amount of subject to be scanned is less than 1000 bytes.
2278
2279 7. A lesser problem with the same optimization is that it was recording the
2280 first character of an anchored pattern as "needed", thus provoking a search
2281 right along the subject, even when the first match of the pattern was going to
2282 fail. The "needed" character is now not set for anchored patterns, unless it
2283 follows something in the pattern that is of non-fixed length. Thus, it still
2284 fulfils its original purpose of finding quick non-matches in cases of nested
2285 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
2286
2287
2288 Version 4.0 17-Feb-03
2289 ---------------------
2290
2291 1. If a comment in an extended regex that started immediately after a meta-item
2292 extended to the end of string, PCRE compiled incorrect data. This could lead to
2293 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
2294
2295 2. Moved to autoconf 2.53 and libtool 1.4.2.
2296
2297 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
2298 the special perltest8 script is no longer needed - all the tests can be run
2299 from a single perltest script.
2300
2301 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
2302 by \s. It has now been removed in PCRE. This means it isn't recognized as
2303 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
2304 class [:space:] *does* include VT, thereby creating a mess.
2305
2306 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
2307 space and tab.
2308
2309 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
2310 its new features into the main test script, reducing the number of scripts.
2311
2312 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
2313 were backward compatible, and made the (?i) apply to the whole pattern, as if
2314 /i were given. Now it behaves more logically, and applies the option setting
2315 only to what follows. PCRE has been changed to follow suit. However, if it
2316 finds options settings right at the start of the pattern, it extracts them into
2317 the global options, as before. Thus, they show up in the info data.
2318
2319 8. Added support for the \Q...\E escape sequence. Characters in between are
2320 treated as literals. This is slightly different from Perl in that $ and @ are
2321 also handled as literals inside the quotes. In Perl, they will cause variable
2322 interpolation. Note the following examples:
2323
2324 Pattern PCRE matches Perl matches
2325
2326 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
2327 \Qabc\$xyz\E abc\$xyz abc\$xyz
2328 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
2329
2330 For compatibility with Perl, \Q...\E sequences are recognized inside character
2331 classes as well as outside them.
2332
2333 9. Re-organized 3 code statements in pcretest to avoid "overflow in
2334 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
2335 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
2336 signed/unsigned warnings.
2337
2338 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
2339 option for pcretest, so I've replaced it by a simple function that does just
2340 that job.
2341
2342 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
2343 "pcregrep -".
2344
2345 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
2346 Java package. This provides some syntactic sugar for simple cases of what my
2347 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
2348 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
2349 item, you can use this simplified notation. Note that only makes sense with
2350 greedy quantifiers. Consequently, the use of the possessive quantifier forces
2351 greediness, whatever the setting of the PCRE_UNGREEDY option.
2352
2353 13. A change of greediness default within a pattern was not taking effect at
2354 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
2355 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
2356 was abstracted outside.
2357
2358 14. PCRE now supports the \G assertion. It is true when the current matching
2359 position is at the start point of the match. This differs from \A when the
2360 starting offset is non-zero. Used with the /g option of pcretest (or similar
2361 code), it works in the same way as it does for Perl's /g option. If all
2362 alternatives of a regex begin with \G, the expression is anchored to the start
2363 match position, and the "anchored" flag is set in the compiled expression.
2364
2365 15. Some bugs concerning the handling of certain option changes within patterns
2366 have been fixed. These applied to options other than (?ims). For example,
2367 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
2368 been the other way round. Some of this was related to change 7 above.
2369
2370 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
2371 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
2372 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
2373 POSIX classes only within a class (e.g. /[[:alpha:]]/).
2374
2375 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
2376 mode. Unlike ".", it always matches newline, whatever the setting of
2377 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
2378 assertions. Perl allows it, but it doesn't (in general) work because it can't
2379 calculate the length of the lookbehind. At least, that's the case for Perl
2380 5.8.0 - I've been told they are going to document that it doesn't work in
2381 future.
2382
2383 18. Added an error diagnosis for escapes that PCRE does not support: these are
2384 \L, \l, \N, \P, \p, \U, \u, and \X.
2385
2386 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
2387 reading past the end of the pattern in cases such as /[abcd/.
2388
2389 20. PCRE was getting more memory than necessary for patterns with classes that
2390 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
2391
2392 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
2393 compiling PCRE for use with Virtual Pascal.
2394
2395 22. Small fix to the Makefile to make it work properly if the build is done
2396 outside the source tree.
2397
2398 23. Added a new extension: a condition to go with recursion. If a conditional
2399 subpattern starts with (?(R) the "true" branch is used if recursion has
2400 happened, whereas the "false" branch is used only at the top level.
2401
2402 24. When there was a very long string of literal characters (over 255 bytes
2403 without UTF support, over 250 bytes with UTF support), the computation of how
2404 much memory was required could be incorrect, leading to segfaults or other
2405 strange effects.
2406
2407 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
2408 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
2409 there was a subsequent back reference to those brackets. This meant that, for
2410 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
2411 possible to check for precisely this case. All we can do is abandon the
2412 optimization if .* occurs inside capturing brackets when there are any back
2413 references whatsoever. (See below for a better fix that came later.)
