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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 418 - (show annotations) (download)
Fri Apr 24 09:17:18 2009 UTC (5 years ago) by ph10
File size: 165868 byte(s)
Fix POSIX error code bug.

1 ChangeLog for PCRE
2 ------------------
3
4 Version 8.00 ??-???-??
5 ----------------------
6
7 1. The table for translating pcre_compile() error codes into POSIX error codes
8 was out-of-date, and there was no check on the pcre_compile() error code
9 being within the table. This could lead to an OK return being given in
10 error.
11
12
13 Version 7.9 11-Apr-09
14 ---------------------
15
16 1. When building with support for bzlib/zlib (pcregrep) and/or readline
17 (pcretest), all targets were linked against these libraries. This included
18 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
19 libraries. This caused unwanted dependencies to be created. This problem
20 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
21 pcretest is linked with readline.
22
23 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
24 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
25 moved outside it again, because FALSE and TRUE are already defined in AIX,
26 but BOOL is not.
27
28 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
29 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
30
31 4. The pcregrep documentation said spaces were inserted as well as colons (or
32 hyphens) following file names and line numbers when outputting matching
33 lines. This is not true; no spaces are inserted. I have also clarified the
34 wording for the --colour (or --color) option.
35
36 5. In pcregrep, when --colour was used with -o, the list of matching strings
37 was not coloured; this is different to GNU grep, so I have changed it to be
38 the same.
39
40 6. When --colo(u)r was used in pcregrep, only the first matching substring in
41 each matching line was coloured. Now it goes on to look for further matches
42 of any of the test patterns, which is the same behaviour as GNU grep.
43
44 7. A pattern that could match an empty string could cause pcregrep to loop; it
45 doesn't make sense to accept an empty string match in pcregrep, so I have
46 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
47 seems to be how GNU grep behaves.
48
49 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
50 start or after a newline", because the conditional assertion was not being
51 correctly handled. The rule now is that both the assertion and what follows
52 in the first alternative must satisfy the test.
53
54 9. If auto-callout was enabled in a pattern with a conditional group whose
55 condition was an assertion, PCRE could crash during matching, both with
56 pcre_exec() and pcre_dfa_exec().
57
58 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
59 used for matching.
60
61 11. Unicode property support in character classes was not working for
62 characters (bytes) greater than 127 when not in UTF-8 mode.
63
64 12. Added the -M command line option to pcretest.
65
66 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
67
68 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
69
70 16. Added comments and documentation about mis-use of no_arg in the C++
71 wrapper.
72
73 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
74 from Martin Jerabek that uses macro names for all relevant character and
75 string constants.
76
77 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
78 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
79 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
80 these, but not everybody uses configure.
81
82 19. A conditional group that had only one branch was not being correctly
83 recognized as an item that could match an empty string. This meant that an
84 enclosing group might also not be so recognized, causing infinite looping
85 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
86 with the subject "ab", where knowledge that the repeated group can match
87 nothing is needed in order to break the loop.
88
89 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
90 exec(), but without supplying a callout function, matching went wrong.
91
92 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
93 leak if the size of the offset vector was greater than 30. When the vector
94 is smaller, the saved offsets during recursion go onto a local stack
95 vector, but for larger vectors malloc() is used. It was failing to free
96 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
97 error, in fact).
98
99 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
100 heapframe that is used only when UTF-8 support is enabled. This caused no
101 problem, but was untidy.
102
103 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
104 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
105 included within another project.
106
107 24. Steven Van Ingelgem's patches to add more options to the CMake support,
108 slightly modified by me:
109
110 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
111 not building pcregrep.
112
113 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
114 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
115
116 25. Forward references, both numeric and by name, in patterns that made use of
117 duplicate group numbers, could behave incorrectly or give incorrect errors,
118 because when scanning forward to find the reference group, PCRE was not
119 taking into account the duplicate group numbers. A pattern such as
120 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
121
122 26. Changed a few more instances of "const unsigned char *" to USPTR, making
123 the feature of a custom pointer more persuasive (as requested by a user).
124
125 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
126 pcretest.c, inside #ifndefs, because it seems they are sometimes already
127 pre-defined.
128
129 28. Added support for (*UTF8) at the start of a pattern.
130
131 29. Arrange for flags added by the "release type" setting in CMake to be shown
132 in the configuration summary.
133
134
135 Version 7.8 05-Sep-08
136 ---------------------
137
138 1. Replaced UCP searching code with optimized version as implemented for Ad
139 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
140 stage table and inline lookup instead of a function, giving speed ups of 2
141 to 5 times on some simple patterns that I tested. Permission was given to
142 distribute the MultiStage2.py script that generates the tables (it's not in
143 the tarball, but is in the Subversion repository).
144
145 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
146 scripts.
147
148 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
149 a group with a zero qualifier. The result of the study could be incorrect,
150 or the function might crash, depending on the pattern.
151
152 4. Caseless matching was not working for non-ASCII characters in back
153 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
154 It now works when Unicode Property Support is available.
155
156 5. In pcretest, an escape such as \x{de} in the data was always generating
157 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
158 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
159 truncation.
160
161 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
162
163 7. Added two (int) casts to pcregrep when printing the difference of two
164 pointers, in case they are 64-bit values.
165
166 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
167 test 2 if it fails.
168
169 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
170 and a #define of that name to empty if it is not externally set. This is to
171 allow users of MSVC to set it if necessary.
172
173 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
174 the convenience functions in the pcre_get.c source file.
175
176 11. An option change at the start of a pattern that had top-level alternatives
177 could cause overwriting and/or a crash. This command provoked a crash in
178 some environments:
179
180 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
181
182 This potential security problem was recorded as CVE-2008-2371.
183
184 12. For a pattern where the match had to start at the beginning or immediately
185 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
186 pcre_dfa_exec() could read past the end of the passed subject if there was
187 no match. To help with detecting such bugs (e.g. with valgrind), I modified
188 pcretest so that it places the subject at the end of its malloc-ed buffer.
189
190 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
191 exec() might read past the end of the data buffer in UTF-8 mode.
192
193 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
194 the data contained the byte 0x85 as part of a UTF-8 character within its
195 first line. This applied both to normal and DFA matching.
196
197 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
198 /^[^d]*?$/8 failed to match "abc".
199
200 16. Added a missing copyright notice to pcrecpp_internal.h.
201
202 17. Make it more clear in the documentation that values returned from
203 pcre_exec() in ovector are byte offsets, not character counts.
204
205 18. Tidied a few places to stop certain compilers from issuing warnings.
206
207 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
208 supplied by Stefan Weber. I made a further small update for 7.8 because
209 there is a change of source arrangements: the pcre_searchfuncs.c module is
210 replaced by pcre_ucd.c.
211
212
213 Version 7.7 07-May-08
214 ---------------------
215
216 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
217 a string to a long long, pretend we don't even have a long long." This is
218 done by checking for the strtoq, strtoll, and _strtoi64 functions.
219
220 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
221 pre-7.6 versions, which defined a global no_arg variable instead of putting
222 it in the RE class. (See also #8 below.)
223
224 3. Remove a line of dead code, identified by coverity and reported by Nuno
225 Lopes.
226
227 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
228
229 (1) The include/exclude patterns were being applied to the whole pathnames
230 of files, instead of just to the final components.
231
232 (2) If there was more than one level of directory, the subdirectories were
233 skipped unless they satisfied the include/exclude conditions. This is
234 inconsistent with GNU grep (and could even be seen as contrary to the
235 pcregrep specification - which I improved to make it absolutely clear).
236 The action now is always to scan all levels of directory, and just
237 apply the include/exclude patterns to regular files.
238
239 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
240 --exclude_dir in the tests to avoid scanning .svn directories.
241
242 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
243 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
244 doesn't support NULs in patterns.
245
246 7. Added some missing "const"s to declarations of static tables in
247 pcre_compile.c and pcre_dfa_exec.c.
248
249 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
250 caused by fix #2 above. (Subsequently also a second patch to fix the
251 first patch. And a third patch - this was a messy problem.)
252
253 9. Applied Craig's patch to remove the use of push_back().
254
255 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
256 matching function regexec().
257
258 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
259 which, however, unlike Perl's \g{...}, are subroutine calls, not back
260 references. PCRE supports relative numbers with this syntax (I don't think
261 Oniguruma does).
262
263 12. Previously, a group with a zero repeat such as (...){0} was completely
264 omitted from the compiled regex. However, this means that if the group
265 was called as a subroutine from elsewhere in the pattern, things went wrong
266 (an internal error was given). Such groups are now left in the compiled
267 pattern, with a new opcode that causes them to be skipped at execution
268 time.
269
270 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
271 to the way PCRE behaves:
272
273 (a) A lone ] character is dis-allowed (Perl treats it as data).
274
275 (b) A back reference to an unmatched subpattern matches an empty string
276 (Perl fails the current match path).
277
278 (c) A data ] in a character class must be notated as \] because if the
279 first data character in a class is ], it defines an empty class. (In
280 Perl it is not possible to have an empty class.) The empty class []
281 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
282 The negative empty class [^] matches any one character, independently
283 of the DOTALL setting.
284
285 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
286 non-existent subpattern following a character class starting with ']' and
287 containing () gave an internal compiling error instead of "reference to
288 non-existent subpattern". Fortunately, when the pattern did exist, the
289 compiled code was correct. (When scanning forwards to check for the
290 existencd of the subpattern, it was treating the data ']' as terminating
291 the class, so got the count wrong. When actually compiling, the reference
292 was subsequently set up correctly.)
293
294 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
295 it was being rejected as not supported by pcre_dfa_exec(), even though
296 other assertions are supported. I have made pcre_dfa_exec() support
297 (*FAIL).
298
299 16. The implementation of 13c above involved the invention of a new opcode,
300 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
301 cannot be changed at match time, I realized I could make a small
302 improvement to matching performance by compiling OP_ALLANY instead of
303 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
304 on the OP_ANY path.
305
306 17. Compiling pcretest on Windows with readline support failed without the
307 following two fixes: (1) Make the unistd.h include conditional on
308 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
309
310 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
311 ncurses library to be included for pcretest when ReadLine support is
312 requested, but also to allow for it to be overridden. This patch came from
313 Daniel Bergström.
314
315 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
316 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
317 any errors with the current Unicode tables. Thanks to Peter Kankowski for
318 spotting this.
319
320
321 Version 7.6 28-Jan-08
322 ---------------------
323
324 1. A character class containing a very large number of characters with
325 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
326 overflow.
327
328 2. Patch to cut out the "long long" test in pcrecpp_unittest when
329 HAVE_LONG_LONG is not defined.
330
331 3. Applied Christian Ehrlicher's patch to update the CMake build files to
332 bring them up to date and include new features. This patch includes:
333
334 - Fixed PH's badly added libz and libbz2 support.
335 - Fixed a problem with static linking.
336 - Added pcredemo. [But later removed - see 7 below.]
337 - Fixed dftables problem and added an option.
338 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
339 HAVE_LONG_LONG.
340 - Added readline support for pcretest.
341 - Added an listing of the option settings after cmake has run.
342
343 4. A user submitted a patch to Makefile that makes it easy to create
344 "pcre.dll" under mingw when using Configure/Make. I added stuff to
345 Makefile.am that cause it to include this special target, without
346 affecting anything else. Note that the same mingw target plus all
347 the other distribution libraries and programs are now supported
348 when configuring with CMake (see 6 below) instead of with
349 Configure/Make.
350
351 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
352 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
353 exported in the Windows port". It has not yet been confirmed that the patch
354 solves the problem, but it does no harm.
355
356 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
357 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
358 with CMake, and also correct the comment about stack recursion.
359
360 7. Remove the automatic building of pcredemo from the ./configure system and
361 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
362 of a program that users should build themselves after PCRE is installed, so
363 building it automatically is not really right. What is more, it gave
364 trouble in some build environments.
365
366 8. Further tidies to CMakeLists.txt from Sheri and Christian.
367
368
369 Version 7.5 10-Jan-08
370 ---------------------
371
372 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
373 values in parens when parsing an RE using the C++ wrapper."
374
375 2. Negative specials like \S did not work in character classes in UTF-8 mode.
376 Characters greater than 255 were excluded from the class instead of being
377 included.
378
379 3. The same bug as (2) above applied to negated POSIX classes such as
380 [:^space:].
381
382 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
383 defined or documented. It seems to have been a typo for PCRE_STATIC, so
384 I have changed it.
385
386 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
387 first named subpattern) and a construct such as (?&a) would reference the
388 first named subpattern whose name started with "a" (in other words, the
389 length check was missing). Both these problems are fixed. "Subpattern name
390 expected" is now given for (?&) (a zero-length name), and this patch also
391 makes it give the same error for \k'' (previously it complained that that
392 was a reference to a non-existent subpattern).
393
394 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
395 this is right because (?- can be followed by option settings as well as by
396 digits. I have, however, made the messages clearer.
397
398 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
399 than the number used in the conditional) now cause a compile-time error.
400 This is actually not compatible with Perl, which accepts such patterns, but
401 treats the conditional as always being FALSE (as PCRE used to), but it
402 seems to me that giving a diagnostic is better.
403
404 8. Change "alphameric" to the more common word "alphanumeric" in comments
405 and messages.
406
407 9. Fix two occurrences of "backslash" in comments that should have been
408 "backspace".
409
410 10. Remove two redundant lines of code that can never be obeyed (their function
411 was moved elsewhere).
412
413 11. The program that makes PCRE's Unicode character property table had a bug
414 which caused it to generate incorrect table entries for sequences of
415 characters that have the same character type, but are in different scripts.
416 It amalgamated them into a single range, with the script of the first of
417 them. In other words, some characters were in the wrong script. There were
418 thirteen such cases, affecting characters in the following ranges:
419
420 U+002b0 - U+002c1
421 U+0060c - U+0060d
422 U+0061e - U+00612
423 U+0064b - U+0065e
424 U+0074d - U+0076d
425 U+01800 - U+01805
426 U+01d00 - U+01d77
427 U+01d9b - U+01dbf
428 U+0200b - U+0200f
429 U+030fc - U+030fe
430 U+03260 - U+0327f
431 U+0fb46 - U+0fbb1
432 U+10450 - U+1049d
433
434 12. The -o option (show only the matching part of a line) for pcregrep was not
435 compatible with GNU grep in that, if there was more than one match in a
436 line, it showed only the first of them. It now behaves in the same way as
437 GNU grep.
438
439 13. If the -o and -v options were combined for pcregrep, it printed a blank
440 line for every non-matching line. GNU grep prints nothing, and pcregrep now
441 does the same. The return code can be used to tell if there were any
442 non-matching lines.
443
444 14. Added --file-offsets and --line-offsets to pcregrep.
445
446 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
447 infinitely looping recursion. The bug was that positive lookaheads were not
448 being skipped when checking for a possible empty match (negative lookaheads
449 and both kinds of lookbehind were skipped).
450
451 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
452 inclusion of <windows.h> to before rather than after the definition of
453 INVALID_FILE_ATTRIBUTES (patch from David Byron).
454
455 17. Specifying a possessive quantifier with a specific limit for a Unicode
456 character property caused pcre_compile() to compile bad code, which led at
457 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
458 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
459 caused the error; without that there was no problem.
460
461 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
462
463 19. Added --enable-pcretest-libreadline.
464
465 20. In pcrecpp.cc, the variable 'count' was incremented twice in
466 RE::GlobalReplace(). As a result, the number of replacements returned was
467 double what it should be. I removed one of the increments, but Craig sent a
468 later patch that removed the other one (the right fix) and added unit tests
469 that check the return values (which was not done before).
