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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 419 - (show annotations) (download)
Wed Aug 12 10:45:33 2009 UTC (5 years, 1 month ago) by ph10
File size: 166063 byte(s)
Open with "rb" in pcregrep.

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8 was out-of-date, and there was no check on the pcre_compile() error code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 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
13 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
14 in a Windows environment.
15
16
17 Version 7.9 11-Apr-09
18 ---------------------
19
20 1. When building with support for bzlib/zlib (pcregrep) and/or readline
21 (pcretest), all targets were linked against these libraries. This included
22 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
23 libraries. This caused unwanted dependencies to be created. This problem
24 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
25 pcretest is linked with readline.
26
27 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
28 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
29 moved outside it again, because FALSE and TRUE are already defined in AIX,
30 but BOOL is not.
31
32 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
33 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
34
35 4. The pcregrep documentation said spaces were inserted as well as colons (or
36 hyphens) following file names and line numbers when outputting matching
37 lines. This is not true; no spaces are inserted. I have also clarified the
38 wording for the --colour (or --color) option.
39
40 5. In pcregrep, when --colour was used with -o, the list of matching strings
41 was not coloured; this is different to GNU grep, so I have changed it to be
42 the same.
43
44 6. When --colo(u)r was used in pcregrep, only the first matching substring in
45 each matching line was coloured. Now it goes on to look for further matches
46 of any of the test patterns, which is the same behaviour as GNU grep.
47
48 7. A pattern that could match an empty string could cause pcregrep to loop; it
49 doesn't make sense to accept an empty string match in pcregrep, so I have
50 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
51 seems to be how GNU grep behaves.
52
53 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
54 start or after a newline", because the conditional assertion was not being
55 correctly handled. The rule now is that both the assertion and what follows
56 in the first alternative must satisfy the test.
57
58 9. If auto-callout was enabled in a pattern with a conditional group whose
59 condition was an assertion, PCRE could crash during matching, both with
60 pcre_exec() and pcre_dfa_exec().
61
62 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
63 used for matching.
64
65 11. Unicode property support in character classes was not working for
66 characters (bytes) greater than 127 when not in UTF-8 mode.
67
68 12. Added the -M command line option to pcretest.
69
70 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
71
72 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
73
74 16. Added comments and documentation about mis-use of no_arg in the C++
75 wrapper.
76
77 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
78 from Martin Jerabek that uses macro names for all relevant character and
79 string constants.
80
81 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
82 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
83 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
84 these, but not everybody uses configure.
85
86 19. A conditional group that had only one branch was not being correctly
87 recognized as an item that could match an empty string. This meant that an
88 enclosing group might also not be so recognized, causing infinite looping
89 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
90 with the subject "ab", where knowledge that the repeated group can match
91 nothing is needed in order to break the loop.
92
93 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
94 exec(), but without supplying a callout function, matching went wrong.
95
96 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
97 leak if the size of the offset vector was greater than 30. When the vector
98 is smaller, the saved offsets during recursion go onto a local stack
99 vector, but for larger vectors malloc() is used. It was failing to free
100 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
101 error, in fact).
102
103 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
104 heapframe that is used only when UTF-8 support is enabled. This caused no
105 problem, but was untidy.
106
107 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
108 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
109 included within another project.
110
111 24. Steven Van Ingelgem's patches to add more options to the CMake support,
112 slightly modified by me:
113
114 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
115 not building pcregrep.
116
117 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
118 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
119
120 25. Forward references, both numeric and by name, in patterns that made use of
121 duplicate group numbers, could behave incorrectly or give incorrect errors,
122 because when scanning forward to find the reference group, PCRE was not
123 taking into account the duplicate group numbers. A pattern such as
124 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
125
126 26. Changed a few more instances of "const unsigned char *" to USPTR, making
127 the feature of a custom pointer more persuasive (as requested by a user).
128
129 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
130 pcretest.c, inside #ifndefs, because it seems they are sometimes already
131 pre-defined.
132
133 28. Added support for (*UTF8) at the start of a pattern.
134
135 29. Arrange for flags added by the "release type" setting in CMake to be shown
136 in the configuration summary.
137
138
139 Version 7.8 05-Sep-08
140 ---------------------
141
142 1. Replaced UCP searching code with optimized version as implemented for Ad
143 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
144 stage table and inline lookup instead of a function, giving speed ups of 2
145 to 5 times on some simple patterns that I tested. Permission was given to
146 distribute the MultiStage2.py script that generates the tables (it's not in
147 the tarball, but is in the Subversion repository).
148
149 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
150 scripts.
151
152 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
153 a group with a zero qualifier. The result of the study could be incorrect,
154 or the function might crash, depending on the pattern.
155
156 4. Caseless matching was not working for non-ASCII characters in back
157 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
158 It now works when Unicode Property Support is available.
159
160 5. In pcretest, an escape such as \x{de} in the data was always generating
161 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
162 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
163 truncation.
164
165 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
166
167 7. Added two (int) casts to pcregrep when printing the difference of two
168 pointers, in case they are 64-bit values.
169
170 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
171 test 2 if it fails.
172
173 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
174 and a #define of that name to empty if it is not externally set. This is to
175 allow users of MSVC to set it if necessary.
176
177 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
178 the convenience functions in the pcre_get.c source file.
179
180 11. An option change at the start of a pattern that had top-level alternatives
181 could cause overwriting and/or a crash. This command provoked a crash in
182 some environments:
183
184 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
185
186 This potential security problem was recorded as CVE-2008-2371.
187
188 12. For a pattern where the match had to start at the beginning or immediately
189 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
190 pcre_dfa_exec() could read past the end of the passed subject if there was
191 no match. To help with detecting such bugs (e.g. with valgrind), I modified
192 pcretest so that it places the subject at the end of its malloc-ed buffer.
193
194 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
195 exec() might read past the end of the data buffer in UTF-8 mode.
196
197 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
198 the data contained the byte 0x85 as part of a UTF-8 character within its
199 first line. This applied both to normal and DFA matching.
200
201 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
202 /^[^d]*?$/8 failed to match "abc".
203
204 16. Added a missing copyright notice to pcrecpp_internal.h.
205
206 17. Make it more clear in the documentation that values returned from
207 pcre_exec() in ovector are byte offsets, not character counts.
208
209 18. Tidied a few places to stop certain compilers from issuing warnings.
210
211 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
212 supplied by Stefan Weber. I made a further small update for 7.8 because
213 there is a change of source arrangements: the pcre_searchfuncs.c module is
214 replaced by pcre_ucd.c.
215
216
217 Version 7.7 07-May-08
218 ---------------------
219
220 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
221 a string to a long long, pretend we don't even have a long long." This is
222 done by checking for the strtoq, strtoll, and _strtoi64 functions.
223
224 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
225 pre-7.6 versions, which defined a global no_arg variable instead of putting
226 it in the RE class. (See also #8 below.)
227
228 3. Remove a line of dead code, identified by coverity and reported by Nuno
229 Lopes.
230
231 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
232
233 (1) The include/exclude patterns were being applied to the whole pathnames
234 of files, instead of just to the final components.
235
236 (2) If there was more than one level of directory, the subdirectories were
237 skipped unless they satisfied the include/exclude conditions. This is
238 inconsistent with GNU grep (and could even be seen as contrary to the
239 pcregrep specification - which I improved to make it absolutely clear).
240 The action now is always to scan all levels of directory, and just
241 apply the include/exclude patterns to regular files.
242
243 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
244 --exclude_dir in the tests to avoid scanning .svn directories.
245
246 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
247 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
248 doesn't support NULs in patterns.
249
250 7. Added some missing "const"s to declarations of static tables in
251 pcre_compile.c and pcre_dfa_exec.c.
252
253 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
254 caused by fix #2 above. (Subsequently also a second patch to fix the
255 first patch. And a third patch - this was a messy problem.)
256
257 9. Applied Craig's patch to remove the use of push_back().
258
259 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
260 matching function regexec().
261
262 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
263 which, however, unlike Perl's \g{...}, are subroutine calls, not back
264 references. PCRE supports relative numbers with this syntax (I don't think
265 Oniguruma does).
266
267 12. Previously, a group with a zero repeat such as (...){0} was completely
268 omitted from the compiled regex. However, this means that if the group
269 was called as a subroutine from elsewhere in the pattern, things went wrong
270 (an internal error was given). Such groups are now left in the compiled
271 pattern, with a new opcode that causes them to be skipped at execution
272 time.
273
274 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
275 to the way PCRE behaves:
276
277 (a) A lone ] character is dis-allowed (Perl treats it as data).
278
279 (b) A back reference to an unmatched subpattern matches an empty string
280 (Perl fails the current match path).
281
282 (c) A data ] in a character class must be notated as \] because if the
283 first data character in a class is ], it defines an empty class. (In
284 Perl it is not possible to have an empty class.) The empty class []
285 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
286 The negative empty class [^] matches any one character, independently
287 of the DOTALL setting.
288
289 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
290 non-existent subpattern following a character class starting with ']' and
291 containing () gave an internal compiling error instead of "reference to
292 non-existent subpattern". Fortunately, when the pattern did exist, the
293 compiled code was correct. (When scanning forwards to check for the
294 existencd of the subpattern, it was treating the data ']' as terminating
295 the class, so got the count wrong. When actually compiling, the reference
296 was subsequently set up correctly.)
297
298 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
299 it was being rejected as not supported by pcre_dfa_exec(), even though
300 other assertions are supported. I have made pcre_dfa_exec() support
301 (*FAIL).
302
303 16. The implementation of 13c above involved the invention of a new opcode,
304 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
305 cannot be changed at match time, I realized I could make a small
306 improvement to matching performance by compiling OP_ALLANY instead of
307 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
308 on the OP_ANY path.
309
310 17. Compiling pcretest on Windows with readline support failed without the
311 following two fixes: (1) Make the unistd.h include conditional on
312 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
313
314 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
315 ncurses library to be included for pcretest when ReadLine support is
316 requested, but also to allow for it to be overridden. This patch came from
317 Daniel Bergström.
318
319 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
320 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
321 any errors with the current Unicode tables. Thanks to Peter Kankowski for
322 spotting this.
323
324
325 Version 7.6 28-Jan-08
326 ---------------------
327
328 1. A character class containing a very large number of characters with
329 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
330 overflow.
331
332 2. Patch to cut out the "long long" test in pcrecpp_unittest when
333 HAVE_LONG_LONG is not defined.
334
335 3. Applied Christian Ehrlicher's patch to update the CMake build files to
336 bring them up to date and include new features. This patch includes:
337
338 - Fixed PH's badly added libz and libbz2 support.
339 - Fixed a problem with static linking.
340 - Added pcredemo. [But later removed - see 7 below.]
341 - Fixed dftables problem and added an option.
342 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
343 HAVE_LONG_LONG.
344 - Added readline support for pcretest.
345 - Added an listing of the option settings after cmake has run.
346
347 4. A user submitted a patch to Makefile that makes it easy to create
348 "pcre.dll" under mingw when using Configure/Make. I added stuff to
349 Makefile.am that cause it to include this special target, without
350 affecting anything else. Note that the same mingw target plus all
351 the other distribution libraries and programs are now supported
352 when configuring with CMake (see 6 below) instead of with
353 Configure/Make.
354
355 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
356 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
357 exported in the Windows port". It has not yet been confirmed that the patch
358 solves the problem, but it does no harm.
359
360 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
361 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
362 with CMake, and also correct the comment about stack recursion.
363
364 7. Remove the automatic building of pcredemo from the ./configure system and
365 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
366 of a program that users should build themselves after PCRE is installed, so
367 building it automatically is not really right. What is more, it gave
368 trouble in some build environments.
369
370 8. Further tidies to CMakeLists.txt from Sheri and Christian.
371
372
373 Version 7.5 10-Jan-08
374 ---------------------
375
376 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
377 values in parens when parsing an RE using the C++ wrapper."
378
379 2. Negative specials like \S did not work in character classes in UTF-8 mode.
380 Characters greater than 255 were excluded from the class instead of being
381 included.
382
383 3. The same bug as (2) above applied to negated POSIX classes such as
384 [:^space:].
385
386 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
387 defined or documented. It seems to have been a typo for PCRE_STATIC, so
388 I have changed it.
389
390 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
391 first named subpattern) and a construct such as (?&a) would reference the
392 first named subpattern whose name started with "a" (in other words, the
393 length check was missing). Both these problems are fixed. "Subpattern name
394 expected" is now given for (?&) (a zero-length name), and this patch also
395 makes it give the same error for \k'' (previously it complained that that
396 was a reference to a non-existent subpattern).
397
398 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
399 this is right because (?- can be followed by option settings as well as by
400 digits. I have, however, made the messages clearer.
401
402 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
403 than the number used in the conditional) now cause a compile-time error.
404 This is actually not compatible with Perl, which accepts such patterns, but
405 treats the conditional as always being FALSE (as PCRE used to), but it
406 seems to me that giving a diagnostic is better.
407
408 8. Change "alphameric" to the more common word "alphanumeric" in comments
409 and messages.
410
411 9. Fix two occurrences of "backslash" in comments that should have been
412 "backspace".
413
414 10. Remove two redundant lines of code that can never be obeyed (their function
415 was moved elsewhere).
416
417 11. The program that makes PCRE's Unicode character property table had a bug
418 which caused it to generate incorrect table entries for sequences of
419 characters that have the same character type, but are in different scripts.
420 It amalgamated them into a single range, with the script of the first of
421 them. In other words, some characters were in the wrong script. There were
422 thirteen such cases, affecting characters in the following ranges:
423
424 U+002b0 - U+002c1
425 U+0060c - U+0060d
426 U+0061e - U+00612
427 U+0064b - U+0065e
428 U+0074d - U+0076d
429 U+01800 - U+01805
430 U+01d00 - U+01d77
431 U+01d9b - U+01dbf
432 U+0200b - U+0200f
433 U+030fc - U+030fe
434 U+03260 - U+0327f
435 U+0fb46 - U+0fbb1
436 U+10450 - U+1049d
437
438 12. The -o option (show only the matching part of a line) for pcregrep was not
439 compatible with GNU grep in that, if there was more than one match in a
440 line, it showed only the first of them. It now behaves in the same way as
441 GNU grep.
442
443 13. If the -o and -v options were combined for pcregrep, it printed a blank
444 line for every non-matching line. GNU grep prints nothing, and pcregrep now
445 does the same. The return code can be used to tell if there were any
446 non-matching lines.
447
448 14. Added --file-offsets and --line-offsets to pcregrep.
449
450 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
451 infinitely looping recursion. The bug was that positive lookaheads were not
452 being skipped when checking for a possible empty match (negative lookaheads
453 and both kinds of lookbehind were skipped).
454
455 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
456 inclusion of <windows.h> to before rather than after the definition of
457 INVALID_FILE_ATTRIBUTES (patch from David Byron).
458
459 17. Specifying a possessive quantifier with a specific limit for a Unicode
460 character property caused pcre_compile() to compile bad code, which led at
461 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
462 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
463 caused the error; without that there was no problem.
464
465 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
466
467 19. Added --enable-pcretest-libreadline.
468
469 20. In pcrecpp.cc, the variable 'count' was incremented twice in
470 RE::GlobalReplace(). As a result, the number of replacements returned was
471 double what it should be. I removed one of the increments, but Craig sent a
472 later patch that removed the other one (the right fix) and added unit tests
473 that check the return values (which was not done before).
474
475 21. Several CMake things:
476
477 (1) Arranged that, when cmake is used on Unix, the libraries end up with
478 the names libpcre and libpcreposix, not just pcre and pcreposix.
