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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 421 - (show annotations) (download)
Fri Aug 14 15:43:27 2009 UTC (5 years, 1 month ago) by ph10
File size: 166863 byte(s)
Fix --fixed-strings.

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8 was out-of-date, and there was no check on the pcre_compile() error code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 3. The pcregrep --count option prints the count for each file even when it is
17 zero, as does GNU grep. However, pcregrep was also printing all files when
18 --files-with-matches was added. Now, when both options are given, it prints
19 counts only for those files that have at least one match. (GNU grep just
20 prints the file name in this circumstance, but including the count seems
21 more useful - otherwise, why use --count?) Also ensured that the
22 combination -clh just lists non-zero counts, with no names.
23
24 4. The long form of the pcregrep -F option was incorrectly implemented as
25 --fixed_strings instead of --fixed-strings. This is an incompatible change,
26 but it seems right to fix it, and I didn't think it was worth preserving
27 the old behaviour.
28
29
30 Version 7.9 11-Apr-09
31 ---------------------
32
33 1. When building with support for bzlib/zlib (pcregrep) and/or readline
34 (pcretest), all targets were linked against these libraries. This included
35 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
36 libraries. This caused unwanted dependencies to be created. This problem
37 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
38 pcretest is linked with readline.
39
40 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
41 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
42 moved outside it again, because FALSE and TRUE are already defined in AIX,
43 but BOOL is not.
44
45 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
46 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
47
48 4. The pcregrep documentation said spaces were inserted as well as colons (or
49 hyphens) following file names and line numbers when outputting matching
50 lines. This is not true; no spaces are inserted. I have also clarified the
51 wording for the --colour (or --color) option.
52
53 5. In pcregrep, when --colour was used with -o, the list of matching strings
54 was not coloured; this is different to GNU grep, so I have changed it to be
55 the same.
56
57 6. When --colo(u)r was used in pcregrep, only the first matching substring in
58 each matching line was coloured. Now it goes on to look for further matches
59 of any of the test patterns, which is the same behaviour as GNU grep.
60
61 7. A pattern that could match an empty string could cause pcregrep to loop; it
62 doesn't make sense to accept an empty string match in pcregrep, so I have
63 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
64 seems to be how GNU grep behaves.
65
66 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
67 start or after a newline", because the conditional assertion was not being
68 correctly handled. The rule now is that both the assertion and what follows
69 in the first alternative must satisfy the test.
70
71 9. If auto-callout was enabled in a pattern with a conditional group whose
72 condition was an assertion, PCRE could crash during matching, both with
73 pcre_exec() and pcre_dfa_exec().
74
75 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
76 used for matching.
77
78 11. Unicode property support in character classes was not working for
79 characters (bytes) greater than 127 when not in UTF-8 mode.
80
81 12. Added the -M command line option to pcretest.
82
83 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
84
85 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
86
87 16. Added comments and documentation about mis-use of no_arg in the C++
88 wrapper.
89
90 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
91 from Martin Jerabek that uses macro names for all relevant character and
92 string constants.
93
94 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
95 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
96 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
97 these, but not everybody uses configure.
98
99 19. A conditional group that had only one branch was not being correctly
100 recognized as an item that could match an empty string. This meant that an
101 enclosing group might also not be so recognized, causing infinite looping
102 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
103 with the subject "ab", where knowledge that the repeated group can match
104 nothing is needed in order to break the loop.
105
106 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
107 exec(), but without supplying a callout function, matching went wrong.
108
109 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
110 leak if the size of the offset vector was greater than 30. When the vector
111 is smaller, the saved offsets during recursion go onto a local stack
112 vector, but for larger vectors malloc() is used. It was failing to free
113 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
114 error, in fact).
115
116 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
117 heapframe that is used only when UTF-8 support is enabled. This caused no
118 problem, but was untidy.
119
120 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
121 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
122 included within another project.
123
124 24. Steven Van Ingelgem's patches to add more options to the CMake support,
125 slightly modified by me:
126
127 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
128 not building pcregrep.
129
130 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
131 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
132
133 25. Forward references, both numeric and by name, in patterns that made use of
134 duplicate group numbers, could behave incorrectly or give incorrect errors,
135 because when scanning forward to find the reference group, PCRE was not
136 taking into account the duplicate group numbers. A pattern such as
137 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
138
139 26. Changed a few more instances of "const unsigned char *" to USPTR, making
140 the feature of a custom pointer more persuasive (as requested by a user).
141
142 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
143 pcretest.c, inside #ifndefs, because it seems they are sometimes already
144 pre-defined.
145
146 28. Added support for (*UTF8) at the start of a pattern.
147
148 29. Arrange for flags added by the "release type" setting in CMake to be shown
149 in the configuration summary.
150
151
152 Version 7.8 05-Sep-08
153 ---------------------
154
155 1. Replaced UCP searching code with optimized version as implemented for Ad
156 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
157 stage table and inline lookup instead of a function, giving speed ups of 2
158 to 5 times on some simple patterns that I tested. Permission was given to
159 distribute the MultiStage2.py script that generates the tables (it's not in
160 the tarball, but is in the Subversion repository).
161
162 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
163 scripts.
164
165 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
166 a group with a zero qualifier. The result of the study could be incorrect,
167 or the function might crash, depending on the pattern.
168
169 4. Caseless matching was not working for non-ASCII characters in back
170 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
171 It now works when Unicode Property Support is available.
172
173 5. In pcretest, an escape such as \x{de} in the data was always generating
174 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
175 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
176 truncation.
177
178 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
179
180 7. Added two (int) casts to pcregrep when printing the difference of two
181 pointers, in case they are 64-bit values.
182
183 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
184 test 2 if it fails.
185
186 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
187 and a #define of that name to empty if it is not externally set. This is to
188 allow users of MSVC to set it if necessary.
189
190 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
191 the convenience functions in the pcre_get.c source file.
192
193 11. An option change at the start of a pattern that had top-level alternatives
194 could cause overwriting and/or a crash. This command provoked a crash in
195 some environments:
196
197 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
198
199 This potential security problem was recorded as CVE-2008-2371.
200
201 12. For a pattern where the match had to start at the beginning or immediately
202 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
203 pcre_dfa_exec() could read past the end of the passed subject if there was
204 no match. To help with detecting such bugs (e.g. with valgrind), I modified
205 pcretest so that it places the subject at the end of its malloc-ed buffer.
206
207 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
208 exec() might read past the end of the data buffer in UTF-8 mode.
209
210 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
211 the data contained the byte 0x85 as part of a UTF-8 character within its
212 first line. This applied both to normal and DFA matching.
213
214 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
215 /^[^d]*?$/8 failed to match "abc".
216
217 16. Added a missing copyright notice to pcrecpp_internal.h.
218
219 17. Make it more clear in the documentation that values returned from
220 pcre_exec() in ovector are byte offsets, not character counts.
221
222 18. Tidied a few places to stop certain compilers from issuing warnings.
223
224 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
225 supplied by Stefan Weber. I made a further small update for 7.8 because
226 there is a change of source arrangements: the pcre_searchfuncs.c module is
227 replaced by pcre_ucd.c.
228
229
230 Version 7.7 07-May-08
231 ---------------------
232
233 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
234 a string to a long long, pretend we don't even have a long long." This is
235 done by checking for the strtoq, strtoll, and _strtoi64 functions.
236
237 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
238 pre-7.6 versions, which defined a global no_arg variable instead of putting
239 it in the RE class. (See also #8 below.)
240
241 3. Remove a line of dead code, identified by coverity and reported by Nuno
242 Lopes.
243
244 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
245
246 (1) The include/exclude patterns were being applied to the whole pathnames
247 of files, instead of just to the final components.
248
249 (2) If there was more than one level of directory, the subdirectories were
250 skipped unless they satisfied the include/exclude conditions. This is
251 inconsistent with GNU grep (and could even be seen as contrary to the
252 pcregrep specification - which I improved to make it absolutely clear).
253 The action now is always to scan all levels of directory, and just
254 apply the include/exclude patterns to regular files.
255
256 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
257 --exclude_dir in the tests to avoid scanning .svn directories.
258
259 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
260 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
261 doesn't support NULs in patterns.
262
263 7. Added some missing "const"s to declarations of static tables in
264 pcre_compile.c and pcre_dfa_exec.c.
265
266 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
267 caused by fix #2 above. (Subsequently also a second patch to fix the
268 first patch. And a third patch - this was a messy problem.)
269
270 9. Applied Craig's patch to remove the use of push_back().
271
272 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
273 matching function regexec().
274
275 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
276 which, however, unlike Perl's \g{...}, are subroutine calls, not back
277 references. PCRE supports relative numbers with this syntax (I don't think
278 Oniguruma does).
279
280 12. Previously, a group with a zero repeat such as (...){0} was completely
281 omitted from the compiled regex. However, this means that if the group
282 was called as a subroutine from elsewhere in the pattern, things went wrong
283 (an internal error was given). Such groups are now left in the compiled
284 pattern, with a new opcode that causes them to be skipped at execution
285 time.
286
287 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
288 to the way PCRE behaves:
289
290 (a) A lone ] character is dis-allowed (Perl treats it as data).
291
292 (b) A back reference to an unmatched subpattern matches an empty string
293 (Perl fails the current match path).
294
295 (c) A data ] in a character class must be notated as \] because if the
296 first data character in a class is ], it defines an empty class. (In
297 Perl it is not possible to have an empty class.) The empty class []
298 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
299 The negative empty class [^] matches any one character, independently
300 of the DOTALL setting.
301
302 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
303 non-existent subpattern following a character class starting with ']' and
304 containing () gave an internal compiling error instead of "reference to
305 non-existent subpattern". Fortunately, when the pattern did exist, the
306 compiled code was correct. (When scanning forwards to check for the
307 existencd of the subpattern, it was treating the data ']' as terminating
308 the class, so got the count wrong. When actually compiling, the reference
309 was subsequently set up correctly.)
310
311 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
312 it was being rejected as not supported by pcre_dfa_exec(), even though
313 other assertions are supported. I have made pcre_dfa_exec() support
314 (*FAIL).
315
316 16. The implementation of 13c above involved the invention of a new opcode,
317 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
318 cannot be changed at match time, I realized I could make a small
319 improvement to matching performance by compiling OP_ALLANY instead of
320 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
321 on the OP_ANY path.
322
323 17. Compiling pcretest on Windows with readline support failed without the
324 following two fixes: (1) Make the unistd.h include conditional on
325 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
326
327 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
328 ncurses library to be included for pcretest when ReadLine support is
329 requested, but also to allow for it to be overridden. This patch came from
330 Daniel Bergström.
331
332 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
333 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
334 any errors with the current Unicode tables. Thanks to Peter Kankowski for
335 spotting this.
336
337
338 Version 7.6 28-Jan-08
339 ---------------------
340
341 1. A character class containing a very large number of characters with
342 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
343 overflow.
344
345 2. Patch to cut out the "long long" test in pcrecpp_unittest when
346 HAVE_LONG_LONG is not defined.
347
348 3. Applied Christian Ehrlicher's patch to update the CMake build files to
349 bring them up to date and include new features. This patch includes:
350
351 - Fixed PH's badly added libz and libbz2 support.
352 - Fixed a problem with static linking.
353 - Added pcredemo. [But later removed - see 7 below.]
354 - Fixed dftables problem and added an option.
355 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
356 HAVE_LONG_LONG.
357 - Added readline support for pcretest.
358 - Added an listing of the option settings after cmake has run.
359
360 4. A user submitted a patch to Makefile that makes it easy to create
361 "pcre.dll" under mingw when using Configure/Make. I added stuff to
362 Makefile.am that cause it to include this special target, without
363 affecting anything else. Note that the same mingw target plus all
364 the other distribution libraries and programs are now supported
365 when configuring with CMake (see 6 below) instead of with
366 Configure/Make.
367
368 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
369 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
370 exported in the Windows port". It has not yet been confirmed that the patch
371 solves the problem, but it does no harm.
372
373 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
374 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
375 with CMake, and also correct the comment about stack recursion.
376
377 7. Remove the automatic building of pcredemo from the ./configure system and
378 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
379 of a program that users should build themselves after PCRE is installed, so
380 building it automatically is not really right. What is more, it gave
381 trouble in some build environments.
382
383 8. Further tidies to CMakeLists.txt from Sheri and Christian.
384
385
386 Version 7.5 10-Jan-08
387 ---------------------
388
389 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
390 values in parens when parsing an RE using the C++ wrapper."
391
392 2. Negative specials like \S did not work in character classes in UTF-8 mode.
393 Characters greater than 255 were excluded from the class instead of being
394 included.
395
396 3. The same bug as (2) above applied to negated POSIX classes such as
397 [:^space:].
398
399 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
400 defined or documented. It seems to have been a typo for PCRE_STATIC, so
401 I have changed it.
402
403 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
404 first named subpattern) and a construct such as (?&a) would reference the
405 first named subpattern whose name started with "a" (in other words, the
406 length check was missing). Both these problems are fixed. "Subpattern name
407 expected" is now given for (?&) (a zero-length name), and this patch also
408 makes it give the same error for \k'' (previously it complained that that
409 was a reference to a non-existent subpattern).
410
411 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
412 this is right because (?- can be followed by option settings as well as by
413 digits. I have, however, made the messages clearer.
414
415 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
416 than the number used in the conditional) now cause a compile-time error.
417 This is actually not compatible with Perl, which accepts such patterns, but
418 treats the conditional as always being FALSE (as PCRE used to), but it
419 seems to me that giving a diagnostic is better.
420
421 8. Change "alphameric" to the more common word "alphanumeric" in comments
422 and messages.
423
424 9. Fix two occurrences of "backslash" in comments that should have been
425 "backspace".
426
427 10. Remove two redundant lines of code that can never be obeyed (their function
428 was moved elsewhere).
429
430 11. The program that makes PCRE's Unicode character property table had a bug
431 which caused it to generate incorrect table entries for sequences of
432 characters that have the same character type, but are in different scripts.
433 It amalgamated them into a single range, with the script of the first of
434 them. In other words, some characters were in the wrong script. There were
435 thirteen such cases, affecting characters in the following ranges:
436
437 U+002b0 - U+002c1
438 U+0060c - U+0060d
439 U+0061e - U+00612
440 U+0064b - U+0065e
441 U+0074d - U+0076d
442 U+01800 - U+01805
443 U+01d00 - U+01d77
444 U+01d9b - U+01dbf
445 U+0200b - U+0200f
446 U+030fc - U+030fe
447 U+03260 - U+0327f
448 U+0fb46 - U+0fbb1
449 U+10450 - U+1049d
450
451 12. The -o option (show only the matching part of a line) for pcregrep was not
452 compatible with GNU grep in that, if there was more than one match in a
453 line, it showed only the first of them. It now behaves in the same way as
454 GNU grep.
455
456 13. If the -o and -v options were combined for pcregrep, it printed a blank
457 line for every non-matching line. GNU grep prints nothing, and pcregrep now
458 does the same. The return code can be used to tell if there were any
459 non-matching lines.
460
461 14. Added --file-offsets and --line-offsets to pcregrep.
462
463 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
464 infinitely looping recursion. The bug was that positive lookaheads were not
465 being skipped when checking for a possible empty match (negative lookaheads
466 and both kinds of lookbehind were skipped).