2414
2415 26. The handling of the optimization for finding the first character of a
2416 non-anchored pattern, and for finding a character that is required later in the
2417 match were failing in some cases. This didn't break the matching; it just
2418 failed to optimize when it could. The way this is done has been re-implemented.
2419
2420 27. Fixed typo in error message for invalid (?R item (it said "(?p").
2421
2422 28. Added a new feature that provides some of the functionality that Perl
2423 provides with (?{...}). The facility is termed a "callout". The way it is done
2424 in PCRE is for the caller to provide an optional function, by setting
2425 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
2426 global variable. By default it is unset, which disables all calling out. To get
2427 the function called, the regex must include (?C) at appropriate points. This
2428 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
2429 This provides a means of identifying different callout points. When PCRE
2430 reaches such a point in the regex, if pcre_callout has been set, the external
2431 function is called. It is provided with data in a structure called
2432 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
2433 matching continues; if it returns a non-zero value, the match at the current
2434 point fails. However, backtracking will occur if possible. [This was changed
2435 later and other features added - see item 49 below.]
2436
2437 29. pcretest is upgraded to test the callout functionality. It provides a
2438 callout function that displays information. By default, it shows the start of
2439 the match and the current position in the text. There are some new data escapes
2440 to vary what happens:
2441
2442 \C+ in addition, show current contents of captured substrings
2443 \C- do not supply a callout function
2444 \C!n return 1 when callout number n is reached
2445 \C!n!m return 1 when callout number n is reached for the mth time
2446
2447 30. If pcregrep was called with the -l option and just a single file name, it
2448 output "<stdin>" if a match was found, instead of the file name.
2449
2450 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
2451 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
2452 pcre_exec(). This saves a malloc/free per call. The default value of
2453 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
2454 when configuring.
2455
2456 32. The default maximum size of a compiled pattern is 64K. There have been a
2457 few cases of people hitting this limit. The code now uses macros to handle the
2458 storing of links as offsets within the compiled pattern. It defaults to 2-byte
2459 links, but this can be changed to 3 or 4 bytes by --with-link-size when
2460 configuring. Tests 2 and 5 work only with 2-byte links because they output
2461 debugging information about compiled patterns.
2462
2463 33. Internal code re-arrangements:
2464
2465 (a) Moved the debugging function for printing out a compiled regex into
2466 its own source file (printint.c) and used #include to pull it into
2467 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
2468 separate copies.
2469
2470 (b) Defined the list of op-code names for debugging as a macro in
2471 internal.h so that it is next to the definition of the opcodes.
2472
2473 (c) Defined a table of op-code lengths for simpler skipping along compiled
2474 code. This is again a macro in internal.h so that it is next to the
2475 definition of the opcodes.
2476
2477 34. Added support for recursive calls to individual subpatterns, along the
2478 lines of Robin Houston's patch (but implemented somewhat differently).
2479
2480 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
2481 allow it to read and process whole directories in Win32. This code was
2482 contributed by Lionel Fourquaux; it has not been tested by me.
2483
2484 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
2485 used to name a group. Names consist of alphanumerics and underscores, and must
2486 be unique. Back references use the syntax (?P=name) and recursive calls use
2487 (?P>name) which is a PCRE extension to the Python extension. Groups still have
2488 numbers. The function pcre_fullinfo() can be used after compilation to extract
2489 a name/number map. There are three relevant calls:
2490
2491 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
2492 PCRE_INFO_NAMECOUNT yields the number of entries
2493 PCRE_INFO_NAMETABLE yields a pointer to the map.
2494
2495 The map is a vector of fixed-size entries. The size of each entry depends on
2496 the length of the longest name used. The first two bytes of each entry are the
2497 group number, most significant byte first. There follows the corresponding
2498 name, zero terminated. The names are in alphabetical order.
2499
2500 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
2501 case instead of 255. Making it the same both with and without UTF-8 support
2502 means that the same test output works with both.
2503
2504 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
2505 calling malloc() with a zero argument.
2506
2507 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
2508 optimization. I've improved things by keeping a bitmap of backreferences with
2509 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
2510 fact referenced, the optimization can be applied. It is unlikely that a
2511 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
2512 the match to follow \n) will appear inside brackets with a number greater than
2513 31, but if it does, any back reference > 31 suppresses the optimization.
2514
2515 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
2516 of disabling numbered capturing parentheses. Any opening parenthesis that is
2517 not followed by ? behaves as if it were followed by ?: but named parentheses
2518 can still be used for capturing (and they will acquire numbers in the usual
2519 way).
2520
2521 41. Redesigned the return codes from the match() function into yes/no/error so
2522 that errors can be passed back from deep inside the nested calls. A malloc
2523 failure while inside a recursive subpattern call now causes the
2524 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
2525
2526 42. It is now possible to set a limit on the number of times the match()
2527 function is called in a call to pcre_exec(). This facility makes it possible to
2528 limit the amount of recursion and backtracking, though not in a directly
2529 obvious way, because the match() function is used in a number of different
2530 circumstances. The count starts from zero for each position in the subject
2531 string (for non-anchored patterns). The default limit is, for compatibility, a
2532 large number, namely 10 000 000. You can change this in two ways:
2533
2534 (a) When configuring PCRE before making, you can use --with-match-limit=n
2535 to set a default value for the compiled library.