479
480 (2) The above change means that pcretest and pcregrep are now correctly
481 linked with the newly-built libraries, not previously installed ones.
482
483 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
484
485 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
486 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
487 UTF-8 newline character). The key issue is that the pattern starts .*;
488 this means that the match must be either at the beginning, or after a
489 newline. The bug was in the code for advancing after a failed match and
490 checking that the new position followed a newline. It was not taking
491 account of UTF-8 characters correctly.
492
493 23. PCRE was behaving differently from Perl in the way it recognized POSIX
494 character classes. PCRE was not treating the sequence [:...:] as a
495 character class unless the ... were all letters. Perl, however, seems to
496 allow any characters between [: and :], though of course it rejects as
497 unknown any "names" that contain non-letters, because all the known class
498 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
499 for example, whereas PCRE did not - it did not recognize a POSIX character
500 class. This seemed a bit dangerous, so the code has been changed to be
501 closer to Perl. The behaviour is not identical to Perl, because PCRE will
502 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
503 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
504 Perl does, and where it didn't before.
505
506 24. Rewrite so as to remove the single use of %n from pcregrep because in some
507 Windows environments %n is disabled by default.
508
509
510 Version 7.4 21-Sep-07
511 ---------------------
512
513 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
514 means that a class such as [\s] counted as "explicit reference to CR or
515 LF". That isn't really right - the whole point of the change was to try to
516 help when there was an actual mention of one of the two characters. So now
517 the change happens only if \r or \n (or a literal CR or LF) character is
518 encountered.
519
520 2. The 32-bit options word was also used for 6 internal flags, but the numbers
521 of both had grown to the point where there were only 3 bits left.
522 Fortunately, there was spare space in the data structure, and so I have
523 moved the internal flags into a new 16-bit field to free up more option
524 bits.
525
526 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
527 but did not set the internal JCHANGED flag - either of these is enough to
528 control the way the "get" function works - but the PCRE_INFO_JCHANGED
529 facility is supposed to tell if (?J) was ever used, so now (?J) at the
530 start sets both bits.
531
532 4. Added options (at build time, compile time, exec time) to change \R from
533 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
534
535 5. doc/pcresyntax.html was missing from the distribution.
536
537 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
538 compatibility, even though it is no longer used.
539
540 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
541 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
542 windows.h file is present (where different names are used). [This was
543 reversed later after testing - see 16 below.]
544
545 8. Changed all #include <config.h> to #include "config.h". There were also
546 some further <pcre.h> cases that I changed to "pcre.h".
547
548 9. When pcregrep was used with the --colour option, it missed the line ending
549 sequence off the lines that it output.
550
551 10. It was pointed out to me that arrays of string pointers cause lots of
552 relocations when a shared library is dynamically loaded. A technique of
553 using a single long string with a table of offsets can drastically reduce
554 these. I have refactored PCRE in four places to do this. The result is
555 dramatic:
556
557 Originally: 290
558 After changing UCP table: 187
559 After changing error message table: 43
560 After changing table of "verbs" 36
561 After changing table of Posix names 22
562
563 Thanks to the folks working on Gregex for glib for this insight.
564
565 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
566 unicode-properties was also set.
567
568 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
569
570 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
571 checked only for CRLF.
572
573 14. Added casts to pcretest.c to avoid compiler warnings.
574
575 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
576
577 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
578 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
579 entirely. This removes changes made in 7 above.
580
581 17. The CMake files have been updated, and there is now more information about
582 building with CMake in the NON-UNIX-USE document.
583
584
585 Version 7.3 28-Aug-07
586 ---------------------
587
588 1. In the rejigging of the build system that eventually resulted in 7.1, the
589 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
590 brackets there is not right, since it causes compilers to look for an
591 installed pcre.h, not the version that is in the source that is being
592 compiled (which of course may be different). I have changed it back to:
593
594 #include "pcre.h"
595
596 I have a vague recollection that the change was concerned with compiling in
597 different directories, but in the new build system, that is taken care of
598 by the VPATH setting the Makefile.
599
600 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
601 when the subject happened to end in the byte 0x85 (e.g. if the last
602 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
603 characters but of course it shouldn't be taken as a newline when it is part
604 of another character. The bug was that, for an unlimited repeat of . in
605 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
606 characters when looking for a newline.
607
608 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
609
610 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
611 in debug output.
612
613 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
614 long printing in the pcrecpp unittest when running under MinGW.
615
616 6. ESC_K was left out of the EBCDIC table.
617
618 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
619 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
620 limit also applies to "virtual nesting" when a pattern is recursive, and in
621 this case 1000 isn't so big. I have been able to remove this limit at the
622 expense of backing off one optimization in certain circumstances. Normally,
623 when pcre_exec() would call its internal match() function recursively and
624 immediately return the result unconditionally, it uses a "tail recursion"
625 feature to save stack. However, when a subpattern that can match an empty
626 string has an unlimited repetition quantifier, it no longer makes this
627 optimization. That gives it a stack frame in which to save the data for
628 checking that an empty string has been matched. Previously this was taken
629 from the 1000-entry workspace that had been reserved. So now there is no
630 explicit limit, but more stack is used.
631
632 8. Applied Daniel's patches to solve problems with the import/export magic
633 syntax that is required for Windows, and which was going wrong for the
634 pcreposix and pcrecpp parts of the library. These were overlooked when this
635 problem was solved for the main library.
636
637 9. There were some crude static tests to avoid integer overflow when computing
638 the size of patterns that contain repeated groups with explicit upper
639 limits. As the maximum quantifier is 65535, the maximum group length was
640 set at 30,000 so that the product of these two numbers did not overflow a
641 32-bit integer. However, it turns out that people want to use groups that
642 are longer than 30,000 bytes (though not repeat them that many times).
643 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
644 made it possible to implement the integer overflow checks in a much more
645 dynamic way, which I have now done. The artificial limitation on group
646 length has been removed - we now have only the limit on the total length of
647 the compiled pattern, which depends on the LINK_SIZE setting.
648
649 10. Fixed a bug in the documentation for get/copy named substring when
650 duplicate names are permitted. If none of the named substrings are set, the
651 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
652 empty string.
653
654 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
655 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
656 because the ] is interpreted as the first data character and the
657 terminating ] is not found. PCRE has been made compatible with Perl in this
658 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
659 cause memory overwriting.
660
661 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
662 string has been matched (to stop an infinite loop). It was not recognizing
663 a conditional subpattern that could match an empty string if that
664 subpattern was within another subpattern. For example, it looped when
665 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
666 condition was not nested. This bug has been fixed.
667
668 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
669 past the start of the subject in the presence of bytes with the top bit
670 set, for example "\x8aBCD".
671
672 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
673 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
674
675 14. Optimized (?!) to (*FAIL).
676
677 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
678 This restricts code points to be within the range 0 to 0x10FFFF, excluding
679 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
680 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
681 does: it's just the validity check that is more restrictive.
682
683 16. Inserted checks for integer overflows during escape sequence (backslash)
684 processing, and also fixed erroneous offset values for syntax errors during
685 backslash processing.
686
687 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
688 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
689
690 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
691 caused an overrun.
692
693 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
694 something other than just ASCII characters) inside a group that had an
695 unlimited repeat caused a loop at compile time (while checking to see
696 whether the group could match an empty string).
697
698 20. Debugging a pattern containing \p or \P could cause a crash. For example,
699 [\P{Any}] did so. (Error in the code for printing property names.)
700
701 21. An orphan \E inside a character class could cause a crash.
702
703 22. A repeated capturing bracket such as (A)? could cause a wild memory
704 reference during compilation.
705
706 23. There are several functions in pcre_compile() that scan along a compiled
707 expression for various reasons (e.g. to see if it's fixed length for look
708 behind). There were bugs in these functions when a repeated \p or \P was
709 present in the pattern. These operators have additional parameters compared
710 with \d, etc, and these were not being taken into account when moving along
711 the compiled data. Specifically:
712
713 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
714 length.
715
716 (b) An item such as \pL+ within a repeated group could cause crashes or
717 loops.
718
719 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
720 "reference to non-existent subpattern" error.
721
722 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
723
724 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
725 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
726
727 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
728
729 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
730 character were causing crashes (broken optimization).
731
732 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
733 \p or \P) caused a compile-time loop.
734
735 28. More problems have arisen in unanchored patterns when CRLF is a valid line
736 break. For example, the unstudied pattern [\r\n]A does not match the string
737 "\r\nA" because change 7.0/46 below moves the current point on by two
738 characters after failing to match at the start. However, the pattern \nA
739 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
740 the same is true. There doesn't seem any very clean way out of this, but
741 what I have chosen to do makes the common cases work: PCRE now takes note
742 of whether there can be an explicit match for \r or \n anywhere in the
743 pattern, and if so, 7.0/46 no longer applies. As part of this change,
744 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
745 pattern has explicit CR or LF references.
746
747 29. Added (*CR) etc for changing newline setting at start of pattern.
748
749
750 Version 7.2 19-Jun-07
751 ---------------------
752
753 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
754 which is apparently normally available under Windows.
755
756 2. Re-jig the pcregrep tests with different newline settings in an attempt
757 to make them independent of the local environment's newline setting.
758
759 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
760
761 4. Some of the "internals" tests were previously cut out when the link size
762 was not 2, because the output contained actual offsets. The recent new
763 "Z" feature of pcretest means that these can be cut out, making the tests
764 usable with all link sizes.
765
766 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
767 stack recursion. This gives a massive performance boost under BSD, but just
768 a small improvement under Linux. However, it saves one field in the frame
769 in all cases.
770
771 6. Added more features from the forthcoming Perl 5.10:
772
773 (a) (?-n) (where n is a string of digits) is a relative subroutine or
774 recursion call. It refers to the nth most recently opened parentheses.
775
776 (b) (?+n) is also a relative subroutine call; it refers to the nth next
777 to be opened parentheses.
778
779 (c) Conditions that refer to capturing parentheses can be specified
780 relatively, for example, (?(-2)... or (?(+3)...
781
782 (d) \K resets the start of the current match so that everything before
783 is not part of it.
784
785 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
786
787 (f) \g{name} is another synonym - part of Perl 5.10's unification of
788 reference syntax.
789
790 (g) (?| introduces a group in which the numbering of parentheses in each
791 alternative starts with the same number.
792
793 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
794
795 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
796 PCRE_INFO_JCHANGED.
797
798 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
799 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
800 for detecting groups that can match an empty string.
801
802 9. A pattern with a very large number of alternatives (more than several
803 hundred) was running out of internal workspace during the pre-compile
804 phase, where pcre_compile() figures out how much memory will be needed. A
805 bit of new cunning has reduced the workspace needed for groups with
806 alternatives. The 1000-alternative test pattern now uses 12 bytes of
807 workspace instead of running out of the 4096 that are available.
808
809 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
810
811 11. Applied patch from Google to remove an optimization that didn't quite work.
812 The report of the bug said:
813
814 pcrecpp::RE("a*").FullMatch("aaa") matches, while
815 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
816 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
817
818 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
819 it matched the wrong number of bytes.
820
821
822 Version 7.1 24-Apr-07
823 ---------------------
824
825 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
826 that is more "standard", making use of automake and other Autotools. There
827 is some re-arrangement of the files and adjustment of comments consequent
828 on this.
829
830 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
831 for recursive directory scanning broke on some systems because the files
832 are not scanned in any specific order and on different systems the order
833 was different. A call to "sort" has been inserted into RunGrepTest for the
834 approprate test as a short-term fix. In the longer term there may be an
835 alternative.
836
837 3. I had an email from Eric Raymond about problems translating some of PCRE's
838 man pages to HTML (despite the fact that I distribute HTML pages, some
839 people do their own conversions for various reasons). The problems
840 concerned the use of low-level troff macros .br and .in. I have therefore
841 removed all such uses from the man pages (some were redundant, some could
842 be replaced by .nf/.fi pairs). The 132html script that I use to generate
843 HTML has been updated to handle .nf/.fi and to complain if it encounters
844 .br or .in.
845
846 4. Updated comments in configure.ac that get placed in config.h.in and also
847 arranged for config.h to be included in the distribution, with the name
848 config.h.generic, for the benefit of those who have to compile without
849 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
850
851 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
852 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
853 makevp.bat for the current PCRE, using the additional files
854 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
855
856 6. A Windows user reported a minor discrepancy with test 2, which turned out
857 to be caused by a trailing space on an input line that had got lost in his
858 copy. The trailing space was an accident, so I've just removed it.
859
860 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
861 that is needed.
862
863 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
864 as "const" (a) because they are and (b) because it helps the PHP
865 maintainers who have recently made a script to detect big data structures
866 in the php code that should be moved to the .rodata section. I remembered
867 to update Builducptable as well, so it won't revert if ucptable.h is ever
868 re-created.
869
870 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
871 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
872 order to be able to cut out the UTF-8 tables in the latter when UTF-8
873 support is not required. This saves 1.5-2K of code, which is important in
874 some applications.
875
876 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
877 so as not to refer to the tables, even though these functions will never be
878 called when UTF-8 support is disabled. Otherwise there are problems with a
879 shared library.
880
881 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
882
883 (a) It was defining its arguments as char * instead of void *.
884
885 (b) It was assuming that all moves were upwards in memory; this was true
886 a long time ago when I wrote it, but is no longer the case.
887
888 The emulated memove() is provided for those environments that have neither
889 memmove() nor bcopy(). I didn't think anyone used it these days, but that
890 is clearly not the case, as these two bugs were recently reported.
891
892 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
893 and Detrail to create the HTML documentation, the .txt form of the man
894 pages, and it removes trailing spaces from listed files. It also creates
895 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
896 case, it wraps all the #defines with #ifndefs. This script should be run
897 before "make dist".
898
899 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
900 with Unicode property support.
901
902 (a) For a maximizing quantifier, if the two different cases of the
903 character were of different lengths in their UTF-8 codings (there are
904 some cases like this - I found 11), and the matching function had to
905 back up over a mixture of the two cases, it incorrectly assumed they
906 were both the same length.
907
908 (b) When PCRE was configured to use the heap rather than the stack for
909 recursion during matching, it was not correctly preserving the data for
910 the other case of a UTF-8 character when checking ahead for a match
911 while processing a minimizing repeat. If the check also involved
912 matching a wide character, but failed, corruption could cause an
913 erroneous result when trying to check for a repeat of the original
914 character.
915
916 13. Some tidying changes to the testing mechanism:
917
918 (a) The RunTest script now detects the internal link size and whether there
919 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
920 values substituted by "configure". (The RunGrepTest script already did
921 this for UTF-8.) The configure.ac script no longer substitutes the
922 relevant variables.
923
924 (b) The debugging options /B and /D in pcretest show the compiled bytecode
925 with length and offset values. This means that the output is different
926 for different internal link sizes. Test 2 is skipped for link sizes
927 other than 2 because of this, bypassing the problem. Unfortunately,
928 there was also a test in test 3 (the locale tests) that used /B and
929 failed for link sizes other than 2. Rather than cut the whole test out,
930 I have added a new /Z option to pcretest that replaces the length and
931 offset values with spaces. This is now used to make test 3 independent
932 of link size. (Test 2 will be tidied up later.)
933
934 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
935 segmentation fault instead of returning the appropriate error message.
936
937 15. In multiline mode when the newline sequence was set to "any", the pattern
938 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
939 This doesn't seem right; it now treats the CRLF combination as the line
940 ending, and so does not match in that case. It's only a pattern such as ^$
941 that would hit this one: something like ^ABC$ would have failed after \r
942 and then tried again after \r\n.
943
944 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
945 in an attempt to make files that differ only in their line terminators
946 compare equal. This works on Linux.