467
468 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
469 inclusion of <windows.h> to before rather than after the definition of
470 INVALID_FILE_ATTRIBUTES (patch from David Byron).
471
472 17. Specifying a possessive quantifier with a specific limit for a Unicode
473 character property caused pcre_compile() to compile bad code, which led at
474 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
475 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
476 caused the error; without that there was no problem.
477
478 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
479
480 19. Added --enable-pcretest-libreadline.
481
482 20. In pcrecpp.cc, the variable 'count' was incremented twice in
483 RE::GlobalReplace(). As a result, the number of replacements returned was
484 double what it should be. I removed one of the increments, but Craig sent a
485 later patch that removed the other one (the right fix) and added unit tests
486 that check the return values (which was not done before).
487
488 21. Several CMake things:
489
490 (1) Arranged that, when cmake is used on Unix, the libraries end up with
491 the names libpcre and libpcreposix, not just pcre and pcreposix.
492
493 (2) The above change means that pcretest and pcregrep are now correctly
494 linked with the newly-built libraries, not previously installed ones.
495
496 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
497
498 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
499 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
500 UTF-8 newline character). The key issue is that the pattern starts .*;
501 this means that the match must be either at the beginning, or after a
502 newline. The bug was in the code for advancing after a failed match and
503 checking that the new position followed a newline. It was not taking
504 account of UTF-8 characters correctly.
505
506 23. PCRE was behaving differently from Perl in the way it recognized POSIX
507 character classes. PCRE was not treating the sequence [:...:] as a
508 character class unless the ... were all letters. Perl, however, seems to
509 allow any characters between [: and :], though of course it rejects as
510 unknown any "names" that contain non-letters, because all the known class
511 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
512 for example, whereas PCRE did not - it did not recognize a POSIX character
513 class. This seemed a bit dangerous, so the code has been changed to be
514 closer to Perl. The behaviour is not identical to Perl, because PCRE will
515 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
516 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
517 Perl does, and where it didn't before.
518
519 24. Rewrite so as to remove the single use of %n from pcregrep because in some
520 Windows environments %n is disabled by default.
521
522
523 Version 7.4 21-Sep-07
524 ---------------------
525
526 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
527 means that a class such as [\s] counted as "explicit reference to CR or
528 LF". That isn't really right - the whole point of the change was to try to
529 help when there was an actual mention of one of the two characters. So now
530 the change happens only if \r or \n (or a literal CR or LF) character is
531 encountered.
532
533 2. The 32-bit options word was also used for 6 internal flags, but the numbers
534 of both had grown to the point where there were only 3 bits left.
535 Fortunately, there was spare space in the data structure, and so I have
536 moved the internal flags into a new 16-bit field to free up more option
537 bits.
538
539 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
540 but did not set the internal JCHANGED flag - either of these is enough to
541 control the way the "get" function works - but the PCRE_INFO_JCHANGED
542 facility is supposed to tell if (?J) was ever used, so now (?J) at the
543 start sets both bits.
544
545 4. Added options (at build time, compile time, exec time) to change \R from
546 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
547
548 5. doc/pcresyntax.html was missing from the distribution.
549
550 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
551 compatibility, even though it is no longer used.
552
553 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
554 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
555 windows.h file is present (where different names are used). [This was
556 reversed later after testing - see 16 below.]
557
558 8. Changed all #include <config.h> to #include "config.h". There were also
559 some further <pcre.h> cases that I changed to "pcre.h".
560
561 9. When pcregrep was used with the --colour option, it missed the line ending
562 sequence off the lines that it output.
563
564 10. It was pointed out to me that arrays of string pointers cause lots of
565 relocations when a shared library is dynamically loaded. A technique of
566 using a single long string with a table of offsets can drastically reduce
567 these. I have refactored PCRE in four places to do this. The result is
568 dramatic:
569
570 Originally: 290
571 After changing UCP table: 187
572 After changing error message table: 43
573 After changing table of "verbs" 36
574 After changing table of Posix names 22
575
576 Thanks to the folks working on Gregex for glib for this insight.
577
578 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
579 unicode-properties was also set.
580
581 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
582
583 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
584 checked only for CRLF.
585
586 14. Added casts to pcretest.c to avoid compiler warnings.
587
588 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
589
590 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
591 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
592 entirely. This removes changes made in 7 above.
593
594 17. The CMake files have been updated, and there is now more information about
595 building with CMake in the NON-UNIX-USE document.
596
597
598 Version 7.3 28-Aug-07
599 ---------------------
600
601 1. In the rejigging of the build system that eventually resulted in 7.1, the
602 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
603 brackets there is not right, since it causes compilers to look for an
604 installed pcre.h, not the version that is in the source that is being
605 compiled (which of course may be different). I have changed it back to:
606
607 #include "pcre.h"
608
609 I have a vague recollection that the change was concerned with compiling in
610 different directories, but in the new build system, that is taken care of
611 by the VPATH setting the Makefile.
612
613 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
614 when the subject happened to end in the byte 0x85 (e.g. if the last
615 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
616 characters but of course it shouldn't be taken as a newline when it is part
617 of another character. The bug was that, for an unlimited repeat of . in
618 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
619 characters when looking for a newline.
620
621 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
622
623 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
624 in debug output.
625
626 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
627 long printing in the pcrecpp unittest when running under MinGW.
628
629 6. ESC_K was left out of the EBCDIC table.
630
631 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
632 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
633 limit also applies to "virtual nesting" when a pattern is recursive, and in
634 this case 1000 isn't so big. I have been able to remove this limit at the
635 expense of backing off one optimization in certain circumstances. Normally,
636 when pcre_exec() would call its internal match() function recursively and
637 immediately return the result unconditionally, it uses a "tail recursion"
638 feature to save stack. However, when a subpattern that can match an empty
639 string has an unlimited repetition quantifier, it no longer makes this
640 optimization. That gives it a stack frame in which to save the data for
641 checking that an empty string has been matched. Previously this was taken
642 from the 1000-entry workspace that had been reserved. So now there is no
643 explicit limit, but more stack is used.
644
645 8. Applied Daniel's patches to solve problems with the import/export magic
646 syntax that is required for Windows, and which was going wrong for the
647 pcreposix and pcrecpp parts of the library. These were overlooked when this
648 problem was solved for the main library.
649
650 9. There were some crude static tests to avoid integer overflow when computing
651 the size of patterns that contain repeated groups with explicit upper
652 limits. As the maximum quantifier is 65535, the maximum group length was
653 set at 30,000 so that the product of these two numbers did not overflow a
654 32-bit integer. However, it turns out that people want to use groups that
655 are longer than 30,000 bytes (though not repeat them that many times).
656 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
657 made it possible to implement the integer overflow checks in a much more
658 dynamic way, which I have now done. The artificial limitation on group
659 length has been removed - we now have only the limit on the total length of
660 the compiled pattern, which depends on the LINK_SIZE setting.
661
662 10. Fixed a bug in the documentation for get/copy named substring when
663 duplicate names are permitted. If none of the named substrings are set, the
664 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
665 empty string.
666
667 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
668 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
669 because the ] is interpreted as the first data character and the
670 terminating ] is not found. PCRE has been made compatible with Perl in this
671 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
672 cause memory overwriting.
673
674 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
675 string has been matched (to stop an infinite loop). It was not recognizing
676 a conditional subpattern that could match an empty string if that
677 subpattern was within another subpattern. For example, it looped when
678 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
679 condition was not nested. This bug has been fixed.
680
681 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
682 past the start of the subject in the presence of bytes with the top bit
683 set, for example "\x8aBCD".
684
685 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
686 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
687
688 14. Optimized (?!) to (*FAIL).
689
690 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
691 This restricts code points to be within the range 0 to 0x10FFFF, excluding
692 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
693 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
694 does: it's just the validity check that is more restrictive.
695
696 16. Inserted checks for integer overflows during escape sequence (backslash)
697 processing, and also fixed erroneous offset values for syntax errors during
698 backslash processing.
699
700 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
701 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
702
703 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
704 caused an overrun.
705
706 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
707 something other than just ASCII characters) inside a group that had an
708 unlimited repeat caused a loop at compile time (while checking to see
709 whether the group could match an empty string).
710
711 20. Debugging a pattern containing \p or \P could cause a crash. For example,
712 [\P{Any}] did so. (Error in the code for printing property names.)
713
714 21. An orphan \E inside a character class could cause a crash.
715
716 22. A repeated capturing bracket such as (A)? could cause a wild memory
717 reference during compilation.
718
719 23. There are several functions in pcre_compile() that scan along a compiled
720 expression for various reasons (e.g. to see if it's fixed length for look
721 behind). There were bugs in these functions when a repeated \p or \P was
722 present in the pattern. These operators have additional parameters compared
723 with \d, etc, and these were not being taken into account when moving along
724 the compiled data. Specifically:
725
726 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
727 length.
728
729 (b) An item such as \pL+ within a repeated group could cause crashes or
730 loops.
731
732 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
733 "reference to non-existent subpattern" error.
734
735 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
736
737 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
738 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
739
740 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
741
742 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
743 character were causing crashes (broken optimization).
744
745 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
746 \p or \P) caused a compile-time loop.
747
748 28. More problems have arisen in unanchored patterns when CRLF is a valid line
749 break. For example, the unstudied pattern [\r\n]A does not match the string
750 "\r\nA" because change 7.0/46 below moves the current point on by two
751 characters after failing to match at the start. However, the pattern \nA
752 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
753 the same is true. There doesn't seem any very clean way out of this, but
754 what I have chosen to do makes the common cases work: PCRE now takes note
755 of whether there can be an explicit match for \r or \n anywhere in the
756 pattern, and if so, 7.0/46 no longer applies. As part of this change,
757 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
758 pattern has explicit CR or LF references.
759
760 29. Added (*CR) etc for changing newline setting at start of pattern.
761
762
763 Version 7.2 19-Jun-07
764 ---------------------
765
766 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
767 which is apparently normally available under Windows.
768
769 2. Re-jig the pcregrep tests with different newline settings in an attempt
770 to make them independent of the local environment's newline setting.
771
772 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
773
774 4. Some of the "internals" tests were previously cut out when the link size
775 was not 2, because the output contained actual offsets. The recent new
776 "Z" feature of pcretest means that these can be cut out, making the tests
777 usable with all link sizes.
778
779 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
780 stack recursion. This gives a massive performance boost under BSD, but just
781 a small improvement under Linux. However, it saves one field in the frame
782 in all cases.
783
784 6. Added more features from the forthcoming Perl 5.10:
785
786 (a) (?-n) (where n is a string of digits) is a relative subroutine or
787 recursion call. It refers to the nth most recently opened parentheses.
788
789 (b) (?+n) is also a relative subroutine call; it refers to the nth next
790 to be opened parentheses.
791
792 (c) Conditions that refer to capturing parentheses can be specified
793 relatively, for example, (?(-2)... or (?(+3)...
794
795 (d) \K resets the start of the current match so that everything before
796 is not part of it.
797
798 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
799
800 (f) \g{name} is another synonym - part of Perl 5.10's unification of
801 reference syntax.
802
803 (g) (?| introduces a group in which the numbering of parentheses in each
804 alternative starts with the same number.
805
806 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
807
808 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
809 PCRE_INFO_JCHANGED.
810
811 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
812 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
813 for detecting groups that can match an empty string.
814
815 9. A pattern with a very large number of alternatives (more than several
816 hundred) was running out of internal workspace during the pre-compile
817 phase, where pcre_compile() figures out how much memory will be needed. A
818 bit of new cunning has reduced the workspace needed for groups with
819 alternatives. The 1000-alternative test pattern now uses 12 bytes of
820 workspace instead of running out of the 4096 that are available.
821
822 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
823
824 11. Applied patch from Google to remove an optimization that didn't quite work.
825 The report of the bug said:
826
827 pcrecpp::RE("a*").FullMatch("aaa") matches, while
828 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
829 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
830
831 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
832 it matched the wrong number of bytes.
833
834
835 Version 7.1 24-Apr-07
836 ---------------------
837
838 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
839 that is more "standard", making use of automake and other Autotools. There
840 is some re-arrangement of the files and adjustment of comments consequent
841 on this.
842
843 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
844 for recursive directory scanning broke on some systems because the files
845 are not scanned in any specific order and on different systems the order
846 was different. A call to "sort" has been inserted into RunGrepTest for the
847 approprate test as a short-term fix. In the longer term there may be an
848 alternative.
849
850 3. I had an email from Eric Raymond about problems translating some of PCRE's
851 man pages to HTML (despite the fact that I distribute HTML pages, some
852 people do their own conversions for various reasons). The problems
853 concerned the use of low-level troff macros .br and .in. I have therefore
854 removed all such uses from the man pages (some were redundant, some could
855 be replaced by .nf/.fi pairs). The 132html script that I use to generate
856 HTML has been updated to handle .nf/.fi and to complain if it encounters
857 .br or .in.
858
859 4. Updated comments in configure.ac that get placed in config.h.in and also
860 arranged for config.h to be included in the distribution, with the name
861 config.h.generic, for the benefit of those who have to compile without
862 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
863
864 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
865 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
866 makevp.bat for the current PCRE, using the additional files
867 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
868
869 6. A Windows user reported a minor discrepancy with test 2, which turned out
870 to be caused by a trailing space on an input line that had got lost in his
871 copy. The trailing space was an accident, so I've just removed it.
872
873 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
874 that is needed.
875
876 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
877 as "const" (a) because they are and (b) because it helps the PHP
878 maintainers who have recently made a script to detect big data structures
879 in the php code that should be moved to the .rodata section. I remembered
880 to update Builducptable as well, so it won't revert if ucptable.h is ever
881 re-created.
882
883 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
884 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
885 order to be able to cut out the UTF-8 tables in the latter when UTF-8
886 support is not required. This saves 1.5-2K of code, which is important in
887 some applications.
888
889 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
890 so as not to refer to the tables, even though these functions will never be
891 called when UTF-8 support is disabled. Otherwise there are problems with a
892 shared library.
893
894 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
895
896 (a) It was defining its arguments as char * instead of void *.
897
898 (b) It was assuming that all moves were upwards in memory; this was true
899 a long time ago when I wrote it, but is no longer the case.
900
901 The emulated memove() is provided for those environments that have neither
902 memmove() nor bcopy(). I didn't think anyone used it these days, but that
903 is clearly not the case, as these two bugs were recently reported.
904
905 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
906 and Detrail to create the HTML documentation, the .txt form of the man
907 pages, and it removes trailing spaces from listed files. It also creates
908 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
909 case, it wraps all the #defines with #ifndefs. This script should be run
910 before "make dist".
911
912 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
913 with Unicode property support.
914
915 (a) For a maximizing quantifier, if the two different cases of the
916 character were of different lengths in their UTF-8 codings (there are
917 some cases like this - I found 11), and the matching function had to
918 back up over a mixture of the two cases, it incorrectly assumed they
919 were both the same length.
920
921 (b) When PCRE was configured to use the heap rather than the stack for
922 recursion during matching, it was not correctly preserving the data for
923 the other case of a UTF-8 character when checking ahead for a match
924 while processing a minimizing repeat. If the check also involved
925 matching a wide character, but failed, corruption could cause an
926 erroneous result when trying to check for a repeat of the original
927 character.