470
471 21. Several CMake things:
472
473 (1) Arranged that, when cmake is used on Unix, the libraries end up with
474 the names libpcre and libpcreposix, not just pcre and pcreposix.
475
476 (2) The above change means that pcretest and pcregrep are now correctly
477 linked with the newly-built libraries, not previously installed ones.
478
479 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
480
481 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
482 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
483 UTF-8 newline character). The key issue is that the pattern starts .*;
484 this means that the match must be either at the beginning, or after a
485 newline. The bug was in the code for advancing after a failed match and
486 checking that the new position followed a newline. It was not taking
487 account of UTF-8 characters correctly.
488
489 23. PCRE was behaving differently from Perl in the way it recognized POSIX
490 character classes. PCRE was not treating the sequence [:...:] as a
491 character class unless the ... were all letters. Perl, however, seems to
492 allow any characters between [: and :], though of course it rejects as
493 unknown any "names" that contain non-letters, because all the known class
494 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
495 for example, whereas PCRE did not - it did not recognize a POSIX character
496 class. This seemed a bit dangerous, so the code has been changed to be
497 closer to Perl. The behaviour is not identical to Perl, because PCRE will
498 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
499 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
500 Perl does, and where it didn't before.
501
502 24. Rewrite so as to remove the single use of %n from pcregrep because in some
503 Windows environments %n is disabled by default.
504
505
506 Version 7.4 21-Sep-07
507 ---------------------
508
509 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
510 means that a class such as [\s] counted as "explicit reference to CR or
511 LF". That isn't really right - the whole point of the change was to try to
512 help when there was an actual mention of one of the two characters. So now
513 the change happens only if \r or \n (or a literal CR or LF) character is
514 encountered.
515
516 2. The 32-bit options word was also used for 6 internal flags, but the numbers
517 of both had grown to the point where there were only 3 bits left.
518 Fortunately, there was spare space in the data structure, and so I have
519 moved the internal flags into a new 16-bit field to free up more option
520 bits.
521
522 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
523 but did not set the internal JCHANGED flag - either of these is enough to
524 control the way the "get" function works - but the PCRE_INFO_JCHANGED
525 facility is supposed to tell if (?J) was ever used, so now (?J) at the
526 start sets both bits.
527
528 4. Added options (at build time, compile time, exec time) to change \R from
529 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
530
531 5. doc/pcresyntax.html was missing from the distribution.
532
533 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
534 compatibility, even though it is no longer used.
535
536 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
537 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
538 windows.h file is present (where different names are used). [This was
539 reversed later after testing - see 16 below.]
540
541 8. Changed all #include <config.h> to #include "config.h". There were also
542 some further <pcre.h> cases that I changed to "pcre.h".
543
544 9. When pcregrep was used with the --colour option, it missed the line ending
545 sequence off the lines that it output.
546
547 10. It was pointed out to me that arrays of string pointers cause lots of
548 relocations when a shared library is dynamically loaded. A technique of
549 using a single long string with a table of offsets can drastically reduce
550 these. I have refactored PCRE in four places to do this. The result is
551 dramatic:
552
553 Originally: 290
554 After changing UCP table: 187
555 After changing error message table: 43
556 After changing table of "verbs" 36
557 After changing table of Posix names 22
558
559 Thanks to the folks working on Gregex for glib for this insight.
560
561 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
562 unicode-properties was also set.
563
564 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
565
566 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
567 checked only for CRLF.
568
569 14. Added casts to pcretest.c to avoid compiler warnings.
570
571 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
572
573 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
574 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
575 entirely. This removes changes made in 7 above.
576
577 17. The CMake files have been updated, and there is now more information about
578 building with CMake in the NON-UNIX-USE document.
579
580
581 Version 7.3 28-Aug-07
582 ---------------------
583
584 1. In the rejigging of the build system that eventually resulted in 7.1, the
585 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
586 brackets there is not right, since it causes compilers to look for an
587 installed pcre.h, not the version that is in the source that is being
588 compiled (which of course may be different). I have changed it back to:
589
590 #include "pcre.h"
591
592 I have a vague recollection that the change was concerned with compiling in
593 different directories, but in the new build system, that is taken care of
594 by the VPATH setting the Makefile.
595
596 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
597 when the subject happened to end in the byte 0x85 (e.g. if the last
598 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
599 characters but of course it shouldn't be taken as a newline when it is part
600 of another character. The bug was that, for an unlimited repeat of . in
601 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
602 characters when looking for a newline.
603
604 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
605
606 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
607 in debug output.
608
609 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
610 long printing in the pcrecpp unittest when running under MinGW.
611
612 6. ESC_K was left out of the EBCDIC table.
613
614 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
615 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
616 limit also applies to "virtual nesting" when a pattern is recursive, and in
617 this case 1000 isn't so big. I have been able to remove this limit at the
618 expense of backing off one optimization in certain circumstances. Normally,
619 when pcre_exec() would call its internal match() function recursively and
620 immediately return the result unconditionally, it uses a "tail recursion"
621 feature to save stack. However, when a subpattern that can match an empty
622 string has an unlimited repetition quantifier, it no longer makes this
623 optimization. That gives it a stack frame in which to save the data for
624 checking that an empty string has been matched. Previously this was taken
625 from the 1000-entry workspace that had been reserved. So now there is no
626 explicit limit, but more stack is used.
627
628 8. Applied Daniel's patches to solve problems with the import/export magic
629 syntax that is required for Windows, and which was going wrong for the
630 pcreposix and pcrecpp parts of the library. These were overlooked when this
631 problem was solved for the main library.
632
633 9. There were some crude static tests to avoid integer overflow when computing
634 the size of patterns that contain repeated groups with explicit upper
635 limits. As the maximum quantifier is 65535, the maximum group length was
636 set at 30,000 so that the product of these two numbers did not overflow a
637 32-bit integer. However, it turns out that people want to use groups that
638 are longer than 30,000 bytes (though not repeat them that many times).
639 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
640 made it possible to implement the integer overflow checks in a much more
641 dynamic way, which I have now done. The artificial limitation on group
642 length has been removed - we now have only the limit on the total length of
643 the compiled pattern, which depends on the LINK_SIZE setting.
644
645 10. Fixed a bug in the documentation for get/copy named substring when
646 duplicate names are permitted. If none of the named substrings are set, the
647 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
648 empty string.
649
650 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
651 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
652 because the ] is interpreted as the first data character and the
653 terminating ] is not found. PCRE has been made compatible with Perl in this
654 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
655 cause memory overwriting.
656
657 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
658 string has been matched (to stop an infinite loop). It was not recognizing
659 a conditional subpattern that could match an empty string if that
660 subpattern was within another subpattern. For example, it looped when
661 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
662 condition was not nested. This bug has been fixed.
663
664 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
665 past the start of the subject in the presence of bytes with the top bit
666 set, for example "\x8aBCD".
667
668 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
669 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
670
671 14. Optimized (?!) to (*FAIL).
672
673 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
674 This restricts code points to be within the range 0 to 0x10FFFF, excluding
675 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
676 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
677 does: it's just the validity check that is more restrictive.
678
679 16. Inserted checks for integer overflows during escape sequence (backslash)
680 processing, and also fixed erroneous offset values for syntax errors during
681 backslash processing.
682
683 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
684 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
685
686 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
687 caused an overrun.
688
689 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
690 something other than just ASCII characters) inside a group that had an
691 unlimited repeat caused a loop at compile time (while checking to see
692 whether the group could match an empty string).
693
694 20. Debugging a pattern containing \p or \P could cause a crash. For example,
695 [\P{Any}] did so. (Error in the code for printing property names.)
696
697 21. An orphan \E inside a character class could cause a crash.
698
699 22. A repeated capturing bracket such as (A)? could cause a wild memory
700 reference during compilation.
701
702 23. There are several functions in pcre_compile() that scan along a compiled
703 expression for various reasons (e.g. to see if it's fixed length for look
704 behind). There were bugs in these functions when a repeated \p or \P was
705 present in the pattern. These operators have additional parameters compared
706 with \d, etc, and these were not being taken into account when moving along
707 the compiled data. Specifically:
708
709 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
710 length.
711
712 (b) An item such as \pL+ within a repeated group could cause crashes or
713 loops.
714
715 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
716 "reference to non-existent subpattern" error.
717
718 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
719
720 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
721 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
722
723 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
724
725 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
726 character were causing crashes (broken optimization).
727
728 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
729 \p or \P) caused a compile-time loop.
730
731 28. More problems have arisen in unanchored patterns when CRLF is a valid line
732 break. For example, the unstudied pattern [\r\n]A does not match the string
733 "\r\nA" because change 7.0/46 below moves the current point on by two
734 characters after failing to match at the start. However, the pattern \nA
735 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
736 the same is true. There doesn't seem any very clean way out of this, but
737 what I have chosen to do makes the common cases work: PCRE now takes note
738 of whether there can be an explicit match for \r or \n anywhere in the
739 pattern, and if so, 7.0/46 no longer applies. As part of this change,
740 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
741 pattern has explicit CR or LF references.
742
743 29. Added (*CR) etc for changing newline setting at start of pattern.
744
745
746 Version 7.2 19-Jun-07
747 ---------------------
748
749 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
750 which is apparently normally available under Windows.
751
752 2. Re-jig the pcregrep tests with different newline settings in an attempt
753 to make them independent of the local environment's newline setting.
754
755 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
756
757 4. Some of the "internals" tests were previously cut out when the link size
758 was not 2, because the output contained actual offsets. The recent new
759 "Z" feature of pcretest means that these can be cut out, making the tests
760 usable with all link sizes.
761
762 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
763 stack recursion. This gives a massive performance boost under BSD, but just
764 a small improvement under Linux. However, it saves one field in the frame
765 in all cases.
766
767 6. Added more features from the forthcoming Perl 5.10:
768
769 (a) (?-n) (where n is a string of digits) is a relative subroutine or
770 recursion call. It refers to the nth most recently opened parentheses.
771
772 (b) (?+n) is also a relative subroutine call; it refers to the nth next
773 to be opened parentheses.
774
775 (c) Conditions that refer to capturing parentheses can be specified
776 relatively, for example, (?(-2)... or (?(+3)...
777
778 (d) \K resets the start of the current match so that everything before
779 is not part of it.
780
781 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
782
783 (f) \g{name} is another synonym - part of Perl 5.10's unification of
784 reference syntax.
785
786 (g) (?| introduces a group in which the numbering of parentheses in each
787 alternative starts with the same number.
788
789 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
790
791 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
792 PCRE_INFO_JCHANGED.
793
794 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
795 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
796 for detecting groups that can match an empty string.
797
798 9. A pattern with a very large number of alternatives (more than several
799 hundred) was running out of internal workspace during the pre-compile
800 phase, where pcre_compile() figures out how much memory will be needed. A
801 bit of new cunning has reduced the workspace needed for groups with
802 alternatives. The 1000-alternative test pattern now uses 12 bytes of
803 workspace instead of running out of the 4096 that are available.
804
805 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
806
807 11. Applied patch from Google to remove an optimization that didn't quite work.
808 The report of the bug said:
809
810 pcrecpp::RE("a*").FullMatch("aaa") matches, while
811 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
812 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
813
814 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
815 it matched the wrong number of bytes.
816
817
818 Version 7.1 24-Apr-07
819 ---------------------
820
821 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
822 that is more "standard", making use of automake and other Autotools. There
823 is some re-arrangement of the files and adjustment of comments consequent
824 on this.
825
826 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
827 for recursive directory scanning broke on some systems because the files
828 are not scanned in any specific order and on different systems the order
829 was different. A call to "sort" has been inserted into RunGrepTest for the
830 approprate test as a short-term fix. In the longer term there may be an
831 alternative.
832
833 3. I had an email from Eric Raymond about problems translating some of PCRE's
834 man pages to HTML (despite the fact that I distribute HTML pages, some
835 people do their own conversions for various reasons). The problems
836 concerned the use of low-level troff macros .br and .in. I have therefore
837 removed all such uses from the man pages (some were redundant, some could
838 be replaced by .nf/.fi pairs). The 132html script that I use to generate
839 HTML has been updated to handle .nf/.fi and to complain if it encounters
840 .br or .in.
841
842 4. Updated comments in configure.ac that get placed in config.h.in and also
843 arranged for config.h to be included in the distribution, with the name
844 config.h.generic, for the benefit of those who have to compile without
845 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
846
847 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
848 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
849 makevp.bat for the current PCRE, using the additional files
850 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
851
852 6. A Windows user reported a minor discrepancy with test 2, which turned out
853 to be caused by a trailing space on an input line that had got lost in his
854 copy. The trailing space was an accident, so I've just removed it.
855
856 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
857 that is needed.
858
859 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
860 as "const" (a) because they are and (b) because it helps the PHP
861 maintainers who have recently made a script to detect big data structures
862 in the php code that should be moved to the .rodata section. I remembered
863 to update Builducptable as well, so it won't revert if ucptable.h is ever
864 re-created.
865
866 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
867 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
868 order to be able to cut out the UTF-8 tables in the latter when UTF-8
869 support is not required. This saves 1.5-2K of code, which is important in
870 some applications.
871
872 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
873 so as not to refer to the tables, even though these functions will never be
874 called when UTF-8 support is disabled. Otherwise there are problems with a
875 shared library.
876
877 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
878
879 (a) It was defining its arguments as char * instead of void *.
880
881 (b) It was assuming that all moves were upwards in memory; this was true
882 a long time ago when I wrote it, but is no longer the case.
883
884 The emulated memove() is provided for those environments that have neither
885 memmove() nor bcopy(). I didn't think anyone used it these days, but that
886 is clearly not the case, as these two bugs were recently reported.
887
888 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
889 and Detrail to create the HTML documentation, the .txt form of the man
890 pages, and it removes trailing spaces from listed files. It also creates
891 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
892 case, it wraps all the #defines with #ifndefs. This script should be run
893 before "make dist".
894
895 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
896 with Unicode property support.
897
898 (a) For a maximizing quantifier, if the two different cases of the
899 character were of different lengths in their UTF-8 codings (there are
900 some cases like this - I found 11), and the matching function had to
901 back up over a mixture of the two cases, it incorrectly assumed they
902 were both the same length.
903
904 (b) When PCRE was configured to use the heap rather than the stack for
905 recursion during matching, it was not correctly preserving the data for
906 the other case of a UTF-8 character when checking ahead for a match
907 while processing a minimizing repeat. If the check also involved
908 matching a wide character, but failed, corruption could cause an
909 erroneous result when trying to check for a repeat of the original
910 character.
911
912 13. Some tidying changes to the testing mechanism:
913
914 (a) The RunTest script now detects the internal link size and whether there
915 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
916 values substituted by "configure". (The RunGrepTest script already did
917 this for UTF-8.) The configure.ac script no longer substitutes the
918 relevant variables.
919
920 (b) The debugging options /B and /D in pcretest show the compiled bytecode
921 with length and offset values. This means that the output is different
922 for different internal link sizes. Test 2 is skipped for link sizes
923 other than 2 because of this, bypassing the problem. Unfortunately,
924 there was also a test in test 3 (the locale tests) that used /B and
925 failed for link sizes other than 2. Rather than cut the whole test out,
926 I have added a new /Z option to pcretest that replaces the length and
927 offset values with spaces. This is now used to make test 3 independent
928 of link size. (Test 2 will be tidied up later.)
929
930 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
931 segmentation fault instead of returning the appropriate error message.
932
933 15. In multiline mode when the newline sequence was set to "any", the pattern
934 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
935 This doesn't seem right; it now treats the CRLF combination as the line
936 ending, and so does not match in that case. It's only a pattern such as ^$
937 that would hit this one: something like ^ABC$ would have failed after \r
938 and then tried again after \r\n.