947
948 17. Under certain error circumstances pcregrep might try to free random memory
949 as it exited. This is now fixed, thanks to valgrind.
950
951 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
952 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
953 was because its rules for how to advance for /g after matching an empty
954 string at the end of a line did not allow for this case. They now check for
955 it specially.
956
957 20. pcretest is supposed to handle patterns and data of any length, by
958 extending its buffers when necessary. It was getting this wrong when the
959 buffer for a data line had to be extended.
960
961 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
962 CRLF as a newline sequence.
963
964 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
965 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
966 I have nevertheless tidied it up.
967
968 23. Added some casts to kill warnings from HP-UX ia64 compiler.
969
970 24. Added a man page for pcre-config.
971
972
973 Version 7.0 19-Dec-06
974 ---------------------
975
976 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
977 moving to gcc 4.1.1.
978
979 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
980 sys/time.h, which is documented as needed for this function. It doesn't
981 seem to matter on Linux, but it showed up on some releases of OS X.
982
983 3. It seems that there are systems where bytes whose values are greater than
984 127 match isprint() in the "C" locale. The "C" locale should be the
985 default when a C program starts up. In most systems, only ASCII printing
986 characters match isprint(). This difference caused the output from pcretest
987 to vary, making some of the tests fail. I have changed pcretest so that:
988
989 (a) When it is outputting text in the compiled version of a pattern, bytes
990 other than 32-126 are always shown as hex escapes.
991
992 (b) When it is outputting text that is a matched part of a subject string,
993 it does the same, unless a different locale has been set for the match
994 (using the /L modifier). In this case, it uses isprint() to decide.
995
996 4. Fixed a major bug that caused incorrect computation of the amount of memory
997 required for a compiled pattern when options that changed within the
998 pattern affected the logic of the preliminary scan that determines the
999 length. The relevant options are -x, and -i in UTF-8 mode. The result was
1000 that the computed length was too small. The symptoms of this bug were
1001 either the PCRE error "internal error: code overflow" from pcre_compile(),
1002 or a glibc crash with a message such as "pcretest: free(): invalid next
1003 size (fast)". Examples of patterns that provoked this bug (shown in
1004 pcretest format) are:
1005
1006 /(?-x: )/x
1007 /(?x)(?-x: \s*#\s*)/
1008 /((?i)[\x{c0}])/8
1009 /(?i:[\x{c0}])/8
1010
1011 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
1012 is now done differently.
1013
1014 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
1015 wrapper classes; (b) implement a new function in the C++ scanner that is
1016 more efficient than the old way of doing things because it avoids levels of
1017 recursion in the regex matching; (c) add a paragraph to the documentation
1018 for the FullMatch() function.
1019
1020 6. The escape sequence \n was being treated as whatever was defined as
1021 "newline". Not only was this contrary to the documentation, which states
1022 that \n is character 10 (hex 0A), but it also went horribly wrong when
1023 "newline" was defined as CRLF. This has been fixed.
1024
1025 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
1026 was being set to -1 for the "end of line" case (supposedly a value that no
1027 character can have). Though this value is never used (the check for end of
1028 line is "zero bytes in current character"), it caused compiler complaints.
1029 I've changed it to 0xffffffff.
1030
1031 8. In pcre_version.c, the version string was being built by a sequence of
1032 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
1033 string (as it is for production releases) called a macro with an empty
1034 argument. The C standard says the result of this is undefined. The gcc
1035 compiler treats it as an empty string (which was what was wanted) but it is
1036 reported that Visual C gives an error. The source has been hacked around to
1037 avoid this problem.
1038
1039 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
1040 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
1041 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
1042 of them did).
1043
1044 10. Originally, pcretest opened its input and output without "b"; then I was
1045 told that "b" was needed in some environments, so it was added for release
1046 5.0 to both the input and output. (It makes no difference on Unix-like
1047 systems.) Later I was told that it is wrong for the input on Windows. I've
1048 now abstracted the modes into two macros, to make it easier to fiddle with
1049 them, and removed "b" from the input mode under Windows.
1050
1051 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
1052
1053 12. Added -help and --help to pcretest as an official way of being reminded
1054 of the options.
1055
1056 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
1057 and pcrecpp.cc because they annoy compilers at high warning levels.
1058
1059 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
1060
1061 15. Fixed an occurrence of == in configure.ac that should have been = (shell
1062 scripts are not C programs :-) and which was not noticed because it works
1063 on Linux.
1064
1065 16. pcretest is supposed to handle any length of pattern and data line (as one
1066 line or as a continued sequence of lines) by extending its input buffer if
1067 necessary. This feature was broken for very long pattern lines, leading to
1068 a string of junk being passed to pcre_compile() if the pattern was longer
1069 than about 50K.
1070
1071 17. I have done a major re-factoring of the way pcre_compile() computes the
1072 amount of memory needed for a compiled pattern. Previously, there was code
1073 that made a preliminary scan of the pattern in order to do this. That was
1074 OK when PCRE was new, but as the facilities have expanded, it has become
1075 harder and harder to keep it in step with the real compile phase, and there
1076 have been a number of bugs (see for example, 4 above). I have now found a
1077 cunning way of running the real compile function in a "fake" mode that
1078 enables it to compute how much memory it would need, while actually only
1079 ever using a few hundred bytes of working memory and without too many
1080 tests of the mode. This should make future maintenance and development
1081 easier. A side effect of this work is that the limit of 200 on the nesting
1082 depth of parentheses has been removed (though this was never a serious
1083 limitation, I suspect). However, there is a downside: pcre_compile() now
1084 runs more slowly than before (30% or more, depending on the pattern). I
1085 hope this isn't a big issue. There is no effect on runtime performance.
1086
1087 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
1088 newline (only possible for the last line of a file) and it was a
1089 pattern that set a locale (followed by /Lsomething), pcretest crashed.
1090
1091 19. Added additional timing features to pcretest. (1) The -tm option now times
1092 matching only, not compiling. (2) Both -t and -tm can be followed, as a
1093 separate command line item, by a number that specifies the number of
1094 repeats to use when timing. The default is 50000; this gives better
1095 precision, but takes uncomfortably long for very large patterns.
1096
1097 20. Extended pcre_study() to be more clever in cases where a branch of a
1098 subpattern has no definite first character. For example, (a*|b*)[cd] would
1099 previously give no result from pcre_study(). Now it recognizes that the
1100 first character must be a, b, c, or d.
1101
1102 21. There was an incorrect error "recursive call could loop indefinitely" if
1103 a subpattern (or the entire pattern) that was being tested for matching an
1104 empty string contained only one non-empty item after a nested subpattern.
1105 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
1106 incorrectly, because the \d was being skipped in the check.
1107
1108 22. The pcretest program now has a new pattern option /B and a command line
1109 option -b, which is equivalent to adding /B to every pattern. This causes
1110 it to show the compiled bytecode, without the additional information that
1111 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
1112 is the same as /B/I).
1113
1114 23. A new optimization is now able automatically to treat some sequences such
1115 as a*b as a*+b. More specifically, if something simple (such as a character
1116 or a simple class like \d) has an unlimited quantifier, and is followed by
1117 something that cannot possibly match the quantified thing, the quantifier
1118 is automatically "possessified".
1119
1120 24. A recursive reference to a subpattern whose number was greater than 39
1121 went wrong under certain circumstances in UTF-8 mode. This bug could also
1122 have affected the operation of pcre_study().
1123
1124 25. Realized that a little bit of performance could be had by replacing
1125 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
1126
1127 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
1128
1129 27. Possessive quantifiers such as a++ were previously implemented by turning
1130 them into atomic groups such as ($>a+). Now they have their own opcodes,
1131 which improves performance. This includes the automatically created ones
1132 from 23 above.
1133
1134 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
1135 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
1136 the first matched character to be a colon. This applied both to named and
1137 numbered groups.
1138
1139 29. The ucpinternal.h header file was missing its idempotency #ifdef.
1140
1141 30. I was sent a "project" file called libpcre.a.dev which I understand makes
1142 building PCRE on Windows easier, so I have included it in the distribution.
1143
1144 31. There is now a check in pcretest against a ridiculously large number being
1145 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
1146 loop, the loop is abandoned.
1147
1148 32. Forward references to subpatterns in conditions such as (?(2)...) where
1149 subpattern 2 is defined later cause pcre_compile() to search forwards in
1150 the pattern for the relevant set of parentheses. This search went wrong
1151 when there were unescaped parentheses in a character class, parentheses
1152 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
1153
1154 33. "Subroutine" calls and backreferences were previously restricted to
1155 referencing subpatterns earlier in the regex. This restriction has now
1156 been removed.
1157
1158 34. Added a number of extra features that are going to be in Perl 5.10. On the
1159 whole, these are just syntactic alternatives for features that PCRE had
1160 previously implemented using the Python syntax or my own invention. The
1161 other formats are all retained for compatibility.
1162
1163 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
1164 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
1165 also .NET compatible.
1166
1167 (b) A recursion or subroutine call to a named group can now be defined as
1168 (?&name) as well as (?P>name).
1169
1170 (c) A backreference to a named group can now be defined as \k<name> or
1171 \k'name' as well as (?P=name). The new forms, as well as being in Perl
1172 5.10, are also .NET compatible.
1173
1174 (d) A conditional reference to a named group can now use the syntax
1175 (?(<name>) or (?('name') as well as (?(name).
1176
1177 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
1178 groups (named and numbered) that are never evaluated inline, but can be
1179 called as "subroutines" from elsewhere. In effect, the DEFINE condition
1180 is always false. There may be only one alternative in such a group.
1181
1182 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
1183 as the simple (?(R). The condition is true only if the most recent
1184 recursion is that of the given number or name. It does not search out
1185 through the entire recursion stack.
1186
1187 (g) The escape \gN or \g{N} has been added, where N is a positive or
1188 negative number, specifying an absolute or relative reference.
1189
1190 35. Tidied to get rid of some further signed/unsigned compiler warnings and
1191 some "unreachable code" warnings.
1192
1193 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
1194 things, this adds five new scripts.
1195
1196 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
1197 There were also incompatibilities regarding the handling of \Q..\E inside
1198 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
1199 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
1200
1201 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
1202 matches an empty string, and forcibly breaks the loop. There were bugs in
1203 this code in non-simple cases. For a pattern such as ^(a()*)* matched
1204 against aaaa the result was just "a" rather than "aaaa", for example. Two
1205 separate and independent bugs (that affected different cases) have been
1206 fixed.
1207
1208 39. Refactored the code to abolish the use of different opcodes for small
1209 capturing bracket numbers. This is a tidy that I avoided doing when I
1210 removed the limit on the number of capturing brackets for 3.5 back in 2001.
1211 The new approach is not only tidier, it makes it possible to reduce the
1212 memory needed to fix the previous bug (38).
1213
1214 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
1215 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
1216 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
1217 mode.
1218
1219 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
1220 report.
1221
1222 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
1223 copy construction and assignment in the C++ wrapper.
1224
1225 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
1226 couple of bugs that could have given wrong results in the "--newline=crlf"
1227 case.
1228
1229 44. Added a number of casts and did some reorganization of signed/unsigned int
1230 variables following suggestions from Dair Grant. Also renamed the variable
1231 "this" as "item" because it is a C++ keyword.
1232
1233 45. Arranged for dftables to add
1234
1235 #include "pcre_internal.h"
1236
1237 to pcre_chartables.c because without it, gcc 4.x may remove the array
1238 definition from the final binary if PCRE is built into a static library and
1239 dead code stripping is activated.
1240
1241 46. For an unanchored pattern, if a match attempt fails at the start of a
1242 newline sequence, and the newline setting is CRLF or ANY, and the next two
1243 characters are CRLF, advance by two characters instead of one.
1244
1245
1246 Version 6.7 04-Jul-06
1247 ---------------------
1248
1249 1. In order to handle tests when input lines are enormously long, pcretest has
1250 been re-factored so that it automatically extends its buffers when
1251 necessary. The code is crude, but this _is_ just a test program. The
1252 default size has been increased from 32K to 50K.
1253
1254 2. The code in pcre_study() was using the value of the re argument before
1255 testing it for NULL. (Of course, in any sensible call of the function, it
1256 won't be NULL.)
1257
1258 3. The memmove() emulation function in pcre_internal.h, which is used on
1259 systems that lack both memmove() and bcopy() - that is, hardly ever -
1260 was missing a "static" storage class specifier.
1261
1262 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
1263 containing an extended class (one that cannot be represented by a bitmap
1264 because it contains high-valued characters or Unicode property items, e.g.
1265 [\pZ]). Almost always one would set UTF-8 mode when processing such a
1266 pattern, but PCRE should not loop if you do not (it no longer does).
1267 [Detail: two cases were found: (a) a repeated subpattern containing an
1268 extended class; (b) a recursive reference to a subpattern that followed a
1269 previous extended class. It wasn't skipping over the extended class
1270 correctly when UTF-8 mode was not set.]
1271
1272 5. A negated single-character class was not being recognized as fixed-length
1273 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
1274 compile error "lookbehind assertion is not fixed length".
1275
1276 6. The RunPerlTest auxiliary script was showing an unexpected difference
1277 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
1278 write a Perl script that can interpret lines of an input file either as
1279 byte characters or as UTF-8, which is what "perltest" was being required to
1280 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
1281 can't do is switch easily at run time between having the "use utf8;" pragma
1282 or not. In the end, I fudged it by using the RunPerlTest script to insert
1283 "use utf8;" explicitly for the UTF-8 tests.
1284
1285 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
1286 the end of the subject string, contrary to the documentation and to what
1287 Perl does. This was true of both matching functions. Now it matches only at
1288 the start of the subject and immediately after *internal* newlines.
1289
1290 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
1291 a pointer to an int instead of a pointer to an unsigned long int. This
1292 caused problems on 64-bit systems.
1293
1294 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
1295 instance of the 'standard' template library not being so standard".
1296
1297 10. There was no check on the number of named subpatterns nor the maximum
1298 length of a subpattern name. The product of these values is used to compute
1299 the size of the memory block for a compiled pattern. By supplying a very
1300 long subpattern name and a large number of named subpatterns, the size
1301 computation could be caused to overflow. This is now prevented by limiting
1302 the length of names to 32 characters, and the number of named subpatterns
1303 to 10,000.
1304
1305 11. Subpatterns that are repeated with specific counts have to be replicated in
1306 the compiled pattern. The size of memory for this was computed from the
1307 length of the subpattern and the repeat count. The latter is limited to
1308 65535, but there was no limit on the former, meaning that integer overflow
1309 could in principle occur. The compiled length of a repeated subpattern is
1310 now limited to 30,000 bytes in order to prevent this.
1311
1312 12. Added the optional facility to have named substrings with the same name.
1313
1314 13. Added the ability to use a named substring as a condition, using the
1315 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
1316 are numbers (not recommended). Forward references are permitted.
1317
1318 14. Added forward references in named backreferences (if you see what I mean).
1319
1320 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
1321 pattern could run off the end of the subject. For example, the pattern
1322 "(?s)(.{1,5})"8 did this with the subject "ab".
1323
1324 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
1325 PCRE_CASELESS was set when matching characters that were quantified with ?
1326 or *.
1327
1328 17. A character class other than a single negated character that had a minimum
1329 but no maximum quantifier - for example [ab]{6,} - was not handled
1330 correctly by pce_dfa_exec(). It would match only one character.
1331
1332 18. A valid (though odd) pattern that looked like a POSIX character
1333 class but used an invalid character after [ (for example [[,abc,]]) caused
1334 pcre_compile() to give the error "Failed: internal error: code overflow" or
1335 in some cases to crash with a glibc free() error. This could even happen if
1336 the pattern terminated after [[ but there just happened to be a sequence of
1337 letters, a binary zero, and a closing ] in the memory that followed.