2536
2537 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
2538 a different value is set. See 45 below.
2539
2540 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
2541
2542 43. Added a new function pcre_config(int, void *) to enable run-time extraction
2543 of things that can be changed at compile time. The first argument specifies
2544 what is wanted and the second points to where the information is to be placed.
2545 The current list of available information is:
2546
2547 PCRE_CONFIG_UTF8
2548
2549 The output is an integer that is set to one if UTF-8 support is available;
2550 otherwise it is set to zero.
2551
2552 PCRE_CONFIG_NEWLINE
2553
2554 The output is an integer that it set to the value of the code that is used for
2555 newline. It is either LF (10) or CR (13).
2556
2557 PCRE_CONFIG_LINK_SIZE
2558
2559 The output is an integer that contains the number of bytes used for internal
2560 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
2561
2562 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
2563
2564 The output is an integer that contains the threshold above which the POSIX
2565 interface uses malloc() for output vectors. See item 31 above.
2566
2567 PCRE_CONFIG_MATCH_LIMIT
2568
2569 The output is an unsigned integer that contains the default limit of the number
2570 of match() calls in a pcre_exec() execution. See 42 above.
2571
2572 44. pcretest has been upgraded by the addition of the -C option. This causes it
2573 to extract all the available output from the new pcre_config() function, and to
2574 output it. The program then exits immediately.
2575
2576 45. A need has arisen to pass over additional data with calls to pcre_exec() in
2577 order to support additional features. One way would have been to define
2578 pcre_exec2() (for example) with extra arguments, but this would not have been
2579 extensible, and would also have required all calls to the original function to
2580 be mapped to the new one. Instead, I have chosen to extend the mechanism that
2581 is used for passing in "extra" data from pcre_study().
2582
2583 The pcre_extra structure is now exposed and defined in pcre.h. It currently
2584 contains the following fields:
2585
2586 flags a bitmap indicating which of the following fields are set
2587 study_data opaque data from pcre_study()
2588 match_limit a way of specifying a limit on match() calls for a specific
2589 call to pcre_exec()
2590 callout_data data for callouts (see 49 below)
2591
2592 The flag bits are also defined in pcre.h, and are
2593
2594 PCRE_EXTRA_STUDY_DATA
2595 PCRE_EXTRA_MATCH_LIMIT
2596 PCRE_EXTRA_CALLOUT_DATA
2597
2598 The pcre_study() function now returns one of these new pcre_extra blocks, with
2599 the actual study data pointed to by the study_data field, and the
2600 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
2601 before. That is, this change is entirely upwards-compatible and requires no
2602 change to existing code.
2603
2604 If you want to pass in additional data to pcre_exec(), you can either place it
2605 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
2606 block.
2607
2608 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
2609 data string contains the escape sequence \M, pcretest calls pcre_exec() several
2610 times with different match limits, until it finds the minimum value needed for
2611 pcre_exec() to complete. The value is then output. This can be instructive; for
2612 most simple matches the number is quite small, but for pathological cases it
2613 gets very large very quickly.
2614
2615 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
2616 returns the size of the data block pointed to by the study_data field in a
2617 pcre_extra block, that is, the value that was passed as the argument to
2618 pcre_malloc() when PCRE was getting memory in which to place the information
2619 created by pcre_study(). The fourth argument should point to a size_t variable.
2620 pcretest has been extended so that this information is shown after a successful
2621 pcre_study() call when information about the compiled regex is being displayed.
2622
2623 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
2624 because what follows is always an absolute path. (Later: it turns out that this
2625 is more than cosmetic for MinGW, because it doesn't like empty path
2626 components.)
2627
2628 49. Some changes have been made to the callout feature (see 28 above):
2629
2630 (i) A callout function now has three choices for what it returns:
2631
2632 0 => success, carry on matching
2633 > 0 => failure at this point, but backtrack if possible
2634 < 0 => serious error, return this value from pcre_exec()
2635
2636 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
2637 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
2638 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
2639 use by callout functions. It will never be used by PCRE itself.
2640
2641 (ii) The pcre_extra structure (see 45 above) has a void * field called
2642 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
2643 pcre_callout_block structure has a field of the same name. The contents of
2644 the field passed in the pcre_extra structure are passed to the callout
2645 function in the corresponding field in the callout block. This makes it
2646 easier to use the same callout-containing regex from multiple threads. For
2647 testing, the pcretest program has a new data escape
2648
2649 \C*n pass the number n (may be negative) as callout_data
2650
2651 If the callout function in pcretest receives a non-zero value as
2652 callout_data, it returns that value.
2653
2654 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
2655 there were some redundant $(CFLAGS) in commands that are now specified as
2656 $(LINK), which already includes $(CFLAGS).
2657
2658 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
2659 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
2660 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
2661 one-byte characters throughout. Note that case-insensitive matching applies
2662 only to characters whose values are less than 256. PCRE doesn't support the
2663 notion of cases for higher-valued characters.