928
929 13. Some tidying changes to the testing mechanism:
930
931 (a) The RunTest script now detects the internal link size and whether there
932 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
933 values substituted by "configure". (The RunGrepTest script already did
934 this for UTF-8.) The configure.ac script no longer substitutes the
935 relevant variables.
936
937 (b) The debugging options /B and /D in pcretest show the compiled bytecode
938 with length and offset values. This means that the output is different
939 for different internal link sizes. Test 2 is skipped for link sizes
940 other than 2 because of this, bypassing the problem. Unfortunately,
941 there was also a test in test 3 (the locale tests) that used /B and
942 failed for link sizes other than 2. Rather than cut the whole test out,
943 I have added a new /Z option to pcretest that replaces the length and
944 offset values with spaces. This is now used to make test 3 independent
945 of link size. (Test 2 will be tidied up later.)
946
947 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
948 segmentation fault instead of returning the appropriate error message.
949
950 15. In multiline mode when the newline sequence was set to "any", the pattern
951 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
952 This doesn't seem right; it now treats the CRLF combination as the line
953 ending, and so does not match in that case. It's only a pattern such as ^$
954 that would hit this one: something like ^ABC$ would have failed after \r
955 and then tried again after \r\n.
956
957 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
958 in an attempt to make files that differ only in their line terminators
959 compare equal. This works on Linux.
960
961 17. Under certain error circumstances pcregrep might try to free random memory
962 as it exited. This is now fixed, thanks to valgrind.
963
964 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
965 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
966 was because its rules for how to advance for /g after matching an empty
967 string at the end of a line did not allow for this case. They now check for
968 it specially.
969
970 20. pcretest is supposed to handle patterns and data of any length, by
971 extending its buffers when necessary. It was getting this wrong when the
972 buffer for a data line had to be extended.
973
974 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
975 CRLF as a newline sequence.
976
977 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
978 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
979 I have nevertheless tidied it up.
980
981 23. Added some casts to kill warnings from HP-UX ia64 compiler.
982
983 24. Added a man page for pcre-config.
984
985
986 Version 7.0 19-Dec-06
987 ---------------------
988
989 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
990 moving to gcc 4.1.1.
991
992 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
993 sys/time.h, which is documented as needed for this function. It doesn't
994 seem to matter on Linux, but it showed up on some releases of OS X.
995
996 3. It seems that there are systems where bytes whose values are greater than
997 127 match isprint() in the "C" locale. The "C" locale should be the
998 default when a C program starts up. In most systems, only ASCII printing
999 characters match isprint(). This difference caused the output from pcretest
1000 to vary, making some of the tests fail. I have changed pcretest so that:
1001
1002 (a) When it is outputting text in the compiled version of a pattern, bytes
1003 other than 32-126 are always shown as hex escapes.
1004
1005 (b) When it is outputting text that is a matched part of a subject string,
1006 it does the same, unless a different locale has been set for the match
1007 (using the /L modifier). In this case, it uses isprint() to decide.
1008
1009 4. Fixed a major bug that caused incorrect computation of the amount of memory
1010 required for a compiled pattern when options that changed within the
1011 pattern affected the logic of the preliminary scan that determines the
1012 length. The relevant options are -x, and -i in UTF-8 mode. The result was
1013 that the computed length was too small. The symptoms of this bug were
1014 either the PCRE error "internal error: code overflow" from pcre_compile(),
1015 or a glibc crash with a message such as "pcretest: free(): invalid next
1016 size (fast)". Examples of patterns that provoked this bug (shown in
1017 pcretest format) are:
1018
1019 /(?-x: )/x
1020 /(?x)(?-x: \s*#\s*)/
1021 /((?i)[\x{c0}])/8
1022 /(?i:[\x{c0}])/8
1023
1024 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
1025 is now done differently.
1026
1027 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
1028 wrapper classes; (b) implement a new function in the C++ scanner that is
1029 more efficient than the old way of doing things because it avoids levels of
1030 recursion in the regex matching; (c) add a paragraph to the documentation
1031 for the FullMatch() function.
1032
1033 6. The escape sequence \n was being treated as whatever was defined as
1034 "newline". Not only was this contrary to the documentation, which states
1035 that \n is character 10 (hex 0A), but it also went horribly wrong when
1036 "newline" was defined as CRLF. This has been fixed.
1037
1038 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
1039 was being set to -1 for the "end of line" case (supposedly a value that no
1040 character can have). Though this value is never used (the check for end of
1041 line is "zero bytes in current character"), it caused compiler complaints.
1042 I've changed it to 0xffffffff.
1043
1044 8. In pcre_version.c, the version string was being built by a sequence of
1045 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
1046 string (as it is for production releases) called a macro with an empty
1047 argument. The C standard says the result of this is undefined. The gcc
1048 compiler treats it as an empty string (which was what was wanted) but it is
1049 reported that Visual C gives an error. The source has been hacked around to
1050 avoid this problem.
1051
1052 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
1053 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
1054 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
1055 of them did).
1056
1057 10. Originally, pcretest opened its input and output without "b"; then I was
1058 told that "b" was needed in some environments, so it was added for release
1059 5.0 to both the input and output. (It makes no difference on Unix-like
1060 systems.) Later I was told that it is wrong for the input on Windows. I've
1061 now abstracted the modes into two macros, to make it easier to fiddle with
1062 them, and removed "b" from the input mode under Windows.
1063
1064 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
1065
1066 12. Added -help and --help to pcretest as an official way of being reminded
1067 of the options.
1068
1069 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
1070 and pcrecpp.cc because they annoy compilers at high warning levels.
1071
1072 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
1073
1074 15. Fixed an occurrence of == in configure.ac that should have been = (shell
1075 scripts are not C programs :-) and which was not noticed because it works
1076 on Linux.
1077
1078 16. pcretest is supposed to handle any length of pattern and data line (as one
1079 line or as a continued sequence of lines) by extending its input buffer if
1080 necessary. This feature was broken for very long pattern lines, leading to
1081 a string of junk being passed to pcre_compile() if the pattern was longer
1082 than about 50K.
1083
1084 17. I have done a major re-factoring of the way pcre_compile() computes the
1085 amount of memory needed for a compiled pattern. Previously, there was code
1086 that made a preliminary scan of the pattern in order to do this. That was
1087 OK when PCRE was new, but as the facilities have expanded, it has become
1088 harder and harder to keep it in step with the real compile phase, and there
1089 have been a number of bugs (see for example, 4 above). I have now found a
1090 cunning way of running the real compile function in a "fake" mode that
1091 enables it to compute how much memory it would need, while actually only
1092 ever using a few hundred bytes of working memory and without too many
1093 tests of the mode. This should make future maintenance and development
1094 easier. A side effect of this work is that the limit of 200 on the nesting
1095 depth of parentheses has been removed (though this was never a serious
1096 limitation, I suspect). However, there is a downside: pcre_compile() now
1097 runs more slowly than before (30% or more, depending on the pattern). I
1098 hope this isn't a big issue. There is no effect on runtime performance.
1099
1100 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
1101 newline (only possible for the last line of a file) and it was a
1102 pattern that set a locale (followed by /Lsomething), pcretest crashed.
1103
1104 19. Added additional timing features to pcretest. (1) The -tm option now times
1105 matching only, not compiling. (2) Both -t and -tm can be followed, as a
1106 separate command line item, by a number that specifies the number of
1107 repeats to use when timing. The default is 50000; this gives better
1108 precision, but takes uncomfortably long for very large patterns.
1109
1110 20. Extended pcre_study() to be more clever in cases where a branch of a
1111 subpattern has no definite first character. For example, (a*|b*)[cd] would
1112 previously give no result from pcre_study(). Now it recognizes that the
1113 first character must be a, b, c, or d.
1114
1115 21. There was an incorrect error "recursive call could loop indefinitely" if
1116 a subpattern (or the entire pattern) that was being tested for matching an
1117 empty string contained only one non-empty item after a nested subpattern.
1118 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
1119 incorrectly, because the \d was being skipped in the check.
1120
1121 22. The pcretest program now has a new pattern option /B and a command line
1122 option -b, which is equivalent to adding /B to every pattern. This causes
1123 it to show the compiled bytecode, without the additional information that
1124 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
1125 is the same as /B/I).
1126
1127 23. A new optimization is now able automatically to treat some sequences such
1128 as a*b as a*+b. More specifically, if something simple (such as a character
1129 or a simple class like \d) has an unlimited quantifier, and is followed by
1130 something that cannot possibly match the quantified thing, the quantifier
1131 is automatically "possessified".
1132
1133 24. A recursive reference to a subpattern whose number was greater than 39
1134 went wrong under certain circumstances in UTF-8 mode. This bug could also
1135 have affected the operation of pcre_study().
1136
1137 25. Realized that a little bit of performance could be had by replacing
1138 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
1139
1140 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
1141
1142 27. Possessive quantifiers such as a++ were previously implemented by turning
1143 them into atomic groups such as ($>a+). Now they have their own opcodes,
1144 which improves performance. This includes the automatically created ones
1145 from 23 above.
1146
1147 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
1148 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
1149 the first matched character to be a colon. This applied both to named and
1150 numbered groups.
1151
1152 29. The ucpinternal.h header file was missing its idempotency #ifdef.
1153
1154 30. I was sent a "project" file called libpcre.a.dev which I understand makes
1155 building PCRE on Windows easier, so I have included it in the distribution.
1156
1157 31. There is now a check in pcretest against a ridiculously large number being
1158 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
1159 loop, the loop is abandoned.
1160
1161 32. Forward references to subpatterns in conditions such as (?(2)...) where
1162 subpattern 2 is defined later cause pcre_compile() to search forwards in
1163 the pattern for the relevant set of parentheses. This search went wrong
1164 when there were unescaped parentheses in a character class, parentheses
1165 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
1166
1167 33. "Subroutine" calls and backreferences were previously restricted to
1168 referencing subpatterns earlier in the regex. This restriction has now
1169 been removed.
1170
1171 34. Added a number of extra features that are going to be in Perl 5.10. On the
1172 whole, these are just syntactic alternatives for features that PCRE had
1173 previously implemented using the Python syntax or my own invention. The
1174 other formats are all retained for compatibility.
1175
1176 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
1177 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
1178 also .NET compatible.
1179
1180 (b) A recursion or subroutine call to a named group can now be defined as
1181 (?&name) as well as (?P>name).
1182
1183 (c) A backreference to a named group can now be defined as \k<name> or
1184 \k'name' as well as (?P=name). The new forms, as well as being in Perl
1185 5.10, are also .NET compatible.
1186
1187 (d) A conditional reference to a named group can now use the syntax
1188 (?(<name>) or (?('name') as well as (?(name).
1189
1190 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
1191 groups (named and numbered) that are never evaluated inline, but can be
1192 called as "subroutines" from elsewhere. In effect, the DEFINE condition
1193 is always false. There may be only one alternative in such a group.
1194
1195 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
1196 as the simple (?(R). The condition is true only if the most recent
1197 recursion is that of the given number or name. It does not search out
1198 through the entire recursion stack.
1199
1200 (g) The escape \gN or \g{N} has been added, where N is a positive or
1201 negative number, specifying an absolute or relative reference.
1202
1203 35. Tidied to get rid of some further signed/unsigned compiler warnings and
1204 some "unreachable code" warnings.
1205
1206 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
1207 things, this adds five new scripts.
1208
1209 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
1210 There were also incompatibilities regarding the handling of \Q..\E inside
1211 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
1212 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
1213
1214 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
1215 matches an empty string, and forcibly breaks the loop. There were bugs in
1216 this code in non-simple cases. For a pattern such as ^(a()*)* matched
1217 against aaaa the result was just "a" rather than "aaaa", for example. Two
1218 separate and independent bugs (that affected different cases) have been
1219 fixed.
1220
1221 39. Refactored the code to abolish the use of different opcodes for small
1222 capturing bracket numbers. This is a tidy that I avoided doing when I
1223 removed the limit on the number of capturing brackets for 3.5 back in 2001.
1224 The new approach is not only tidier, it makes it possible to reduce the
1225 memory needed to fix the previous bug (38).
1226
1227 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
1228 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
1229 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
1230 mode.
1231
1232 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
1233 report.
1234
1235 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
1236 copy construction and assignment in the C++ wrapper.
1237
1238 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
1239 couple of bugs that could have given wrong results in the "--newline=crlf"
1240 case.
1241
1242 44. Added a number of casts and did some reorganization of signed/unsigned int
1243 variables following suggestions from Dair Grant. Also renamed the variable
1244 "this" as "item" because it is a C++ keyword.
1245
1246 45. Arranged for dftables to add
1247
1248 #include "pcre_internal.h"
1249
1250 to pcre_chartables.c because without it, gcc 4.x may remove the array
1251 definition from the final binary if PCRE is built into a static library and
1252 dead code stripping is activated.
1253
1254 46. For an unanchored pattern, if a match attempt fails at the start of a
1255 newline sequence, and the newline setting is CRLF or ANY, and the next two
1256 characters are CRLF, advance by two characters instead of one.
1257
1258
1259 Version 6.7 04-Jul-06
1260 ---------------------
1261
1262 1. In order to handle tests when input lines are enormously long, pcretest has
1263 been re-factored so that it automatically extends its buffers when
1264 necessary. The code is crude, but this _is_ just a test program. The
1265 default size has been increased from 32K to 50K.
1266
1267 2. The code in pcre_study() was using the value of the re argument before
1268 testing it for NULL. (Of course, in any sensible call of the function, it
1269 won't be NULL.)
1270
1271 3. The memmove() emulation function in pcre_internal.h, which is used on
1272 systems that lack both memmove() and bcopy() - that is, hardly ever -
1273 was missing a "static" storage class specifier.
1274
1275 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
1276 containing an extended class (one that cannot be represented by a bitmap
1277 because it contains high-valued characters or Unicode property items, e.g.
1278 [\pZ]). Almost always one would set UTF-8 mode when processing such a
1279 pattern, but PCRE should not loop if you do not (it no longer does).
1280 [Detail: two cases were found: (a) a repeated subpattern containing an
1281 extended class; (b) a recursive reference to a subpattern that followed a
1282 previous extended class. It wasn't skipping over the extended class
1283 correctly when UTF-8 mode was not set.]
1284
1285 5. A negated single-character class was not being recognized as fixed-length
1286 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
1287 compile error "lookbehind assertion is not fixed length".
1288
1289 6. The RunPerlTest auxiliary script was showing an unexpected difference
1290 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
1291 write a Perl script that can interpret lines of an input file either as
1292 byte characters or as UTF-8, which is what "perltest" was being required to
1293 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
1294 can't do is switch easily at run time between having the "use utf8;" pragma
1295 or not. In the end, I fudged it by using the RunPerlTest script to insert
1296 "use utf8;" explicitly for the UTF-8 tests.
1297
1298 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
1299 the end of the subject string, contrary to the documentation and to what
1300 Perl does. This was true of both matching functions. Now it matches only at
1301 the start of the subject and immediately after *internal* newlines.
1302
1303 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
1304 a pointer to an int instead of a pointer to an unsigned long int. This
1305 caused problems on 64-bit systems.