1338
1339 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
1340 over the years. Originally (before any Unicode support), just the bottom 8
1341 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
1342 output from "man perlunicode" includes this:
1343
1344 The regular expression compiler produces polymorphic opcodes. That
1345 is, the pattern adapts to the data and automatically switches to
1346 the Unicode character scheme when presented with Unicode data--or
1347 instead uses a traditional byte scheme when presented with byte
1348 data.
1349
1350 Sadly, a wide octal escape does not cause a switch, and in a string with
1351 no other multibyte characters, these octal escapes are treated as before.
1352 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
1353 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
1354 Unicode string.
1355
1356 I have not perpetrated such confusion in PCRE. Up till now, it took just
1357 the bottom 8 bits, as in old Perl. I have now made octal escapes with
1358 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
1359 translate to the appropriate multibyte character.
1360
1361 29. Applied some refactoring to reduce the number of warnings from Microsoft
1362 and Borland compilers. This has included removing the fudge introduced
1363 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
1364 a warning about an unused variable.
1365
1366 21. PCRE has not included VT (character 0x0b) in the set of whitespace
1367 characters since release 4.0, because Perl (from release 5.004) does not.
1368 [Or at least, is documented not to: some releases seem to be in conflict
1369 with the documentation.] However, when a pattern was studied with
1370 pcre_study() and all its branches started with \s, PCRE still included VT
1371 as a possible starting character. Of course, this did no harm; it just
1372 caused an unnecessary match attempt.
1373
1374 22. Removed a now-redundant internal flag bit that recorded the fact that case
1375 dependency changed within the pattern. This was once needed for "required
1376 byte" processing, but is no longer used. This recovers a now-scarce options
1377 bit. Also moved the least significant internal flag bit to the most-
1378 significant bit of the word, which was not previously used (hangover from
1379 the days when it was an int rather than a uint) to free up another bit for
1380 the future.
1381
1382 23. Added support for CRLF line endings as well as CR and LF. As well as the
1383 default being selectable at build time, it can now be changed at runtime
1384 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
1385 specify that it is scanning data with non-default line endings.
1386
1387 24. Changed the definition of CXXLINK to make it agree with the definition of
1388 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
1389
1390 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
1391 recursions. This makes a big different to stack usage for some patterns.
1392
1393 26. If a subpattern containing a named recursion or subroutine reference such
1394 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
1395 the space required for the compiled pattern went wrong and gave too small a
1396 value. Depending on the environment, this could lead to "Failed: internal
1397 error: code overflow at offset 49" or "glibc detected double free or
1398 corruption" errors.
1399
1400 27. Applied patches from Google (a) to support the new newline modes and (b) to
1401 advance over multibyte UTF-8 characters in GlobalReplace.
1402
1403 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
1404 difference for some implementation of PCRE in some Windows version.
1405
1406 29. Added some extra testing facilities to pcretest:
1407
1408 \q<number> in a data line sets the "match limit" value
1409 \Q<number> in a data line sets the "match recursion limt" value
1410 -S <number> sets the stack size, where <number> is in megabytes
1411
1412 The -S option isn't available for Windows.
1413
1414
1415 Version 6.6 06-Feb-06
1416 ---------------------
1417
1418 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
1419 in pcreposix.h. I have copied the definition from pcre.h.
1420
1421 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
1422 because pcre.h is no longer a built file.
1423
1424 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
1425 not normally included in the compiled code.
1426
1427
1428 Version 6.5 01-Feb-06
1429 ---------------------
1430
1431 1. When using the partial match feature with pcre_dfa_exec(), it was not
1432 anchoring the second and subsequent partial matches at the new starting
1433 point. This could lead to incorrect results. For example, with the pattern
1434 /1234/, partially matching against "123" and then "a4" gave a match.
1435
1436 2. Changes to pcregrep:
1437
1438 (a) All non-match returns from pcre_exec() were being treated as failures
1439 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
1440 error message is output. Some extra information is given for the
1441 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
1442 probably the only errors that are likely to be caused by users (by
1443 specifying a regex that has nested indefinite repeats, for instance).
1444 If there are more than 20 of these errors, pcregrep is abandoned.
1445
1446 (b) A binary zero was treated as data while matching, but terminated the
1447 output line if it was written out. This has been fixed: binary zeroes
1448 are now no different to any other data bytes.
1449
1450 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
1451 used to set a locale for matching. The --locale=xxxx long option has
1452 been added (no short equivalent) to specify a locale explicitly on the
1453 pcregrep command, overriding the environment variables.
1454
1455 (d) When -B was used with -n, some line numbers in the output were one less
1456 than they should have been.
1457
1458 (e) Added the -o (--only-matching) option.
1459
1460 (f) If -A or -C was used with -c (count only), some lines of context were
1461 accidentally printed for the final match.
1462
1463 (g) Added the -H (--with-filename) option.
1464
1465 (h) The combination of options -rh failed to suppress file names for files
1466 that were found from directory arguments.
1467
1468 (i) Added the -D (--devices) and -d (--directories) options.
1469
1470 (j) Added the -F (--fixed-strings) option.
1471
1472 (k) Allow "-" to be used as a file name for -f as well as for a data file.
1473
1474 (l) Added the --colo(u)r option.
1475
1476 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
1477 is not present by default.
1478
1479 3. A nasty bug was discovered in the handling of recursive patterns, that is,
1480 items such as (?R) or (?1), when the recursion could match a number of
1481 alternatives. If it matched one of the alternatives, but subsequently,
1482 outside the recursion, there was a failure, the code tried to back up into
1483 the recursion. However, because of the way PCRE is implemented, this is not
1484 possible, and the result was an incorrect result from the match.
1485
1486 In order to prevent this happening, the specification of recursion has
1487 been changed so that all such subpatterns are automatically treated as
1488 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
1489
1490 4. I had overlooked the fact that, in some locales, there are characters for
1491 which isalpha() is true but neither isupper() nor islower() are true. In
1492 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
1493 and ordfeminine) are like this. This affected the treatment of \w and \W
1494 when they appeared in character classes, but not when they appeared outside
1495 a character class. The bit map for "word" characters is now created
1496 separately from the results of isalnum() instead of just taking it from the
1497 upper, lower, and digit maps. (Plus the underscore character, of course.)
1498
1499 5. The above bug also affected the handling of POSIX character classes such as
1500 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
1501 permanent tables. Instead, the bit maps for such a class were previously
1502 created as the appropriate unions of the upper, lower, and digit bitmaps.
1503 Now they are created by subtraction from the [[:word:]] class, which has
1504 its own bitmap.
1505
1506 6. The [[:blank:]] character class matches horizontal, but not vertical space.
1507 It is created by subtracting the vertical space characters (\x09, \x0a,
1508 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
1509 subtraction was done in the overall bitmap for a character class, meaning
1510 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
1511 be recognized. This bug has been fixed.
1512
1513 7. Patches from the folks at Google:
1514
1515 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
1516 real life, but is still worth protecting against".
1517
1518 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
1519 regular expressions".
1520
1521 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
1522 have it.
1523
1524 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
1525 "configure" and the latter not, in order to fix a problem somebody had
1526 with compiling the Arg class on HP-UX.
1527
1528 (e) Improve the error-handling of the C++ wrapper a little bit.
1529
1530 (f) New tests for checking recursion limiting.
1531
1532 8. The pcre_memmove() function, which is used only if the environment does not
1533 have a standard memmove() function (and is therefore rarely compiled),
1534 contained two bugs: (a) use of int instead of size_t, and (b) it was not
1535 returning a result (though PCRE never actually uses the result).
1536
1537 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
1538 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
1539 returned instead of calling malloc() with an overflowing number that would
1540 most likely cause subsequent chaos.
1541
1542 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
1543
1544 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
1545 with this option is matched, the nmatch and pmatch options of regexec() are
1546 ignored.
1547
1548 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
1549 provided in case anyone wants to the the POSIX interface with UTF-8
1550 strings.
1551
1552 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
1553 C++ linking (needed for some HP-UX environments).
1554
1555 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
1556 (unused parameter) and in the pcre_printint() function (omitted "default"
1557 switch label when the default is to do nothing).
1558
1559 15. Added some code to make it possible, when PCRE is compiled as a C++
1560 library, to replace subject pointers for pcre_exec() with a smart pointer
1561 class, thus making it possible to process discontinuous strings.
1562
1563 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
1564 much the same function. They were added by different people who were trying
1565 to make PCRE easy to compile on non-Unix systems. It has been suggested
1566 that PCRE_EXPORT be abolished now that there is more automatic apparatus
1567 for compiling on Windows systems. I have therefore replaced it with
1568 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
1569 defaults to "extern" for C or "extern C" for C++, which works fine on
1570 Unix-like systems. It is now possible to override the value of PCRE_DATA_
1571 SCOPE with something explicit in config.h. In addition:
1572
1573 (a) pcreposix.h still had just "extern" instead of either of these macros;
1574 I have replaced it with PCRE_DATA_SCOPE.
1575
1576 (b) Functions such as _pcre_xclass(), which are internal to the library,
1577 but external in the C sense, all had PCRE_EXPORT in their definitions.
1578 This is apparently wrong for the Windows case, so I have removed it.
1579 (It makes no difference on Unix-like systems.)
1580
1581 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
1582 of recursive calls to match(). This is different to MATCH_LIMIT because
1583 that limits the total number of calls to match(), not all of which increase
1584 the depth of recursion. Limiting the recursion depth limits the amount of
1585 stack (or heap if NO_RECURSE is set) that is used. The default can be set
1586 when PCRE is compiled, and changed at run time. A patch from Google adds
1587 this functionality to the C++ interface.
1588
1589 18. Changes to the handling of Unicode character properties:
1590
1591 (a) Updated the table to Unicode 4.1.0.
1592
1593 (b) Recognize characters that are not in the table as "Cn" (undefined).
1594
1595 (c) I revised the way the table is implemented to a much improved format
1596 which includes recognition of ranges. It now supports the ranges that
1597 are defined in UnicodeData.txt, and it also amalgamates other
1598 characters into ranges. This has reduced the number of entries in the
1599 table from around 16,000 to around 3,000, thus reducing its size
1600 considerably. I realized I did not need to use a tree structure after
1601 all - a binary chop search is just as efficient. Having reduced the
1602 number of entries, I extended their size from 6 bytes to 8 bytes to
1603 allow for more data.
1604
1605 (d) Added support for Unicode script names via properties such as \p{Han}.
1606
1607 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
1608 matching that character.
1609
1610 20. When matching a repeated Unicode property with a minimum greater than zero,
1611 (for example \pL{2,}), PCRE could look past the end of the subject if it
1612 reached it while seeking the minimum number of characters. This could
1613 happen only if some of the characters were more than one byte long, because
1614 there is a check for at least the minimum number of bytes.
1615
1616 21. Refactored the implementation of \p and \P so as to be more general, to
1617 allow for more different types of property in future. This has changed the
1618 compiled form incompatibly. Anybody with saved compiled patterns that use
1619 \p or \P will have to recompile them.
1620
1621 22. Added "Any" and "L&" to the supported property types.
1622
1623 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
1624 but give a compile time error if the value is greater than 0xff.
1625
1626 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
1627 accidentally not being installed or uninstalled.
1628
1629 25. The pcre.h file was built from pcre.h.in, but the only changes that were
1630 made were to insert the current release number. This seemed silly, because
1631 it made things harder for people building PCRE on systems that don't run
1632 "configure". I have turned pcre.h into a distributed file, no longer built
1633 by "configure", with the version identification directly included. There is
1634 no longer a pcre.h.in file.
1635
1636 However, this change necessitated a change to the pcre-config script as
1637 well. It is built from pcre-config.in, and one of the substitutions was the
1638 release number. I have updated configure.ac so that ./configure now finds
1639 the release number by grepping pcre.h.
1640
1641 26. Added the ability to run the tests under valgrind.
1642
1643
1644 Version 6.4 05-Sep-05
1645 ---------------------
1646
1647 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
1648 "--" to be printed when multiple files were scanned, even when none of the
1649 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
1650 consider it to be a bug, and have restored the previous behaviour.
1651
1652 2. A couple of code tidies to get rid of compiler warnings.
1653
1654 3. The pcretest program used to cheat by referring to symbols in the library
1655 whose names begin with _pcre_. These are internal symbols that are not
1656 really supposed to be visible externally, and in some environments it is
1657 possible to suppress them. The cheating is now confined to including
1658 certain files from the library's source, which is a bit cleaner.
1659
1660 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
1661 file's purpose clearer.
1662
1663 5. Reorganized pcre_ucp_findchar().
1664
1665
1666 Version 6.3 15-Aug-05
1667 ---------------------
1668
1669 1. The file libpcre.pc.in did not have general read permission in the tarball.
1670
1671 2. There were some problems when building without C++ support:
1672
1673 (a) If C++ support was not built, "make install" and "make test" still
1674 tried to test it.
1675
1676 (b) There were problems when the value of CXX was explicitly set. Some
1677 changes have been made to try to fix these, and ...
1678
1679 (c) --disable-cpp can now be used to explicitly disable C++ support.
1680
1681 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
1682 backslash in a target when C++ was disabled. This confuses some
1683 versions of "make", apparently. Using an intermediate variable solves
1684 this. (Same for CPP_LOBJ.)
1685
1686 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
1687 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
1688 necessary on certain architectures.
1689
1690 4. Added a setting of -export-symbols-regex to the link command to remove
1691 those symbols that are exported in the C sense, but actually are local
1692 within the library, and not documented. Their names all begin with
1693 "_pcre_". This is not a perfect job, because (a) we have to except some
1694 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
1695 available (and never for static libraries). I have made a note to try to
1696 find a way round (a) in the future.
1697
1698
1699 Version 6.2 01-Aug-05
1700 ---------------------
1701
1702 1. There was no test for integer overflow of quantifier values. A construction
1703 such as {1111111111111111} would give undefined results. What is worse, if
1704 a minimum quantifier for a parenthesized subpattern overflowed and became
1705 negative, the calculation of the memory size went wrong. This could have
1706 led to memory overwriting.
1707
1708 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
1709
1710 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
1711 operating environments where this matters.
1712
1713 4. Applied Giuseppe Maxia's patch to add additional features for controlling
1714 PCRE options from within the C++ wrapper.
1715
1716 5. Named capturing subpatterns were not being correctly counted when a pattern
1717 was compiled. This caused two problems: (a) If there were more than 100
1718 such subpatterns, the calculation of the memory needed for the whole
1719 compiled pattern went wrong, leading to an overflow error. (b) Numerical
1720 back references of the form \12, where the number was greater than 9, were
1721 not recognized as back references, even though there were sufficient
1722 previous subpatterns.
1723
1724 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
1725 versions of gcc, e.g. 2.95.4.
1726
1727
1728 Version 6.1 21-Jun-05
1729 ---------------------
1730
1731 1. There was one reference to the variable "posix" in pcretest.c that was not
1732 surrounded by "#if !defined NOPOSIX".
1733
1734 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
1735 the cross-check on the old pcre_info() function, for the benefit of the
1736 cut-down version of PCRE that is currently imported into Exim.
1737
1738 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
1739 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
1740 patterns that start (?i) but I don't think that matters. The use of (?i) is
1741 just an example; this all applies to the other options as well.
1742
1743 4. Since libtool seems to echo the compile commands it is issuing, the output
1744 from "make" can be reduced a bit by putting "@" in front of each libtool
1745 compile command.