2664
2665 (i) A character class whose characters are all within 0-255 is handled as
2666 a bit map, and the map is inverted for negative classes. Previously, a
2667 character > 255 always failed to match such a class; however it should
2668 match if the class was a negative one (e.g. [^ab]). This has been fixed.
2669
2670 (ii) A negated character class with a single character < 255 is coded as
2671 "not this character" (OP_NOT). This wasn't working properly when the test
2672 character was multibyte, either singly or repeated.
2673
2674 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
2675 mode, for example: \x{100}{2,3}.
2676
2677 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
2678 singly or repeated) now correctly test multibyte characters. However,
2679 PCRE doesn't recognize any characters with values greater than 255 as
2680 digits, spaces, or word characters. Such characters always match \D, \S,
2681 and \W, and never match \d, \s, or \w.
2682
2683 (v) Classes may now contain characters and character ranges with values
2684 greater than 255. For example: [ab\x{100}-\x{400}].
2685
2686 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
2687 PCRE in UTF-8 mode.
2688
2689 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
2690 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
2691 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
2692 value.)
2693
2694 53. The single man page has become too large. I have therefore split it up into
2695 a number of separate man pages. These also give rise to individual HTML pages;
2696 these are now put in a separate directory, and there is an index.html page that
2697 lists them all. Some hyperlinking between the pages has been installed.
2698
2699 54. Added convenience functions for handling named capturing parentheses.
2700
2701 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
2702 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
2703 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
2704 are faulted.
2705
2706 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
2707 calling configure. These values are used when compiling the dftables.c program
2708 which is run to generate the source of the default character tables. They
2709 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
2710 you will need to set these values.
2711
2712 57. Updated the building process for Windows DLL, as provided by Fred Cox.
2713
2714
2715 Version 3.9 02-Jan-02
2716 ---------------------
2717
2718 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
2719
2720 2. If --disable-static was given, the building process failed when trying to
2721 build pcretest and pcregrep. (For some reason it was using libtool to compile
2722 them, which is not right, as they aren't part of the library.)
2723
2724
2725 Version 3.8 18-Dec-01
2726 ---------------------
2727
2728 1. The experimental UTF-8 code was completely screwed up. It was packing the
2729 bytes in the wrong order. How dumb can you get?
2730
2731
2732 Version 3.7 29-Oct-01
2733 ---------------------
2734
2735 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
2736 This caused pcretest, when used on the test data, to segfault. Unfortunately,
2737 this didn't happen under Solaris 8, where I normally test things.
2738
2739 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
2740 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
2741 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
2742 this fix an hour or so after the initial 3.7 release.)
2743
2744
2745 Version 3.6 23-Oct-01
2746 ---------------------
2747
2748 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
2749 offsets passed as NULL with zero offset count.
2750
2751 2. The config.guess and config.sub files had not been updated when I moved to
2752 the latest autoconf.
2753
2754
2755 Version 3.5 15-Aug-01
2756 ---------------------
2757
2758 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
2759 had been forgotten.
2760
2761 2. By using declared but undefined structures, we can avoid using "void"
2762 definitions in pcre.h while keeping the internal definitions of the structures
2763 private.
2764
2765 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
2766 user point of view, this means that both static and shared libraries are built
2767 by default, but this can be individually controlled. More of the work of
2768 handling this static/shared cases is now inside libtool instead of PCRE's make
2769 file.
2770
2771 4. The pcretest utility is now installed along with pcregrep because it is
2772 useful for users (to test regexs) and by doing this, it automatically gets
2773 relinked by libtool. The documentation has been turned into a man page, so
2774 there are now .1, .txt, and .html versions in /doc.
2775
2776 5. Upgrades to pcregrep:
2777 (i) Added long-form option names like gnu grep.
2778 (ii) Added --help to list all options with an explanatory phrase.
2779 (iii) Added -r, --recursive to recurse into sub-directories.
2780 (iv) Added -f, --file to read patterns from a file.
2781
2782 6. pcre_exec() was referring to its "code" argument before testing that
2783 argument for NULL (and giving an error if it was NULL).
2784
2785 7. Upgraded Makefile.in to allow for compiling in a different directory from
2786 the source directory.
2787
2788 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
2789 options bits, the pointer it was passed was to an int instead of to an unsigned
2790 long int. This mattered only on 64-bit systems.
2791
2792 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
2793 generated) instead of pcre.in, which it its source. Also made the same change
2794 in several of the .c files.
2795
2796 10. A new release of gcc defines printf() as a macro, which broke pcretest
2797 because it had an ifdef in the middle of a string argument for printf(). Fixed
2798 by using separate calls to printf().
2799
2800 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
2801 script, to force use of CR or LF instead of \n in the source. On non-Unix
2802 systems, the value can be set in config.h.
2803
2804 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
2805 absolute limit. Changed the text of the error message to make this clear, and
2806 likewise updated the man page.
2807
2808 13. The limit of 99 on the number of capturing subpatterns has been removed.
2809 The new limit is 65535, which I hope will not be a "real" limit.
2810
2811
2812 Version 3.4 22-Aug-00
2813 ---------------------
2814
2815 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
2816
2817 2. Diagnose condition (?(0) as an error instead of crashing on matching.