939
940 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
941 in an attempt to make files that differ only in their line terminators
942 compare equal. This works on Linux.
943
944 17. Under certain error circumstances pcregrep might try to free random memory
945 as it exited. This is now fixed, thanks to valgrind.
946
947 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
948 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
949 was because its rules for how to advance for /g after matching an empty
950 string at the end of a line did not allow for this case. They now check for
951 it specially.
952
953 20. pcretest is supposed to handle patterns and data of any length, by
954 extending its buffers when necessary. It was getting this wrong when the
955 buffer for a data line had to be extended.
956
957 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
958 CRLF as a newline sequence.
959
960 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
961 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
962 I have nevertheless tidied it up.
963
964 23. Added some casts to kill warnings from HP-UX ia64 compiler.
965
966 24. Added a man page for pcre-config.
967
968
969 Version 7.0 19-Dec-06
970 ---------------------
971
972 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
973 moving to gcc 4.1.1.
974
975 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
976 sys/time.h, which is documented as needed for this function. It doesn't
977 seem to matter on Linux, but it showed up on some releases of OS X.
978
979 3. It seems that there are systems where bytes whose values are greater than
980 127 match isprint() in the "C" locale. The "C" locale should be the
981 default when a C program starts up. In most systems, only ASCII printing
982 characters match isprint(). This difference caused the output from pcretest
983 to vary, making some of the tests fail. I have changed pcretest so that:
984
985 (a) When it is outputting text in the compiled version of a pattern, bytes
986 other than 32-126 are always shown as hex escapes.
987
988 (b) When it is outputting text that is a matched part of a subject string,
989 it does the same, unless a different locale has been set for the match
990 (using the /L modifier). In this case, it uses isprint() to decide.
991
992 4. Fixed a major bug that caused incorrect computation of the amount of memory
993 required for a compiled pattern when options that changed within the
994 pattern affected the logic of the preliminary scan that determines the
995 length. The relevant options are -x, and -i in UTF-8 mode. The result was
996 that the computed length was too small. The symptoms of this bug were
997 either the PCRE error "internal error: code overflow" from pcre_compile(),
998 or a glibc crash with a message such as "pcretest: free(): invalid next
999 size (fast)". Examples of patterns that provoked this bug (shown in
1000 pcretest format) are:
1001
1002 /(?-x: )/x
1003 /(?x)(?-x: \s*#\s*)/
1004 /((?i)[\x{c0}])/8
1005 /(?i:[\x{c0}])/8
1006
1007 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
1008 is now done differently.
1009
1010 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
1011 wrapper classes; (b) implement a new function in the C++ scanner that is
1012 more efficient than the old way of doing things because it avoids levels of
1013 recursion in the regex matching; (c) add a paragraph to the documentation
1014 for the FullMatch() function.
1015
1016 6. The escape sequence \n was being treated as whatever was defined as
1017 "newline". Not only was this contrary to the documentation, which states
1018 that \n is character 10 (hex 0A), but it also went horribly wrong when
1019 "newline" was defined as CRLF. This has been fixed.
1020
1021 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
1022 was being set to -1 for the "end of line" case (supposedly a value that no
1023 character can have). Though this value is never used (the check for end of
1024 line is "zero bytes in current character"), it caused compiler complaints.
1025 I've changed it to 0xffffffff.
1026
1027 8. In pcre_version.c, the version string was being built by a sequence of
1028 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
1029 string (as it is for production releases) called a macro with an empty
1030 argument. The C standard says the result of this is undefined. The gcc
1031 compiler treats it as an empty string (which was what was wanted) but it is
1032 reported that Visual C gives an error. The source has been hacked around to
1033 avoid this problem.
1034
1035 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
1036 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
1037 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
1038 of them did).
1039
1040 10. Originally, pcretest opened its input and output without "b"; then I was
1041 told that "b" was needed in some environments, so it was added for release
1042 5.0 to both the input and output. (It makes no difference on Unix-like
1043 systems.) Later I was told that it is wrong for the input on Windows. I've
1044 now abstracted the modes into two macros, to make it easier to fiddle with
1045 them, and removed "b" from the input mode under Windows.
1046
1047 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
1048
1049 12. Added -help and --help to pcretest as an official way of being reminded
1050 of the options.
1051
1052 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
1053 and pcrecpp.cc because they annoy compilers at high warning levels.
1054
1055 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
1056
1057 15. Fixed an occurrence of == in configure.ac that should have been = (shell
1058 scripts are not C programs :-) and which was not noticed because it works
1059 on Linux.
1060
1061 16. pcretest is supposed to handle any length of pattern and data line (as one
1062 line or as a continued sequence of lines) by extending its input buffer if
1063 necessary. This feature was broken for very long pattern lines, leading to
1064 a string of junk being passed to pcre_compile() if the pattern was longer
1065 than about 50K.
1066
1067 17. I have done a major re-factoring of the way pcre_compile() computes the
1068 amount of memory needed for a compiled pattern. Previously, there was code
1069 that made a preliminary scan of the pattern in order to do this. That was
1070 OK when PCRE was new, but as the facilities have expanded, it has become
1071 harder and harder to keep it in step with the real compile phase, and there
1072 have been a number of bugs (see for example, 4 above). I have now found a
1073 cunning way of running the real compile function in a "fake" mode that
1074 enables it to compute how much memory it would need, while actually only
1075 ever using a few hundred bytes of working memory and without too many
1076 tests of the mode. This should make future maintenance and development
1077 easier. A side effect of this work is that the limit of 200 on the nesting
1078 depth of parentheses has been removed (though this was never a serious
1079 limitation, I suspect). However, there is a downside: pcre_compile() now
1080 runs more slowly than before (30% or more, depending on the pattern). I
1081 hope this isn't a big issue. There is no effect on runtime performance.
1082
1083 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
1084 newline (only possible for the last line of a file) and it was a
1085 pattern that set a locale (followed by /Lsomething), pcretest crashed.
1086
1087 19. Added additional timing features to pcretest. (1) The -tm option now times
1088 matching only, not compiling. (2) Both -t and -tm can be followed, as a
1089 separate command line item, by a number that specifies the number of
1090 repeats to use when timing. The default is 50000; this gives better
1091 precision, but takes uncomfortably long for very large patterns.
1092
1093 20. Extended pcre_study() to be more clever in cases where a branch of a
1094 subpattern has no definite first character. For example, (a*|b*)[cd] would
1095 previously give no result from pcre_study(). Now it recognizes that the
1096 first character must be a, b, c, or d.
1097
1098 21. There was an incorrect error "recursive call could loop indefinitely" if
1099 a subpattern (or the entire pattern) that was being tested for matching an
1100 empty string contained only one non-empty item after a nested subpattern.
1101 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
1102 incorrectly, because the \d was being skipped in the check.
1103
1104 22. The pcretest program now has a new pattern option /B and a command line
1105 option -b, which is equivalent to adding /B to every pattern. This causes
1106 it to show the compiled bytecode, without the additional information that
1107 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
1108 is the same as /B/I).
1109
1110 23. A new optimization is now able automatically to treat some sequences such
1111 as a*b as a*+b. More specifically, if something simple (such as a character
1112 or a simple class like \d) has an unlimited quantifier, and is followed by
1113 something that cannot possibly match the quantified thing, the quantifier
1114 is automatically "possessified".
1115
1116 24. A recursive reference to a subpattern whose number was greater than 39
1117 went wrong under certain circumstances in UTF-8 mode. This bug could also
1118 have affected the operation of pcre_study().
1119
1120 25. Realized that a little bit of performance could be had by replacing
1121 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
1122
1123 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
1124
1125 27. Possessive quantifiers such as a++ were previously implemented by turning
1126 them into atomic groups such as ($>a+). Now they have their own opcodes,
1127 which improves performance. This includes the automatically created ones
1128 from 23 above.
1129
1130 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
1131 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
1132 the first matched character to be a colon. This applied both to named and
1133 numbered groups.
1134
1135 29. The ucpinternal.h header file was missing its idempotency #ifdef.
1136
1137 30. I was sent a "project" file called libpcre.a.dev which I understand makes
1138 building PCRE on Windows easier, so I have included it in the distribution.
1139
1140 31. There is now a check in pcretest against a ridiculously large number being
1141 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
1142 loop, the loop is abandoned.
1143
1144 32. Forward references to subpatterns in conditions such as (?(2)...) where
1145 subpattern 2 is defined later cause pcre_compile() to search forwards in
1146 the pattern for the relevant set of parentheses. This search went wrong
1147 when there were unescaped parentheses in a character class, parentheses
1148 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
1149
1150 33. "Subroutine" calls and backreferences were previously restricted to
1151 referencing subpatterns earlier in the regex. This restriction has now
1152 been removed.
1153
1154 34. Added a number of extra features that are going to be in Perl 5.10. On the
1155 whole, these are just syntactic alternatives for features that PCRE had
1156 previously implemented using the Python syntax or my own invention. The
1157 other formats are all retained for compatibility.
1158
1159 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
1160 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
1161 also .NET compatible.
1162
1163 (b) A recursion or subroutine call to a named group can now be defined as
1164 (?&name) as well as (?P>name).
1165
1166 (c) A backreference to a named group can now be defined as \k<name> or
1167 \k'name' as well as (?P=name). The new forms, as well as being in Perl
1168 5.10, are also .NET compatible.
1169
1170 (d) A conditional reference to a named group can now use the syntax
1171 (?(<name>) or (?('name') as well as (?(name).
1172
1173 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
1174 groups (named and numbered) that are never evaluated inline, but can be
1175 called as "subroutines" from elsewhere. In effect, the DEFINE condition
1176 is always false. There may be only one alternative in such a group.
1177
1178 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
1179 as the simple (?(R). The condition is true only if the most recent
1180 recursion is that of the given number or name. It does not search out
1181 through the entire recursion stack.
1182
1183 (g) The escape \gN or \g{N} has been added, where N is a positive or
1184 negative number, specifying an absolute or relative reference.
1185
1186 35. Tidied to get rid of some further signed/unsigned compiler warnings and
1187 some "unreachable code" warnings.
1188
1189 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
1190 things, this adds five new scripts.
1191
1192 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
1193 There were also incompatibilities regarding the handling of \Q..\E inside
1194 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
1195 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
1196
1197 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
1198 matches an empty string, and forcibly breaks the loop. There were bugs in
1199 this code in non-simple cases. For a pattern such as ^(a()*)* matched
1200 against aaaa the result was just "a" rather than "aaaa", for example. Two
1201 separate and independent bugs (that affected different cases) have been
1202 fixed.
1203
1204 39. Refactored the code to abolish the use of different opcodes for small
1205 capturing bracket numbers. This is a tidy that I avoided doing when I
1206 removed the limit on the number of capturing brackets for 3.5 back in 2001.
1207 The new approach is not only tidier, it makes it possible to reduce the
1208 memory needed to fix the previous bug (38).
1209
1210 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
1211 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
1212 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
1213 mode.
1214
1215 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
1216 report.
1217
1218 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
1219 copy construction and assignment in the C++ wrapper.
1220
1221 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
1222 couple of bugs that could have given wrong results in the "--newline=crlf"
1223 case.
1224
1225 44. Added a number of casts and did some reorganization of signed/unsigned int
1226 variables following suggestions from Dair Grant. Also renamed the variable
1227 "this" as "item" because it is a C++ keyword.
1228
1229 45. Arranged for dftables to add
1230
1231 #include "pcre_internal.h"
1232
1233 to pcre_chartables.c because without it, gcc 4.x may remove the array
1234 definition from the final binary if PCRE is built into a static library and
1235 dead code stripping is activated.
1236
1237 46. For an unanchored pattern, if a match attempt fails at the start of a
1238 newline sequence, and the newline setting is CRLF or ANY, and the next two
1239 characters are CRLF, advance by two characters instead of one.
1240
1241
1242 Version 6.7 04-Jul-06
1243 ---------------------
1244
1245 1. In order to handle tests when input lines are enormously long, pcretest has
1246 been re-factored so that it automatically extends its buffers when
1247 necessary. The code is crude, but this _is_ just a test program. The
1248 default size has been increased from 32K to 50K.
1249
1250 2. The code in pcre_study() was using the value of the re argument before
1251 testing it for NULL. (Of course, in any sensible call of the function, it
1252 won't be NULL.)
1253
1254 3. The memmove() emulation function in pcre_internal.h, which is used on
1255 systems that lack both memmove() and bcopy() - that is, hardly ever -
1256 was missing a "static" storage class specifier.
1257
1258 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
1259 containing an extended class (one that cannot be represented by a bitmap
1260 because it contains high-valued characters or Unicode property items, e.g.
1261 [\pZ]). Almost always one would set UTF-8 mode when processing such a
1262 pattern, but PCRE should not loop if you do not (it no longer does).
1263 [Detail: two cases were found: (a) a repeated subpattern containing an
1264 extended class; (b) a recursive reference to a subpattern that followed a
1265 previous extended class. It wasn't skipping over the extended class
1266 correctly when UTF-8 mode was not set.]
1267
1268 5. A negated single-character class was not being recognized as fixed-length
1269 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
1270 compile error "lookbehind assertion is not fixed length".
1271
1272 6. The RunPerlTest auxiliary script was showing an unexpected difference
1273 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
1274 write a Perl script that can interpret lines of an input file either as
1275 byte characters or as UTF-8, which is what "perltest" was being required to
1276 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
1277 can't do is switch easily at run time between having the "use utf8;" pragma
1278 or not. In the end, I fudged it by using the RunPerlTest script to insert
1279 "use utf8;" explicitly for the UTF-8 tests.
1280
1281 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
1282 the end of the subject string, contrary to the documentation and to what
1283 Perl does. This was true of both matching functions. Now it matches only at
1284 the start of the subject and immediately after *internal* newlines.
1285
1286 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
1287 a pointer to an int instead of a pointer to an unsigned long int. This
1288 caused problems on 64-bit systems.
1289
1290 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
1291 instance of the 'standard' template library not being so standard".
1292
1293 10. There was no check on the number of named subpatterns nor the maximum
1294 length of a subpattern name. The product of these values is used to compute
1295 the size of the memory block for a compiled pattern. By supplying a very
1296 long subpattern name and a large number of named subpatterns, the size
1297 computation could be caused to overflow. This is now prevented by limiting
1298 the length of names to 32 characters, and the number of named subpatterns
1299 to 10,000.
1300
1301 11. Subpatterns that are repeated with specific counts have to be replicated in
1302 the compiled pattern. The size of memory for this was computed from the
1303 length of the subpattern and the repeat count. The latter is limited to
1304 65535, but there was no limit on the former, meaning that integer overflow
1305 could in principle occur. The compiled length of a repeated subpattern is
1306 now limited to 30,000 bytes in order to prevent this.
1307
1308 12. Added the optional facility to have named substrings with the same name.
1309
1310 13. Added the ability to use a named substring as a condition, using the
1311 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
1312 are numbers (not recommended). Forward references are permitted.
1313
1314 14. Added forward references in named backreferences (if you see what I mean).
1315
1316 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
1317 pattern could run off the end of the subject. For example, the pattern
1318 "(?s)(.{1,5})"8 did this with the subject "ab".
1319
1320 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
1321 PCRE_CASELESS was set when matching characters that were quantified with ?
1322 or *.
1323
1324 17. A character class other than a single negated character that had a minimum
1325 but no maximum quantifier - for example [ab]{6,} - was not handled
1326 correctly by pce_dfa_exec(). It would match only one character.