1306
1307 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
1308 instance of the 'standard' template library not being so standard".
1309
1310 10. There was no check on the number of named subpatterns nor the maximum
1311 length of a subpattern name. The product of these values is used to compute
1312 the size of the memory block for a compiled pattern. By supplying a very
1313 long subpattern name and a large number of named subpatterns, the size
1314 computation could be caused to overflow. This is now prevented by limiting
1315 the length of names to 32 characters, and the number of named subpatterns
1316 to 10,000.
1317
1318 11. Subpatterns that are repeated with specific counts have to be replicated in
1319 the compiled pattern. The size of memory for this was computed from the
1320 length of the subpattern and the repeat count. The latter is limited to
1321 65535, but there was no limit on the former, meaning that integer overflow
1322 could in principle occur. The compiled length of a repeated subpattern is
1323 now limited to 30,000 bytes in order to prevent this.
1324
1325 12. Added the optional facility to have named substrings with the same name.
1326
1327 13. Added the ability to use a named substring as a condition, using the
1328 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
1329 are numbers (not recommended). Forward references are permitted.
1330
1331 14. Added forward references in named backreferences (if you see what I mean).
1332
1333 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
1334 pattern could run off the end of the subject. For example, the pattern
1335 "(?s)(.{1,5})"8 did this with the subject "ab".
1336
1337 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
1338 PCRE_CASELESS was set when matching characters that were quantified with ?
1339 or *.
1340
1341 17. A character class other than a single negated character that had a minimum
1342 but no maximum quantifier - for example [ab]{6,} - was not handled
1343 correctly by pce_dfa_exec(). It would match only one character.
1344
1345 18. A valid (though odd) pattern that looked like a POSIX character
1346 class but used an invalid character after [ (for example [[,abc,]]) caused
1347 pcre_compile() to give the error "Failed: internal error: code overflow" or
1348 in some cases to crash with a glibc free() error. This could even happen if
1349 the pattern terminated after [[ but there just happened to be a sequence of
1350 letters, a binary zero, and a closing ] in the memory that followed.
1351
1352 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
1353 over the years. Originally (before any Unicode support), just the bottom 8
1354 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
1355 output from "man perlunicode" includes this:
1356
1357 The regular expression compiler produces polymorphic opcodes. That
1358 is, the pattern adapts to the data and automatically switches to
1359 the Unicode character scheme when presented with Unicode data--or
1360 instead uses a traditional byte scheme when presented with byte
1361 data.
1362
1363 Sadly, a wide octal escape does not cause a switch, and in a string with
1364 no other multibyte characters, these octal escapes are treated as before.
1365 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
1366 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
1367 Unicode string.
1368
1369 I have not perpetrated such confusion in PCRE. Up till now, it took just
1370 the bottom 8 bits, as in old Perl. I have now made octal escapes with
1371 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
1372 translate to the appropriate multibyte character.
1373
1374 29. Applied some refactoring to reduce the number of warnings from Microsoft
1375 and Borland compilers. This has included removing the fudge introduced
1376 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
1377 a warning about an unused variable.
1378
1379 21. PCRE has not included VT (character 0x0b) in the set of whitespace
1380 characters since release 4.0, because Perl (from release 5.004) does not.
1381 [Or at least, is documented not to: some releases seem to be in conflict
1382 with the documentation.] However, when a pattern was studied with
1383 pcre_study() and all its branches started with \s, PCRE still included VT
1384 as a possible starting character. Of course, this did no harm; it just
1385 caused an unnecessary match attempt.
1386
1387 22. Removed a now-redundant internal flag bit that recorded the fact that case
1388 dependency changed within the pattern. This was once needed for "required
1389 byte" processing, but is no longer used. This recovers a now-scarce options
1390 bit. Also moved the least significant internal flag bit to the most-
1391 significant bit of the word, which was not previously used (hangover from
1392 the days when it was an int rather than a uint) to free up another bit for
1393 the future.
1394
1395 23. Added support for CRLF line endings as well as CR and LF. As well as the
1396 default being selectable at build time, it can now be changed at runtime
1397 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
1398 specify that it is scanning data with non-default line endings.
1399
1400 24. Changed the definition of CXXLINK to make it agree with the definition of
1401 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
1402
1403 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
1404 recursions. This makes a big different to stack usage for some patterns.
1405
1406 26. If a subpattern containing a named recursion or subroutine reference such
1407 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
1408 the space required for the compiled pattern went wrong and gave too small a
1409 value. Depending on the environment, this could lead to "Failed: internal
1410 error: code overflow at offset 49" or "glibc detected double free or
1411 corruption" errors.
1412
1413 27. Applied patches from Google (a) to support the new newline modes and (b) to
1414 advance over multibyte UTF-8 characters in GlobalReplace.
1415
1416 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
1417 difference for some implementation of PCRE in some Windows version.
1418
1419 29. Added some extra testing facilities to pcretest:
1420
1421 \q<number> in a data line sets the "match limit" value
1422 \Q<number> in a data line sets the "match recursion limt" value
1423 -S <number> sets the stack size, where <number> is in megabytes
1424
1425 The -S option isn't available for Windows.
1426
1427
1428 Version 6.6 06-Feb-06
1429 ---------------------
1430
1431 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
1432 in pcreposix.h. I have copied the definition from pcre.h.
1433
1434 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
1435 because pcre.h is no longer a built file.
1436
1437 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
1438 not normally included in the compiled code.
1439
1440
1441 Version 6.5 01-Feb-06
1442 ---------------------
1443
1444 1. When using the partial match feature with pcre_dfa_exec(), it was not
1445 anchoring the second and subsequent partial matches at the new starting
1446 point. This could lead to incorrect results. For example, with the pattern
1447 /1234/, partially matching against "123" and then "a4" gave a match.
1448
1449 2. Changes to pcregrep:
1450
1451 (a) All non-match returns from pcre_exec() were being treated as failures
1452 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
1453 error message is output. Some extra information is given for the
1454 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
1455 probably the only errors that are likely to be caused by users (by
1456 specifying a regex that has nested indefinite repeats, for instance).
1457 If there are more than 20 of these errors, pcregrep is abandoned.
1458
1459 (b) A binary zero was treated as data while matching, but terminated the
1460 output line if it was written out. This has been fixed: binary zeroes
1461 are now no different to any other data bytes.
1462
1463 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
1464 used to set a locale for matching. The --locale=xxxx long option has
1465 been added (no short equivalent) to specify a locale explicitly on the
1466 pcregrep command, overriding the environment variables.
1467
1468 (d) When -B was used with -n, some line numbers in the output were one less
1469 than they should have been.
1470
1471 (e) Added the -o (--only-matching) option.
1472
1473 (f) If -A or -C was used with -c (count only), some lines of context were
1474 accidentally printed for the final match.
1475
1476 (g) Added the -H (--with-filename) option.
1477
1478 (h) The combination of options -rh failed to suppress file names for files
1479 that were found from directory arguments.
1480
1481 (i) Added the -D (--devices) and -d (--directories) options.
1482
1483 (j) Added the -F (--fixed-strings) option.
1484
1485 (k) Allow "-" to be used as a file name for -f as well as for a data file.
1486
1487 (l) Added the --colo(u)r option.
1488
1489 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
1490 is not present by default.
1491
1492 3. A nasty bug was discovered in the handling of recursive patterns, that is,
1493 items such as (?R) or (?1), when the recursion could match a number of
1494 alternatives. If it matched one of the alternatives, but subsequently,
1495 outside the recursion, there was a failure, the code tried to back up into
1496 the recursion. However, because of the way PCRE is implemented, this is not
1497 possible, and the result was an incorrect result from the match.
1498
1499 In order to prevent this happening, the specification of recursion has
1500 been changed so that all such subpatterns are automatically treated as
1501 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
1502
1503 4. I had overlooked the fact that, in some locales, there are characters for
1504 which isalpha() is true but neither isupper() nor islower() are true. In
1505 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
1506 and ordfeminine) are like this. This affected the treatment of \w and \W
1507 when they appeared in character classes, but not when they appeared outside
1508 a character class. The bit map for "word" characters is now created
1509 separately from the results of isalnum() instead of just taking it from the
1510 upper, lower, and digit maps. (Plus the underscore character, of course.)
1511
1512 5. The above bug also affected the handling of POSIX character classes such as
1513 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
1514 permanent tables. Instead, the bit maps for such a class were previously
1515 created as the appropriate unions of the upper, lower, and digit bitmaps.
1516 Now they are created by subtraction from the [[:word:]] class, which has
1517 its own bitmap.
1518
1519 6. The [[:blank:]] character class matches horizontal, but not vertical space.
1520 It is created by subtracting the vertical space characters (\x09, \x0a,
1521 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
1522 subtraction was done in the overall bitmap for a character class, meaning
1523 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
1524 be recognized. This bug has been fixed.
1525
1526 7. Patches from the folks at Google:
1527
1528 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
1529 real life, but is still worth protecting against".
1530
1531 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
1532 regular expressions".
1533
1534 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
1535 have it.
1536
1537 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
1538 "configure" and the latter not, in order to fix a problem somebody had
1539 with compiling the Arg class on HP-UX.
1540
1541 (e) Improve the error-handling of the C++ wrapper a little bit.
1542
1543 (f) New tests for checking recursion limiting.
1544
1545 8. The pcre_memmove() function, which is used only if the environment does not
1546 have a standard memmove() function (and is therefore rarely compiled),
1547 contained two bugs: (a) use of int instead of size_t, and (b) it was not
1548 returning a result (though PCRE never actually uses the result).
1549
1550 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
1551 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
1552 returned instead of calling malloc() with an overflowing number that would
1553 most likely cause subsequent chaos.
1554
1555 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
1556
1557 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
1558 with this option is matched, the nmatch and pmatch options of regexec() are
1559 ignored.
1560
1561 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
1562 provided in case anyone wants to the the POSIX interface with UTF-8
1563 strings.
1564
1565 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
1566 C++ linking (needed for some HP-UX environments).
1567
1568 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
1569 (unused parameter) and in the pcre_printint() function (omitted "default"
1570 switch label when the default is to do nothing).
1571
1572 15. Added some code to make it possible, when PCRE is compiled as a C++
1573 library, to replace subject pointers for pcre_exec() with a smart pointer
1574 class, thus making it possible to process discontinuous strings.
1575
1576 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
1577 much the same function. They were added by different people who were trying
1578 to make PCRE easy to compile on non-Unix systems. It has been suggested
1579 that PCRE_EXPORT be abolished now that there is more automatic apparatus
1580 for compiling on Windows systems. I have therefore replaced it with
1581 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
1582 defaults to "extern" for C or "extern C" for C++, which works fine on
1583 Unix-like systems. It is now possible to override the value of PCRE_DATA_
1584 SCOPE with something explicit in config.h. In addition:
1585
1586 (a) pcreposix.h still had just "extern" instead of either of these macros;
1587 I have replaced it with PCRE_DATA_SCOPE.
1588
1589 (b) Functions such as _pcre_xclass(), which are internal to the library,
1590 but external in the C sense, all had PCRE_EXPORT in their definitions.
1591 This is apparently wrong for the Windows case, so I have removed it.
1592 (It makes no difference on Unix-like systems.)
1593
1594 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
1595 of recursive calls to match(). This is different to MATCH_LIMIT because
1596 that limits the total number of calls to match(), not all of which increase
1597 the depth of recursion. Limiting the recursion depth limits the amount of
1598 stack (or heap if NO_RECURSE is set) that is used. The default can be set
1599 when PCRE is compiled, and changed at run time. A patch from Google adds
1600 this functionality to the C++ interface.
1601
1602 18. Changes to the handling of Unicode character properties:
1603
1604 (a) Updated the table to Unicode 4.1.0.
1605
1606 (b) Recognize characters that are not in the table as "Cn" (undefined).
1607
1608 (c) I revised the way the table is implemented to a much improved format
1609 which includes recognition of ranges. It now supports the ranges that
1610 are defined in UnicodeData.txt, and it also amalgamates other
1611 characters into ranges. This has reduced the number of entries in the
1612 table from around 16,000 to around 3,000, thus reducing its size
1613 considerably. I realized I did not need to use a tree structure after
1614 all - a binary chop search is just as efficient. Having reduced the
1615 number of entries, I extended their size from 6 bytes to 8 bytes to
1616 allow for more data.
1617
1618 (d) Added support for Unicode script names via properties such as \p{Han}.
1619
1620 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
1621 matching that character.
1622
1623 20. When matching a repeated Unicode property with a minimum greater than zero,
1624 (for example \pL{2,}), PCRE could look past the end of the subject if it
1625 reached it while seeking the minimum number of characters. This could
1626 happen only if some of the characters were more than one byte long, because
1627 there is a check for at least the minimum number of bytes.
1628
1629 21. Refactored the implementation of \p and \P so as to be more general, to
1630 allow for more different types of property in future. This has changed the
1631 compiled form incompatibly. Anybody with saved compiled patterns that use
1632 \p or \P will have to recompile them.
1633
1634 22. Added "Any" and "L&" to the supported property types.
1635
1636 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
1637 but give a compile time error if the value is greater than 0xff.
1638
1639 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
1640 accidentally not being installed or uninstalled.
1641
1642 25. The pcre.h file was built from pcre.h.in, but the only changes that were
1643 made were to insert the current release number. This seemed silly, because
1644 it made things harder for people building PCRE on systems that don't run
1645 "configure". I have turned pcre.h into a distributed file, no longer built
1646 by "configure", with the version identification directly included. There is
1647 no longer a pcre.h.in file.
1648
1649 However, this change necessitated a change to the pcre-config script as
1650 well. It is built from pcre-config.in, and one of the substitutions was the
1651 release number. I have updated configure.ac so that ./configure now finds
1652 the release number by grepping pcre.h.
1653
1654 26. Added the ability to run the tests under valgrind.
1655
1656
1657 Version 6.4 05-Sep-05
1658 ---------------------
1659
1660 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
1661 "--" to be printed when multiple files were scanned, even when none of the
1662 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
1663 consider it to be a bug, and have restored the previous behaviour.
1664
1665 2. A couple of code tidies to get rid of compiler warnings.
1666
1667 3. The pcretest program used to cheat by referring to symbols in the library
1668 whose names begin with _pcre_. These are internal symbols that are not
1669 really supposed to be visible externally, and in some environments it is
1670 possible to suppress them. The cheating is now confined to including
1671 certain files from the library's source, which is a bit cleaner.
1672
1673 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
1674 file's purpose clearer.
1675
1676 5. Reorganized pcre_ucp_findchar().
1677
1678
1679 Version 6.3 15-Aug-05
1680 ---------------------
1681
1682 1. The file libpcre.pc.in did not have general read permission in the tarball.
1683
1684 2. There were some problems when building without C++ support:
1685
1686 (a) If C++ support was not built, "make install" and "make test" still
1687 tried to test it.
1688
1689 (b) There were problems when the value of CXX was explicitly set. Some
1690 changes have been made to try to fix these, and ...
1691
1692 (c) --disable-cpp can now be used to explicitly disable C++ support.
1693
1694 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
1695 backslash in a target when C++ was disabled. This confuses some
1696 versions of "make", apparently. Using an intermediate variable solves
1697 this. (Same for CPP_LOBJ.)