2818
2819
2820 Version 3.3 01-Aug-00
2821 ---------------------
2822
2823 1. If an octal character was given, but the value was greater than \377, it
2824 was not getting masked to the least significant bits, as documented. This could
2825 lead to crashes in some systems.
2826
2827 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
2828 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
2829
2830 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
2831 These just pass their arguments on to (pcre_free)(), but they are provided
2832 because some uses of PCRE bind it to non-C systems that can call its functions,
2833 but cannot call free() or pcre_free() directly.
2834
2835 4. Add "make test" as a synonym for "make check". Corrected some comments in
2836 the Makefile.
2837
2838 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
2839 Makefile.
2840
2841 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
2842 command called pgrep for grepping around the active processes.
2843
2844 7. Added the beginnings of support for UTF-8 character strings.
2845
2846 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
2847 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
2848 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
2849 out for the ar command.)
2850
2851
2852 Version 3.2 12-May-00
2853 ---------------------
2854
2855 This is purely a bug fixing release.
2856
2857 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
2858 of ZA. This was just one example of several cases that could provoke this bug,
2859 which was introduced by change 9 of version 2.00. The code for breaking
2860 infinite loops after an iteration that matches an empty string was't working
2861 correctly.
2862
2863 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
2864 when matched against abbab (for example). After matching an empty string, it
2865 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
2866 caused it to match further down the string than it should.
2867
2868 3. The code contained an inclusion of sys/types.h. It isn't clear why this
2869 was there because it doesn't seem to be needed, and it causes trouble on some
2870 systems, as it is not a Standard C header. It has been removed.
2871
2872 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
2873 were reported on the Macintosh. The changes were from
2874
2875 while ((c = *(++ptr)) != 0 && c != '\n');
2876 to
2877 while ((c = *(++ptr)) != 0 && c != '\n') ;
2878
2879 Totally extraordinary, but if that's what it takes...
2880
2881 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
2882 available. Added HAVE_BCOPY and an autoconf test for it; if neither
2883 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
2884 assumes the way PCRE uses memmove() (always moving upwards).
2885
2886 6. PCRE is being used in one environment where strchr() is not available. There
2887 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
2888 faster code anyway.
2889
2890
2891 Version 3.1 09-Feb-00
2892 ---------------------
2893
2894 The only change in this release is the fixing of some bugs in Makefile.in for
2895 the "install" target:
2896
2897 (1) It was failing to install pcreposix.h.
2898
2899 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
2900
2901
2902 Version 3.0 01-Feb-00
2903 ---------------------
2904
2905 1. Add support for the /+ modifier to perltest (to output $` like it does in
2906 pcretest).
2907
2908 2. Add support for the /g modifier to perltest.
2909
2910 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
2911 matches null strings.
2912
2913 4. Fix perltest so that it doesn't do unwanted things when fed an empty
2914 pattern. Perl treats empty patterns specially - it reuses the most recent
2915 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
2916 effect.
2917
2918 5. The POSIX interface was broken in that it was just handing over the POSIX
2919 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
2920 required a bigger vector, with some working space on the end. This means that
2921 the POSIX wrapper now has to get and free some memory, and copy the results.
2922
2923 6. Added some simple autoconf support, placing the test data and the
2924 documentation in separate directories, re-organizing some of the
2925 information files, and making it build pcre-config (a GNU standard). Also added
2926 libtool support for building PCRE as a shared library, which is now the
2927 default.
2928
2929 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
2930 09 are not valid octal constants. Single digits will be used for minor values
2931 less than 10.
2932
2933 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
2934 existing programs that set these in the POSIX interface can use PCRE without
2935 modification.
2936
2937 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
2938 return all that pcre_info() returns, plus additional data. The pcre_info()
2939 function is retained for compatibility, but is considered to be obsolete.
2940
2941 10. Added experimental recursion feature (?R) to handle one common case that
2942 Perl 5.6 will be able to do with (?p{...}).
2943
2944 11. Added support for POSIX character classes like [:alpha:], which Perl is
2945 adopting.
2946
2947
2948 Version 2.08 31-Aug-99
2949 ----------------------
2950
2951 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
2952 trying to match at the startoffset position, but instead was moving forward to
2953 the next newline as if a previous match had failed.
2954
2955 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
2956 and could get into a loop if a null string was matched other than at the start
2957 of the subject.
2958
2959 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
2960 be distinguished at compile time, and for completeness also added PCRE_DATE.
2961
2962 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
2963 in GnuWin32 environments.
2964
2965
2966 Version 2.07 29-Jul-99
2967 ----------------------
2968
2969 1. The documentation is now supplied in plain text form and HTML as well as in
2970 the form of man page sources.
2971
2972 2. C++ compilers don't like assigning (void *) values to other pointer types.
2973 In particular this affects malloc(). Although there is no problem in Standard
2974 C, I've put in casts to keep C++ compilers happy.
2975
2976 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
2977 should be (const char *).
2978
2979 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
2980 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
2981 However, I haven't made this a standard facility. The documentation doesn't
2982 mention it, and the Makefile doesn't support it.