1327
1328 18. A valid (though odd) pattern that looked like a POSIX character
1329 class but used an invalid character after [ (for example [[,abc,]]) caused
1330 pcre_compile() to give the error "Failed: internal error: code overflow" or
1331 in some cases to crash with a glibc free() error. This could even happen if
1332 the pattern terminated after [[ but there just happened to be a sequence of
1333 letters, a binary zero, and a closing ] in the memory that followed.
1334
1335 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
1336 over the years. Originally (before any Unicode support), just the bottom 8
1337 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
1338 output from "man perlunicode" includes this:
1339
1340 The regular expression compiler produces polymorphic opcodes. That
1341 is, the pattern adapts to the data and automatically switches to
1342 the Unicode character scheme when presented with Unicode data--or
1343 instead uses a traditional byte scheme when presented with byte
1344 data.
1345
1346 Sadly, a wide octal escape does not cause a switch, and in a string with
1347 no other multibyte characters, these octal escapes are treated as before.
1348 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
1349 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
1350 Unicode string.
1351
1352 I have not perpetrated such confusion in PCRE. Up till now, it took just
1353 the bottom 8 bits, as in old Perl. I have now made octal escapes with
1354 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
1355 translate to the appropriate multibyte character.
1356
1357 29. Applied some refactoring to reduce the number of warnings from Microsoft
1358 and Borland compilers. This has included removing the fudge introduced
1359 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
1360 a warning about an unused variable.
1361
1362 21. PCRE has not included VT (character 0x0b) in the set of whitespace
1363 characters since release 4.0, because Perl (from release 5.004) does not.
1364 [Or at least, is documented not to: some releases seem to be in conflict
1365 with the documentation.] However, when a pattern was studied with
1366 pcre_study() and all its branches started with \s, PCRE still included VT
1367 as a possible starting character. Of course, this did no harm; it just
1368 caused an unnecessary match attempt.
1369
1370 22. Removed a now-redundant internal flag bit that recorded the fact that case
1371 dependency changed within the pattern. This was once needed for "required
1372 byte" processing, but is no longer used. This recovers a now-scarce options
1373 bit. Also moved the least significant internal flag bit to the most-
1374 significant bit of the word, which was not previously used (hangover from
1375 the days when it was an int rather than a uint) to free up another bit for
1376 the future.
1377
1378 23. Added support for CRLF line endings as well as CR and LF. As well as the
1379 default being selectable at build time, it can now be changed at runtime
1380 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
1381 specify that it is scanning data with non-default line endings.
1382
1383 24. Changed the definition of CXXLINK to make it agree with the definition of
1384 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
1385
1386 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
1387 recursions. This makes a big different to stack usage for some patterns.
1388
1389 26. If a subpattern containing a named recursion or subroutine reference such
1390 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
1391 the space required for the compiled pattern went wrong and gave too small a
1392 value. Depending on the environment, this could lead to "Failed: internal
1393 error: code overflow at offset 49" or "glibc detected double free or
1394 corruption" errors.
1395
1396 27. Applied patches from Google (a) to support the new newline modes and (b) to
1397 advance over multibyte UTF-8 characters in GlobalReplace.
1398
1399 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
1400 difference for some implementation of PCRE in some Windows version.
1401
1402 29. Added some extra testing facilities to pcretest:
1403
1404 \q<number> in a data line sets the "match limit" value
1405 \Q<number> in a data line sets the "match recursion limt" value
1406 -S <number> sets the stack size, where <number> is in megabytes
1407
1408 The -S option isn't available for Windows.
1409
1410
1411 Version 6.6 06-Feb-06
1412 ---------------------
1413
1414 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
1415 in pcreposix.h. I have copied the definition from pcre.h.
1416
1417 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
1418 because pcre.h is no longer a built file.
1419
1420 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
1421 not normally included in the compiled code.
1422
1423
1424 Version 6.5 01-Feb-06
1425 ---------------------
1426
1427 1. When using the partial match feature with pcre_dfa_exec(), it was not
1428 anchoring the second and subsequent partial matches at the new starting
1429 point. This could lead to incorrect results. For example, with the pattern
1430 /1234/, partially matching against "123" and then "a4" gave a match.
1431
1432 2. Changes to pcregrep:
1433
1434 (a) All non-match returns from pcre_exec() were being treated as failures
1435 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
1436 error message is output. Some extra information is given for the
1437 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
1438 probably the only errors that are likely to be caused by users (by
1439 specifying a regex that has nested indefinite repeats, for instance).
1440 If there are more than 20 of these errors, pcregrep is abandoned.
1441
1442 (b) A binary zero was treated as data while matching, but terminated the
1443 output line if it was written out. This has been fixed: binary zeroes
1444 are now no different to any other data bytes.
1445
1446 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
1447 used to set a locale for matching. The --locale=xxxx long option has
1448 been added (no short equivalent) to specify a locale explicitly on the
1449 pcregrep command, overriding the environment variables.
1450
1451 (d) When -B was used with -n, some line numbers in the output were one less
1452 than they should have been.
1453
1454 (e) Added the -o (--only-matching) option.
1455
1456 (f) If -A or -C was used with -c (count only), some lines of context were
1457 accidentally printed for the final match.
1458
1459 (g) Added the -H (--with-filename) option.
1460
1461 (h) The combination of options -rh failed to suppress file names for files
1462 that were found from directory arguments.
1463
1464 (i) Added the -D (--devices) and -d (--directories) options.
1465
1466 (j) Added the -F (--fixed-strings) option.
1467
1468 (k) Allow "-" to be used as a file name for -f as well as for a data file.
1469
1470 (l) Added the --colo(u)r option.
1471
1472 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
1473 is not present by default.
1474
1475 3. A nasty bug was discovered in the handling of recursive patterns, that is,
1476 items such as (?R) or (?1), when the recursion could match a number of
1477 alternatives. If it matched one of the alternatives, but subsequently,
1478 outside the recursion, there was a failure, the code tried to back up into
1479 the recursion. However, because of the way PCRE is implemented, this is not
1480 possible, and the result was an incorrect result from the match.
1481
1482 In order to prevent this happening, the specification of recursion has
1483 been changed so that all such subpatterns are automatically treated as
1484 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
1485
1486 4. I had overlooked the fact that, in some locales, there are characters for
1487 which isalpha() is true but neither isupper() nor islower() are true. In
1488 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
1489 and ordfeminine) are like this. This affected the treatment of \w and \W
1490 when they appeared in character classes, but not when they appeared outside
1491 a character class. The bit map for "word" characters is now created
1492 separately from the results of isalnum() instead of just taking it from the
1493 upper, lower, and digit maps. (Plus the underscore character, of course.)
1494
1495 5. The above bug also affected the handling of POSIX character classes such as
1496 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
1497 permanent tables. Instead, the bit maps for such a class were previously
1498 created as the appropriate unions of the upper, lower, and digit bitmaps.
1499 Now they are created by subtraction from the [[:word:]] class, which has
1500 its own bitmap.
1501
1502 6. The [[:blank:]] character class matches horizontal, but not vertical space.
1503 It is created by subtracting the vertical space characters (\x09, \x0a,
1504 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
1505 subtraction was done in the overall bitmap for a character class, meaning
1506 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
1507 be recognized. This bug has been fixed.
1508
1509 7. Patches from the folks at Google:
1510
1511 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
1512 real life, but is still worth protecting against".
1513
1514 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
1515 regular expressions".
1516
1517 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
1518 have it.
1519
1520 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
1521 "configure" and the latter not, in order to fix a problem somebody had
1522 with compiling the Arg class on HP-UX.
1523
1524 (e) Improve the error-handling of the C++ wrapper a little bit.
1525
1526 (f) New tests for checking recursion limiting.
1527
1528 8. The pcre_memmove() function, which is used only if the environment does not
1529 have a standard memmove() function (and is therefore rarely compiled),
1530 contained two bugs: (a) use of int instead of size_t, and (b) it was not
1531 returning a result (though PCRE never actually uses the result).
1532
1533 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
1534 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
1535 returned instead of calling malloc() with an overflowing number that would
1536 most likely cause subsequent chaos.
1537
1538 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
1539
1540 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
1541 with this option is matched, the nmatch and pmatch options of regexec() are
1542 ignored.
1543
1544 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
1545 provided in case anyone wants to the the POSIX interface with UTF-8
1546 strings.
1547
1548 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
1549 C++ linking (needed for some HP-UX environments).
1550
1551 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
1552 (unused parameter) and in the pcre_printint() function (omitted "default"
1553 switch label when the default is to do nothing).
1554
1555 15. Added some code to make it possible, when PCRE is compiled as a C++
1556 library, to replace subject pointers for pcre_exec() with a smart pointer
1557 class, thus making it possible to process discontinuous strings.
1558
1559 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
1560 much the same function. They were added by different people who were trying
1561 to make PCRE easy to compile on non-Unix systems. It has been suggested
1562 that PCRE_EXPORT be abolished now that there is more automatic apparatus
1563 for compiling on Windows systems. I have therefore replaced it with
1564 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
1565 defaults to "extern" for C or "extern C" for C++, which works fine on
1566 Unix-like systems. It is now possible to override the value of PCRE_DATA_
1567 SCOPE with something explicit in config.h. In addition:
1568
1569 (a) pcreposix.h still had just "extern" instead of either of these macros;
1570 I have replaced it with PCRE_DATA_SCOPE.
1571
1572 (b) Functions such as _pcre_xclass(), which are internal to the library,
1573 but external in the C sense, all had PCRE_EXPORT in their definitions.
1574 This is apparently wrong for the Windows case, so I have removed it.
1575 (It makes no difference on Unix-like systems.)
1576
1577 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
1578 of recursive calls to match(). This is different to MATCH_LIMIT because
1579 that limits the total number of calls to match(), not all of which increase
1580 the depth of recursion. Limiting the recursion depth limits the amount of
1581 stack (or heap if NO_RECURSE is set) that is used. The default can be set
1582 when PCRE is compiled, and changed at run time. A patch from Google adds
1583 this functionality to the C++ interface.
1584
1585 18. Changes to the handling of Unicode character properties:
1586
1587 (a) Updated the table to Unicode 4.1.0.
1588
1589 (b) Recognize characters that are not in the table as "Cn" (undefined).
1590
1591 (c) I revised the way the table is implemented to a much improved format
1592 which includes recognition of ranges. It now supports the ranges that
1593 are defined in UnicodeData.txt, and it also amalgamates other
1594 characters into ranges. This has reduced the number of entries in the
1595 table from around 16,000 to around 3,000, thus reducing its size
1596 considerably. I realized I did not need to use a tree structure after
1597 all - a binary chop search is just as efficient. Having reduced the
1598 number of entries, I extended their size from 6 bytes to 8 bytes to
1599 allow for more data.
1600
1601 (d) Added support for Unicode script names via properties such as \p{Han}.
1602
1603 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
1604 matching that character.
1605
1606 20. When matching a repeated Unicode property with a minimum greater than zero,
1607 (for example \pL{2,}), PCRE could look past the end of the subject if it
1608 reached it while seeking the minimum number of characters. This could
1609 happen only if some of the characters were more than one byte long, because
1610 there is a check for at least the minimum number of bytes.
1611
1612 21. Refactored the implementation of \p and \P so as to be more general, to
1613 allow for more different types of property in future. This has changed the
1614 compiled form incompatibly. Anybody with saved compiled patterns that use
1615 \p or \P will have to recompile them.
1616
1617 22. Added "Any" and "L&" to the supported property types.
1618
1619 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
1620 but give a compile time error if the value is greater than 0xff.
1621
1622 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
1623 accidentally not being installed or uninstalled.
1624
1625 25. The pcre.h file was built from pcre.h.in, but the only changes that were
1626 made were to insert the current release number. This seemed silly, because
1627 it made things harder for people building PCRE on systems that don't run
1628 "configure". I have turned pcre.h into a distributed file, no longer built
1629 by "configure", with the version identification directly included. There is
1630 no longer a pcre.h.in file.
1631
1632 However, this change necessitated a change to the pcre-config script as
1633 well. It is built from pcre-config.in, and one of the substitutions was the
1634 release number. I have updated configure.ac so that ./configure now finds
1635 the release number by grepping pcre.h.
1636
1637 26. Added the ability to run the tests under valgrind.
1638
1639
1640 Version 6.4 05-Sep-05
1641 ---------------------
1642
1643 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
1644 "--" to be printed when multiple files were scanned, even when none of the
1645 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
1646 consider it to be a bug, and have restored the previous behaviour.
1647
1648 2. A couple of code tidies to get rid of compiler warnings.
1649
1650 3. The pcretest program used to cheat by referring to symbols in the library
1651 whose names begin with _pcre_. These are internal symbols that are not
1652 really supposed to be visible externally, and in some environments it is
1653 possible to suppress them. The cheating is now confined to including
1654 certain files from the library's source, which is a bit cleaner.
1655
1656 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
1657 file's purpose clearer.
1658
1659 5. Reorganized pcre_ucp_findchar().
1660
1661
1662 Version 6.3 15-Aug-05
1663 ---------------------
1664
1665 1. The file libpcre.pc.in did not have general read permission in the tarball.
1666
1667 2. There were some problems when building without C++ support:
1668
1669 (a) If C++ support was not built, "make install" and "make test" still
1670 tried to test it.
1671
1672 (b) There were problems when the value of CXX was explicitly set. Some
1673 changes have been made to try to fix these, and ...
1674
1675 (c) --disable-cpp can now be used to explicitly disable C++ support.
1676
1677 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
1678 backslash in a target when C++ was disabled. This confuses some
1679 versions of "make", apparently. Using an intermediate variable solves
1680 this. (Same for CPP_LOBJ.)
1681
1682 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
1683 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
1684 necessary on certain architectures.
1685
1686 4. Added a setting of -export-symbols-regex to the link command to remove
1687 those symbols that are exported in the C sense, but actually are local
1688 within the library, and not documented. Their names all begin with
1689 "_pcre_". This is not a perfect job, because (a) we have to except some
1690 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
1691 available (and never for static libraries). I have made a note to try to
1692 find a way round (a) in the future.
1693
1694
1695 Version 6.2 01-Aug-05
1696 ---------------------
1697
1698 1. There was no test for integer overflow of quantifier values. A construction
1699 such as {1111111111111111} would give undefined results. What is worse, if
1700 a minimum quantifier for a parenthesized subpattern overflowed and became
1701 negative, the calculation of the memory size went wrong. This could have
1702 led to memory overwriting.
1703
1704 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
1705
1706 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
1707 operating environments where this matters.
1708
1709 4. Applied Giuseppe Maxia's patch to add additional features for controlling
1710 PCRE options from within the C++ wrapper.
1711
1712 5. Named capturing subpatterns were not being correctly counted when a pattern
1713 was compiled. This caused two problems: (a) If there were more than 100
1714 such subpatterns, the calculation of the memory needed for the whole
1715 compiled pattern went wrong, leading to an overflow error. (b) Numerical
1716 back references of the form \12, where the number was greater than 9, were
1717 not recognized as back references, even though there were sufficient
1718 previous subpatterns.
1719
1720 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
1721 versions of gcc, e.g. 2.95.4.
1722
1723
1724 Version 6.1 21-Jun-05
1725 ---------------------
1726
1727 1. There was one reference to the variable "posix" in pcretest.c that was not
1728 surrounded by "#if !defined NOPOSIX".
1729
1730 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
1731 the cross-check on the old pcre_info() function, for the benefit of the
1732 cut-down version of PCRE that is currently imported into Exim.
1733
1734 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
1735 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
1736 patterns that start (?i) but I don't think that matters. The use of (?i) is
1737 just an example; this all applies to the other options as well.