1746
1747 5. Patch from the folks at Google for configure.in to be a bit more thorough
1748 in checking for a suitable C++ installation before trying to compile the
1749 C++ stuff. This should fix a reported problem when a compiler was present,
1750 but no suitable headers.
1751
1752 6. The man pages all had just "PCRE" as their title. I have changed them to
1753 be the relevant file name. I have also arranged that these names are
1754 retained in the file doc/pcre.txt, which is a concatenation in text format
1755 of all the man pages except the little individual ones for each function.
1756
1757 7. The NON-UNIX-USE file had not been updated for the different set of source
1758 files that come with release 6. I also added a few comments about the C++
1759 wrapper.
1760
1761
1762 Version 6.0 07-Jun-05
1763 ---------------------
1764
1765 1. Some minor internal re-organization to help with my DFA experiments.
1766
1767 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
1768 didn't matter for the library itself when fully configured, but did matter
1769 when compiling without UCP support, or within Exim, where the ucp files are
1770 not imported.
1771
1772 3. Refactoring of the library code to split up the various functions into
1773 different source modules. The addition of the new DFA matching code (see
1774 below) to a single monolithic source would have made it really too
1775 unwieldy, quite apart from causing all the code to be include in a
1776 statically linked application, when only some functions are used. This is
1777 relevant even without the DFA addition now that patterns can be compiled in
1778 one application and matched in another.
1779
1780 The downside of splitting up is that there have to be some external
1781 functions and data tables that are used internally in different modules of
1782 the library but which are not part of the API. These have all had their
1783 names changed to start with "_pcre_" so that they are unlikely to clash
1784 with other external names.
1785
1786 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
1787 a different (DFA) algorithm. Although it is slower than the original
1788 function, it does have some advantages for certain types of matching
1789 problem.
1790
1791 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
1792 including restarting after a partial match.
1793
1794 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
1795 defined when compiling for Windows was sent to me. I have put it into the
1796 code, though I have no means of testing or verifying it.
1797
1798 7. Added the pcre_refcount() auxiliary function.
1799
1800 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
1801 match before or at the first newline in the subject string. In pcretest,
1802 the /f option on a pattern can be used to set this.
1803
1804 9. A repeated \w when used in UTF-8 mode with characters greater than 256
1805 would behave wrongly. This has been present in PCRE since release 4.0.
1806
1807 10. A number of changes to the pcregrep command:
1808
1809 (a) Refactored how -x works; insert ^(...)$ instead of setting
1810 PCRE_ANCHORED and checking the length, in preparation for adding
1811 something similar for -w.
1812
1813 (b) Added the -w (match as a word) option.
1814
1815 (c) Refactored the way lines are read and buffered so as to have more
1816 than one at a time available.
1817
1818 (d) Implemented a pcregrep test script.
1819
1820 (e) Added the -M (multiline match) option. This allows patterns to match
1821 over several lines of the subject. The buffering ensures that at least
1822 8K, or the rest of the document (whichever is the shorter) is available
1823 for matching (and similarly the previous 8K for lookbehind assertions).
1824
1825 (f) Changed the --help output so that it now says
1826
1827 -w, --word-regex(p)
1828
1829 instead of two lines, one with "regex" and the other with "regexp"
1830 because that confused at least one person since the short forms are the
1831 same. (This required a bit of code, as the output is generated
1832 automatically from a table. It wasn't just a text change.)
1833
1834 (g) -- can be used to terminate pcregrep options if the next thing isn't an
1835 option but starts with a hyphen. Could be a pattern or a path name
1836 starting with a hyphen, for instance.
1837
1838 (h) "-" can be given as a file name to represent stdin.
1839
1840 (i) When file names are being printed, "(standard input)" is used for
1841 the standard input, for compatibility with GNU grep. Previously
1842 "<stdin>" was used.
1843
1844 (j) The option --label=xxx can be used to supply a name to be used for
1845 stdin when file names are being printed. There is no short form.
1846
1847 (k) Re-factored the options decoding logic because we are going to add
1848 two more options that take data. Such options can now be given in four
1849 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
1850
1851 (l) Added the -A, -B, and -C options for requesting that lines of context
1852 around matches be printed.
1853
1854 (m) Added the -L option to print the names of files that do not contain
1855 any matching lines, that is, the complement of -l.
1856
1857 (n) The return code is 2 if any file cannot be opened, but pcregrep does
1858 continue to scan other files.
1859
1860 (o) The -s option was incorrectly implemented. For compatibility with other
1861 greps, it now suppresses the error message for a non-existent or non-
1862 accessible file (but not the return code). There is a new option called
1863 -q that suppresses the output of matching lines, which was what -s was
1864 previously doing.
1865
1866 (p) Added --include and --exclude options to specify files for inclusion
1867 and exclusion when recursing.
1868
1869 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
1870 Hopefully, it now does.
1871
1872 12. Missing cast in pcre_study().
1873
1874 13. Added an "uninstall" target to the makefile.
1875
1876 14. Replaced "extern" in the function prototypes in Makefile.in with
1877 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
1878 world, but is set differently for Windows.
1879
1880 15. Added a second compiling function called pcre_compile2(). The only
1881 difference is that it has an extra argument, which is a pointer to an
1882 integer error code. When there is a compile-time failure, this is set
1883 non-zero, in addition to the error test pointer being set to point to an
1884 error message. The new argument may be NULL if no error number is required
1885 (but then you may as well call pcre_compile(), which is now just a
1886 wrapper). This facility is provided because some applications need a
1887 numeric error indication, but it has also enabled me to tidy up the way
1888 compile-time errors are handled in the POSIX wrapper.
1889
1890 16. Added VPATH=.libs to the makefile; this should help when building with one
1891 prefix path and installing with another. (Or so I'm told by someone who
1892 knows more about this stuff than I do.)
1893
1894 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
1895 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
1896 match everything, including newlines. This is not POSIX-compatible, but
1897 somebody wanted the feature. From pcretest it can be activated by using
1898 both the P and the s flags.
1899
1900 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
1901
1902 19. libpcre.pc was being incorrectly installed as executable.
1903
1904 20. A couple of places in pcretest check for end-of-line by looking for '\n';
1905 it now also looks for '\r' so that it will work unmodified on Windows.
1906
1907 21. Added Google's contributed C++ wrapper to the distribution.
1908
1909 22. Added some untidy missing memory free() calls in pcretest, to keep
1910 Electric Fence happy when testing.
1911
1912
1913
1914 Version 5.0 13-Sep-04
1915 ---------------------
1916
1917 1. Internal change: literal characters are no longer packed up into items
1918 containing multiple characters in a single byte-string. Each character
1919 is now matched using a separate opcode. However, there may be more than one
1920 byte in the character in UTF-8 mode.
1921
1922 2. The pcre_callout_block structure has two new fields: pattern_position and
1923 next_item_length. These contain the offset in the pattern to the next match
1924 item, and its length, respectively.
1925
1926 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
1927 insertion of callouts before each pattern item. Added the /C option to
1928 pcretest to make use of this.
1929
1930 4. On the advice of a Windows user, the lines
1931
1932 #if defined(_WIN32) || defined(WIN32)
1933 _setmode( _fileno( stdout ), 0x8000 );
1934 #endif /* defined(_WIN32) || defined(WIN32) */
1935
1936 have been added to the source of pcretest. This apparently does useful
1937 magic in relation to line terminators.
1938
1939 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
1940 for the benefit of those environments where the "b" makes a difference.
1941
1942 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
1943 to know about it. I have put a hack into configure.in that adds in code
1944 to set GCC=yes if CC=icc. This seems to end up at a point in the
1945 generated configure script that is early enough to affect the setting of
1946 compiler options, which is what is needed, but I have no means of testing
1947 whether it really works. (The user who reported this had patched the
1948 generated configure script, which of course I cannot do.)
1949
1950 LATER: After change 22 below (new libtool files), the configure script
1951 seems to know about icc (and also ecc). Therefore, I have commented out
1952 this hack in configure.in.
1953
1954 7. Added support for pkg-config (2 patches were sent in).
1955
1956 8. Negated POSIX character classes that used a combination of internal tables
1957 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
1958 [[:^ascii]]. Typically, they would match almost any characters. The other
1959 POSIX classes were not broken in this way.
1960
1961 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
1962 to find the match, as PCRE was deluded into thinking that the match had to
1963 start at the start point or following a newline. The same bug applied to
1964 patterns with negative forward assertions or any backward assertions
1965 preceding ".*" at the start, unless the pattern required a fixed first
1966 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
1967
1968 10. In UTF-8 mode, when moving forwards in the subject after a failed match
1969 starting at the last subject character, bytes beyond the end of the subject
1970 string were read.
1971
1972 11. Renamed the variable "class" as "classbits" to make life easier for C++
1973 users. (Previously there was a macro definition, but it apparently wasn't
1974 enough.)
1975
1976 12. Added the new field "tables" to the extra data so that tables can be passed
1977 in at exec time, or the internal tables can be re-selected. This allows
1978 a compiled regex to be saved and re-used at a later time by a different
1979 program that might have everything at different addresses.
1980
1981 13. Modified the pcre-config script so that, when run on Solaris, it shows a
1982 -R library as well as a -L library.
1983
1984 14. The debugging options of pcretest (-d on the command line or D on a
1985 pattern) showed incorrect output for anything following an extended class
1986 that contained multibyte characters and which was followed by a quantifier.
1987
1988 15. Added optional support for general category Unicode character properties
1989 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
1990 support. It adds about 90K to the size of the library. The meanings of the
1991 inbuilt class escapes such as \d and \s have NOT been changed.
1992
1993 16. Updated pcredemo.c to include calls to free() to release the memory for the
1994 compiled pattern.
1995
1996 17. The generated file chartables.c was being created in the source directory
1997 instead of in the building directory. This caused the build to fail if the
1998 source directory was different from the building directory, and was
1999 read-only.
2000
2001 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
2002 file. No doubt somebody will tell me if they don't make sense... Also added
2003 Dan Mooney's comments about building on OpenVMS.
2004
2005 19. Added support for partial matching via the PCRE_PARTIAL option for
2006 pcre_exec() and the \P data escape in pcretest.
2007
2008 20. Extended pcretest with 3 new pattern features:
2009
2010 (i) A pattern option of the form ">rest-of-line" causes pcretest to
2011 write the compiled pattern to the file whose name is "rest-of-line".
2012 This is a straight binary dump of the data, with the saved pointer to
2013 the character tables forced to be NULL. The study data, if any, is
2014 written too. After writing, pcretest reads a new pattern.
2015
2016 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
2017 compiled pattern from the given file. There must not be any
2018 occurrences of "<" in the file name (pretty unlikely); if there are,
2019 pcretest will instead treat the initial "<" as a pattern delimiter.
2020 After reading in the pattern, pcretest goes on to read data lines as
2021 usual.
2022
2023 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
2024 and 16-bit fields in a compiled pattern, to simulate a pattern that
2025 was compiled on a host of opposite endianness.
2026
2027 21. The pcre-exec() function can now cope with patterns that were compiled on
2028 hosts of opposite endianness, with this restriction:
2029
2030 As for any compiled expression that is saved and used later, the tables
2031 pointer field cannot be preserved; the extra_data field in the arguments
2032 to pcre_exec() should be used to pass in a tables address if a value
2033 other than the default internal tables were used at compile time.
2034
2035 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
2036 now diagnosed as an error. Previously, most of the time, a negative number
2037 would have been treated as zero, but if in addition "ovector" was passed as
2038 NULL, a crash could occur.
2039
2040 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
2041 new versions from the libtool 1.5 distribution (the last one is a copy of
2042 a file called libtool.m4). This seems to have fixed the need to patch
2043 "configure" to support Darwin 1.3 (which I used to do). However, I still
2044 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
2045 workstation).
2046
2047 24. Changed the PCRE licence to be the more standard "BSD" licence.
2048
2049
2050 Version 4.5 01-Dec-03
2051 ---------------------
2052
2053 1. There has been some re-arrangement of the code for the match() function so
2054 that it can be compiled in a version that does not call itself recursively.
2055 Instead, it keeps those local variables that need separate instances for
2056 each "recursion" in a frame on the heap, and gets/frees frames whenever it
2057 needs to "recurse". Keeping track of where control must go is done by means
2058 of setjmp/longjmp. The whole thing is implemented by a set of macros that
2059 hide most of the details from the main code, and operates only if
2060 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
2061 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
2062 operating.
2063
2064 To make it easier for callers to provide specially tailored get/free
2065 functions for this usage, two new functions, pcre_stack_malloc, and
2066 pcre_stack_free, are used. They are always called in strict stacking order,
2067 and the size of block requested is always the same.
2068
2069 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
2070 PCRE has been compiled to use the stack or the heap for recursion. The
2071 -C option of pcretest uses this to show which version is compiled.
2072
2073 A new data escape \S, is added to pcretest; it causes the amounts of store
2074 obtained and freed by both kinds of malloc/free at match time to be added
2075 to the output.
2076
2077 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
2078 what's available on my current Linux desktop machine.
2079
2080 3. When matching a UTF-8 string, the test for a valid string at the start has
2081 been extended. If start_offset is not zero, PCRE now checks that it points
2082 to a byte that is the start of a UTF-8 character. If not, it returns
2083 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
2084 this is necessary because there may be backward assertions in the pattern.
2085 When matching the same subject several times, it may save resources to use
2086 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
2087
2088 4. The code for checking the validity of UTF-8 strings has been tightened so
2089 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
2090 containing "overlong sequences".
2091
2092 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
2093 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
2094 should have been "&", but it just so happened that all the cases this let
2095 through by mistake were picked up later in the function.
2096
2097 6. I had used a variable called "isblank" - this is a C99 function, causing
2098 some compilers to warn. To avoid this, I renamed it (as "blankclass").
2099
2100 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
2101 prompting; (b) run "./pcretest /dev/null" at the start of the test script
2102 so the version is shown; (c) stop "make test" echoing "./RunTest".
2103
2104 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
2105
2106 9. The prototype for memmove() for systems that don't have it was using
2107 size_t, but the inclusion of the header that defines size_t was later. I've
2108 moved the #includes for the C headers earlier to avoid this.
2109
2110 10. Added some adjustments to the code to make it easier to compiler on certain
2111 special systems:
2112
2113 (a) Some "const" qualifiers were missing.
2114 (b) Added the macro EXPORT before all exported functions; by default this
2115 is defined to be empty.
2116 (c) Changed the dftables auxiliary program (that builds chartables.c) so
2117 that it reads its output file name as an argument instead of writing
2118 to the standard output and assuming this can be redirected.
2119
2120 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
2121 class containing characters with values greater than 255, PCRE compilation
2122 went into a loop.
2123
2124 12. A recursive reference to a subpattern that was within another subpattern
2125 that had a minimum quantifier of zero caused PCRE to crash. For example,
2126 (x(y(?2))z)? provoked this bug with a subject that got as far as the
2127 recursion. If the recursively-called subpattern itself had a zero repeat,
2128 that was OK.
2129
2130 13. In pcretest, the buffer for reading a data line was set at 30K, but the
2131 buffer into which it was copied (for escape processing) was still set at
2132 1024, so long lines caused crashes.
2133
2134 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
2135 "internal error: code overflow...". This applied to any character class
2136 that was followed by a possessive quantifier.
2137
2138 15. Modified the Makefile to add libpcre.la as a prerequisite for
2139 libpcreposix.la because I was told this is needed for a parallel build to
2140 work.
2141
2142 16. If a pattern that contained .* following optional items at the start was
2143 studied, the wrong optimizing data was generated, leading to matching
2144 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
2145 matching string must start with a or b or c. The correct conclusion for
2146 this pattern is that a match can start with any character.