2983
2984 5. The Makefile now contains an "install" target, with editable destinations at
2985 the top of the file. The pcretest program is not installed.
2986
2987 6. pgrep -V now gives the PCRE version number and date.
2988
2989 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
2990 causing the entire string to be ignored, instead of just the last character.
2991
2992 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
2993 non-matching string, it can take a very, very long time, even for strings of
2994 quite modest length, because of the nested recursion. PCRE now does better in
2995 some of these cases. It does this by remembering the last required literal
2996 character in the pattern, and pre-searching the subject to ensure it is present
2997 before running the real match. In other words, it applies a heuristic to detect
2998 some types of certain failure quickly, and in the above example, if presented
2999 with a string that has no trailing " it gives "no match" very quickly.
3000
3001 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
3002 other alternatives are tried instead.
3003
3004
3005 Version 2.06 09-Jun-99
3006 ----------------------
3007
3008 1. Change pcretest's output for amount of store used to show just the code
3009 space, because the remainder (the data block) varies in size between 32-bit and
3010 64-bit systems.
3011
3012 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
3013 start matching at. This allows lookbehinds to work when searching for multiple
3014 occurrences in a string.
3015
3016 3. Added additional options to pcretest for testing multiple occurrences:
3017
3018 /+ outputs the rest of the string that follows a match
3019 /g loops for multiple occurrences, using the new startoffset argument
3020 /G loops for multiple occurrences by passing an incremented pointer
3021
3022 4. PCRE wasn't doing the "first character" optimization for patterns starting
3023 with \b or \B, though it was doing it for other lookbehind assertions. That is,
3024 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
3025 the letter 'x'. On long subject strings, this gives a significant speed-up.
3026
3027
3028 Version 2.05 21-Apr-99
3029 ----------------------
3030
3031 1. Changed the type of magic_number from int to long int so that it works
3032 properly on 16-bit systems.
3033
3034 2. Fixed a bug which caused patterns starting with .* not to work correctly
3035 when the subject string contained newline characters. PCRE was assuming
3036 anchoring for such patterns in all cases, which is not correct because .* will
3037 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
3038 DOTALL is set at top level; otherwise it knows that patterns starting with .*
3039 must be retried after every newline in the subject.
3040
3041
3042 Version 2.04 18-Feb-99
3043 ----------------------
3044
3045 1. For parenthesized subpatterns with repeats whose minimum was zero, the
3046 computation of the store needed to hold the pattern was incorrect (too large).
3047 If such patterns were nested a few deep, this could multiply and become a real
3048 problem.
3049
3050 2. Added /M option to pcretest to show the memory requirement of a specific
3051 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
3052
3053 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
3054 compiled in such a way that the backtracking after subsequent failure was
3055 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
3056 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
3057
3058
3059 Version 2.03 02-Feb-99
3060 ----------------------
3061
3062 1. Fixed typo and small mistake in man page.
3063
3064 2. Added 4th condition (GPL supersedes if conflict) and created separate
3065 LICENCE file containing the conditions.
3066
3067 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
3068 Perl, that is the internal \ allows the delimiter to be included in the
3069 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
3070 the final delimiter, add \ to the end of the pattern (to test the error).
3071
3072 4. Added the convenience functions for extracting substrings after a successful
3073 match. Updated pcretest to make it able to test these functions.
3074
3075
3076 Version 2.02 14-Jan-99
3077 ----------------------
3078
3079 1. Initialized the working variables associated with each extraction so that
3080 their saving and restoring doesn't refer to uninitialized store.
3081
3082 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
3083 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
3084 fix the problem.
3085
3086 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
3087 calls, and wasn't printing the correct value for compiling calls. Increased the
3088 default value of LOOPREPEAT, and the number of significant figures in the
3089 times.
3090
3091 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
3092
3093 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
3094 a building problem on Windows NT with a FAT file system.
3095
3096
3097 Version 2.01 21-Oct-98
3098 ----------------------
3099
3100 1. Changed the API for pcre_compile() to allow for the provision of a pointer
3101 to character tables built by pcre_maketables() in the current locale. If NULL
3102 is passed, the default tables are used.
3103
3104
3105 Version 2.00 24-Sep-98
3106 ----------------------
3107
3108 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
3109 it any more.
3110
3111 2. Allow quantification of (?>) groups, and make it work correctly.
3112
3113 3. The first character computation wasn't working for (?>) groups.
3114
3115 4. Correct the implementation of \Z (it is permitted to match on the \n at the
3116 end of the subject) and add 5.005's \z, which really does match only at the
3117 very end of the subject.
3118
3119 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
3120
3121 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
3122 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
3123 localized options. All options to pcre_study() were also removed.
3124
3125 7. Add other new features from 5.005:
3126
3127 $(?<= positive lookbehind
3128 $(?<! negative lookbehind
3129 (?imsx-imsx) added the unsetting capability
3130 such a setting is global if at outer level; local otherwise
3131 (?imsx-imsx:) non-capturing groups with option setting
3132 (?(cond)re|re) conditional pattern matching
3133
3134 A backreference to itself in a repeated group matches the previous
3135 captured string.