1738
1739 4. Since libtool seems to echo the compile commands it is issuing, the output
1740 from "make" can be reduced a bit by putting "@" in front of each libtool
1741 compile command.
1742
1743 5. Patch from the folks at Google for configure.in to be a bit more thorough
1744 in checking for a suitable C++ installation before trying to compile the
1745 C++ stuff. This should fix a reported problem when a compiler was present,
1746 but no suitable headers.
1747
1748 6. The man pages all had just "PCRE" as their title. I have changed them to
1749 be the relevant file name. I have also arranged that these names are
1750 retained in the file doc/pcre.txt, which is a concatenation in text format
1751 of all the man pages except the little individual ones for each function.
1752
1753 7. The NON-UNIX-USE file had not been updated for the different set of source
1754 files that come with release 6. I also added a few comments about the C++
1755 wrapper.
1756
1757
1758 Version 6.0 07-Jun-05
1759 ---------------------
1760
1761 1. Some minor internal re-organization to help with my DFA experiments.
1762
1763 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
1764 didn't matter for the library itself when fully configured, but did matter
1765 when compiling without UCP support, or within Exim, where the ucp files are
1766 not imported.
1767
1768 3. Refactoring of the library code to split up the various functions into
1769 different source modules. The addition of the new DFA matching code (see
1770 below) to a single monolithic source would have made it really too
1771 unwieldy, quite apart from causing all the code to be include in a
1772 statically linked application, when only some functions are used. This is
1773 relevant even without the DFA addition now that patterns can be compiled in
1774 one application and matched in another.
1775
1776 The downside of splitting up is that there have to be some external
1777 functions and data tables that are used internally in different modules of
1778 the library but which are not part of the API. These have all had their
1779 names changed to start with "_pcre_" so that they are unlikely to clash
1780 with other external names.
1781
1782 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
1783 a different (DFA) algorithm. Although it is slower than the original
1784 function, it does have some advantages for certain types of matching
1785 problem.
1786
1787 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
1788 including restarting after a partial match.
1789
1790 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
1791 defined when compiling for Windows was sent to me. I have put it into the
1792 code, though I have no means of testing or verifying it.
1793
1794 7. Added the pcre_refcount() auxiliary function.
1795
1796 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
1797 match before or at the first newline in the subject string. In pcretest,
1798 the /f option on a pattern can be used to set this.
1799
1800 9. A repeated \w when used in UTF-8 mode with characters greater than 256
1801 would behave wrongly. This has been present in PCRE since release 4.0.
1802
1803 10. A number of changes to the pcregrep command:
1804
1805 (a) Refactored how -x works; insert ^(...)$ instead of setting
1806 PCRE_ANCHORED and checking the length, in preparation for adding
1807 something similar for -w.
1808
1809 (b) Added the -w (match as a word) option.
1810
1811 (c) Refactored the way lines are read and buffered so as to have more
1812 than one at a time available.
1813
1814 (d) Implemented a pcregrep test script.
1815
1816 (e) Added the -M (multiline match) option. This allows patterns to match
1817 over several lines of the subject. The buffering ensures that at least
1818 8K, or the rest of the document (whichever is the shorter) is available
1819 for matching (and similarly the previous 8K for lookbehind assertions).
1820
1821 (f) Changed the --help output so that it now says
1822
1823 -w, --word-regex(p)
1824
1825 instead of two lines, one with "regex" and the other with "regexp"
1826 because that confused at least one person since the short forms are the
1827 same. (This required a bit of code, as the output is generated
1828 automatically from a table. It wasn't just a text change.)
1829
1830 (g) -- can be used to terminate pcregrep options if the next thing isn't an
1831 option but starts with a hyphen. Could be a pattern or a path name
1832 starting with a hyphen, for instance.
1833
1834 (h) "-" can be given as a file name to represent stdin.
1835
1836 (i) When file names are being printed, "(standard input)" is used for
1837 the standard input, for compatibility with GNU grep. Previously
1838 "<stdin>" was used.
1839
1840 (j) The option --label=xxx can be used to supply a name to be used for
1841 stdin when file names are being printed. There is no short form.
1842
1843 (k) Re-factored the options decoding logic because we are going to add
1844 two more options that take data. Such options can now be given in four
1845 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
1846
1847 (l) Added the -A, -B, and -C options for requesting that lines of context
1848 around matches be printed.
1849
1850 (m) Added the -L option to print the names of files that do not contain
1851 any matching lines, that is, the complement of -l.
1852
1853 (n) The return code is 2 if any file cannot be opened, but pcregrep does
1854 continue to scan other files.
1855
1856 (o) The -s option was incorrectly implemented. For compatibility with other
1857 greps, it now suppresses the error message for a non-existent or non-
1858 accessible file (but not the return code). There is a new option called
1859 -q that suppresses the output of matching lines, which was what -s was
1860 previously doing.
1861
1862 (p) Added --include and --exclude options to specify files for inclusion
1863 and exclusion when recursing.
1864
1865 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
1866 Hopefully, it now does.
1867
1868 12. Missing cast in pcre_study().
1869
1870 13. Added an "uninstall" target to the makefile.
1871
1872 14. Replaced "extern" in the function prototypes in Makefile.in with
1873 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
1874 world, but is set differently for Windows.
1875
1876 15. Added a second compiling function called pcre_compile2(). The only
1877 difference is that it has an extra argument, which is a pointer to an
1878 integer error code. When there is a compile-time failure, this is set
1879 non-zero, in addition to the error test pointer being set to point to an
1880 error message. The new argument may be NULL if no error number is required
1881 (but then you may as well call pcre_compile(), which is now just a
1882 wrapper). This facility is provided because some applications need a
1883 numeric error indication, but it has also enabled me to tidy up the way
1884 compile-time errors are handled in the POSIX wrapper.
1885
1886 16. Added VPATH=.libs to the makefile; this should help when building with one
1887 prefix path and installing with another. (Or so I'm told by someone who
1888 knows more about this stuff than I do.)
1889
1890 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
1891 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
1892 match everything, including newlines. This is not POSIX-compatible, but
1893 somebody wanted the feature. From pcretest it can be activated by using
1894 both the P and the s flags.
1895
1896 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
1897
1898 19. libpcre.pc was being incorrectly installed as executable.
1899
1900 20. A couple of places in pcretest check for end-of-line by looking for '\n';
1901 it now also looks for '\r' so that it will work unmodified on Windows.
1902
1903 21. Added Google's contributed C++ wrapper to the distribution.
1904
1905 22. Added some untidy missing memory free() calls in pcretest, to keep
1906 Electric Fence happy when testing.
1907
1908
1909
1910 Version 5.0 13-Sep-04
1911 ---------------------
1912
1913 1. Internal change: literal characters are no longer packed up into items
1914 containing multiple characters in a single byte-string. Each character
1915 is now matched using a separate opcode. However, there may be more than one
1916 byte in the character in UTF-8 mode.
1917
1918 2. The pcre_callout_block structure has two new fields: pattern_position and
1919 next_item_length. These contain the offset in the pattern to the next match
1920 item, and its length, respectively.
1921
1922 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
1923 insertion of callouts before each pattern item. Added the /C option to
1924 pcretest to make use of this.
1925
1926 4. On the advice of a Windows user, the lines
1927
1928 #if defined(_WIN32) || defined(WIN32)
1929 _setmode( _fileno( stdout ), 0x8000 );
1930 #endif /* defined(_WIN32) || defined(WIN32) */
1931
1932 have been added to the source of pcretest. This apparently does useful
1933 magic in relation to line terminators.
1934
1935 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
1936 for the benefit of those environments where the "b" makes a difference.
1937
1938 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
1939 to know about it. I have put a hack into configure.in that adds in code
1940 to set GCC=yes if CC=icc. This seems to end up at a point in the
1941 generated configure script that is early enough to affect the setting of
1942 compiler options, which is what is needed, but I have no means of testing
1943 whether it really works. (The user who reported this had patched the
1944 generated configure script, which of course I cannot do.)
1945
1946 LATER: After change 22 below (new libtool files), the configure script
1947 seems to know about icc (and also ecc). Therefore, I have commented out
1948 this hack in configure.in.
1949
1950 7. Added support for pkg-config (2 patches were sent in).
1951
1952 8. Negated POSIX character classes that used a combination of internal tables
1953 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
1954 [[:^ascii]]. Typically, they would match almost any characters. The other
1955 POSIX classes were not broken in this way.
1956
1957 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
1958 to find the match, as PCRE was deluded into thinking that the match had to
1959 start at the start point or following a newline. The same bug applied to
1960 patterns with negative forward assertions or any backward assertions
1961 preceding ".*" at the start, unless the pattern required a fixed first
1962 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
1963
1964 10. In UTF-8 mode, when moving forwards in the subject after a failed match
1965 starting at the last subject character, bytes beyond the end of the subject
1966 string were read.
1967
1968 11. Renamed the variable "class" as "classbits" to make life easier for C++
1969 users. (Previously there was a macro definition, but it apparently wasn't
1970 enough.)
1971
1972 12. Added the new field "tables" to the extra data so that tables can be passed
1973 in at exec time, or the internal tables can be re-selected. This allows
1974 a compiled regex to be saved and re-used at a later time by a different
1975 program that might have everything at different addresses.
1976
1977 13. Modified the pcre-config script so that, when run on Solaris, it shows a
1978 -R library as well as a -L library.
1979
1980 14. The debugging options of pcretest (-d on the command line or D on a
1981 pattern) showed incorrect output for anything following an extended class
1982 that contained multibyte characters and which was followed by a quantifier.
1983
1984 15. Added optional support for general category Unicode character properties
1985 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
1986 support. It adds about 90K to the size of the library. The meanings of the
1987 inbuilt class escapes such as \d and \s have NOT been changed.
1988
1989 16. Updated pcredemo.c to include calls to free() to release the memory for the
1990 compiled pattern.
1991
1992 17. The generated file chartables.c was being created in the source directory
1993 instead of in the building directory. This caused the build to fail if the
1994 source directory was different from the building directory, and was
1995 read-only.
1996
1997 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
1998 file. No doubt somebody will tell me if they don't make sense... Also added
1999 Dan Mooney's comments about building on OpenVMS.
2000
2001 19. Added support for partial matching via the PCRE_PARTIAL option for
2002 pcre_exec() and the \P data escape in pcretest.
2003
2004 20. Extended pcretest with 3 new pattern features:
2005
2006 (i) A pattern option of the form ">rest-of-line" causes pcretest to
2007 write the compiled pattern to the file whose name is "rest-of-line".
2008 This is a straight binary dump of the data, with the saved pointer to
2009 the character tables forced to be NULL. The study data, if any, is
2010 written too. After writing, pcretest reads a new pattern.
2011
2012 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
2013 compiled pattern from the given file. There must not be any
2014 occurrences of "<" in the file name (pretty unlikely); if there are,
2015 pcretest will instead treat the initial "<" as a pattern delimiter.
2016 After reading in the pattern, pcretest goes on to read data lines as
2017 usual.
2018
2019 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
2020 and 16-bit fields in a compiled pattern, to simulate a pattern that
2021 was compiled on a host of opposite endianness.
2022
2023 21. The pcre-exec() function can now cope with patterns that were compiled on
2024 hosts of opposite endianness, with this restriction:
2025
2026 As for any compiled expression that is saved and used later, the tables
2027 pointer field cannot be preserved; the extra_data field in the arguments
2028 to pcre_exec() should be used to pass in a tables address if a value
2029 other than the default internal tables were used at compile time.
2030
2031 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
2032 now diagnosed as an error. Previously, most of the time, a negative number
2033 would have been treated as zero, but if in addition "ovector" was passed as
2034 NULL, a crash could occur.
2035
2036 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
2037 new versions from the libtool 1.5 distribution (the last one is a copy of
2038 a file called libtool.m4). This seems to have fixed the need to patch
2039 "configure" to support Darwin 1.3 (which I used to do). However, I still
2040 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
2041 workstation).
2042
2043 24. Changed the PCRE licence to be the more standard "BSD" licence.
2044
2045
2046 Version 4.5 01-Dec-03
2047 ---------------------
2048
2049 1. There has been some re-arrangement of the code for the match() function so
2050 that it can be compiled in a version that does not call itself recursively.
2051 Instead, it keeps those local variables that need separate instances for
2052 each "recursion" in a frame on the heap, and gets/frees frames whenever it
2053 needs to "recurse". Keeping track of where control must go is done by means
2054 of setjmp/longjmp. The whole thing is implemented by a set of macros that
2055 hide most of the details from the main code, and operates only if
2056 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
2057 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
2058 operating.
2059
2060 To make it easier for callers to provide specially tailored get/free
2061 functions for this usage, two new functions, pcre_stack_malloc, and
2062 pcre_stack_free, are used. They are always called in strict stacking order,
2063 and the size of block requested is always the same.
2064
2065 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
2066 PCRE has been compiled to use the stack or the heap for recursion. The
2067 -C option of pcretest uses this to show which version is compiled.
2068
2069 A new data escape \S, is added to pcretest; it causes the amounts of store
2070 obtained and freed by both kinds of malloc/free at match time to be added
2071 to the output.
2072
2073 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
2074 what's available on my current Linux desktop machine.
2075
2076 3. When matching a UTF-8 string, the test for a valid string at the start has
2077 been extended. If start_offset is not zero, PCRE now checks that it points
2078 to a byte that is the start of a UTF-8 character. If not, it returns
2079 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
2080 this is necessary because there may be backward assertions in the pattern.
2081 When matching the same subject several times, it may save resources to use
2082 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
2083
2084 4. The code for checking the validity of UTF-8 strings has been tightened so
2085 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
2086 containing "overlong sequences".
2087
2088 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
2089 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
2090 should have been "&", but it just so happened that all the cases this let
2091 through by mistake were picked up later in the function.
2092
2093 6. I had used a variable called "isblank" - this is a C99 function, causing
2094 some compilers to warn. To avoid this, I renamed it (as "blankclass").
2095
2096 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
2097 prompting; (b) run "./pcretest /dev/null" at the start of the test script
2098 so the version is shown; (c) stop "make test" echoing "./RunTest".
2099
2100 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
2101
2102 9. The prototype for memmove() for systems that don't have it was using
2103 size_t, but the inclusion of the header that defines size_t was later. I've
2104 moved the #includes for the C headers earlier to avoid this.
2105
2106 10. Added some adjustments to the code to make it easier to compiler on certain
2107 special systems:
2108
2109 (a) Some "const" qualifiers were missing.
2110 (b) Added the macro EXPORT before all exported functions; by default this
2111 is defined to be empty.
2112 (c) Changed the dftables auxiliary program (that builds chartables.c) so
2113 that it reads its output file name as an argument instead of writing
2114 to the standard output and assuming this can be redirected.
2115
2116 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
2117 class containing characters with values greater than 255, PCRE compilation
2118 went into a loop.
2119
2120 12. A recursive reference to a subpattern that was within another subpattern
2121 that had a minimum quantifier of zero caused PCRE to crash. For example,
2122 (x(y(?2))z)? provoked this bug with a subject that got as far as the
2123 recursion. If the recursively-called subpattern itself had a zero repeat,
2124 that was OK.
2125
2126 13. In pcretest, the buffer for reading a data line was set at 30K, but the
2127 buffer into which it was copied (for escape processing) was still set at
2128 1024, so long lines caused crashes.
2129
2130 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
2131 "internal error: code overflow...". This applied to any character class
2132 that was followed by a possessive quantifier.
2133
2134 15. Modified the Makefile to add libpcre.la as a prerequisite for
2135 libpcreposix.la because I was told this is needed for a parallel build to
2136 work.