1698
1699 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
1700 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
1701 necessary on certain architectures.
1702
1703 4. Added a setting of -export-symbols-regex to the link command to remove
1704 those symbols that are exported in the C sense, but actually are local
1705 within the library, and not documented. Their names all begin with
1706 "_pcre_". This is not a perfect job, because (a) we have to except some
1707 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
1708 available (and never for static libraries). I have made a note to try to
1709 find a way round (a) in the future.
1710
1711
1712 Version 6.2 01-Aug-05
1713 ---------------------
1714
1715 1. There was no test for integer overflow of quantifier values. A construction
1716 such as {1111111111111111} would give undefined results. What is worse, if
1717 a minimum quantifier for a parenthesized subpattern overflowed and became
1718 negative, the calculation of the memory size went wrong. This could have
1719 led to memory overwriting.
1720
1721 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
1722
1723 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
1724 operating environments where this matters.
1725
1726 4. Applied Giuseppe Maxia's patch to add additional features for controlling
1727 PCRE options from within the C++ wrapper.
1728
1729 5. Named capturing subpatterns were not being correctly counted when a pattern
1730 was compiled. This caused two problems: (a) If there were more than 100
1731 such subpatterns, the calculation of the memory needed for the whole
1732 compiled pattern went wrong, leading to an overflow error. (b) Numerical
1733 back references of the form \12, where the number was greater than 9, were
1734 not recognized as back references, even though there were sufficient
1735 previous subpatterns.
1736
1737 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
1738 versions of gcc, e.g. 2.95.4.
1739
1740
1741 Version 6.1 21-Jun-05
1742 ---------------------
1743
1744 1. There was one reference to the variable "posix" in pcretest.c that was not
1745 surrounded by "#if !defined NOPOSIX".
1746
1747 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
1748 the cross-check on the old pcre_info() function, for the benefit of the
1749 cut-down version of PCRE that is currently imported into Exim.
1750
1751 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
1752 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
1753 patterns that start (?i) but I don't think that matters. The use of (?i) is
1754 just an example; this all applies to the other options as well.
1755
1756 4. Since libtool seems to echo the compile commands it is issuing, the output
1757 from "make" can be reduced a bit by putting "@" in front of each libtool
1758 compile command.
1759
1760 5. Patch from the folks at Google for configure.in to be a bit more thorough
1761 in checking for a suitable C++ installation before trying to compile the
1762 C++ stuff. This should fix a reported problem when a compiler was present,
1763 but no suitable headers.
1764
1765 6. The man pages all had just "PCRE" as their title. I have changed them to
1766 be the relevant file name. I have also arranged that these names are
1767 retained in the file doc/pcre.txt, which is a concatenation in text format
1768 of all the man pages except the little individual ones for each function.
1769
1770 7. The NON-UNIX-USE file had not been updated for the different set of source
1771 files that come with release 6. I also added a few comments about the C++
1772 wrapper.
1773
1774
1775 Version 6.0 07-Jun-05
1776 ---------------------
1777
1778 1. Some minor internal re-organization to help with my DFA experiments.
1779
1780 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
1781 didn't matter for the library itself when fully configured, but did matter
1782 when compiling without UCP support, or within Exim, where the ucp files are
1783 not imported.
1784
1785 3. Refactoring of the library code to split up the various functions into
1786 different source modules. The addition of the new DFA matching code (see
1787 below) to a single monolithic source would have made it really too
1788 unwieldy, quite apart from causing all the code to be include in a
1789 statically linked application, when only some functions are used. This is
1790 relevant even without the DFA addition now that patterns can be compiled in
1791 one application and matched in another.
1792
1793 The downside of splitting up is that there have to be some external
1794 functions and data tables that are used internally in different modules of
1795 the library but which are not part of the API. These have all had their
1796 names changed to start with "_pcre_" so that they are unlikely to clash
1797 with other external names.
1798
1799 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
1800 a different (DFA) algorithm. Although it is slower than the original
1801 function, it does have some advantages for certain types of matching
1802 problem.
1803
1804 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
1805 including restarting after a partial match.
1806
1807 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
1808 defined when compiling for Windows was sent to me. I have put it into the
1809 code, though I have no means of testing or verifying it.
1810
1811 7. Added the pcre_refcount() auxiliary function.
1812
1813 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
1814 match before or at the first newline in the subject string. In pcretest,
1815 the /f option on a pattern can be used to set this.
1816
1817 9. A repeated \w when used in UTF-8 mode with characters greater than 256
1818 would behave wrongly. This has been present in PCRE since release 4.0.
1819
1820 10. A number of changes to the pcregrep command:
1821
1822 (a) Refactored how -x works; insert ^(...)$ instead of setting
1823 PCRE_ANCHORED and checking the length, in preparation for adding
1824 something similar for -w.
1825
1826 (b) Added the -w (match as a word) option.
1827
1828 (c) Refactored the way lines are read and buffered so as to have more
1829 than one at a time available.
1830
1831 (d) Implemented a pcregrep test script.
1832
1833 (e) Added the -M (multiline match) option. This allows patterns to match
1834 over several lines of the subject. The buffering ensures that at least
1835 8K, or the rest of the document (whichever is the shorter) is available
1836 for matching (and similarly the previous 8K for lookbehind assertions).
1837
1838 (f) Changed the --help output so that it now says
1839
1840 -w, --word-regex(p)
1841
1842 instead of two lines, one with "regex" and the other with "regexp"
1843 because that confused at least one person since the short forms are the
1844 same. (This required a bit of code, as the output is generated
1845 automatically from a table. It wasn't just a text change.)
1846
1847 (g) -- can be used to terminate pcregrep options if the next thing isn't an
1848 option but starts with a hyphen. Could be a pattern or a path name
1849 starting with a hyphen, for instance.
1850
1851 (h) "-" can be given as a file name to represent stdin.
1852
1853 (i) When file names are being printed, "(standard input)" is used for
1854 the standard input, for compatibility with GNU grep. Previously
1855 "<stdin>" was used.
1856
1857 (j) The option --label=xxx can be used to supply a name to be used for
1858 stdin when file names are being printed. There is no short form.
1859
1860 (k) Re-factored the options decoding logic because we are going to add
1861 two more options that take data. Such options can now be given in four
1862 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
1863
1864 (l) Added the -A, -B, and -C options for requesting that lines of context
1865 around matches be printed.
1866
1867 (m) Added the -L option to print the names of files that do not contain
1868 any matching lines, that is, the complement of -l.
1869
1870 (n) The return code is 2 if any file cannot be opened, but pcregrep does
1871 continue to scan other files.
1872
1873 (o) The -s option was incorrectly implemented. For compatibility with other
1874 greps, it now suppresses the error message for a non-existent or non-
1875 accessible file (but not the return code). There is a new option called
1876 -q that suppresses the output of matching lines, which was what -s was
1877 previously doing.
1878
1879 (p) Added --include and --exclude options to specify files for inclusion
1880 and exclusion when recursing.
1881
1882 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
1883 Hopefully, it now does.
1884
1885 12. Missing cast in pcre_study().
1886
1887 13. Added an "uninstall" target to the makefile.
1888
1889 14. Replaced "extern" in the function prototypes in Makefile.in with
1890 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
1891 world, but is set differently for Windows.
1892
1893 15. Added a second compiling function called pcre_compile2(). The only
1894 difference is that it has an extra argument, which is a pointer to an
1895 integer error code. When there is a compile-time failure, this is set
1896 non-zero, in addition to the error test pointer being set to point to an
1897 error message. The new argument may be NULL if no error number is required
1898 (but then you may as well call pcre_compile(), which is now just a
1899 wrapper). This facility is provided because some applications need a
1900 numeric error indication, but it has also enabled me to tidy up the way
1901 compile-time errors are handled in the POSIX wrapper.
1902
1903 16. Added VPATH=.libs to the makefile; this should help when building with one
1904 prefix path and installing with another. (Or so I'm told by someone who
1905 knows more about this stuff than I do.)
1906
1907 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
1908 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
1909 match everything, including newlines. This is not POSIX-compatible, but
1910 somebody wanted the feature. From pcretest it can be activated by using
1911 both the P and the s flags.
1912
1913 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
1914
1915 19. libpcre.pc was being incorrectly installed as executable.
1916
1917 20. A couple of places in pcretest check for end-of-line by looking for '\n';
1918 it now also looks for '\r' so that it will work unmodified on Windows.
1919
1920 21. Added Google's contributed C++ wrapper to the distribution.
1921
1922 22. Added some untidy missing memory free() calls in pcretest, to keep
1923 Electric Fence happy when testing.
1924
1925
1926
1927 Version 5.0 13-Sep-04
1928 ---------------------
1929
1930 1. Internal change: literal characters are no longer packed up into items
1931 containing multiple characters in a single byte-string. Each character
1932 is now matched using a separate opcode. However, there may be more than one
1933 byte in the character in UTF-8 mode.
1934
1935 2. The pcre_callout_block structure has two new fields: pattern_position and
1936 next_item_length. These contain the offset in the pattern to the next match
1937 item, and its length, respectively.
1938
1939 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
1940 insertion of callouts before each pattern item. Added the /C option to
1941 pcretest to make use of this.
1942
1943 4. On the advice of a Windows user, the lines
1944
1945 #if defined(_WIN32) || defined(WIN32)
1946 _setmode( _fileno( stdout ), 0x8000 );
1947 #endif /* defined(_WIN32) || defined(WIN32) */
1948
1949 have been added to the source of pcretest. This apparently does useful
1950 magic in relation to line terminators.
1951
1952 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
1953 for the benefit of those environments where the "b" makes a difference.
1954
1955 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
1956 to know about it. I have put a hack into configure.in that adds in code
1957 to set GCC=yes if CC=icc. This seems to end up at a point in the
1958 generated configure script that is early enough to affect the setting of
1959 compiler options, which is what is needed, but I have no means of testing
1960 whether it really works. (The user who reported this had patched the
1961 generated configure script, which of course I cannot do.)
1962
1963 LATER: After change 22 below (new libtool files), the configure script
1964 seems to know about icc (and also ecc). Therefore, I have commented out
1965 this hack in configure.in.
1966
1967 7. Added support for pkg-config (2 patches were sent in).
1968
1969 8. Negated POSIX character classes that used a combination of internal tables
1970 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
1971 [[:^ascii]]. Typically, they would match almost any characters. The other
1972 POSIX classes were not broken in this way.
1973
1974 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
1975 to find the match, as PCRE was deluded into thinking that the match had to
1976 start at the start point or following a newline. The same bug applied to
1977 patterns with negative forward assertions or any backward assertions
1978 preceding ".*" at the start, unless the pattern required a fixed first
1979 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
1980
1981 10. In UTF-8 mode, when moving forwards in the subject after a failed match
1982 starting at the last subject character, bytes beyond the end of the subject
1983 string were read.
1984
1985 11. Renamed the variable "class" as "classbits" to make life easier for C++
1986 users. (Previously there was a macro definition, but it apparently wasn't
1987 enough.)
1988
1989 12. Added the new field "tables" to the extra data so that tables can be passed
1990 in at exec time, or the internal tables can be re-selected. This allows
1991 a compiled regex to be saved and re-used at a later time by a different
1992 program that might have everything at different addresses.
1993
1994 13. Modified the pcre-config script so that, when run on Solaris, it shows a
1995 -R library as well as a -L library.
1996
1997 14. The debugging options of pcretest (-d on the command line or D on a
1998 pattern) showed incorrect output for anything following an extended class
1999 that contained multibyte characters and which was followed by a quantifier.
2000
2001 15. Added optional support for general category Unicode character properties
2002 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
2003 support. It adds about 90K to the size of the library. The meanings of the
2004 inbuilt class escapes such as \d and \s have NOT been changed.
2005
2006 16. Updated pcredemo.c to include calls to free() to release the memory for the
2007 compiled pattern.
2008
2009 17. The generated file chartables.c was being created in the source directory
2010 instead of in the building directory. This caused the build to fail if the
2011 source directory was different from the building directory, and was
2012 read-only.
2013
2014 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
2015 file. No doubt somebody will tell me if they don't make sense... Also added
2016 Dan Mooney's comments about building on OpenVMS.
2017
2018 19. Added support for partial matching via the PCRE_PARTIAL option for
2019 pcre_exec() and the \P data escape in pcretest.
2020
2021 20. Extended pcretest with 3 new pattern features:
2022
2023 (i) A pattern option of the form ">rest-of-line" causes pcretest to
2024 write the compiled pattern to the file whose name is "rest-of-line".
2025 This is a straight binary dump of the data, with the saved pointer to
2026 the character tables forced to be NULL. The study data, if any, is
2027 written too. After writing, pcretest reads a new pattern.
2028
2029 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
2030 compiled pattern from the given file. There must not be any
2031 occurrences of "<" in the file name (pretty unlikely); if there are,
2032 pcretest will instead treat the initial "<" as a pattern delimiter.
2033 After reading in the pattern, pcretest goes on to read data lines as
2034 usual.
2035
2036 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
2037 and 16-bit fields in a compiled pattern, to simulate a pattern that
2038 was compiled on a host of opposite endianness.
2039
2040 21. The pcre-exec() function can now cope with patterns that were compiled on
2041 hosts of opposite endianness, with this restriction:
2042
2043 As for any compiled expression that is saved and used later, the tables
2044 pointer field cannot be preserved; the extra_data field in the arguments
2045 to pcre_exec() should be used to pass in a tables address if a value
2046 other than the default internal tables were used at compile time.
2047
2048 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
2049 now diagnosed as an error. Previously, most of the time, a negative number
2050 would have been treated as zero, but if in addition "ovector" was passed as
2051 NULL, a crash could occur.
2052
2053 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
2054 new versions from the libtool 1.5 distribution (the last one is a copy of
2055 a file called libtool.m4). This seems to have fixed the need to patch
2056 "configure" to support Darwin 1.3 (which I used to do). However, I still
2057 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
2058 workstation).
2059
2060 24. Changed the PCRE licence to be the more standard "BSD" licence.
2061
2062
2063 Version 4.5 01-Dec-03
2064 ---------------------
2065
2066 1. There has been some re-arrangement of the code for the match() function so
2067 that it can be compiled in a version that does not call itself recursively.
2068 Instead, it keeps those local variables that need separate instances for
2069 each "recursion" in a frame on the heap, and gets/frees frames whenever it
2070 needs to "recurse". Keeping track of where control must go is done by means
2071 of setjmp/longjmp. The whole thing is implemented by a set of macros that
2072 hide most of the details from the main code, and operates only if
2073 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
2074 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
2075 operating.
2076
2077 To make it easier for callers to provide specially tailored get/free
2078 functions for this usage, two new functions, pcre_stack_malloc, and
2079 pcre_stack_free, are used. They are always called in strict stacking order,
2080 and the size of block requested is always the same.
2081
2082 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
2083 PCRE has been compiled to use the stack or the heap for recursion. The
2084 -C option of pcretest uses this to show which version is compiled.
2085
2086 A new data escape \S, is added to pcretest; it causes the amounts of store
2087 obtained and freed by both kinds of malloc/free at match time to be added
2088 to the output.
2089
2090 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
2091 what's available on my current Linux desktop machine.