3136
3137 8. General tidying up of studying (both automatic and via "study")
3138 consequential on the addition of new assertions.
3139
3140 9. As in 5.005, unlimited repeated groups that could match an empty substring
3141 are no longer faulted at compile time. Instead, the loop is forcibly broken at
3142 runtime if any iteration does actually match an empty substring.
3143
3144 10. Include the RunTest script in the distribution.
3145
3146 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
3147 discrepancies, some of which were old and were also with respect to 5.004. They
3148 have now been fixed.
3149
3150
3151 Version 1.09 28-Apr-98
3152 ----------------------
3153
3154 1. A negated single character class followed by a quantifier with a minimum
3155 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
3156 program crashes, or just wrong answers. This did not apply to negated classes
3157 containing more than one character, or to minima other than one.
3158
3159
3160 Version 1.08 27-Mar-98
3161 ----------------------
3162
3163 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
3164
3165 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
3166 latter must appear before anything that relies on it in the pattern.
3167
3168
3169 Version 1.07 16-Feb-98
3170 ----------------------
3171
3172 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
3173 repeat of a potentially empty string).
3174
3175
3176 Version 1.06 23-Jan-98
3177 ----------------------
3178
3179 1. Added Markus Oberhumer's little patches for C++.
3180
3181 2. Literal strings longer than 255 characters were broken.
3182
3183
3184 Version 1.05 23-Dec-97
3185 ----------------------
3186
3187 1. Negated character classes containing more than one character were failing if
3188 PCRE_CASELESS was set at run time.
3189
3190
3191 Version 1.04 19-Dec-97
3192 ----------------------
3193
3194 1. Corrected the man page, where some "const" qualifiers had been omitted.
3195
3196 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
3197 input syntax.
3198
3199 3. Fixed memory leak which occurred when a regex with back references was
3200 matched with an offsets vector that wasn't big enough. The temporary memory
3201 that is used in this case wasn't being freed if the match failed.
3202
3203 4. Tidied pcretest to ensure it frees memory that it gets.
3204
3205 5. Temporary memory was being obtained in the case where the passed offsets
3206 vector was exactly big enough.
3207
3208 6. Corrected definition of offsetof() from change 5 below.
3209
3210 7. I had screwed up change 6 below and broken the rules for the use of
3211 setjmp(). Now fixed.
3212
3213
3214 Version 1.03 18-Dec-97
3215 ----------------------
3216
3217 1. A erroneous regex with a missing opening parenthesis was correctly
3218 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
3219 on some systems.
3220
3221 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
3222 it was reported that one broken compiler failed on the former because "code" is
3223 also an independent variable.
3224
3225 3. The erroneous regex a[]b caused an array overrun reference.
3226
3227 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
3228 fail on data ending with that character. (It was going on too far, and checking
3229 the next character, typically a binary zero.) This was specific to the
3230 optimized code for single-character negative classes.
3231
3232 5. Added a contributed patch from the TIN world which does the following:
3233
3234 + Add an undef for memmove, in case the the system defines a macro for it.
3235
3236 + Add a definition of offsetof(), in case there isn't one. (I don't know
3237 the reason behind this - offsetof() is part of the ANSI standard - but
3238 it does no harm).
3239
3240 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
3241 most of the places where whitespace preceded '#'. I have given up and
3242 allowed the remaining 2 cases to be at the margin.
3243
3244 + Rename some variables in pcre to eliminate shadowing. This seems very
3245 pedantic, but does no harm, of course.
3246
3247 6. Moved the call to setjmp() into its own function, to get rid of warnings
3248 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
3249
3250 7. Constructs such as \d{8,} were compiling into the equivalent of
3251 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
3252 outcome, but in this particular case used more store than had been allocated,
3253 which caused the bug to be discovered because it threw up an internal error.
3254
3255 8. The debugging code in both pcre and pcretest for outputting the compiled
3256 form of a regex was going wrong in the case of back references followed by
3257 curly-bracketed repeats.
3258
3259
3260 Version 1.02 12-Dec-97
3261 ----------------------
3262
3263 1. Typos in pcre.3 and comments in the source fixed.
3264
3265 2. Applied a contributed patch to get rid of places where it used to remove
3266 'const' from variables, and fixed some signed/unsigned and uninitialized
3267 variable warnings.
3268
3269 3. Added the "runtest" target to Makefile.
3270
3271 4. Set default compiler flag to -O2 rather than just -O.
3272
3273
3274 Version 1.01 19-Nov-97
3275 ----------------------
3276
3277 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
3278 like /([ab]*)*/, that is, for classes with more than one character in them.
3279
3280 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
3281 as /((?>a*))*/ (a PCRE_EXTRA facility).
3282
3283
3284 Version 1.00 18-Nov-97
3285 ----------------------
3286
3287 1. Added compile-time macros to support systems such as SunOS4 which don't have
3288 memmove() or strerror() but have other things that can be used instead.
3289
3290 2. Arranged that "make clean" removes the executables.
3291
3292
3293 Version 0.99 27-Oct-97
3294 ----------------------
3295
3296 1. Fixed bug in code for optimizing classes with only one character. It was
3297 initializing a 32-byte map regardless, which could cause it to run off the end
3298 of the memory it had got.