2137
2138 16. If a pattern that contained .* following optional items at the start was
2139 studied, the wrong optimizing data was generated, leading to matching
2140 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
2141 matching string must start with a or b or c. The correct conclusion for
2142 this pattern is that a match can start with any character.
2143
2144
2145 Version 4.4 13-Aug-03
2146 ---------------------
2147
2148 1. In UTF-8 mode, a character class containing characters with values between
2149 127 and 255 was not handled correctly if the compiled pattern was studied.
2150 In fixing this, I have also improved the studying algorithm for such
2151 classes (slightly).
2152
2153 2. Three internal functions had redundant arguments passed to them. Removal
2154 might give a very teeny performance improvement.
2155
2156 3. Documentation bug: the value of the capture_top field in a callout is *one
2157 more than* the number of the hightest numbered captured substring.
2158
2159 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
2160 in incorrectly linking with a previously installed version. They now link
2161 explicitly with libpcre.la.
2162
2163 5. configure.in no longer needs to recognize Cygwin specially.
2164
2165 6. A problem in pcre.in for Windows platforms is fixed.
2166
2167 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
2168 pcretest, it used to include the size of the study block as part of its
2169 output. Unfortunately, the structure contains a field that has a different
2170 size on different hardware architectures. This meant that the tests that
2171 showed this size failed. As the block is currently always of a fixed size,
2172 this information isn't actually particularly useful in pcretest output, so
2173 I have just removed it.
2174
2175 8. Three pre-processor statements accidentally did not start in column 1.
2176 Sadly, there are *still* compilers around that complain, even though
2177 standard C has not required this for well over a decade. Sigh.
2178
2179 9. In pcretest, the code for checking callouts passed small integers in the
2180 callout_data field, which is a void * field. However, some picky compilers
2181 complained about the casts involved for this on 64-bit systems. Now
2182 pcretest passes the address of the small integer instead, which should get
2183 rid of the warnings.
2184
2185 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
2186 both compile and run time, and gives an error if an invalid UTF-8 sequence
2187 is found. There is a option for disabling this check in cases where the
2188 string is known to be correct and/or the maximum performance is wanted.
2189
2190 11. In response to a bug report, I changed one line in Makefile.in from
2191
2192 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
2193 to
2194 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
2195
2196 to look similar to other lines, but I have no way of telling whether this
2197 is the right thing to do, as I do not use Windows. No doubt I'll get told
2198 if it's wrong...
2199
2200
2201 Version 4.3 21-May-03
2202 ---------------------
2203
2204 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
2205 Makefile.
2206
2207 2. Some refactoring to improve the quality of the code:
2208
2209 (i) The utf8_table... variables are now declared "const".
2210
2211 (ii) The code for \cx, which used the "case flipping" table to upper case
2212 lower case letters, now just substracts 32. This is ASCII-specific,
2213 but the whole concept of \cx is ASCII-specific, so it seems
2214 reasonable.
2215
2216 (iii) PCRE was using its character types table to recognize decimal and
2217 hexadecimal digits in the pattern. This is silly, because it handles
2218 only 0-9, a-f, and A-F, but the character types table is locale-
2219 specific, which means strange things might happen. A private
2220 table is now used for this - though it costs 256 bytes, a table is
2221 much faster than multiple explicit tests. Of course, the standard
2222 character types table is still used for matching digits in subject
2223 strings against \d.
2224
2225 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
2226 ending in _t are). So I've renamed it as ESC_tee.
2227
2228 3. The first argument for regexec() in the POSIX wrapper should have been
2229 defined as "const".
2230
2231 4. Changed pcretest to use malloc() for its buffers so that they can be
2232 Electric Fenced for debugging.
2233
2234 5. There were several places in the code where, in UTF-8 mode, PCRE would try
2235 to read one or more bytes before the start of the subject string. Often this
2236 had no effect on PCRE's behaviour, but in some circumstances it could
2237 provoke a segmentation fault.
2238
2239 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
2240 to try to read one or more bytes before the start of the subject string.
2241
2242 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
2243 UTF-8 support could misbehave in various ways if the subject string
2244 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
2245 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
2246 back over UTF-8 characters.)
2247
2248
2249 Version 4.2 14-Apr-03
2250 ---------------------
2251
2252 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
2253
2254 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
2255 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
2256 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
2257 [WIN_PREFIX]: new variable, "cyg" for Cygwin
2258 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
2259 and BUILD_EXEEXT
2260 Note: automatic setting of the BUILD variables is not yet working
2261 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
2262 compile-time but not at link-time
2263 [LINK]: use for linking executables only
2264 make different versions for Windows and non-Windows
2265 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
2266 libraries
2267 [LINK_FOR_BUILD]: new variable
2268 [OBJEXT]: use throughout
2269 [EXEEXT]: use throughout
2270 <winshared>: new target
2271 <wininstall>: new target
2272 <dftables.o>: use native compiler
2273 <dftables>: use native linker
2274 <install>: handle Windows platform correctly
2275 <clean>: ditto
2276 <check>: ditto
2277 copy DLL to top builddir before testing
2278
2279 As part of these changes, -no-undefined was removed again. This was reported
2280 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
2281 in any case.
2282
2283 3. Some tidies to get rid of compiler warnings:
2284
2285 . In the match_data structure, match_limit was an unsigned long int, whereas
2286 match_call_count was an int. I've made them both unsigned long ints.
2287
2288 . In pcretest the fact that a const uschar * doesn't automatically cast to
2289 a void * provoked a warning.
2290
2291 . Turning on some more compiler warnings threw up some "shadow" variables
2292 and a few more missing casts.
2293
2294 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
2295 option, a class that contained a single character with a value between 128
2296 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
2297
2298 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
2299 option, a class that contained several characters, but with at least one
2300 whose value was between 128 and 255 caused PCRE to crash.
2301
2302
2303 Version 4.1 12-Mar-03
2304 ---------------------
2305
2306 1. Compiling with gcc -pedantic found a couple of places where casts were
2307 needed, and a string in dftables.c that was longer than standard compilers are
2308 required to support.
2309
2310 2. Compiling with Sun's compiler found a few more places where the code could
2311 be tidied up in order to avoid warnings.
2312
2313 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
2314 first of these names is deprecated in the latest Autoconf in favour of the name
2315 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
2316 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
2317 analogy I have changed it to CFLAGS_FOR_BUILD.
2318
2319 4. Added -no-undefined to the linking command in the Makefile, because this is
2320 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
2321 linking step for the pcreposix library.
2322
2323 5. PCRE was failing to diagnose the case of two named groups with the same
2324 name.
2325
2326 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
2327 literal character that is needed in the subject for a match, and scans along to
2328 ensure that it is present before embarking on the full matching process. This
2329 saves time in cases of nested unlimited repeats that are never going to match.
2330 Problem: the scan can take a lot of time if the subject is very long (e.g.
2331 megabytes), thus penalizing straightforward matches. It is now done only if the
2332 amount of subject to be scanned is less than 1000 bytes.
2333
2334 7. A lesser problem with the same optimization is that it was recording the
2335 first character of an anchored pattern as "needed", thus provoking a search
2336 right along the subject, even when the first match of the pattern was going to
2337 fail. The "needed" character is now not set for anchored patterns, unless it
2338 follows something in the pattern that is of non-fixed length. Thus, it still
2339 fulfils its original purpose of finding quick non-matches in cases of nested
2340 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
2341
2342
2343 Version 4.0 17-Feb-03
2344 ---------------------
2345
2346 1. If a comment in an extended regex that started immediately after a meta-item
2347 extended to the end of string, PCRE compiled incorrect data. This could lead to
2348 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
2349
2350 2. Moved to autoconf 2.53 and libtool 1.4.2.
2351
2352 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
2353 the special perltest8 script is no longer needed - all the tests can be run
2354 from a single perltest script.
2355
2356 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
2357 by \s. It has now been removed in PCRE. This means it isn't recognized as
2358 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
2359 class [:space:] *does* include VT, thereby creating a mess.
2360
2361 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
2362 space and tab.
2363
2364 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
2365 its new features into the main test script, reducing the number of scripts.
2366
2367 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
2368 were backward compatible, and made the (?i) apply to the whole pattern, as if
2369 /i were given. Now it behaves more logically, and applies the option setting
2370 only to what follows. PCRE has been changed to follow suit. However, if it
2371 finds options settings right at the start of the pattern, it extracts them into
2372 the global options, as before. Thus, they show up in the info data.
2373
2374 8. Added support for the \Q...\E escape sequence. Characters in between are
2375 treated as literals. This is slightly different from Perl in that $ and @ are
2376 also handled as literals inside the quotes. In Perl, they will cause variable
2377 interpolation. Note the following examples:
2378
2379 Pattern PCRE matches Perl matches
2380
2381 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
2382 \Qabc\$xyz\E abc\$xyz abc\$xyz
2383 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
2384
2385 For compatibility with Perl, \Q...\E sequences are recognized inside character
2386 classes as well as outside them.
2387
2388 9. Re-organized 3 code statements in pcretest to avoid "overflow in
2389 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
2390 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
2391 signed/unsigned warnings.
2392
2393 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
2394 option for pcretest, so I've replaced it by a simple function that does just
2395 that job.
2396
2397 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
2398 "pcregrep -".
2399
2400 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
2401 Java package. This provides some syntactic sugar for simple cases of what my
2402 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
2403 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
2404 item, you can use this simplified notation. Note that only makes sense with
2405 greedy quantifiers. Consequently, the use of the possessive quantifier forces
2406 greediness, whatever the setting of the PCRE_UNGREEDY option.
2407
2408 13. A change of greediness default within a pattern was not taking effect at
2409 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
2410 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
2411 was abstracted outside.
2412
2413 14. PCRE now supports the \G assertion. It is true when the current matching
2414 position is at the start point of the match. This differs from \A when the
2415 starting offset is non-zero. Used with the /g option of pcretest (or similar
2416 code), it works in the same way as it does for Perl's /g option. If all
2417 alternatives of a regex begin with \G, the expression is anchored to the start
2418 match position, and the "anchored" flag is set in the compiled expression.
2419
2420 15. Some bugs concerning the handling of certain option changes within patterns
2421 have been fixed. These applied to options other than (?ims). For example,
2422 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
2423 been the other way round. Some of this was related to change 7 above.
2424
2425 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
2426 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
2427 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
2428 POSIX classes only within a class (e.g. /[[:alpha:]]/).
2429
2430 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
2431 mode. Unlike ".", it always matches newline, whatever the setting of
2432 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
2433 assertions. Perl allows it, but it doesn't (in general) work because it can't
2434 calculate the length of the lookbehind. At least, that's the case for Perl
2435 5.8.0 - I've been told they are going to document that it doesn't work in
2436 future.
2437
2438 18. Added an error diagnosis for escapes that PCRE does not support: these are
2439 \L, \l, \N, \P, \p, \U, \u, and \X.
2440
2441 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
2442 reading past the end of the pattern in cases such as /[abcd/.
2443
2444 20. PCRE was getting more memory than necessary for patterns with classes that
2445 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
2446
2447 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
2448 compiling PCRE for use with Virtual Pascal.
2449
2450 22. Small fix to the Makefile to make it work properly if the build is done
2451 outside the source tree.
2452
2453 23. Added a new extension: a condition to go with recursion. If a conditional
2454 subpattern starts with (?(R) the "true" branch is used if recursion has
2455 happened, whereas the "false" branch is used only at the top level.
2456
2457 24. When there was a very long string of literal characters (over 255 bytes
2458 without UTF support, over 250 bytes with UTF support), the computation of how
2459 much memory was required could be incorrect, leading to segfaults or other
2460 strange effects.
2461
2462 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
2463 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
2464 there was a subsequent back reference to those brackets. This meant that, for
2465 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
2466 possible to check for precisely this case. All we can do is abandon the
2467 optimization if .* occurs inside capturing brackets when there are any back
2468 references whatsoever. (See below for a better fix that came later.)
2469
2470 26. The handling of the optimization for finding the first character of a
2471 non-anchored pattern, and for finding a character that is required later in the
2472 match were failing in some cases. This didn't break the matching; it just
2473 failed to optimize when it could. The way this is done has been re-implemented.
2474
2475 27. Fixed typo in error message for invalid (?R item (it said "(?p").
2476
2477 28. Added a new feature that provides some of the functionality that Perl
2478 provides with (?{...}). The facility is termed a "callout". The way it is done
2479 in PCRE is for the caller to provide an optional function, by setting
2480 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
2481 global variable. By default it is unset, which disables all calling out. To get
2482 the function called, the regex must include (?C) at appropriate points. This
2483 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
2484 This provides a means of identifying different callout points. When PCRE
2485 reaches such a point in the regex, if pcre_callout has been set, the external
2486 function is called. It is provided with data in a structure called
2487 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
2488 matching continues; if it returns a non-zero value, the match at the current
2489 point fails. However, backtracking will occur if possible. [This was changed
2490 later and other features added - see item 49 below.]
2491
2492 29. pcretest is upgraded to test the callout functionality. It provides a
2493 callout function that displays information. By default, it shows the start of
2494 the match and the current position in the text. There are some new data escapes
2495 to vary what happens:
2496
2497 \C+ in addition, show current contents of captured substrings
2498 \C- do not supply a callout function
2499 \C!n return 1 when callout number n is reached
2500 \C!n!m return 1 when callout number n is reached for the mth time
2501
2502 30. If pcregrep was called with the -l option and just a single file name, it
2503 output "<stdin>" if a match was found, instead of the file name.
2504
2505 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
2506 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
2507 pcre_exec(). This saves a malloc/free per call. The default value of
2508 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
2509 when configuring.
2510
2511 32. The default maximum size of a compiled pattern is 64K. There have been a
2512 few cases of people hitting this limit. The code now uses macros to handle the
2513 storing of links as offsets within the compiled pattern. It defaults to 2-byte
2514 links, but this can be changed to 3 or 4 bytes by --with-link-size when
2515 configuring. Tests 2 and 5 work only with 2-byte links because they output
2516 debugging information about compiled patterns.
2517
2518 33. Internal code re-arrangements:
2519
2520 (a) Moved the debugging function for printing out a compiled regex into
2521 its own source file (printint.c) and used #include to pull it into
2522 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
2523 separate copies.
2524
2525 (b) Defined the list of op-code names for debugging as a macro in
2526 internal.h so that it is next to the definition of the opcodes.
2527
2528 (c) Defined a table of op-code lengths for simpler skipping along compiled
2529 code. This is again a macro in internal.h so that it is next to the
2530 definition of the opcodes.
2531
2532 34. Added support for recursive calls to individual subpatterns, along the
2533 lines of Robin Houston's patch (but implemented somewhat differently).
2534
2535 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
2536 allow it to read and process whole directories in Win32. This code was
2537 contributed by Lionel Fourquaux; it has not been tested by me.
2538
2539 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
2540 used to name a group. Names consist of alphanumerics and underscores, and must
2541 be unique. Back references use the syntax (?P=name) and recursive calls use
2542 (?P>name) which is a PCRE extension to the Python extension. Groups still have
2543 numbers. The function pcre_fullinfo() can be used after compilation to extract
2544 a name/number map. There are three relevant calls:
2545
2546 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
2547 PCRE_INFO_NAMECOUNT yields the number of entries
2548 PCRE_INFO_NAMETABLE yields a pointer to the map.
2549
2550 The map is a vector of fixed-size entries. The size of each entry depends on
2551 the length of the longest name used. The first two bytes of each entry are the
2552 group number, most significant byte first. There follows the corresponding
2553 name, zero terminated. The names are in alphabetical order.
2554
2555 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
2556 case instead of 255. Making it the same both with and without UTF-8 support
2557 means that the same test output works with both.