2147
2148
2149 Version 4.4 13-Aug-03
2150 ---------------------
2151
2152 1. In UTF-8 mode, a character class containing characters with values between
2153 127 and 255 was not handled correctly if the compiled pattern was studied.
2154 In fixing this, I have also improved the studying algorithm for such
2155 classes (slightly).
2156
2157 2. Three internal functions had redundant arguments passed to them. Removal
2158 might give a very teeny performance improvement.
2159
2160 3. Documentation bug: the value of the capture_top field in a callout is *one
2161 more than* the number of the hightest numbered captured substring.
2162
2163 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
2164 in incorrectly linking with a previously installed version. They now link
2165 explicitly with libpcre.la.
2166
2167 5. configure.in no longer needs to recognize Cygwin specially.
2168
2169 6. A problem in pcre.in for Windows platforms is fixed.
2170
2171 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
2172 pcretest, it used to include the size of the study block as part of its
2173 output. Unfortunately, the structure contains a field that has a different
2174 size on different hardware architectures. This meant that the tests that
2175 showed this size failed. As the block is currently always of a fixed size,
2176 this information isn't actually particularly useful in pcretest output, so
2177 I have just removed it.
2178
2179 8. Three pre-processor statements accidentally did not start in column 1.
2180 Sadly, there are *still* compilers around that complain, even though
2181 standard C has not required this for well over a decade. Sigh.
2182
2183 9. In pcretest, the code for checking callouts passed small integers in the
2184 callout_data field, which is a void * field. However, some picky compilers
2185 complained about the casts involved for this on 64-bit systems. Now
2186 pcretest passes the address of the small integer instead, which should get
2187 rid of the warnings.
2188
2189 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
2190 both compile and run time, and gives an error if an invalid UTF-8 sequence
2191 is found. There is a option for disabling this check in cases where the
2192 string is known to be correct and/or the maximum performance is wanted.
2193
2194 11. In response to a bug report, I changed one line in Makefile.in from
2195
2196 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
2197 to
2198 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
2199
2200 to look similar to other lines, but I have no way of telling whether this
2201 is the right thing to do, as I do not use Windows. No doubt I'll get told
2202 if it's wrong...
2203
2204
2205 Version 4.3 21-May-03
2206 ---------------------
2207
2208 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
2209 Makefile.
2210
2211 2. Some refactoring to improve the quality of the code:
2212
2213 (i) The utf8_table... variables are now declared "const".
2214
2215 (ii) The code for \cx, which used the "case flipping" table to upper case
2216 lower case letters, now just substracts 32. This is ASCII-specific,
2217 but the whole concept of \cx is ASCII-specific, so it seems
2218 reasonable.
2219
2220 (iii) PCRE was using its character types table to recognize decimal and
2221 hexadecimal digits in the pattern. This is silly, because it handles
2222 only 0-9, a-f, and A-F, but the character types table is locale-
2223 specific, which means strange things might happen. A private
2224 table is now used for this - though it costs 256 bytes, a table is
2225 much faster than multiple explicit tests. Of course, the standard
2226 character types table is still used for matching digits in subject
2227 strings against \d.
2228
2229 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
2230 ending in _t are). So I've renamed it as ESC_tee.
2231
2232 3. The first argument for regexec() in the POSIX wrapper should have been
2233 defined as "const".
2234
2235 4. Changed pcretest to use malloc() for its buffers so that they can be
2236 Electric Fenced for debugging.
2237
2238 5. There were several places in the code where, in UTF-8 mode, PCRE would try
2239 to read one or more bytes before the start of the subject string. Often this
2240 had no effect on PCRE's behaviour, but in some circumstances it could
2241 provoke a segmentation fault.
2242
2243 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
2244 to try to read one or more bytes before the start of the subject string.
2245
2246 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
2247 UTF-8 support could misbehave in various ways if the subject string
2248 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
2249 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
2250 back over UTF-8 characters.)
2251
2252
2253 Version 4.2 14-Apr-03
2254 ---------------------
2255
2256 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
2257
2258 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
2259 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
2260 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
2261 [WIN_PREFIX]: new variable, "cyg" for Cygwin
2262 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
2263 and BUILD_EXEEXT
2264 Note: automatic setting of the BUILD variables is not yet working
2265 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
2266 compile-time but not at link-time
2267 [LINK]: use for linking executables only
2268 make different versions for Windows and non-Windows
2269 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
2270 libraries
2271 [LINK_FOR_BUILD]: new variable
2272 [OBJEXT]: use throughout
2273 [EXEEXT]: use throughout
2274 <winshared>: new target
2275 <wininstall>: new target
2276 <dftables.o>: use native compiler
2277 <dftables>: use native linker
2278 <install>: handle Windows platform correctly
2279 <clean>: ditto
2280 <check>: ditto
2281 copy DLL to top builddir before testing
2282
2283 As part of these changes, -no-undefined was removed again. This was reported
2284 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
2285 in any case.
2286
2287 3. Some tidies to get rid of compiler warnings:
2288
2289 . In the match_data structure, match_limit was an unsigned long int, whereas
2290 match_call_count was an int. I've made them both unsigned long ints.
2291
2292 . In pcretest the fact that a const uschar * doesn't automatically cast to
2293 a void * provoked a warning.
2294
2295 . Turning on some more compiler warnings threw up some "shadow" variables
2296 and a few more missing casts.
2297
2298 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
2299 option, a class that contained a single character with a value between 128
2300 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
2301
2302 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
2303 option, a class that contained several characters, but with at least one
2304 whose value was between 128 and 255 caused PCRE to crash.
2305
2306
2307 Version 4.1 12-Mar-03
2308 ---------------------
2309
2310 1. Compiling with gcc -pedantic found a couple of places where casts were
2311 needed, and a string in dftables.c that was longer than standard compilers are
2312 required to support.
2313
2314 2. Compiling with Sun's compiler found a few more places where the code could
2315 be tidied up in order to avoid warnings.
2316
2317 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
2318 first of these names is deprecated in the latest Autoconf in favour of the name
2319 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
2320 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
2321 analogy I have changed it to CFLAGS_FOR_BUILD.
2322
2323 4. Added -no-undefined to the linking command in the Makefile, because this is
2324 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
2325 linking step for the pcreposix library.
2326
2327 5. PCRE was failing to diagnose the case of two named groups with the same
2328 name.
2329
2330 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
2331 literal character that is needed in the subject for a match, and scans along to
2332 ensure that it is present before embarking on the full matching process. This
2333 saves time in cases of nested unlimited repeats that are never going to match.
2334 Problem: the scan can take a lot of time if the subject is very long (e.g.
2335 megabytes), thus penalizing straightforward matches. It is now done only if the
2336 amount of subject to be scanned is less than 1000 bytes.
2337
2338 7. A lesser problem with the same optimization is that it was recording the
2339 first character of an anchored pattern as "needed", thus provoking a search
2340 right along the subject, even when the first match of the pattern was going to
2341 fail. The "needed" character is now not set for anchored patterns, unless it
2342 follows something in the pattern that is of non-fixed length. Thus, it still
2343 fulfils its original purpose of finding quick non-matches in cases of nested
2344 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
2345
2346
2347 Version 4.0 17-Feb-03
2348 ---------------------
2349
2350 1. If a comment in an extended regex that started immediately after a meta-item
2351 extended to the end of string, PCRE compiled incorrect data. This could lead to
2352 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
2353
2354 2. Moved to autoconf 2.53 and libtool 1.4.2.
2355
2356 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
2357 the special perltest8 script is no longer needed - all the tests can be run
2358 from a single perltest script.
2359
2360 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
2361 by \s. It has now been removed in PCRE. This means it isn't recognized as
2362 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
2363 class [:space:] *does* include VT, thereby creating a mess.
2364
2365 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
2366 space and tab.
2367
2368 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
2369 its new features into the main test script, reducing the number of scripts.
2370
2371 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
2372 were backward compatible, and made the (?i) apply to the whole pattern, as if
2373 /i were given. Now it behaves more logically, and applies the option setting
2374 only to what follows. PCRE has been changed to follow suit. However, if it
2375 finds options settings right at the start of the pattern, it extracts them into
2376 the global options, as before. Thus, they show up in the info data.
2377
2378 8. Added support for the \Q...\E escape sequence. Characters in between are
2379 treated as literals. This is slightly different from Perl in that $ and @ are
2380 also handled as literals inside the quotes. In Perl, they will cause variable
2381 interpolation. Note the following examples:
2382
2383 Pattern PCRE matches Perl matches
2384
2385 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
2386 \Qabc\$xyz\E abc\$xyz abc\$xyz
2387 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
2388
2389 For compatibility with Perl, \Q...\E sequences are recognized inside character
2390 classes as well as outside them.
2391
2392 9. Re-organized 3 code statements in pcretest to avoid "overflow in
2393 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
2394 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
2395 signed/unsigned warnings.
2396
2397 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
2398 option for pcretest, so I've replaced it by a simple function that does just
2399 that job.
2400
2401 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
2402 "pcregrep -".
2403
2404 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
2405 Java package. This provides some syntactic sugar for simple cases of what my
2406 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
2407 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
2408 item, you can use this simplified notation. Note that only makes sense with
2409 greedy quantifiers. Consequently, the use of the possessive quantifier forces
2410 greediness, whatever the setting of the PCRE_UNGREEDY option.
2411
2412 13. A change of greediness default within a pattern was not taking effect at
2413 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
2414 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
2415 was abstracted outside.
2416
2417 14. PCRE now supports the \G assertion. It is true when the current matching
2418 position is at the start point of the match. This differs from \A when the
2419 starting offset is non-zero. Used with the /g option of pcretest (or similar
2420 code), it works in the same way as it does for Perl's /g option. If all
2421 alternatives of a regex begin with \G, the expression is anchored to the start
2422 match position, and the "anchored" flag is set in the compiled expression.
2423
2424 15. Some bugs concerning the handling of certain option changes within patterns
2425 have been fixed. These applied to options other than (?ims). For example,
2426 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
2427 been the other way round. Some of this was related to change 7 above.
2428
2429 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
2430 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
2431 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
2432 POSIX classes only within a class (e.g. /[[:alpha:]]/).
2433
2434 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
2435 mode. Unlike ".", it always matches newline, whatever the setting of
2436 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
2437 assertions. Perl allows it, but it doesn't (in general) work because it can't
2438 calculate the length of the lookbehind. At least, that's the case for Perl
2439 5.8.0 - I've been told they are going to document that it doesn't work in
2440 future.
2441
2442 18. Added an error diagnosis for escapes that PCRE does not support: these are
2443 \L, \l, \N, \P, \p, \U, \u, and \X.
2444
2445 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
2446 reading past the end of the pattern in cases such as /[abcd/.
2447
2448 20. PCRE was getting more memory than necessary for patterns with classes that
2449 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
2450
2451 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
2452 compiling PCRE for use with Virtual Pascal.
2453
2454 22. Small fix to the Makefile to make it work properly if the build is done
2455 outside the source tree.
2456
2457 23. Added a new extension: a condition to go with recursion. If a conditional
2458 subpattern starts with (?(R) the "true" branch is used if recursion has
2459 happened, whereas the "false" branch is used only at the top level.
2460
2461 24. When there was a very long string of literal characters (over 255 bytes
2462 without UTF support, over 250 bytes with UTF support), the computation of how
2463 much memory was required could be incorrect, leading to segfaults or other
2464 strange effects.
2465
2466 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
2467 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
2468 there was a subsequent back reference to those brackets. This meant that, for
2469 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
2470 possible to check for precisely this case. All we can do is abandon the
2471 optimization if .* occurs inside capturing brackets when there are any back
2472 references whatsoever. (See below for a better fix that came later.)
2473
2474 26. The handling of the optimization for finding the first character of a
2475 non-anchored pattern, and for finding a character that is required later in the
2476 match were failing in some cases. This didn't break the matching; it just
2477 failed to optimize when it could. The way this is done has been re-implemented.
2478
2479 27. Fixed typo in error message for invalid (?R item (it said "(?p").
2480
2481 28. Added a new feature that provides some of the functionality that Perl
2482 provides with (?{...}). The facility is termed a "callout". The way it is done
2483 in PCRE is for the caller to provide an optional function, by setting
2484 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
2485 global variable. By default it is unset, which disables all calling out. To get
2486 the function called, the regex must include (?C) at appropriate points. This
2487 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
2488 This provides a means of identifying different callout points. When PCRE
2489 reaches such a point in the regex, if pcre_callout has been set, the external
2490 function is called. It is provided with data in a structure called
2491 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
2492 matching continues; if it returns a non-zero value, the match at the current
2493 point fails. However, backtracking will occur if possible. [This was changed
2494 later and other features added - see item 49 below.]
2495
2496 29. pcretest is upgraded to test the callout functionality. It provides a
2497 callout function that displays information. By default, it shows the start of
2498 the match and the current position in the text. There are some new data escapes
2499 to vary what happens:
2500
2501 \C+ in addition, show current contents of captured substrings
2502 \C- do not supply a callout function
2503 \C!n return 1 when callout number n is reached
2504 \C!n!m return 1 when callout number n is reached for the mth time
2505
2506 30. If pcregrep was called with the -l option and just a single file name, it
2507 output "<stdin>" if a match was found, instead of the file name.
2508
2509 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
2510 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
2511 pcre_exec(). This saves a malloc/free per call. The default value of
2512 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
2513 when configuring.
2514
2515 32. The default maximum size of a compiled pattern is 64K. There have been a
2516 few cases of people hitting this limit. The code now uses macros to handle the
2517 storing of links as offsets within the compiled pattern. It defaults to 2-byte
2518 links, but this can be changed to 3 or 4 bytes by --with-link-size when
2519 configuring. Tests 2 and 5 work only with 2-byte links because they output
2520 debugging information about compiled patterns.
2521
2522 33. Internal code re-arrangements:
2523
2524 (a) Moved the debugging function for printing out a compiled regex into
2525 its own source file (printint.c) and used #include to pull it into
2526 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
2527 separate copies.
2528
2529 (b) Defined the list of op-code names for debugging as a macro in
2530 internal.h so that it is next to the definition of the opcodes.
2531
2532 (c) Defined a table of op-code lengths for simpler skipping along compiled
2533 code. This is again a macro in internal.h so that it is next to the
2534 definition of the opcodes.
2535
2536 34. Added support for recursive calls to individual subpatterns, along the
2537 lines of Robin Houston's patch (but implemented somewhat differently).
2538
2539 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
2540 allow it to read and process whole directories in Win32. This code was
2541 contributed by Lionel Fourquaux; it has not been tested by me.
2542
2543 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
2544 used to name a group. Names consist of alphanumerics and underscores, and must
2545 be unique. Back references use the syntax (?P=name) and recursive calls use
2546 (?P>name) which is a PCRE extension to the Python extension. Groups still have
2547 numbers. The function pcre_fullinfo() can be used after compilation to extract
2548 a name/number map. There are three relevant calls:
2549
2550 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
2551 PCRE_INFO_NAMECOUNT yields the number of entries
2552 PCRE_INFO_NAMETABLE yields a pointer to the map.
2553
2554 The map is a vector of fixed-size entries. The size of each entry depends on
2555 the length of the longest name used. The first two bytes of each entry are the
2556 group number, most significant byte first. There follows the corresponding
2557 name, zero terminated. The names are in alphabetical order.
2558
2559 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
2560 case instead of 255. Making it the same both with and without UTF-8 support
2561 means that the same test output works with both.
2562
2563 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
2564 calling malloc() with a zero argument.