2092
2093 3. When matching a UTF-8 string, the test for a valid string at the start has
2094 been extended. If start_offset is not zero, PCRE now checks that it points
2095 to a byte that is the start of a UTF-8 character. If not, it returns
2096 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
2097 this is necessary because there may be backward assertions in the pattern.
2098 When matching the same subject several times, it may save resources to use
2099 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
2100
2101 4. The code for checking the validity of UTF-8 strings has been tightened so
2102 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
2103 containing "overlong sequences".
2104
2105 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
2106 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
2107 should have been "&", but it just so happened that all the cases this let
2108 through by mistake were picked up later in the function.
2109
2110 6. I had used a variable called "isblank" - this is a C99 function, causing
2111 some compilers to warn. To avoid this, I renamed it (as "blankclass").
2112
2113 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
2114 prompting; (b) run "./pcretest /dev/null" at the start of the test script
2115 so the version is shown; (c) stop "make test" echoing "./RunTest".
2116
2117 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
2118
2119 9. The prototype for memmove() for systems that don't have it was using
2120 size_t, but the inclusion of the header that defines size_t was later. I've
2121 moved the #includes for the C headers earlier to avoid this.
2122
2123 10. Added some adjustments to the code to make it easier to compiler on certain
2124 special systems:
2125
2126 (a) Some "const" qualifiers were missing.
2127 (b) Added the macro EXPORT before all exported functions; by default this
2128 is defined to be empty.
2129 (c) Changed the dftables auxiliary program (that builds chartables.c) so
2130 that it reads its output file name as an argument instead of writing
2131 to the standard output and assuming this can be redirected.
2132
2133 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
2134 class containing characters with values greater than 255, PCRE compilation
2135 went into a loop.
2136
2137 12. A recursive reference to a subpattern that was within another subpattern
2138 that had a minimum quantifier of zero caused PCRE to crash. For example,
2139 (x(y(?2))z)? provoked this bug with a subject that got as far as the
2140 recursion. If the recursively-called subpattern itself had a zero repeat,
2141 that was OK.
2142
2143 13. In pcretest, the buffer for reading a data line was set at 30K, but the
2144 buffer into which it was copied (for escape processing) was still set at
2145 1024, so long lines caused crashes.
2146
2147 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
2148 "internal error: code overflow...". This applied to any character class
2149 that was followed by a possessive quantifier.
2150
2151 15. Modified the Makefile to add libpcre.la as a prerequisite for
2152 libpcreposix.la because I was told this is needed for a parallel build to
2153 work.
2154
2155 16. If a pattern that contained .* following optional items at the start was
2156 studied, the wrong optimizing data was generated, leading to matching
2157 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
2158 matching string must start with a or b or c. The correct conclusion for
2159 this pattern is that a match can start with any character.
2160
2161
2162 Version 4.4 13-Aug-03
2163 ---------------------
2164
2165 1. In UTF-8 mode, a character class containing characters with values between
2166 127 and 255 was not handled correctly if the compiled pattern was studied.
2167 In fixing this, I have also improved the studying algorithm for such
2168 classes (slightly).
2169
2170 2. Three internal functions had redundant arguments passed to them. Removal
2171 might give a very teeny performance improvement.
2172
2173 3. Documentation bug: the value of the capture_top field in a callout is *one
2174 more than* the number of the hightest numbered captured substring.
2175
2176 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
2177 in incorrectly linking with a previously installed version. They now link
2178 explicitly with libpcre.la.
2179
2180 5. configure.in no longer needs to recognize Cygwin specially.
2181
2182 6. A problem in pcre.in for Windows platforms is fixed.
2183
2184 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
2185 pcretest, it used to include the size of the study block as part of its
2186 output. Unfortunately, the structure contains a field that has a different
2187 size on different hardware architectures. This meant that the tests that
2188 showed this size failed. As the block is currently always of a fixed size,
2189 this information isn't actually particularly useful in pcretest output, so
2190 I have just removed it.
2191
2192 8. Three pre-processor statements accidentally did not start in column 1.
2193 Sadly, there are *still* compilers around that complain, even though
2194 standard C has not required this for well over a decade. Sigh.
2195
2196 9. In pcretest, the code for checking callouts passed small integers in the
2197 callout_data field, which is a void * field. However, some picky compilers
2198 complained about the casts involved for this on 64-bit systems. Now
2199 pcretest passes the address of the small integer instead, which should get
2200 rid of the warnings.
2201
2202 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
2203 both compile and run time, and gives an error if an invalid UTF-8 sequence
2204 is found. There is a option for disabling this check in cases where the
2205 string is known to be correct and/or the maximum performance is wanted.
2206
2207 11. In response to a bug report, I changed one line in Makefile.in from
2208
2209 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
2210 to
2211 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
2212
2213 to look similar to other lines, but I have no way of telling whether this
2214 is the right thing to do, as I do not use Windows. No doubt I'll get told
2215 if it's wrong...
2216
2217
2218 Version 4.3 21-May-03
2219 ---------------------
2220
2221 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
2222 Makefile.
2223
2224 2. Some refactoring to improve the quality of the code:
2225
2226 (i) The utf8_table... variables are now declared "const".
2227
2228 (ii) The code for \cx, which used the "case flipping" table to upper case
2229 lower case letters, now just substracts 32. This is ASCII-specific,
2230 but the whole concept of \cx is ASCII-specific, so it seems
2231 reasonable.
2232
2233 (iii) PCRE was using its character types table to recognize decimal and
2234 hexadecimal digits in the pattern. This is silly, because it handles
2235 only 0-9, a-f, and A-F, but the character types table is locale-
2236 specific, which means strange things might happen. A private
2237 table is now used for this - though it costs 256 bytes, a table is
2238 much faster than multiple explicit tests. Of course, the standard
2239 character types table is still used for matching digits in subject
2240 strings against \d.
2241
2242 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
2243 ending in _t are). So I've renamed it as ESC_tee.
2244
2245 3. The first argument for regexec() in the POSIX wrapper should have been
2246 defined as "const".
2247
2248 4. Changed pcretest to use malloc() for its buffers so that they can be
2249 Electric Fenced for debugging.
2250
2251 5. There were several places in the code where, in UTF-8 mode, PCRE would try
2252 to read one or more bytes before the start of the subject string. Often this
2253 had no effect on PCRE's behaviour, but in some circumstances it could
2254 provoke a segmentation fault.
2255
2256 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
2257 to try to read one or more bytes before the start of the subject string.
2258
2259 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
2260 UTF-8 support could misbehave in various ways if the subject string
2261 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
2262 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
2263 back over UTF-8 characters.)
2264
2265
2266 Version 4.2 14-Apr-03
2267 ---------------------
2268
2269 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
2270
2271 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
2272 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
2273 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
2274 [WIN_PREFIX]: new variable, "cyg" for Cygwin
2275 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
2276 and BUILD_EXEEXT
2277 Note: automatic setting of the BUILD variables is not yet working
2278 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
2279 compile-time but not at link-time
2280 [LINK]: use for linking executables only
2281 make different versions for Windows and non-Windows
2282 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
2283 libraries
2284 [LINK_FOR_BUILD]: new variable
2285 [OBJEXT]: use throughout
2286 [EXEEXT]: use throughout
2287 <winshared>: new target
2288 <wininstall>: new target
2289 <dftables.o>: use native compiler
2290 <dftables>: use native linker
2291 <install>: handle Windows platform correctly
2292 <clean>: ditto
2293 <check>: ditto
2294 copy DLL to top builddir before testing
2295
2296 As part of these changes, -no-undefined was removed again. This was reported
2297 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
2298 in any case.
2299
2300 3. Some tidies to get rid of compiler warnings:
2301
2302 . In the match_data structure, match_limit was an unsigned long int, whereas
2303 match_call_count was an int. I've made them both unsigned long ints.
2304
2305 . In pcretest the fact that a const uschar * doesn't automatically cast to
2306 a void * provoked a warning.
2307
2308 . Turning on some more compiler warnings threw up some "shadow" variables
2309 and a few more missing casts.
2310
2311 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
2312 option, a class that contained a single character with a value between 128
2313 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
2314
2315 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
2316 option, a class that contained several characters, but with at least one
2317 whose value was between 128 and 255 caused PCRE to crash.
2318
2319
2320 Version 4.1 12-Mar-03
2321 ---------------------
2322
2323 1. Compiling with gcc -pedantic found a couple of places where casts were
2324 needed, and a string in dftables.c that was longer than standard compilers are
2325 required to support.
2326
2327 2. Compiling with Sun's compiler found a few more places where the code could
2328 be tidied up in order to avoid warnings.
2329
2330 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
2331 first of these names is deprecated in the latest Autoconf in favour of the name
2332 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
2333 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
2334 analogy I have changed it to CFLAGS_FOR_BUILD.
2335
2336 4. Added -no-undefined to the linking command in the Makefile, because this is
2337 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
2338 linking step for the pcreposix library.
2339
2340 5. PCRE was failing to diagnose the case of two named groups with the same
2341 name.
2342
2343 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
2344 literal character that is needed in the subject for a match, and scans along to
2345 ensure that it is present before embarking on the full matching process. This
2346 saves time in cases of nested unlimited repeats that are never going to match.
2347 Problem: the scan can take a lot of time if the subject is very long (e.g.
2348 megabytes), thus penalizing straightforward matches. It is now done only if the
2349 amount of subject to be scanned is less than 1000 bytes.
2350
2351 7. A lesser problem with the same optimization is that it was recording the
2352 first character of an anchored pattern as "needed", thus provoking a search
2353 right along the subject, even when the first match of the pattern was going to
2354 fail. The "needed" character is now not set for anchored patterns, unless it
2355 follows something in the pattern that is of non-fixed length. Thus, it still
2356 fulfils its original purpose of finding quick non-matches in cases of nested
2357 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
2358
2359
2360 Version 4.0 17-Feb-03
2361 ---------------------
2362
2363 1. If a comment in an extended regex that started immediately after a meta-item
2364 extended to the end of string, PCRE compiled incorrect data. This could lead to
2365 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
2366
2367 2. Moved to autoconf 2.53 and libtool 1.4.2.
2368
2369 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
2370 the special perltest8 script is no longer needed - all the tests can be run
2371 from a single perltest script.
2372
2373 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
2374 by \s. It has now been removed in PCRE. This means it isn't recognized as
2375 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
2376 class [:space:] *does* include VT, thereby creating a mess.
2377
2378 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
2379 space and tab.
2380
2381 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
2382 its new features into the main test script, reducing the number of scripts.
2383
2384 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
2385 were backward compatible, and made the (?i) apply to the whole pattern, as if
2386 /i were given. Now it behaves more logically, and applies the option setting
2387 only to what follows. PCRE has been changed to follow suit. However, if it
2388 finds options settings right at the start of the pattern, it extracts them into
2389 the global options, as before. Thus, they show up in the info data.
2390
2391 8. Added support for the \Q...\E escape sequence. Characters in between are
2392 treated as literals. This is slightly different from Perl in that $ and @ are
2393 also handled as literals inside the quotes. In Perl, they will cause variable
2394 interpolation. Note the following examples:
2395
2396 Pattern PCRE matches Perl matches
2397
2398 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
2399 \Qabc\$xyz\E abc\$xyz abc\$xyz
2400 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
2401
2402 For compatibility with Perl, \Q...\E sequences are recognized inside character
2403 classes as well as outside them.
2404
2405 9. Re-organized 3 code statements in pcretest to avoid "overflow in
2406 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
2407 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
2408 signed/unsigned warnings.
2409
2410 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
2411 option for pcretest, so I've replaced it by a simple function that does just
2412 that job.
2413
2414 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
2415 "pcregrep -".
2416
2417 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
2418 Java package. This provides some syntactic sugar for simple cases of what my
2419 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
2420 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
2421 item, you can use this simplified notation. Note that only makes sense with
2422 greedy quantifiers. Consequently, the use of the possessive quantifier forces
2423 greediness, whatever the setting of the PCRE_UNGREEDY option.
2424
2425 13. A change of greediness default within a pattern was not taking effect at
2426 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
2427 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
2428 was abstracted outside.
2429
2430 14. PCRE now supports the \G assertion. It is true when the current matching
2431 position is at the start point of the match. This differs from \A when the
2432 starting offset is non-zero. Used with the /g option of pcretest (or similar
2433 code), it works in the same way as it does for Perl's /g option. If all
2434 alternatives of a regex begin with \G, the expression is anchored to the start
2435 match position, and the "anchored" flag is set in the compiled expression.
2436
2437 15. Some bugs concerning the handling of certain option changes within patterns
2438 have been fixed. These applied to options other than (?ims). For example,
2439 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
2440 been the other way round. Some of this was related to change 7 above.
2441
2442 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
2443 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
2444 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
2445 POSIX classes only within a class (e.g. /[[:alpha:]]/).
2446
2447 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
2448 mode. Unlike ".", it always matches newline, whatever the setting of
2449 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
2450 assertions. Perl allows it, but it doesn't (in general) work because it can't
2451 calculate the length of the lookbehind. At least, that's the case for Perl
2452 5.8.0 - I've been told they are going to document that it doesn't work in
2453 future.
2454
2455 18. Added an error diagnosis for escapes that PCRE does not support: these are
2456 \L, \l, \N, \P, \p, \U, \u, and \X.
2457
2458 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
2459 reading past the end of the pattern in cases such as /[abcd/.
2460
2461 20. PCRE was getting more memory than necessary for patterns with classes that
2462 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
2463
2464 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
2465 compiling PCRE for use with Virtual Pascal.
2466
2467 22. Small fix to the Makefile to make it work properly if the build is done
2468 outside the source tree.
2469
2470 23. Added a new extension: a condition to go with recursion. If a conditional
2471 subpattern starts with (?(R) the "true" branch is used if recursion has
2472 happened, whereas the "false" branch is used only at the top level.
2473
2474 24. When there was a very long string of literal characters (over 255 bytes
2475 without UTF support, over 250 bytes with UTF support), the computation of how
2476 much memory was required could be incorrect, leading to segfaults or other
2477 strange effects.
2478
2479 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
2480 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
2481 there was a subsequent back reference to those brackets. This meant that, for
2482 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
2483 possible to check for precisely this case. All we can do is abandon the
2484 optimization if .* occurs inside capturing brackets when there are any back
2485 references whatsoever. (See below for a better fix that came later.)
2486
2487 26. The handling of the optimization for finding the first character of a
2488 non-anchored pattern, and for finding a character that is required later in the
2489 match were failing in some cases. This didn't break the matching; it just
2490 failed to optimize when it could. The way this is done has been re-implemented.
2491
2492 27. Fixed typo in error message for invalid (?R item (it said "(?p").
2493
2494 28. Added a new feature that provides some of the functionality that Perl
2495 provides with (?{...}). The facility is termed a "callout". The way it is done
2496 in PCRE is for the caller to provide an optional function, by setting
2497 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
2498 global variable. By default it is unset, which disables all calling out. To get
2499 the function called, the regex must include (?C) at appropriate points. This
2500 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
2501 This provides a means of identifying different callout points. When PCRE
2502 reaches such a point in the regex, if pcre_callout has been set, the external
2503 function is called. It is provided with data in a structure called
2504 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
2505 matching continues; if it returns a non-zero value, the match at the current
2506 point fails. However, backtracking will occur if possible. [This was changed
2507 later and other features added - see item 49 below.]