2558
2559 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
2560 calling malloc() with a zero argument.
2561
2562 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
2563 optimization. I've improved things by keeping a bitmap of backreferences with
2564 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
2565 fact referenced, the optimization can be applied. It is unlikely that a
2566 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
2567 the match to follow \n) will appear inside brackets with a number greater than
2568 31, but if it does, any back reference > 31 suppresses the optimization.
2569
2570 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
2571 of disabling numbered capturing parentheses. Any opening parenthesis that is
2572 not followed by ? behaves as if it were followed by ?: but named parentheses
2573 can still be used for capturing (and they will acquire numbers in the usual
2574 way).
2575
2576 41. Redesigned the return codes from the match() function into yes/no/error so
2577 that errors can be passed back from deep inside the nested calls. A malloc
2578 failure while inside a recursive subpattern call now causes the
2579 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
2580
2581 42. It is now possible to set a limit on the number of times the match()
2582 function is called in a call to pcre_exec(). This facility makes it possible to
2583 limit the amount of recursion and backtracking, though not in a directly
2584 obvious way, because the match() function is used in a number of different
2585 circumstances. The count starts from zero for each position in the subject
2586 string (for non-anchored patterns). The default limit is, for compatibility, a
2587 large number, namely 10 000 000. You can change this in two ways:
2588
2589 (a) When configuring PCRE before making, you can use --with-match-limit=n
2590 to set a default value for the compiled library.
2591
2592 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
2593 a different value is set. See 45 below.
2594
2595 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
2596
2597 43. Added a new function pcre_config(int, void *) to enable run-time extraction
2598 of things that can be changed at compile time. The first argument specifies
2599 what is wanted and the second points to where the information is to be placed.
2600 The current list of available information is:
2601
2602 PCRE_CONFIG_UTF8
2603
2604 The output is an integer that is set to one if UTF-8 support is available;
2605 otherwise it is set to zero.
2606
2607 PCRE_CONFIG_NEWLINE
2608
2609 The output is an integer that it set to the value of the code that is used for
2610 newline. It is either LF (10) or CR (13).
2611
2612 PCRE_CONFIG_LINK_SIZE
2613
2614 The output is an integer that contains the number of bytes used for internal
2615 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
2616
2617 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
2618
2619 The output is an integer that contains the threshold above which the POSIX
2620 interface uses malloc() for output vectors. See item 31 above.
2621
2622 PCRE_CONFIG_MATCH_LIMIT
2623
2624 The output is an unsigned integer that contains the default limit of the number
2625 of match() calls in a pcre_exec() execution. See 42 above.
2626
2627 44. pcretest has been upgraded by the addition of the -C option. This causes it
2628 to extract all the available output from the new pcre_config() function, and to
2629 output it. The program then exits immediately.
2630
2631 45. A need has arisen to pass over additional data with calls to pcre_exec() in
2632 order to support additional features. One way would have been to define
2633 pcre_exec2() (for example) with extra arguments, but this would not have been
2634 extensible, and would also have required all calls to the original function to
2635 be mapped to the new one. Instead, I have chosen to extend the mechanism that
2636 is used for passing in "extra" data from pcre_study().
2637
2638 The pcre_extra structure is now exposed and defined in pcre.h. It currently
2639 contains the following fields:
2640
2641 flags a bitmap indicating which of the following fields are set
2642 study_data opaque data from pcre_study()
2643 match_limit a way of specifying a limit on match() calls for a specific
2644 call to pcre_exec()
2645 callout_data data for callouts (see 49 below)
2646
2647 The flag bits are also defined in pcre.h, and are
2648
2649 PCRE_EXTRA_STUDY_DATA
2650 PCRE_EXTRA_MATCH_LIMIT
2651 PCRE_EXTRA_CALLOUT_DATA
2652
2653 The pcre_study() function now returns one of these new pcre_extra blocks, with
2654 the actual study data pointed to by the study_data field, and the
2655 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
2656 before. That is, this change is entirely upwards-compatible and requires no
2657 change to existing code.
2658
2659 If you want to pass in additional data to pcre_exec(), you can either place it
2660 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
2661 block.
2662
2663 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
2664 data string contains the escape sequence \M, pcretest calls pcre_exec() several
2665 times with different match limits, until it finds the minimum value needed for
2666 pcre_exec() to complete. The value is then output. This can be instructive; for
2667 most simple matches the number is quite small, but for pathological cases it
2668 gets very large very quickly.
2669
2670 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
2671 returns the size of the data block pointed to by the study_data field in a
2672 pcre_extra block, that is, the value that was passed as the argument to
2673 pcre_malloc() when PCRE was getting memory in which to place the information
2674 created by pcre_study(). The fourth argument should point to a size_t variable.
2675 pcretest has been extended so that this information is shown after a successful
2676 pcre_study() call when information about the compiled regex is being displayed.
2677
2678 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
2679 because what follows is always an absolute path. (Later: it turns out that this
2680 is more than cosmetic for MinGW, because it doesn't like empty path
2681 components.)
2682
2683 49. Some changes have been made to the callout feature (see 28 above):
2684
2685 (i) A callout function now has three choices for what it returns:
2686
2687 0 => success, carry on matching
2688 > 0 => failure at this point, but backtrack if possible
2689 < 0 => serious error, return this value from pcre_exec()
2690
2691 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
2692 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
2693 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
2694 use by callout functions. It will never be used by PCRE itself.
2695
2696 (ii) The pcre_extra structure (see 45 above) has a void * field called
2697 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
2698 pcre_callout_block structure has a field of the same name. The contents of
2699 the field passed in the pcre_extra structure are passed to the callout
2700 function in the corresponding field in the callout block. This makes it
2701 easier to use the same callout-containing regex from multiple threads. For
2702 testing, the pcretest program has a new data escape
2703
2704 \C*n pass the number n (may be negative) as callout_data
2705
2706 If the callout function in pcretest receives a non-zero value as
2707 callout_data, it returns that value.
2708
2709 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
2710 there were some redundant $(CFLAGS) in commands that are now specified as
2711 $(LINK), which already includes $(CFLAGS).
2712
2713 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
2714 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
2715 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
2716 one-byte characters throughout. Note that case-insensitive matching applies
2717 only to characters whose values are less than 256. PCRE doesn't support the
2718 notion of cases for higher-valued characters.
2719
2720 (i) A character class whose characters are all within 0-255 is handled as
2721 a bit map, and the map is inverted for negative classes. Previously, a
2722 character > 255 always failed to match such a class; however it should
2723 match if the class was a negative one (e.g. [^ab]). This has been fixed.
2724
2725 (ii) A negated character class with a single character < 255 is coded as
2726 "not this character" (OP_NOT). This wasn't working properly when the test
2727 character was multibyte, either singly or repeated.
2728
2729 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
2730 mode, for example: \x{100}{2,3}.
2731
2732 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
2733 singly or repeated) now correctly test multibyte characters. However,
2734 PCRE doesn't recognize any characters with values greater than 255 as
2735 digits, spaces, or word characters. Such characters always match \D, \S,
2736 and \W, and never match \d, \s, or \w.
2737
2738 (v) Classes may now contain characters and character ranges with values
2739 greater than 255. For example: [ab\x{100}-\x{400}].
2740
2741 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
2742 PCRE in UTF-8 mode.
2743
2744 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
2745 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
2746 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
2747 value.)
2748
2749 53. The single man page has become too large. I have therefore split it up into
2750 a number of separate man pages. These also give rise to individual HTML pages;
2751 these are now put in a separate directory, and there is an index.html page that
2752 lists them all. Some hyperlinking between the pages has been installed.
2753
2754 54. Added convenience functions for handling named capturing parentheses.
2755
2756 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
2757 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
2758 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
2759 are faulted.
2760
2761 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
2762 calling configure. These values are used when compiling the dftables.c program
2763 which is run to generate the source of the default character tables. They
2764 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
2765 you will need to set these values.
2766
2767 57. Updated the building process for Windows DLL, as provided by Fred Cox.
2768
2769
2770 Version 3.9 02-Jan-02
2771 ---------------------
2772
2773 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
2774
2775 2. If --disable-static was given, the building process failed when trying to
2776 build pcretest and pcregrep. (For some reason it was using libtool to compile
2777 them, which is not right, as they aren't part of the library.)
2778
2779
2780 Version 3.8 18-Dec-01
2781 ---------------------
2782
2783 1. The experimental UTF-8 code was completely screwed up. It was packing the
2784 bytes in the wrong order. How dumb can you get?
2785
2786
2787 Version 3.7 29-Oct-01
2788 ---------------------
2789
2790 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
2791 This caused pcretest, when used on the test data, to segfault. Unfortunately,
2792 this didn't happen under Solaris 8, where I normally test things.
2793
2794 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
2795 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
2796 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
2797 this fix an hour or so after the initial 3.7 release.)
2798
2799
2800 Version 3.6 23-Oct-01
2801 ---------------------
2802
2803 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
2804 offsets passed as NULL with zero offset count.
2805
2806 2. The config.guess and config.sub files had not been updated when I moved to
2807 the latest autoconf.
2808
2809
2810 Version 3.5 15-Aug-01
2811 ---------------------
2812
2813 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
2814 had been forgotten.
2815
2816 2. By using declared but undefined structures, we can avoid using "void"
2817 definitions in pcre.h while keeping the internal definitions of the structures
2818 private.
2819
2820 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
2821 user point of view, this means that both static and shared libraries are built
2822 by default, but this can be individually controlled. More of the work of
2823 handling this static/shared cases is now inside libtool instead of PCRE's make
2824 file.
2825
2826 4. The pcretest utility is now installed along with pcregrep because it is
2827 useful for users (to test regexs) and by doing this, it automatically gets
2828 relinked by libtool. The documentation has been turned into a man page, so
2829 there are now .1, .txt, and .html versions in /doc.
2830
2831 5. Upgrades to pcregrep:
2832 (i) Added long-form option names like gnu grep.
2833 (ii) Added --help to list all options with an explanatory phrase.
2834 (iii) Added -r, --recursive to recurse into sub-directories.
2835 (iv) Added -f, --file to read patterns from a file.
2836
2837 6. pcre_exec() was referring to its "code" argument before testing that
2838 argument for NULL (and giving an error if it was NULL).
2839
2840 7. Upgraded Makefile.in to allow for compiling in a different directory from
2841 the source directory.
2842
2843 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
2844 options bits, the pointer it was passed was to an int instead of to an unsigned
2845 long int. This mattered only on 64-bit systems.
2846
2847 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
2848 generated) instead of pcre.in, which it its source. Also made the same change
2849 in several of the .c files.
2850
2851 10. A new release of gcc defines printf() as a macro, which broke pcretest
2852 because it had an ifdef in the middle of a string argument for printf(). Fixed
2853 by using separate calls to printf().
2854
2855 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
2856 script, to force use of CR or LF instead of \n in the source. On non-Unix
2857 systems, the value can be set in config.h.
2858
2859 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
2860 absolute limit. Changed the text of the error message to make this clear, and
2861 likewise updated the man page.
2862
2863 13. The limit of 99 on the number of capturing subpatterns has been removed.
2864 The new limit is 65535, which I hope will not be a "real" limit.
2865
2866
2867 Version 3.4 22-Aug-00
2868 ---------------------
2869
2870 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
2871
2872 2. Diagnose condition (?(0) as an error instead of crashing on matching.
2873
2874
2875 Version 3.3 01-Aug-00
2876 ---------------------
2877
2878 1. If an octal character was given, but the value was greater than \377, it
2879 was not getting masked to the least significant bits, as documented. This could
2880 lead to crashes in some systems.
2881
2882 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
2883 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
2884
2885 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
2886 These just pass their arguments on to (pcre_free)(), but they are provided
2887 because some uses of PCRE bind it to non-C systems that can call its functions,
2888 but cannot call free() or pcre_free() directly.
2889
2890 4. Add "make test" as a synonym for "make check". Corrected some comments in
2891 the Makefile.
2892
2893 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
2894 Makefile.
2895
2896 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
2897 command called pgrep for grepping around the active processes.
2898
2899 7. Added the beginnings of support for UTF-8 character strings.
2900
2901 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
2902 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
2903 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
2904 out for the ar command.)
2905
2906
2907 Version 3.2 12-May-00
2908 ---------------------
2909
2910 This is purely a bug fixing release.
2911
2912 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
2913 of ZA. This was just one example of several cases that could provoke this bug,
2914 which was introduced by change 9 of version 2.00. The code for breaking
2915 infinite loops after an iteration that matches an empty string was't working
2916 correctly.
2917
2918 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
2919 when matched against abbab (for example). After matching an empty string, it
2920 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
2921 caused it to match further down the string than it should.
2922
2923 3. The code contained an inclusion of sys/types.h. It isn't clear why this
2924 was there because it doesn't seem to be needed, and it causes trouble on some
2925 systems, as it is not a Standard C header. It has been removed.
2926
2927 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
2928 were reported on the Macintosh. The changes were from
2929
2930 while ((c = *(++ptr)) != 0 && c != '\n');
2931 to
2932 while ((c = *(++ptr)) != 0 && c != '\n') ;
2933
2934 Totally extraordinary, but if that's what it takes...
2935
2936 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
2937 available. Added HAVE_BCOPY and an autoconf test for it; if neither
2938 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
2939 assumes the way PCRE uses memmove() (always moving upwards).
2940
2941 6. PCRE is being used in one environment where strchr() is not available. There
2942 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
2943 faster code anyway.
2944
2945
2946 Version 3.1 09-Feb-00
2947 ---------------------
2948
2949 The only change in this release is the fixing of some bugs in Makefile.in for
2950 the "install" target:
2951
2952 (1) It was failing to install pcreposix.h.
2953
2954 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
2955
2956
2957 Version 3.0 01-Feb-00
2958 ---------------------
2959
2960 1. Add support for the /+ modifier to perltest (to output $` like it does in
2961 pcretest).
2962
2963 2. Add support for the /g modifier to perltest.
2964
2965 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
2966 matches null strings.
2967
2968 4. Fix perltest so that it doesn't do unwanted things when fed an empty
2969 pattern. Perl treats empty patterns specially - it reuses the most recent
2970 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
2971 effect.
2972
2973 5. The POSIX interface was broken in that it was just handing over the POSIX
2974 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
2975 required a bigger vector, with some working space on the end. This means that
2976 the POSIX wrapper now has to get and free some memory, and copy the results.
2977
2978 6. Added some simple autoconf support, placing the test data and the
2979 documentation in separate directories, re-organizing some of the
2980 information files, and making it build pcre-config (a GNU standard). Also added
2981 libtool support for building PCRE as a shared library, which is now the
2982 default.
2983
2984 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
2985 09 are not valid octal constants. Single digits will be used for minor values
2986 less than 10.
2987
2988 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
2989 existing programs that set these in the POSIX interface can use PCRE without
2990 modification.
2991
2992 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
2993 return all that pcre_info() returns, plus additional data. The pcre_info()
2994 function is retained for compatibility, but is considered to be obsolete.
2995
2996 10. Added experimental recursion feature (?R) to handle one common case that
2997 Perl 5.6 will be able to do with (?p{...}).
2998
2999 11. Added support for POSIX character classes like [:alpha:], which Perl is
3000 adopting.
3001
3002
3003 Version 2.08 31-Aug-99
3004 ----------------------
3005
3006 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
3007 trying to match at the startoffset position, but instead was moving forward to
3008 the next newline as if a previous match had failed.
3009
3010 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
3011 and could get into a loop if a null string was matched other than at the start
3012 of the subject.
3013
3014 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
3015 be distinguished at compile time, and for completeness also added PCRE_DATE.