3299
3300 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
3301
3302
3303 Version 0.98 22-Oct-97
3304 ----------------------
3305
3306 1. Fixed bug in code for handling temporary memory usage when there are more
3307 back references than supplied space in the ovector. This could cause segfaults.
3308
3309
3310 Version 0.97 21-Oct-97
3311 ----------------------
3312
3313 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
3314
3315 2. Optimized negated single characters not to use a bit map.
3316
3317 3. Brought error texts together as macro definitions; clarified some of them;
3318 fixed one that was wrong - it said "range out of order" when it meant "invalid
3319 escape sequence".
3320
3321 4. Changed some char * arguments to const char *.
3322
3323 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
3324
3325 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
3326 pcretest.
3327
3328
3329 Version 0.96 16-Oct-97
3330 ----------------------
3331
3332 1. Added a simple "pgrep" utility to the distribution.
3333
3334 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
3335 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
3336 where "ddd" means "one or more decimal digits".
3337
3338 3. Fixed serious bug. If a pattern had a back reference, but the call to
3339 pcre_exec() didn't supply a large enough ovector to record the related
3340 identifying subpattern, the match always failed. PCRE now remembers the number
3341 of the largest back reference, and gets some temporary memory in which to save
3342 the offsets during matching if necessary, in order to ensure that
3343 backreferences always work.
3344
3345 4. Increased the compatibility with Perl in a number of ways:
3346
3347 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
3348 to request this handling. The option can be set at compile or exec time.
3349
3350 (b) $ matches before a terminating newline by default; an option
3351 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
3352 mode). The option can be set at compile or exec time.
3353
3354 (c) The handling of \ followed by a digit other than 0 is now supposed to be
3355 the same as Perl's. If the decimal number it represents is less than 10
3356 or there aren't that many previous left capturing parentheses, an octal
3357 escape is read. Inside a character class, it's always an octal escape,
3358 even if it is a single digit.
3359
3360 (d) An escaped but undefined alphabetic character is taken as a literal,
3361 unless PCRE_EXTRA is set. Currently this just reserves the remaining
3362 escapes.
3363
3364 (e) {0} is now permitted. (The previous item is removed from the compiled
3365 pattern).
3366
3367 5. Changed all the names of code files so that the basic parts are no longer
3368 than 10 characters, and abolished the teeny "globals.c" file.
3369
3370 6. Changed the handling of character classes; they are now done with a 32-byte
3371 bit map always.
3372
3373 7. Added the -d and /D options to pcretest to make it possible to look at the
3374 internals of compilation without having to recompile pcre.
3375
3376
3377 Version 0.95 23-Sep-97
3378 ----------------------
3379
3380 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
3381 \x20 at the start of a run of normal characters. These were being treated as
3382 real characters, instead of the source characters being re-checked.
3383
3384
3385 Version 0.94 18-Sep-97
3386 ----------------------
3387
3388 1. The functions are now thread-safe, with the caveat that the global variables
3389 containing pointers to malloc() and free() or alternative functions are the
3390 same for all threads.
3391
3392 2. Get pcre_study() to generate a bitmap of initial characters for non-
3393 anchored patterns when this is possible, and use it if passed to pcre_exec().
3394
3395
3396 Version 0.93 15-Sep-97
3397 ----------------------
3398
3399 1. /(b)|(:+)/ was computing an incorrect first character.
3400
3401 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
3402 but not actually doing anything yet.
3403
3404 3. Treat "-" characters in classes that cannot be part of ranges as literals,
3405 as Perl does (e.g. [-az] or [az-]).
3406
3407 4. Set the anchored flag if a branch starts with .* or .*? because that tests
3408 all possible positions.
3409
3410 5. Split up into different modules to avoid including unneeded functions in a
3411 compiled binary. However, compile and exec are still in one module. The "study"
3412 function is split off.
3413
3414 6. The character tables are now in a separate module whose source is generated
3415 by an auxiliary program - but can then be edited by hand if required. There are
3416 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
3417 toupper() in the code.
3418
3419 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
3420 make them global. Abolish the function for setting them, as the caller can now
3421 set them directly.
3422
3423
3424 Version 0.92 11-Sep-97
3425 ----------------------
3426
3427 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
3428 (e.g. /a{1,3}/) was broken (I mis-optimized it).
3429
3430 2. Caseless matching was not working in character classes if the characters in
3431 the pattern were in upper case.
3432
3433 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
3434
3435 4. Make PCRE_ANCHORED public and accept as a compile option.
3436
3437 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
3438 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
3439 pass them.
3440
3441 6. Give an error if bad option bits passed at compile or run time.
3442
3443 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
3444 pcretest to cause it to pass that flag.
3445
3446 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
3447 options, and the first character, if set.
3448
3449 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
3450
3451
3452 Version 0.91 10-Sep-97
3453 ----------------------
3454
3455 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
3456 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
3457
3458 2. PCRE was looping on encountering an indefinitely repeated back reference to
3459 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
3460 Perl does - treats the match as successful.
3461
3462 ****

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