2508
2509 29. pcretest is upgraded to test the callout functionality. It provides a
2510 callout function that displays information. By default, it shows the start of
2511 the match and the current position in the text. There are some new data escapes
2512 to vary what happens:
2513
2514 \C+ in addition, show current contents of captured substrings
2515 \C- do not supply a callout function
2516 \C!n return 1 when callout number n is reached
2517 \C!n!m return 1 when callout number n is reached for the mth time
2518
2519 30. If pcregrep was called with the -l option and just a single file name, it
2520 output "<stdin>" if a match was found, instead of the file name.
2521
2522 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
2523 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
2524 pcre_exec(). This saves a malloc/free per call. The default value of
2525 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
2526 when configuring.
2527
2528 32. The default maximum size of a compiled pattern is 64K. There have been a
2529 few cases of people hitting this limit. The code now uses macros to handle the
2530 storing of links as offsets within the compiled pattern. It defaults to 2-byte
2531 links, but this can be changed to 3 or 4 bytes by --with-link-size when
2532 configuring. Tests 2 and 5 work only with 2-byte links because they output
2533 debugging information about compiled patterns.
2534
2535 33. Internal code re-arrangements:
2536
2537 (a) Moved the debugging function for printing out a compiled regex into
2538 its own source file (printint.c) and used #include to pull it into
2539 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
2540 separate copies.
2541
2542 (b) Defined the list of op-code names for debugging as a macro in
2543 internal.h so that it is next to the definition of the opcodes.
2544
2545 (c) Defined a table of op-code lengths for simpler skipping along compiled
2546 code. This is again a macro in internal.h so that it is next to the
2547 definition of the opcodes.
2548
2549 34. Added support for recursive calls to individual subpatterns, along the
2550 lines of Robin Houston's patch (but implemented somewhat differently).
2551
2552 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
2553 allow it to read and process whole directories in Win32. This code was
2554 contributed by Lionel Fourquaux; it has not been tested by me.
2555
2556 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
2557 used to name a group. Names consist of alphanumerics and underscores, and must
2558 be unique. Back references use the syntax (?P=name) and recursive calls use
2559 (?P>name) which is a PCRE extension to the Python extension. Groups still have
2560 numbers. The function pcre_fullinfo() can be used after compilation to extract
2561 a name/number map. There are three relevant calls:
2562
2563 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
2564 PCRE_INFO_NAMECOUNT yields the number of entries
2565 PCRE_INFO_NAMETABLE yields a pointer to the map.
2566
2567 The map is a vector of fixed-size entries. The size of each entry depends on
2568 the length of the longest name used. The first two bytes of each entry are the
2569 group number, most significant byte first. There follows the corresponding
2570 name, zero terminated. The names are in alphabetical order.
2571
2572 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
2573 case instead of 255. Making it the same both with and without UTF-8 support
2574 means that the same test output works with both.
2575
2576 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
2577 calling malloc() with a zero argument.
2578
2579 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
2580 optimization. I've improved things by keeping a bitmap of backreferences with
2581 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
2582 fact referenced, the optimization can be applied. It is unlikely that a
2583 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
2584 the match to follow \n) will appear inside brackets with a number greater than
2585 31, but if it does, any back reference > 31 suppresses the optimization.
2586
2587 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
2588 of disabling numbered capturing parentheses. Any opening parenthesis that is
2589 not followed by ? behaves as if it were followed by ?: but named parentheses
2590 can still be used for capturing (and they will acquire numbers in the usual
2591 way).
2592
2593 41. Redesigned the return codes from the match() function into yes/no/error so
2594 that errors can be passed back from deep inside the nested calls. A malloc
2595 failure while inside a recursive subpattern call now causes the
2596 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
2597
2598 42. It is now possible to set a limit on the number of times the match()
2599 function is called in a call to pcre_exec(). This facility makes it possible to
2600 limit the amount of recursion and backtracking, though not in a directly
2601 obvious way, because the match() function is used in a number of different
2602 circumstances. The count starts from zero for each position in the subject
2603 string (for non-anchored patterns). The default limit is, for compatibility, a
2604 large number, namely 10 000 000. You can change this in two ways:
2605
2606 (a) When configuring PCRE before making, you can use --with-match-limit=n
2607 to set a default value for the compiled library.
2608
2609 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
2610 a different value is set. See 45 below.
2611
2612 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
2613
2614 43. Added a new function pcre_config(int, void *) to enable run-time extraction
2615 of things that can be changed at compile time. The first argument specifies
2616 what is wanted and the second points to where the information is to be placed.
2617 The current list of available information is:
2618
2619 PCRE_CONFIG_UTF8
2620
2621 The output is an integer that is set to one if UTF-8 support is available;
2622 otherwise it is set to zero.
2623
2624 PCRE_CONFIG_NEWLINE
2625
2626 The output is an integer that it set to the value of the code that is used for
2627 newline. It is either LF (10) or CR (13).
2628
2629 PCRE_CONFIG_LINK_SIZE
2630
2631 The output is an integer that contains the number of bytes used for internal
2632 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
2633
2634 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
2635
2636 The output is an integer that contains the threshold above which the POSIX
2637 interface uses malloc() for output vectors. See item 31 above.
2638
2639 PCRE_CONFIG_MATCH_LIMIT
2640
2641 The output is an unsigned integer that contains the default limit of the number
2642 of match() calls in a pcre_exec() execution. See 42 above.
2643
2644 44. pcretest has been upgraded by the addition of the -C option. This causes it
2645 to extract all the available output from the new pcre_config() function, and to
2646 output it. The program then exits immediately.
2647
2648 45. A need has arisen to pass over additional data with calls to pcre_exec() in
2649 order to support additional features. One way would have been to define
2650 pcre_exec2() (for example) with extra arguments, but this would not have been
2651 extensible, and would also have required all calls to the original function to
2652 be mapped to the new one. Instead, I have chosen to extend the mechanism that
2653 is used for passing in "extra" data from pcre_study().
2654
2655 The pcre_extra structure is now exposed and defined in pcre.h. It currently
2656 contains the following fields:
2657
2658 flags a bitmap indicating which of the following fields are set
2659 study_data opaque data from pcre_study()
2660 match_limit a way of specifying a limit on match() calls for a specific
2661 call to pcre_exec()
2662 callout_data data for callouts (see 49 below)
2663
2664 The flag bits are also defined in pcre.h, and are
2665
2666 PCRE_EXTRA_STUDY_DATA
2667 PCRE_EXTRA_MATCH_LIMIT
2668 PCRE_EXTRA_CALLOUT_DATA
2669
2670 The pcre_study() function now returns one of these new pcre_extra blocks, with
2671 the actual study data pointed to by the study_data field, and the
2672 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
2673 before. That is, this change is entirely upwards-compatible and requires no
2674 change to existing code.
2675
2676 If you want to pass in additional data to pcre_exec(), you can either place it
2677 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
2678 block.
2679
2680 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
2681 data string contains the escape sequence \M, pcretest calls pcre_exec() several
2682 times with different match limits, until it finds the minimum value needed for
2683 pcre_exec() to complete. The value is then output. This can be instructive; for
2684 most simple matches the number is quite small, but for pathological cases it
2685 gets very large very quickly.
2686
2687 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
2688 returns the size of the data block pointed to by the study_data field in a
2689 pcre_extra block, that is, the value that was passed as the argument to
2690 pcre_malloc() when PCRE was getting memory in which to place the information
2691 created by pcre_study(). The fourth argument should point to a size_t variable.
2692 pcretest has been extended so that this information is shown after a successful
2693 pcre_study() call when information about the compiled regex is being displayed.
2694
2695 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
2696 because what follows is always an absolute path. (Later: it turns out that this
2697 is more than cosmetic for MinGW, because it doesn't like empty path
2698 components.)
2699
2700 49. Some changes have been made to the callout feature (see 28 above):
2701
2702 (i) A callout function now has three choices for what it returns:
2703
2704 0 => success, carry on matching
2705 > 0 => failure at this point, but backtrack if possible
2706 < 0 => serious error, return this value from pcre_exec()
2707
2708 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
2709 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
2710 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
2711 use by callout functions. It will never be used by PCRE itself.
2712
2713 (ii) The pcre_extra structure (see 45 above) has a void * field called
2714 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
2715 pcre_callout_block structure has a field of the same name. The contents of
2716 the field passed in the pcre_extra structure are passed to the callout
2717 function in the corresponding field in the callout block. This makes it
2718 easier to use the same callout-containing regex from multiple threads. For
2719 testing, the pcretest program has a new data escape
2720
2721 \C*n pass the number n (may be negative) as callout_data
2722
2723 If the callout function in pcretest receives a non-zero value as
2724 callout_data, it returns that value.
2725
2726 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
2727 there were some redundant $(CFLAGS) in commands that are now specified as
2728 $(LINK), which already includes $(CFLAGS).
2729
2730 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
2731 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
2732 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
2733 one-byte characters throughout. Note that case-insensitive matching applies
2734 only to characters whose values are less than 256. PCRE doesn't support the
2735 notion of cases for higher-valued characters.
2736
2737 (i) A character class whose characters are all within 0-255 is handled as
2738 a bit map, and the map is inverted for negative classes. Previously, a
2739 character > 255 always failed to match such a class; however it should
2740 match if the class was a negative one (e.g. [^ab]). This has been fixed.
2741
2742 (ii) A negated character class with a single character < 255 is coded as
2743 "not this character" (OP_NOT). This wasn't working properly when the test
2744 character was multibyte, either singly or repeated.
2745
2746 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
2747 mode, for example: \x{100}{2,3}.
2748
2749 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
2750 singly or repeated) now correctly test multibyte characters. However,
2751 PCRE doesn't recognize any characters with values greater than 255 as
2752 digits, spaces, or word characters. Such characters always match \D, \S,
2753 and \W, and never match \d, \s, or \w.
2754
2755 (v) Classes may now contain characters and character ranges with values
2756 greater than 255. For example: [ab\x{100}-\x{400}].
2757
2758 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
2759 PCRE in UTF-8 mode.
2760
2761 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
2762 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
2763 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
2764 value.)
2765
2766 53. The single man page has become too large. I have therefore split it up into
2767 a number of separate man pages. These also give rise to individual HTML pages;
2768 these are now put in a separate directory, and there is an index.html page that
2769 lists them all. Some hyperlinking between the pages has been installed.
2770
2771 54. Added convenience functions for handling named capturing parentheses.
2772
2773 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
2774 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
2775 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
2776 are faulted.
2777
2778 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
2779 calling configure. These values are used when compiling the dftables.c program
2780 which is run to generate the source of the default character tables. They
2781 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
2782 you will need to set these values.
2783
2784 57. Updated the building process for Windows DLL, as provided by Fred Cox.
2785
2786
2787 Version 3.9 02-Jan-02
2788 ---------------------
2789
2790 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
2791
2792 2. If --disable-static was given, the building process failed when trying to
2793 build pcretest and pcregrep. (For some reason it was using libtool to compile
2794 them, which is not right, as they aren't part of the library.)
2795
2796
2797 Version 3.8 18-Dec-01
2798 ---------------------
2799
2800 1. The experimental UTF-8 code was completely screwed up. It was packing the
2801 bytes in the wrong order. How dumb can you get?
2802
2803
2804 Version 3.7 29-Oct-01
2805 ---------------------
2806
2807 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
2808 This caused pcretest, when used on the test data, to segfault. Unfortunately,
2809 this didn't happen under Solaris 8, where I normally test things.
2810
2811 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
2812 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
2813 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
2814 this fix an hour or so after the initial 3.7 release.)
2815
2816
2817 Version 3.6 23-Oct-01
2818 ---------------------
2819
2820 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
2821 offsets passed as NULL with zero offset count.
2822
2823 2. The config.guess and config.sub files had not been updated when I moved to
2824 the latest autoconf.
2825
2826
2827 Version 3.5 15-Aug-01
2828 ---------------------
2829
2830 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
2831 had been forgotten.
2832
2833 2. By using declared but undefined structures, we can avoid using "void"
2834 definitions in pcre.h while keeping the internal definitions of the structures
2835 private.
2836
2837 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
2838 user point of view, this means that both static and shared libraries are built
2839 by default, but this can be individually controlled. More of the work of
2840 handling this static/shared cases is now inside libtool instead of PCRE's make
2841 file.
2842
2843 4. The pcretest utility is now installed along with pcregrep because it is
2844 useful for users (to test regexs) and by doing this, it automatically gets
2845 relinked by libtool. The documentation has been turned into a man page, so
2846 there are now .1, .txt, and .html versions in /doc.
2847
2848 5. Upgrades to pcregrep:
2849 (i) Added long-form option names like gnu grep.
2850 (ii) Added --help to list all options with an explanatory phrase.
2851 (iii) Added -r, --recursive to recurse into sub-directories.
2852 (iv) Added -f, --file to read patterns from a file.
2853
2854 6. pcre_exec() was referring to its "code" argument before testing that
2855 argument for NULL (and giving an error if it was NULL).
2856
2857 7. Upgraded Makefile.in to allow for compiling in a different directory from
2858 the source directory.
2859
2860 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
2861 options bits, the pointer it was passed was to an int instead of to an unsigned
2862 long int. This mattered only on 64-bit systems.
2863
2864 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
2865 generated) instead of pcre.in, which it its source. Also made the same change
2866 in several of the .c files.
2867
2868 10. A new release of gcc defines printf() as a macro, which broke pcretest
2869 because it had an ifdef in the middle of a string argument for printf(). Fixed
2870 by using separate calls to printf().
2871
2872 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
2873 script, to force use of CR or LF instead of \n in the source. On non-Unix
2874 systems, the value can be set in config.h.
2875
2876 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
2877 absolute limit. Changed the text of the error message to make this clear, and
2878 likewise updated the man page.
2879
2880 13. The limit of 99 on the number of capturing subpatterns has been removed.
2881 The new limit is 65535, which I hope will not be a "real" limit.
2882
2883
2884 Version 3.4 22-Aug-00
2885 ---------------------
2886
2887 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
2888
2889 2. Diagnose condition (?(0) as an error instead of crashing on matching.
2890
2891
2892 Version 3.3 01-Aug-00
2893 ---------------------
2894
2895 1. If an octal character was given, but the value was greater than \377, it
2896 was not getting masked to the least significant bits, as documented. This could
2897 lead to crashes in some systems.
2898
2899 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
2900 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
2901
2902 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
2903 These just pass their arguments on to (pcre_free)(), but they are provided
2904 because some uses of PCRE bind it to non-C systems that can call its functions,
2905 but cannot call free() or pcre_free() directly.
2906
2907 4. Add "make test" as a synonym for "make check". Corrected some comments in
2908 the Makefile.
2909
2910 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
2911 Makefile.
2912
2913 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
2914 command called pgrep for grepping around the active processes.
2915
2916 7. Added the beginnings of support for UTF-8 character strings.
2917
2918 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
2919 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
2920 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
2921 out for the ar command.)
2922
2923
2924 Version 3.2 12-May-00
2925 ---------------------
2926
2927 This is purely a bug fixing release.
2928
2929 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
2930 of ZA. This was just one example of several cases that could provoke this bug,
2931 which was introduced by change 9 of version 2.00. The code for breaking
2932 infinite loops after an iteration that matches an empty string was't working
2933 correctly.