3016
3017 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
3018 in GnuWin32 environments.
3019
3020
3021 Version 2.07 29-Jul-99
3022 ----------------------
3023
3024 1. The documentation is now supplied in plain text form and HTML as well as in
3025 the form of man page sources.
3026
3027 2. C++ compilers don't like assigning (void *) values to other pointer types.
3028 In particular this affects malloc(). Although there is no problem in Standard
3029 C, I've put in casts to keep C++ compilers happy.
3030
3031 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
3032 should be (const char *).
3033
3034 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
3035 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
3036 However, I haven't made this a standard facility. The documentation doesn't
3037 mention it, and the Makefile doesn't support it.
3038
3039 5. The Makefile now contains an "install" target, with editable destinations at
3040 the top of the file. The pcretest program is not installed.
3041
3042 6. pgrep -V now gives the PCRE version number and date.
3043
3044 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
3045 causing the entire string to be ignored, instead of just the last character.
3046
3047 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
3048 non-matching string, it can take a very, very long time, even for strings of
3049 quite modest length, because of the nested recursion. PCRE now does better in
3050 some of these cases. It does this by remembering the last required literal
3051 character in the pattern, and pre-searching the subject to ensure it is present
3052 before running the real match. In other words, it applies a heuristic to detect
3053 some types of certain failure quickly, and in the above example, if presented
3054 with a string that has no trailing " it gives "no match" very quickly.
3055
3056 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
3057 other alternatives are tried instead.
3058
3059
3060 Version 2.06 09-Jun-99
3061 ----------------------
3062
3063 1. Change pcretest's output for amount of store used to show just the code
3064 space, because the remainder (the data block) varies in size between 32-bit and
3065 64-bit systems.
3066
3067 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
3068 start matching at. This allows lookbehinds to work when searching for multiple
3069 occurrences in a string.
3070
3071 3. Added additional options to pcretest for testing multiple occurrences:
3072
3073 /+ outputs the rest of the string that follows a match
3074 /g loops for multiple occurrences, using the new startoffset argument
3075 /G loops for multiple occurrences by passing an incremented pointer
3076
3077 4. PCRE wasn't doing the "first character" optimization for patterns starting
3078 with \b or \B, though it was doing it for other lookbehind assertions. That is,
3079 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
3080 the letter 'x'. On long subject strings, this gives a significant speed-up.
3081
3082
3083 Version 2.05 21-Apr-99
3084 ----------------------
3085
3086 1. Changed the type of magic_number from int to long int so that it works
3087 properly on 16-bit systems.
3088
3089 2. Fixed a bug which caused patterns starting with .* not to work correctly
3090 when the subject string contained newline characters. PCRE was assuming
3091 anchoring for such patterns in all cases, which is not correct because .* will
3092 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
3093 DOTALL is set at top level; otherwise it knows that patterns starting with .*
3094 must be retried after every newline in the subject.
3095
3096
3097 Version 2.04 18-Feb-99
3098 ----------------------
3099
3100 1. For parenthesized subpatterns with repeats whose minimum was zero, the
3101 computation of the store needed to hold the pattern was incorrect (too large).
3102 If such patterns were nested a few deep, this could multiply and become a real
3103 problem.
3104
3105 2. Added /M option to pcretest to show the memory requirement of a specific
3106 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
3107
3108 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
3109 compiled in such a way that the backtracking after subsequent failure was
3110 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
3111 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
3112
3113
3114 Version 2.03 02-Feb-99
3115 ----------------------
3116
3117 1. Fixed typo and small mistake in man page.
3118
3119 2. Added 4th condition (GPL supersedes if conflict) and created separate
3120 LICENCE file containing the conditions.
3121
3122 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
3123 Perl, that is the internal \ allows the delimiter to be included in the
3124 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
3125 the final delimiter, add \ to the end of the pattern (to test the error).
3126
3127 4. Added the convenience functions for extracting substrings after a successful
3128 match. Updated pcretest to make it able to test these functions.
3129
3130
3131 Version 2.02 14-Jan-99
3132 ----------------------
3133
3134 1. Initialized the working variables associated with each extraction so that
3135 their saving and restoring doesn't refer to uninitialized store.
3136
3137 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
3138 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
3139 fix the problem.
3140
3141 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
3142 calls, and wasn't printing the correct value for compiling calls. Increased the
3143 default value of LOOPREPEAT, and the number of significant figures in the
3144 times.
3145
3146 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
3147
3148 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
3149 a building problem on Windows NT with a FAT file system.
3150
3151
3152 Version 2.01 21-Oct-98
3153 ----------------------
3154
3155 1. Changed the API for pcre_compile() to allow for the provision of a pointer
3156 to character tables built by pcre_maketables() in the current locale. If NULL
3157 is passed, the default tables are used.
3158
3159
3160 Version 2.00 24-Sep-98
3161 ----------------------
3162
3163 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
3164 it any more.
3165
3166 2. Allow quantification of (?>) groups, and make it work correctly.
3167
3168 3. The first character computation wasn't working for (?>) groups.
3169
3170 4. Correct the implementation of \Z (it is permitted to match on the \n at the
3171 end of the subject) and add 5.005's \z, which really does match only at the
3172 very end of the subject.
3173
3174 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
3175
3176 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
3177 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
3178 localized options. All options to pcre_study() were also removed.
3179
3180 7. Add other new features from 5.005:
3181
3182 $(?<= positive lookbehind
3183 $(?<! negative lookbehind
3184 (?imsx-imsx) added the unsetting capability
3185 such a setting is global if at outer level; local otherwise
3186 (?imsx-imsx:) non-capturing groups with option setting
3187 (?(cond)re|re) conditional pattern matching
3188
3189 A backreference to itself in a repeated group matches the previous
3190 captured string.
3191
3192 8. General tidying up of studying (both automatic and via "study")
3193 consequential on the addition of new assertions.
3194
3195 9. As in 5.005, unlimited repeated groups that could match an empty substring
3196 are no longer faulted at compile time. Instead, the loop is forcibly broken at
3197 runtime if any iteration does actually match an empty substring.
3198
3199 10. Include the RunTest script in the distribution.
3200
3201 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
3202 discrepancies, some of which were old and were also with respect to 5.004. They
3203 have now been fixed.
3204
3205
3206 Version 1.09 28-Apr-98
3207 ----------------------
3208
3209 1. A negated single character class followed by a quantifier with a minimum
3210 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
3211 program crashes, or just wrong answers. This did not apply to negated classes
3212 containing more than one character, or to minima other than one.
3213
3214
3215 Version 1.08 27-Mar-98
3216 ----------------------
3217
3218 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
3219
3220 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
3221 latter must appear before anything that relies on it in the pattern.
3222
3223
3224 Version 1.07 16-Feb-98
3225 ----------------------
3226
3227 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
3228 repeat of a potentially empty string).
3229
3230
3231 Version 1.06 23-Jan-98
3232 ----------------------
3233
3234 1. Added Markus Oberhumer's little patches for C++.
3235
3236 2. Literal strings longer than 255 characters were broken.
3237
3238
3239 Version 1.05 23-Dec-97
3240 ----------------------
3241
3242 1. Negated character classes containing more than one character were failing if
3243 PCRE_CASELESS was set at run time.
3244
3245
3246 Version 1.04 19-Dec-97
3247 ----------------------
3248
3249 1. Corrected the man page, where some "const" qualifiers had been omitted.
3250
3251 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
3252 input syntax.
3253
3254 3. Fixed memory leak which occurred when a regex with back references was
3255 matched with an offsets vector that wasn't big enough. The temporary memory
3256 that is used in this case wasn't being freed if the match failed.
3257
3258 4. Tidied pcretest to ensure it frees memory that it gets.
3259
3260 5. Temporary memory was being obtained in the case where the passed offsets
3261 vector was exactly big enough.
3262
3263 6. Corrected definition of offsetof() from change 5 below.
3264
3265 7. I had screwed up change 6 below and broken the rules for the use of
3266 setjmp(). Now fixed.
3267
3268
3269 Version 1.03 18-Dec-97
3270 ----------------------
3271
3272 1. A erroneous regex with a missing opening parenthesis was correctly
3273 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
3274 on some systems.
3275
3276 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
3277 it was reported that one broken compiler failed on the former because "code" is
3278 also an independent variable.
3279
3280 3. The erroneous regex a[]b caused an array overrun reference.
3281
3282 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
3283 fail on data ending with that character. (It was going on too far, and checking
3284 the next character, typically a binary zero.) This was specific to the
3285 optimized code for single-character negative classes.
3286
3287 5. Added a contributed patch from the TIN world which does the following:
3288
3289 + Add an undef for memmove, in case the the system defines a macro for it.
3290
3291 + Add a definition of offsetof(), in case there isn't one. (I don't know
3292 the reason behind this - offsetof() is part of the ANSI standard - but
3293 it does no harm).
3294
3295 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
3296 most of the places where whitespace preceded '#'. I have given up and
3297 allowed the remaining 2 cases to be at the margin.
3298
3299 + Rename some variables in pcre to eliminate shadowing. This seems very
3300 pedantic, but does no harm, of course.
3301
3302 6. Moved the call to setjmp() into its own function, to get rid of warnings
3303 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
3304
3305 7. Constructs such as \d{8,} were compiling into the equivalent of
3306 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
3307 outcome, but in this particular case used more store than had been allocated,
3308 which caused the bug to be discovered because it threw up an internal error.
3309
3310 8. The debugging code in both pcre and pcretest for outputting the compiled
3311 form of a regex was going wrong in the case of back references followed by
3312 curly-bracketed repeats.
3313
3314
3315 Version 1.02 12-Dec-97
3316 ----------------------
3317
3318 1. Typos in pcre.3 and comments in the source fixed.
3319
3320 2. Applied a contributed patch to get rid of places where it used to remove
3321 'const' from variables, and fixed some signed/unsigned and uninitialized
3322 variable warnings.
3323
3324 3. Added the "runtest" target to Makefile.
3325
3326 4. Set default compiler flag to -O2 rather than just -O.
3327
3328
3329 Version 1.01 19-Nov-97
3330 ----------------------
3331
3332 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
3333 like /([ab]*)*/, that is, for classes with more than one character in them.
3334
3335 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
3336 as /((?>a*))*/ (a PCRE_EXTRA facility).
3337
3338
3339 Version 1.00 18-Nov-97
3340 ----------------------
3341
3342 1. Added compile-time macros to support systems such as SunOS4 which don't have
3343 memmove() or strerror() but have other things that can be used instead.
3344
3345 2. Arranged that "make clean" removes the executables.
3346
3347
3348 Version 0.99 27-Oct-97
3349 ----------------------
3350
3351 1. Fixed bug in code for optimizing classes with only one character. It was
3352 initializing a 32-byte map regardless, which could cause it to run off the end
3353 of the memory it had got.
3354
3355 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
3356
3357
3358 Version 0.98 22-Oct-97
3359 ----------------------
3360
3361 1. Fixed bug in code for handling temporary memory usage when there are more
3362 back references than supplied space in the ovector. This could cause segfaults.
3363
3364
3365 Version 0.97 21-Oct-97
3366 ----------------------
3367
3368 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
3369
3370 2. Optimized negated single characters not to use a bit map.
3371
3372 3. Brought error texts together as macro definitions; clarified some of them;
3373 fixed one that was wrong - it said "range out of order" when it meant "invalid
3374 escape sequence".
3375
3376 4. Changed some char * arguments to const char *.
3377
3378 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
3379
3380 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
3381 pcretest.
3382
3383
3384 Version 0.96 16-Oct-97
3385 ----------------------
3386
3387 1. Added a simple "pgrep" utility to the distribution.
3388
3389 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
3390 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
3391 where "ddd" means "one or more decimal digits".
3392
3393 3. Fixed serious bug. If a pattern had a back reference, but the call to
3394 pcre_exec() didn't supply a large enough ovector to record the related
3395 identifying subpattern, the match always failed. PCRE now remembers the number
3396 of the largest back reference, and gets some temporary memory in which to save
3397 the offsets during matching if necessary, in order to ensure that
3398 backreferences always work.
3399
3400 4. Increased the compatibility with Perl in a number of ways:
3401
3402 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
3403 to request this handling. The option can be set at compile or exec time.
3404
3405 (b) $ matches before a terminating newline by default; an option
3406 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
3407 mode). The option can be set at compile or exec time.
3408
3409 (c) The handling of \ followed by a digit other than 0 is now supposed to be
3410 the same as Perl's. If the decimal number it represents is less than 10
3411 or there aren't that many previous left capturing parentheses, an octal
3412 escape is read. Inside a character class, it's always an octal escape,
3413 even if it is a single digit.
3414
3415 (d) An escaped but undefined alphabetic character is taken as a literal,
3416 unless PCRE_EXTRA is set. Currently this just reserves the remaining
3417 escapes.
3418
3419 (e) {0} is now permitted. (The previous item is removed from the compiled
3420 pattern).
3421
3422 5. Changed all the names of code files so that the basic parts are no longer
3423 than 10 characters, and abolished the teeny "globals.c" file.
3424
3425 6. Changed the handling of character classes; they are now done with a 32-byte
3426 bit map always.
3427
3428 7. Added the -d and /D options to pcretest to make it possible to look at the
3429 internals of compilation without having to recompile pcre.
3430
3431
3432 Version 0.95 23-Sep-97
3433 ----------------------
3434
3435 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
3436 \x20 at the start of a run of normal characters. These were being treated as
3437 real characters, instead of the source characters being re-checked.
3438
3439
3440 Version 0.94 18-Sep-97
3441 ----------------------
3442
3443 1. The functions are now thread-safe, with the caveat that the global variables
3444 containing pointers to malloc() and free() or alternative functions are the
3445 same for all threads.
3446
3447 2. Get pcre_study() to generate a bitmap of initial characters for non-
3448 anchored patterns when this is possible, and use it if passed to pcre_exec().
3449
3450
3451 Version 0.93 15-Sep-97
3452 ----------------------
3453
3454 1. /(b)|(:+)/ was computing an incorrect first character.
3455
3456 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
3457 but not actually doing anything yet.
3458
3459 3. Treat "-" characters in classes that cannot be part of ranges as literals,
3460 as Perl does (e.g. [-az] or [az-]).
3461
3462 4. Set the anchored flag if a branch starts with .* or .*? because that tests
3463 all possible positions.
3464
3465 5. Split up into different modules to avoid including unneeded functions in a
3466 compiled binary. However, compile and exec are still in one module. The "study"
3467 function is split off.
3468
3469 6. The character tables are now in a separate module whose source is generated
3470 by an auxiliary program - but can then be edited by hand if required. There are
3471 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
3472 toupper() in the code.
3473
3474 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
3475 make them global. Abolish the function for setting them, as the caller can now
3476 set them directly.
3477
3478
3479 Version 0.92 11-Sep-97
3480 ----------------------
3481
3482 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
3483 (e.g. /a{1,3}/) was broken (I mis-optimized it).
3484
3485 2. Caseless matching was not working in character classes if the characters in
3486 the pattern were in upper case.
3487
3488 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
3489
3490 4. Make PCRE_ANCHORED public and accept as a compile option.
3491
3492 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
3493 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
3494 pass them.
3495
3496 6. Give an error if bad option bits passed at compile or run time.
3497
3498 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
3499 pcretest to cause it to pass that flag.
3500
3501 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
3502 options, and the first character, if set.
3503
3504 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
3505
3506
3507 Version 0.91 10-Sep-97
3508 ----------------------
3509
3510 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
3511 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
3512
3513 2. PCRE was looping on encountering an indefinitely repeated back reference to
3514 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
3515 Perl does - treats the match as successful.
3516
3517 ****

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