2565
2566 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
2567 optimization. I've improved things by keeping a bitmap of backreferences with
2568 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
2569 fact referenced, the optimization can be applied. It is unlikely that a
2570 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
2571 the match to follow \n) will appear inside brackets with a number greater than
2572 31, but if it does, any back reference > 31 suppresses the optimization.
2573
2574 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
2575 of disabling numbered capturing parentheses. Any opening parenthesis that is
2576 not followed by ? behaves as if it were followed by ?: but named parentheses
2577 can still be used for capturing (and they will acquire numbers in the usual
2578 way).
2579
2580 41. Redesigned the return codes from the match() function into yes/no/error so
2581 that errors can be passed back from deep inside the nested calls. A malloc
2582 failure while inside a recursive subpattern call now causes the
2583 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
2584
2585 42. It is now possible to set a limit on the number of times the match()
2586 function is called in a call to pcre_exec(). This facility makes it possible to
2587 limit the amount of recursion and backtracking, though not in a directly
2588 obvious way, because the match() function is used in a number of different
2589 circumstances. The count starts from zero for each position in the subject
2590 string (for non-anchored patterns). The default limit is, for compatibility, a
2591 large number, namely 10 000 000. You can change this in two ways:
2592
2593 (a) When configuring PCRE before making, you can use --with-match-limit=n
2594 to set a default value for the compiled library.
2595
2596 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
2597 a different value is set. See 45 below.
2598
2599 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
2600
2601 43. Added a new function pcre_config(int, void *) to enable run-time extraction
2602 of things that can be changed at compile time. The first argument specifies
2603 what is wanted and the second points to where the information is to be placed.
2604 The current list of available information is:
2605
2606 PCRE_CONFIG_UTF8
2607
2608 The output is an integer that is set to one if UTF-8 support is available;
2609 otherwise it is set to zero.
2610
2611 PCRE_CONFIG_NEWLINE
2612
2613 The output is an integer that it set to the value of the code that is used for
2614 newline. It is either LF (10) or CR (13).
2615
2616 PCRE_CONFIG_LINK_SIZE
2617
2618 The output is an integer that contains the number of bytes used for internal
2619 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
2620
2621 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
2622
2623 The output is an integer that contains the threshold above which the POSIX
2624 interface uses malloc() for output vectors. See item 31 above.
2625
2626 PCRE_CONFIG_MATCH_LIMIT
2627
2628 The output is an unsigned integer that contains the default limit of the number
2629 of match() calls in a pcre_exec() execution. See 42 above.
2630
2631 44. pcretest has been upgraded by the addition of the -C option. This causes it
2632 to extract all the available output from the new pcre_config() function, and to
2633 output it. The program then exits immediately.
2634
2635 45. A need has arisen to pass over additional data with calls to pcre_exec() in
2636 order to support additional features. One way would have been to define
2637 pcre_exec2() (for example) with extra arguments, but this would not have been
2638 extensible, and would also have required all calls to the original function to
2639 be mapped to the new one. Instead, I have chosen to extend the mechanism that
2640 is used for passing in "extra" data from pcre_study().
2641
2642 The pcre_extra structure is now exposed and defined in pcre.h. It currently
2643 contains the following fields:
2644
2645 flags a bitmap indicating which of the following fields are set
2646 study_data opaque data from pcre_study()
2647 match_limit a way of specifying a limit on match() calls for a specific
2648 call to pcre_exec()
2649 callout_data data for callouts (see 49 below)
2650
2651 The flag bits are also defined in pcre.h, and are
2652
2653 PCRE_EXTRA_STUDY_DATA
2654 PCRE_EXTRA_MATCH_LIMIT
2655 PCRE_EXTRA_CALLOUT_DATA
2656
2657 The pcre_study() function now returns one of these new pcre_extra blocks, with
2658 the actual study data pointed to by the study_data field, and the
2659 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
2660 before. That is, this change is entirely upwards-compatible and requires no
2661 change to existing code.
2662
2663 If you want to pass in additional data to pcre_exec(), you can either place it
2664 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
2665 block.
2666
2667 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
2668 data string contains the escape sequence \M, pcretest calls pcre_exec() several
2669 times with different match limits, until it finds the minimum value needed for
2670 pcre_exec() to complete. The value is then output. This can be instructive; for
2671 most simple matches the number is quite small, but for pathological cases it
2672 gets very large very quickly.
2673
2674 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
2675 returns the size of the data block pointed to by the study_data field in a
2676 pcre_extra block, that is, the value that was passed as the argument to
2677 pcre_malloc() when PCRE was getting memory in which to place the information
2678 created by pcre_study(). The fourth argument should point to a size_t variable.
2679 pcretest has been extended so that this information is shown after a successful
2680 pcre_study() call when information about the compiled regex is being displayed.
2681
2682 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
2683 because what follows is always an absolute path. (Later: it turns out that this
2684 is more than cosmetic for MinGW, because it doesn't like empty path
2685 components.)
2686
2687 49. Some changes have been made to the callout feature (see 28 above):
2688
2689 (i) A callout function now has three choices for what it returns:
2690
2691 0 => success, carry on matching
2692 > 0 => failure at this point, but backtrack if possible
2693 < 0 => serious error, return this value from pcre_exec()
2694
2695 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
2696 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
2697 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
2698 use by callout functions. It will never be used by PCRE itself.
2699
2700 (ii) The pcre_extra structure (see 45 above) has a void * field called
2701 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
2702 pcre_callout_block structure has a field of the same name. The contents of
2703 the field passed in the pcre_extra structure are passed to the callout
2704 function in the corresponding field in the callout block. This makes it
2705 easier to use the same callout-containing regex from multiple threads. For
2706 testing, the pcretest program has a new data escape
2707
2708 \C*n pass the number n (may be negative) as callout_data
2709
2710 If the callout function in pcretest receives a non-zero value as
2711 callout_data, it returns that value.
2712
2713 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
2714 there were some redundant $(CFLAGS) in commands that are now specified as
2715 $(LINK), which already includes $(CFLAGS).
2716
2717 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
2718 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
2719 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
2720 one-byte characters throughout. Note that case-insensitive matching applies
2721 only to characters whose values are less than 256. PCRE doesn't support the
2722 notion of cases for higher-valued characters.
2723
2724 (i) A character class whose characters are all within 0-255 is handled as
2725 a bit map, and the map is inverted for negative classes. Previously, a
2726 character > 255 always failed to match such a class; however it should
2727 match if the class was a negative one (e.g. [^ab]). This has been fixed.
2728
2729 (ii) A negated character class with a single character < 255 is coded as
2730 "not this character" (OP_NOT). This wasn't working properly when the test
2731 character was multibyte, either singly or repeated.
2732
2733 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
2734 mode, for example: \x{100}{2,3}.
2735
2736 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
2737 singly or repeated) now correctly test multibyte characters. However,
2738 PCRE doesn't recognize any characters with values greater than 255 as
2739 digits, spaces, or word characters. Such characters always match \D, \S,
2740 and \W, and never match \d, \s, or \w.
2741
2742 (v) Classes may now contain characters and character ranges with values
2743 greater than 255. For example: [ab\x{100}-\x{400}].
2744
2745 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
2746 PCRE in UTF-8 mode.
2747
2748 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
2749 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
2750 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
2751 value.)
2752
2753 53. The single man page has become too large. I have therefore split it up into
2754 a number of separate man pages. These also give rise to individual HTML pages;
2755 these are now put in a separate directory, and there is an index.html page that
2756 lists them all. Some hyperlinking between the pages has been installed.
2757
2758 54. Added convenience functions for handling named capturing parentheses.
2759
2760 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
2761 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
2762 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
2763 are faulted.
2764
2765 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
2766 calling configure. These values are used when compiling the dftables.c program
2767 which is run to generate the source of the default character tables. They
2768 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
2769 you will need to set these values.
2770
2771 57. Updated the building process for Windows DLL, as provided by Fred Cox.
2772
2773
2774 Version 3.9 02-Jan-02
2775 ---------------------
2776
2777 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
2778
2779 2. If --disable-static was given, the building process failed when trying to
2780 build pcretest and pcregrep. (For some reason it was using libtool to compile
2781 them, which is not right, as they aren't part of the library.)
2782
2783
2784 Version 3.8 18-Dec-01
2785 ---------------------
2786
2787 1. The experimental UTF-8 code was completely screwed up. It was packing the
2788 bytes in the wrong order. How dumb can you get?
2789
2790
2791 Version 3.7 29-Oct-01
2792 ---------------------
2793
2794 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
2795 This caused pcretest, when used on the test data, to segfault. Unfortunately,
2796 this didn't happen under Solaris 8, where I normally test things.
2797
2798 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
2799 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
2800 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
2801 this fix an hour or so after the initial 3.7 release.)
2802
2803
2804 Version 3.6 23-Oct-01
2805 ---------------------
2806
2807 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
2808 offsets passed as NULL with zero offset count.
2809
2810 2. The config.guess and config.sub files had not been updated when I moved to
2811 the latest autoconf.
2812
2813
2814 Version 3.5 15-Aug-01
2815 ---------------------
2816
2817 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
2818 had been forgotten.
2819
2820 2. By using declared but undefined structures, we can avoid using "void"
2821 definitions in pcre.h while keeping the internal definitions of the structures
2822 private.
2823
2824 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
2825 user point of view, this means that both static and shared libraries are built
2826 by default, but this can be individually controlled. More of the work of
2827 handling this static/shared cases is now inside libtool instead of PCRE's make
2828 file.
2829
2830 4. The pcretest utility is now installed along with pcregrep because it is
2831 useful for users (to test regexs) and by doing this, it automatically gets
2832 relinked by libtool. The documentation has been turned into a man page, so
2833 there are now .1, .txt, and .html versions in /doc.
2834
2835 5. Upgrades to pcregrep:
2836 (i) Added long-form option names like gnu grep.
2837 (ii) Added --help to list all options with an explanatory phrase.
2838 (iii) Added -r, --recursive to recurse into sub-directories.
2839 (iv) Added -f, --file to read patterns from a file.
2840
2841 6. pcre_exec() was referring to its "code" argument before testing that
2842 argument for NULL (and giving an error if it was NULL).
2843
2844 7. Upgraded Makefile.in to allow for compiling in a different directory from
2845 the source directory.
2846
2847 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
2848 options bits, the pointer it was passed was to an int instead of to an unsigned
2849 long int. This mattered only on 64-bit systems.
2850
2851 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
2852 generated) instead of pcre.in, which it its source. Also made the same change
2853 in several of the .c files.
2854
2855 10. A new release of gcc defines printf() as a macro, which broke pcretest
2856 because it had an ifdef in the middle of a string argument for printf(). Fixed
2857 by using separate calls to printf().
2858
2859 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
2860 script, to force use of CR or LF instead of \n in the source. On non-Unix
2861 systems, the value can be set in config.h.
2862
2863 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
2864 absolute limit. Changed the text of the error message to make this clear, and
2865 likewise updated the man page.
2866
2867 13. The limit of 99 on the number of capturing subpatterns has been removed.
2868 The new limit is 65535, which I hope will not be a "real" limit.
2869
2870
2871 Version 3.4 22-Aug-00
2872 ---------------------
2873
2874 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
2875
2876 2. Diagnose condition (?(0) as an error instead of crashing on matching.
2877
2878
2879 Version 3.3 01-Aug-00
2880 ---------------------
2881
2882 1. If an octal character was given, but the value was greater than \377, it
2883 was not getting masked to the least significant bits, as documented. This could
2884 lead to crashes in some systems.
2885
2886 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
2887 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
2888
2889 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
2890 These just pass their arguments on to (pcre_free)(), but they are provided
2891 because some uses of PCRE bind it to non-C systems that can call its functions,
2892 but cannot call free() or pcre_free() directly.
2893
2894 4. Add "make test" as a synonym for "make check". Corrected some comments in
2895 the Makefile.
2896
2897 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
2898 Makefile.
2899
2900 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
2901 command called pgrep for grepping around the active processes.
2902
2903 7. Added the beginnings of support for UTF-8 character strings.
2904
2905 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
2906 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
2907 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
2908 out for the ar command.)
2909
2910
2911 Version 3.2 12-May-00
2912 ---------------------
2913
2914 This is purely a bug fixing release.
2915
2916 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
2917 of ZA. This was just one example of several cases that could provoke this bug,
2918 which was introduced by change 9 of version 2.00. The code for breaking
2919 infinite loops after an iteration that matches an empty string was't working
2920 correctly.
2921
2922 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
2923 when matched against abbab (for example). After matching an empty string, it
2924 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
2925 caused it to match further down the string than it should.
2926
2927 3. The code contained an inclusion of sys/types.h. It isn't clear why this
2928 was there because it doesn't seem to be needed, and it causes trouble on some
2929 systems, as it is not a Standard C header. It has been removed.
2930
2931 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
2932 were reported on the Macintosh. The changes were from
2933
2934 while ((c = *(++ptr)) != 0 && c != '\n');
2935 to
2936 while ((c = *(++ptr)) != 0 && c != '\n') ;
2937
2938 Totally extraordinary, but if that's what it takes...
2939
2940 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
2941 available. Added HAVE_BCOPY and an autoconf test for it; if neither
2942 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
2943 assumes the way PCRE uses memmove() (always moving upwards).
2944
2945 6. PCRE is being used in one environment where strchr() is not available. There
2946 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
2947 faster code anyway.
2948
2949
2950 Version 3.1 09-Feb-00
2951 ---------------------
2952
2953 The only change in this release is the fixing of some bugs in Makefile.in for
2954 the "install" target:
2955
2956 (1) It was failing to install pcreposix.h.
2957
2958 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
2959
2960
2961 Version 3.0 01-Feb-00
2962 ---------------------
2963
2964 1. Add support for the /+ modifier to perltest (to output $` like it does in
2965 pcretest).
2966
2967 2. Add support for the /g modifier to perltest.
2968
2969 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
2970 matches null strings.
2971
2972 4. Fix perltest so that it doesn't do unwanted things when fed an empty
2973 pattern. Perl treats empty patterns specially - it reuses the most recent
2974 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
2975 effect.
2976
2977 5. The POSIX interface was broken in that it was just handing over the POSIX
2978 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
2979 required a bigger vector, with some working space on the end. This means that
2980 the POSIX wrapper now has to get and free some memory, and copy the results.
2981
2982 6. Added some simple autoconf support, placing the test data and the
2983 documentation in separate directories, re-organizing some of the
2984 information files, and making it build pcre-config (a GNU standard). Also added
2985 libtool support for building PCRE as a shared library, which is now the
2986 default.
2987
2988 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
2989 09 are not valid octal constants. Single digits will be used for minor values
2990 less than 10.
2991
2992 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
2993 existing programs that set these in the POSIX interface can use PCRE without
2994 modification.
2995
2996 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
2997 return all that pcre_info() returns, plus additional data. The pcre_info()
2998 function is retained for compatibility, but is considered to be obsolete.
2999
3000 10. Added experimental recursion feature (?R) to handle one common case that
3001 Perl 5.6 will be able to do with (?p{...}).
3002
3003 11. Added support for POSIX character classes like [:alpha:], which Perl is
3004 adopting.
3005
3006
3007 Version 2.08 31-Aug-99
3008 ----------------------
3009
3010 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
3011 trying to match at the startoffset position, but instead was moving forward to
3012 the next newline as if a previous match had failed.
3013
3014 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
3015 and could get into a loop if a null string was matched other than at the start
3016 of the subject.
3017
3018 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
3019 be distinguished at compile time, and for completeness also added PCRE_DATE.
3020
3021 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
3022 in GnuWin32 environments.
3023
3024
3025 Version 2.07 29-Jul-99
3026 ----------------------
3027
3028 1. The documentation is now supplied in plain text form and HTML as well as in
3029 the form of man page sources.
3030
3031 2. C++ compilers don't like assigning (void *) values to other pointer types.
3032 In particular this affects malloc(). Although there is no problem in Standard
3033 C, I've put in casts to keep C++ compilers happy.
3034
3035 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
3036 should be (const char *).
3037
3038 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
3039 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
3040 However, I haven't made this a standard facility. The documentation doesn't
3041 mention it, and the Makefile doesn't support it.
3042
3043 5. The Makefile now contains an "install" target, with editable destinations at
3044 the top of the file. The pcretest program is not installed.
3045
3046 6. pgrep -V now gives the PCRE version number and date.
3047
3048 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
3049 causing the entire string to be ignored, instead of just the last character.
3050
3051 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
3052 non-matching string, it can take a very, very long time, even for strings of
3053 quite modest length, because of the nested recursion. PCRE now does better in
3054 some of these cases. It does this by remembering the last required literal
3055 character in the pattern, and pre-searching the subject to ensure it is present
3056 before running the real match. In other words, it applies a heuristic to detect
3057 some types of certain failure quickly, and in the above example, if presented
3058 with a string that has no trailing " it gives "no match" very quickly.
3059
3060 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
3061 other alternatives are tried instead.
3062
3063
3064 Version 2.06 09-Jun-99
3065 ----------------------
3066
3067 1. Change pcretest's output for amount of store used to show just the code
3068 space, because the remainder (the data block) varies in size between 32-bit and
3069 64-bit systems.
3070
3071 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
3072 start matching at. This allows lookbehinds to work when searching for multiple
3073 occurrences in a string.
3074
3075 3. Added additional options to pcretest for testing multiple occurrences:
3076
3077 /+ outputs the rest of the string that follows a match
3078 /g loops for multiple occurrences, using the new startoffset argument
3079 /G loops for multiple occurrences by passing an incremented pointer
3080
3081 4. PCRE wasn't doing the "first character" optimization for patterns starting
3082 with \b or \B, though it was doing it for other lookbehind assertions. That is,
3083 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
3084 the letter 'x'. On long subject strings, this gives a significant speed-up.
3085
3086
3087 Version 2.05 21-Apr-99
3088 ----------------------
3089
3090 1. Changed the type of magic_number from int to long int so that it works
3091 properly on 16-bit systems.
3092
3093 2. Fixed a bug which caused patterns starting with .* not to work correctly
3094 when the subject string contained newline characters. PCRE was assuming
3095 anchoring for such patterns in all cases, which is not correct because .* will
3096 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
3097 DOTALL is set at top level; otherwise it knows that patterns starting with .*
3098 must be retried after every newline in the subject.
3099
3100
3101 Version 2.04 18-Feb-99
3102 ----------------------
3103
3104 1. For parenthesized subpatterns with repeats whose minimum was zero, the
3105 computation of the store needed to hold the pattern was incorrect (too large).
3106 If such patterns were nested a few deep, this could multiply and become a real
3107 problem.
3108
3109 2. Added /M option to pcretest to show the memory requirement of a specific
3110 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
3111
3112 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
3113 compiled in such a way that the backtracking after subsequent failure was
3114 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
3115 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
3116
3117
3118 Version 2.03 02-Feb-99
3119 ----------------------
3120
3121 1. Fixed typo and small mistake in man page.
3122
3123 2. Added 4th condition (GPL supersedes if conflict) and created separate
3124 LICENCE file containing the conditions.
3125
3126 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
3127 Perl, that is the internal \ allows the delimiter to be included in the
3128 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
3129 the final delimiter, add \ to the end of the pattern (to test the error).
3130
3131 4. Added the convenience functions for extracting substrings after a successful
3132 match. Updated pcretest to make it able to test these functions.
3133
3134
3135 Version 2.02 14-Jan-99
3136 ----------------------
3137
3138 1. Initialized the working variables associated with each extraction so that
3139 their saving and restoring doesn't refer to uninitialized store.
3140
3141 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
3142 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
3143 fix the problem.
3144
3145 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
3146 calls, and wasn't printing the correct value for compiling calls. Increased the
3147 default value of LOOPREPEAT, and the number of significant figures in the
3148 times.
3149
3150 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
3151
3152 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
3153 a building problem on Windows NT with a FAT file system.
3154
3155
3156 Version 2.01 21-Oct-98
3157 ----------------------
3158
3159 1. Changed the API for pcre_compile() to allow for the provision of a pointer
3160 to character tables built by pcre_maketables() in the current locale. If NULL
3161 is passed, the default tables are used.
3162
3163
3164 Version 2.00 24-Sep-98
3165 ----------------------
3166
3167 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
3168 it any more.
3169
3170 2. Allow quantification of (?>) groups, and make it work correctly.
3171
3172 3. The first character computation wasn't working for (?>) groups.
3173
3174 4. Correct the implementation of \Z (it is permitted to match on the \n at the
3175 end of the subject) and add 5.005's \z, which really does match only at the
3176 very end of the subject.
3177
3178 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
3179
3180 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
3181 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
3182 localized options. All options to pcre_study() were also removed.
3183
3184 7. Add other new features from 5.005:
3185
3186 $(?<= positive lookbehind
3187 $(?<! negative lookbehind
3188 (?imsx-imsx) added the unsetting capability
3189 such a setting is global if at outer level; local otherwise
3190 (?imsx-imsx:) non-capturing groups with option setting
3191 (?(cond)re|re) conditional pattern matching
3192
3193 A backreference to itself in a repeated group matches the previous
3194 captured string.
3195
3196 8. General tidying up of studying (both automatic and via "study")
3197 consequential on the addition of new assertions.
3198
3199 9. As in 5.005, unlimited repeated groups that could match an empty substring
3200 are no longer faulted at compile time. Instead, the loop is forcibly broken at
3201 runtime if any iteration does actually match an empty substring.
3202
3203 10. Include the RunTest script in the distribution.
3204
3205 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
3206 discrepancies, some of which were old and were also with respect to 5.004. They
3207 have now been fixed.
3208
3209
3210 Version 1.09 28-Apr-98
3211 ----------------------
3212
3213 1. A negated single character class followed by a quantifier with a minimum
3214 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
3215 program crashes, or just wrong answers. This did not apply to negated classes
3216 containing more than one character, or to minima other than one.
3217
3218
3219 Version 1.08 27-Mar-98
3220 ----------------------
3221
3222 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
3223
3224 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
3225 latter must appear before anything that relies on it in the pattern.
3226
3227
3228 Version 1.07 16-Feb-98
3229 ----------------------
3230
3231 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
3232 repeat of a potentially empty string).
3233
3234
3235 Version 1.06 23-Jan-98
3236 ----------------------
3237
3238 1. Added Markus Oberhumer's little patches for C++.
3239
3240 2. Literal strings longer than 255 characters were broken.
3241
3242
3243 Version 1.05 23-Dec-97
3244 ----------------------
3245
3246 1. Negated character classes containing more than one character were failing if
3247 PCRE_CASELESS was set at run time.
3248
3249
3250 Version 1.04 19-Dec-97
3251 ----------------------
3252
3253 1. Corrected the man page, where some "const" qualifiers had been omitted.
3254
3255 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
3256 input syntax.
3257
3258 3. Fixed memory leak which occurred when a regex with back references was
3259 matched with an offsets vector that wasn't big enough. The temporary memory
3260 that is used in this case wasn't being freed if the match failed.
3261
3262 4. Tidied pcretest to ensure it frees memory that it gets.
3263
3264 5. Temporary memory was being obtained in the case where the passed offsets
3265 vector was exactly big enough.
3266
3267 6. Corrected definition of offsetof() from change 5 below.
3268
3269 7. I had screwed up change 6 below and broken the rules for the use of
3270 setjmp(). Now fixed.
3271
3272
3273 Version 1.03 18-Dec-97
3274 ----------------------
3275
3276 1. A erroneous regex with a missing opening parenthesis was correctly
3277 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
3278 on some systems.
3279
3280 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
3281 it was reported that one broken compiler failed on the former because "code" is
3282 also an independent variable.
3283
3284 3. The erroneous regex a[]b caused an array overrun reference.
3285
3286 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
3287 fail on data ending with that character. (It was going on too far, and checking
3288 the next character, typically a binary zero.) This was specific to the
3289 optimized code for single-character negative classes.
3290
3291 5. Added a contributed patch from the TIN world which does the following:
3292
3293 + Add an undef for memmove, in case the the system defines a macro for it.
3294
3295 + Add a definition of offsetof(), in case there isn't one. (I don't know
3296 the reason behind this - offsetof() is part of the ANSI standard - but
3297 it does no harm).
3298
3299 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
3300 most of the places where whitespace preceded '#'. I have given up and
3301 allowed the remaining 2 cases to be at the margin.
3302
3303 + Rename some variables in pcre to eliminate shadowing. This seems very
3304 pedantic, but does no harm, of course.
3305
3306 6. Moved the call to setjmp() into its own function, to get rid of warnings
3307 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
3308
3309 7. Constructs such as \d{8,} were compiling into the equivalent of
3310 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
3311 outcome, but in this particular case used more store than had been allocated,
3312 which caused the bug to be discovered because it threw up an internal error.
3313
3314 8. The debugging code in both pcre and pcretest for outputting the compiled
3315 form of a regex was going wrong in the case of back references followed by
3316 curly-bracketed repeats.
3317
3318
3319 Version 1.02 12-Dec-97
3320 ----------------------
3321
3322 1. Typos in pcre.3 and comments in the source fixed.
3323
3324 2. Applied a contributed patch to get rid of places where it used to remove
3325 'const' from variables, and fixed some signed/unsigned and uninitialized
3326 variable warnings.
3327
3328 3. Added the "runtest" target to Makefile.
3329
3330 4. Set default compiler flag to -O2 rather than just -O.
3331
3332
3333 Version 1.01 19-Nov-97
3334 ----------------------
3335
3336 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
3337 like /([ab]*)*/, that is, for classes with more than one character in them.
3338
3339 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
3340 as /((?>a*))*/ (a PCRE_EXTRA facility).
3341
3342
3343 Version 1.00 18-Nov-97
3344 ----------------------
3345
3346 1. Added compile-time macros to support systems such as SunOS4 which don't have
3347 memmove() or strerror() but have other things that can be used instead.
3348
3349 2. Arranged that "make clean" removes the executables.
3350
3351
3352 Version 0.99 27-Oct-97
3353 ----------------------
3354
3355 1. Fixed bug in code for optimizing classes with only one character. It was
3356 initializing a 32-byte map regardless, which could cause it to run off the end
3357 of the memory it had got.
3358
3359 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
3360
3361
3362 Version 0.98 22-Oct-97
3363 ----------------------
3364
3365 1. Fixed bug in code for handling temporary memory usage when there are more
3366 back references than supplied space in the ovector. This could cause segfaults.
3367
3368
3369 Version 0.97 21-Oct-97
3370 ----------------------
3371
3372 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
3373
3374 2. Optimized negated single characters not to use a bit map.
3375
3376 3. Brought error texts together as macro definitions; clarified some of them;
3377 fixed one that was wrong - it said "range out of order" when it meant "invalid
3378 escape sequence".
3379
3380 4. Changed some char * arguments to const char *.
3381
3382 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
3383
3384 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
3385 pcretest.
3386
3387
3388 Version 0.96 16-Oct-97
3389 ----------------------
3390
3391 1. Added a simple "pgrep" utility to the distribution.
3392
3393 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
3394 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
3395 where "ddd" means "one or more decimal digits".
3396
3397 3. Fixed serious bug. If a pattern had a back reference, but the call to
3398 pcre_exec() didn't supply a large enough ovector to record the related
3399 identifying subpattern, the match always failed. PCRE now remembers the number
3400 of the largest back reference, and gets some temporary memory in which to save
3401 the offsets during matching if necessary, in order to ensure that
3402 backreferences always work.
3403
3404 4. Increased the compatibility with Perl in a number of ways:
3405
3406 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
3407 to request this handling. The option can be set at compile or exec time.
3408
3409 (b) $ matches before a terminating newline by default; an option
3410 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
3411 mode). The option can be set at compile or exec time.
3412
3413 (c) The handling of \ followed by a digit other than 0 is now supposed to be
3414 the same as Perl's. If the decimal number it represents is less than 10
3415 or there aren't that many previous left capturing parentheses, an octal
3416 escape is read. Inside a character class, it's always an octal escape,
3417 even if it is a single digit.
3418
3419 (d) An escaped but undefined alphabetic character is taken as a literal,
3420 unless PCRE_EXTRA is set. Currently this just reserves the remaining
3421 escapes.
3422
3423 (e) {0} is now permitted. (The previous item is removed from the compiled
3424 pattern).
3425
3426 5. Changed all the names of code files so that the basic parts are no longer
3427 than 10 characters, and abolished the teeny "globals.c" file.
3428
3429 6. Changed the handling of character classes; they are now done with a 32-byte
3430 bit map always.
3431
3432 7. Added the -d and /D options to pcretest to make it possible to look at the
3433 internals of compilation without having to recompile pcre.
3434
3435
3436 Version 0.95 23-Sep-97
3437 ----------------------
3438
3439 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
3440 \x20 at the start of a run of normal characters. These were being treated as
3441 real characters, instead of the source characters being re-checked.
3442
3443
3444 Version 0.94 18-Sep-97
3445 ----------------------
3446
3447 1. The functions are now thread-safe, with the caveat that the global variables
3448 containing pointers to malloc() and free() or alternative functions are the
3449 same for all threads.
3450
3451 2. Get pcre_study() to generate a bitmap of initial characters for non-
3452 anchored patterns when this is possible, and use it if passed to pcre_exec().
3453
3454
3455 Version 0.93 15-Sep-97
3456 ----------------------
3457
3458 1. /(b)|(:+)/ was computing an incorrect first character.
3459
3460 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
3461 but not actually doing anything yet.
3462
3463 3. Treat "-" characters in classes that cannot be part of ranges as literals,
3464 as Perl does (e.g. [-az] or [az-]).
3465
3466 4. Set the anchored flag if a branch starts with .* or .*? because that tests
3467 all possible positions.
3468
3469 5. Split up into different modules to avoid including unneeded functions in a
3470 compiled binary. However, compile and exec are still in one module. The "study"
3471 function is split off.
3472
3473 6. The character tables are now in a separate module whose source is generated
3474 by an auxiliary program - but can then be edited by hand if required. There are
3475 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
3476 toupper() in the code.
3477
3478 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
3479 make them global. Abolish the function for setting them, as the caller can now
3480 set them directly.
3481
3482
3483 Version 0.92 11-Sep-97
3484 ----------------------
3485
3486 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
3487 (e.g. /a{1,3}/) was broken (I mis-optimized it).
3488
3489 2. Caseless matching was not working in character classes if the characters in
3490 the pattern were in upper case.
3491
3492 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
3493
3494 4. Make PCRE_ANCHORED public and accept as a compile option.
3495
3496 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
3497 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
3498 pass them.
3499
3500 6. Give an error if bad option bits passed at compile or run time.
3501
3502 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
3503 pcretest to cause it to pass that flag.
3504
3505 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
3506 options, and the first character, if set.
3507
3508 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
3509
3510
3511 Version 0.91 10-Sep-97
3512 ----------------------
3513
3514 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
3515 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
3516
3517 2. PCRE was looping on encountering an indefinitely repeated back reference to
3518 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
3519 Perl does - treats the match as successful.
3520
3521 ****

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