2934
2935 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
2936 when matched against abbab (for example). After matching an empty string, it
2937 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
2938 caused it to match further down the string than it should.
2939
2940 3. The code contained an inclusion of sys/types.h. It isn't clear why this
2941 was there because it doesn't seem to be needed, and it causes trouble on some
2942 systems, as it is not a Standard C header. It has been removed.
2943
2944 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
2945 were reported on the Macintosh. The changes were from
2946
2947 while ((c = *(++ptr)) != 0 && c != '\n');
2948 to
2949 while ((c = *(++ptr)) != 0 && c != '\n') ;
2950
2951 Totally extraordinary, but if that's what it takes...
2952
2953 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
2954 available. Added HAVE_BCOPY and an autoconf test for it; if neither
2955 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
2956 assumes the way PCRE uses memmove() (always moving upwards).
2957
2958 6. PCRE is being used in one environment where strchr() is not available. There
2959 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
2960 faster code anyway.
2961
2962
2963 Version 3.1 09-Feb-00
2964 ---------------------
2965
2966 The only change in this release is the fixing of some bugs in Makefile.in for
2967 the "install" target:
2968
2969 (1) It was failing to install pcreposix.h.
2970
2971 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
2972
2973
2974 Version 3.0 01-Feb-00
2975 ---------------------
2976
2977 1. Add support for the /+ modifier to perltest (to output $` like it does in
2978 pcretest).
2979
2980 2. Add support for the /g modifier to perltest.
2981
2982 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
2983 matches null strings.
2984
2985 4. Fix perltest so that it doesn't do unwanted things when fed an empty
2986 pattern. Perl treats empty patterns specially - it reuses the most recent
2987 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
2988 effect.
2989
2990 5. The POSIX interface was broken in that it was just handing over the POSIX
2991 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
2992 required a bigger vector, with some working space on the end. This means that
2993 the POSIX wrapper now has to get and free some memory, and copy the results.
2994
2995 6. Added some simple autoconf support, placing the test data and the
2996 documentation in separate directories, re-organizing some of the
2997 information files, and making it build pcre-config (a GNU standard). Also added
2998 libtool support for building PCRE as a shared library, which is now the
2999 default.
3000
3001 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
3002 09 are not valid octal constants. Single digits will be used for minor values
3003 less than 10.
3004
3005 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
3006 existing programs that set these in the POSIX interface can use PCRE without
3007 modification.
3008
3009 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
3010 return all that pcre_info() returns, plus additional data. The pcre_info()
3011 function is retained for compatibility, but is considered to be obsolete.
3012
3013 10. Added experimental recursion feature (?R) to handle one common case that
3014 Perl 5.6 will be able to do with (?p{...}).
3015
3016 11. Added support for POSIX character classes like [:alpha:], which Perl is
3017 adopting.
3018
3019
3020 Version 2.08 31-Aug-99
3021 ----------------------
3022
3023 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
3024 trying to match at the startoffset position, but instead was moving forward to
3025 the next newline as if a previous match had failed.
3026
3027 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
3028 and could get into a loop if a null string was matched other than at the start
3029 of the subject.
3030
3031 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
3032 be distinguished at compile time, and for completeness also added PCRE_DATE.
3033
3034 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
3035 in GnuWin32 environments.
3036
3037
3038 Version 2.07 29-Jul-99
3039 ----------------------
3040
3041 1. The documentation is now supplied in plain text form and HTML as well as in
3042 the form of man page sources.
3043
3044 2. C++ compilers don't like assigning (void *) values to other pointer types.
3045 In particular this affects malloc(). Although there is no problem in Standard
3046 C, I've put in casts to keep C++ compilers happy.
3047
3048 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
3049 should be (const char *).
3050
3051 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
3052 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
3053 However, I haven't made this a standard facility. The documentation doesn't
3054 mention it, and the Makefile doesn't support it.
3055
3056 5. The Makefile now contains an "install" target, with editable destinations at
3057 the top of the file. The pcretest program is not installed.
3058
3059 6. pgrep -V now gives the PCRE version number and date.
3060
3061 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
3062 causing the entire string to be ignored, instead of just the last character.
3063
3064 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
3065 non-matching string, it can take a very, very long time, even for strings of
3066 quite modest length, because of the nested recursion. PCRE now does better in
3067 some of these cases. It does this by remembering the last required literal
3068 character in the pattern, and pre-searching the subject to ensure it is present
3069 before running the real match. In other words, it applies a heuristic to detect
3070 some types of certain failure quickly, and in the above example, if presented
3071 with a string that has no trailing " it gives "no match" very quickly.
3072
3073 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
3074 other alternatives are tried instead.
3075
3076
3077 Version 2.06 09-Jun-99
3078 ----------------------
3079
3080 1. Change pcretest's output for amount of store used to show just the code
3081 space, because the remainder (the data block) varies in size between 32-bit and
3082 64-bit systems.
3083
3084 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
3085 start matching at. This allows lookbehinds to work when searching for multiple
3086 occurrences in a string.
3087
3088 3. Added additional options to pcretest for testing multiple occurrences:
3089
3090 /+ outputs the rest of the string that follows a match
3091 /g loops for multiple occurrences, using the new startoffset argument
3092 /G loops for multiple occurrences by passing an incremented pointer
3093
3094 4. PCRE wasn't doing the "first character" optimization for patterns starting
3095 with \b or \B, though it was doing it for other lookbehind assertions. That is,
3096 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
3097 the letter 'x'. On long subject strings, this gives a significant speed-up.
3098
3099
3100 Version 2.05 21-Apr-99
3101 ----------------------
3102
3103 1. Changed the type of magic_number from int to long int so that it works
3104 properly on 16-bit systems.
3105
3106 2. Fixed a bug which caused patterns starting with .* not to work correctly
3107 when the subject string contained newline characters. PCRE was assuming
3108 anchoring for such patterns in all cases, which is not correct because .* will
3109 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
3110 DOTALL is set at top level; otherwise it knows that patterns starting with .*
3111 must be retried after every newline in the subject.
3112
3113
3114 Version 2.04 18-Feb-99
3115 ----------------------
3116
3117 1. For parenthesized subpatterns with repeats whose minimum was zero, the
3118 computation of the store needed to hold the pattern was incorrect (too large).
3119 If such patterns were nested a few deep, this could multiply and become a real
3120 problem.
3121
3122 2. Added /M option to pcretest to show the memory requirement of a specific
3123 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
3124
3125 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
3126 compiled in such a way that the backtracking after subsequent failure was
3127 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
3128 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
3129
3130
3131 Version 2.03 02-Feb-99
3132 ----------------------
3133
3134 1. Fixed typo and small mistake in man page.
3135
3136 2. Added 4th condition (GPL supersedes if conflict) and created separate
3137 LICENCE file containing the conditions.
3138
3139 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
3140 Perl, that is the internal \ allows the delimiter to be included in the
3141 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
3142 the final delimiter, add \ to the end of the pattern (to test the error).
3143
3144 4. Added the convenience functions for extracting substrings after a successful
3145 match. Updated pcretest to make it able to test these functions.
3146
3147
3148 Version 2.02 14-Jan-99
3149 ----------------------
3150
3151 1. Initialized the working variables associated with each extraction so that
3152 their saving and restoring doesn't refer to uninitialized store.
3153
3154 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
3155 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
3156 fix the problem.
3157
3158 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
3159 calls, and wasn't printing the correct value for compiling calls. Increased the
3160 default value of LOOPREPEAT, and the number of significant figures in the
3161 times.
3162
3163 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
3164
3165 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
3166 a building problem on Windows NT with a FAT file system.
3167
3168
3169 Version 2.01 21-Oct-98
3170 ----------------------
3171
3172 1. Changed the API for pcre_compile() to allow for the provision of a pointer
3173 to character tables built by pcre_maketables() in the current locale. If NULL
3174 is passed, the default tables are used.
3175
3176
3177 Version 2.00 24-Sep-98
3178 ----------------------
3179
3180 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
3181 it any more.
3182
3183 2. Allow quantification of (?>) groups, and make it work correctly.
3184
3185 3. The first character computation wasn't working for (?>) groups.
3186
3187 4. Correct the implementation of \Z (it is permitted to match on the \n at the
3188 end of the subject) and add 5.005's \z, which really does match only at the
3189 very end of the subject.
3190
3191 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
3192
3193 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
3194 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
3195 localized options. All options to pcre_study() were also removed.
3196
3197 7. Add other new features from 5.005:
3198
3199 $(?<= positive lookbehind
3200 $(?<! negative lookbehind
3201 (?imsx-imsx) added the unsetting capability
3202 such a setting is global if at outer level; local otherwise
3203 (?imsx-imsx:) non-capturing groups with option setting
3204 (?(cond)re|re) conditional pattern matching
3205
3206 A backreference to itself in a repeated group matches the previous
3207 captured string.
3208
3209 8. General tidying up of studying (both automatic and via "study")
3210 consequential on the addition of new assertions.
3211
3212 9. As in 5.005, unlimited repeated groups that could match an empty substring
3213 are no longer faulted at compile time. Instead, the loop is forcibly broken at
3214 runtime if any iteration does actually match an empty substring.
3215
3216 10. Include the RunTest script in the distribution.
3217
3218 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
3219 discrepancies, some of which were old and were also with respect to 5.004. They
3220 have now been fixed.
3221
3222
3223 Version 1.09 28-Apr-98
3224 ----------------------
3225
3226 1. A negated single character class followed by a quantifier with a minimum
3227 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
3228 program crashes, or just wrong answers. This did not apply to negated classes
3229 containing more than one character, or to minima other than one.
3230
3231
3232 Version 1.08 27-Mar-98
3233 ----------------------
3234
3235 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
3236
3237 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
3238 latter must appear before anything that relies on it in the pattern.
3239
3240
3241 Version 1.07 16-Feb-98
3242 ----------------------
3243
3244 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
3245 repeat of a potentially empty string).
3246
3247
3248 Version 1.06 23-Jan-98
3249 ----------------------
3250
3251 1. Added Markus Oberhumer's little patches for C++.
3252
3253 2. Literal strings longer than 255 characters were broken.
3254
3255
3256 Version 1.05 23-Dec-97
3257 ----------------------
3258
3259 1. Negated character classes containing more than one character were failing if
3260 PCRE_CASELESS was set at run time.
3261
3262
3263 Version 1.04 19-Dec-97
3264 ----------------------
3265
3266 1. Corrected the man page, where some "const" qualifiers had been omitted.
3267
3268 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
3269 input syntax.
3270
3271 3. Fixed memory leak which occurred when a regex with back references was
3272 matched with an offsets vector that wasn't big enough. The temporary memory
3273 that is used in this case wasn't being freed if the match failed.
3274
3275 4. Tidied pcretest to ensure it frees memory that it gets.
3276
3277 5. Temporary memory was being obtained in the case where the passed offsets
3278 vector was exactly big enough.
3279
3280 6. Corrected definition of offsetof() from change 5 below.
3281
3282 7. I had screwed up change 6 below and broken the rules for the use of
3283 setjmp(). Now fixed.
3284
3285
3286 Version 1.03 18-Dec-97
3287 ----------------------
3288
3289 1. A erroneous regex with a missing opening parenthesis was correctly
3290 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
3291 on some systems.
3292
3293 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
3294 it was reported that one broken compiler failed on the former because "code" is
3295 also an independent variable.
3296
3297 3. The erroneous regex a[]b caused an array overrun reference.
3298
3299 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
3300 fail on data ending with that character. (It was going on too far, and checking
3301 the next character, typically a binary zero.) This was specific to the
3302 optimized code for single-character negative classes.
3303
3304 5. Added a contributed patch from the TIN world which does the following:
3305
3306 + Add an undef for memmove, in case the the system defines a macro for it.
3307
3308 + Add a definition of offsetof(), in case there isn't one. (I don't know
3309 the reason behind this - offsetof() is part of the ANSI standard - but
3310 it does no harm).
3311
3312 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
3313 most of the places where whitespace preceded '#'. I have given up and
3314 allowed the remaining 2 cases to be at the margin.
3315
3316 + Rename some variables in pcre to eliminate shadowing. This seems very
3317 pedantic, but does no harm, of course.
3318
3319 6. Moved the call to setjmp() into its own function, to get rid of warnings
3320 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
3321
3322 7. Constructs such as \d{8,} were compiling into the equivalent of
3323 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
3324 outcome, but in this particular case used more store than had been allocated,
3325 which caused the bug to be discovered because it threw up an internal error.
3326
3327 8. The debugging code in both pcre and pcretest for outputting the compiled
3328 form of a regex was going wrong in the case of back references followed by
3329 curly-bracketed repeats.
3330
3331
3332 Version 1.02 12-Dec-97
3333 ----------------------
3334
3335 1. Typos in pcre.3 and comments in the source fixed.
3336
3337 2. Applied a contributed patch to get rid of places where it used to remove
3338 'const' from variables, and fixed some signed/unsigned and uninitialized
3339 variable warnings.
3340
3341 3. Added the "runtest" target to Makefile.
3342
3343 4. Set default compiler flag to -O2 rather than just -O.
3344
3345
3346 Version 1.01 19-Nov-97
3347 ----------------------
3348
3349 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
3350 like /([ab]*)*/, that is, for classes with more than one character in them.
3351
3352 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
3353 as /((?>a*))*/ (a PCRE_EXTRA facility).
3354
3355
3356 Version 1.00 18-Nov-97
3357 ----------------------
3358
3359 1. Added compile-time macros to support systems such as SunOS4 which don't have
3360 memmove() or strerror() but have other things that can be used instead.
3361
3362 2. Arranged that "make clean" removes the executables.
3363
3364
3365 Version 0.99 27-Oct-97
3366 ----------------------
3367
3368 1. Fixed bug in code for optimizing classes with only one character. It was
3369 initializing a 32-byte map regardless, which could cause it to run off the end
3370 of the memory it had got.
3371
3372 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
3373
3374
3375 Version 0.98 22-Oct-97
3376 ----------------------
3377
3378 1. Fixed bug in code for handling temporary memory usage when there are more
3379 back references than supplied space in the ovector. This could cause segfaults.
3380
3381
3382 Version 0.97 21-Oct-97
3383 ----------------------
3384
3385 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
3386
3387 2. Optimized negated single characters not to use a bit map.
3388
3389 3. Brought error texts together as macro definitions; clarified some of them;
3390 fixed one that was wrong - it said "range out of order" when it meant "invalid
3391 escape sequence".
3392
3393 4. Changed some char * arguments to const char *.
3394
3395 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
3396
3397 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
3398 pcretest.
3399
3400
3401 Version 0.96 16-Oct-97
3402 ----------------------
3403
3404 1. Added a simple "pgrep" utility to the distribution.
3405
3406 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
3407 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
3408 where "ddd" means "one or more decimal digits".
3409
3410 3. Fixed serious bug. If a pattern had a back reference, but the call to
3411 pcre_exec() didn't supply a large enough ovector to record the related
3412 identifying subpattern, the match always failed. PCRE now remembers the number
3413 of the largest back reference, and gets some temporary memory in which to save
3414 the offsets during matching if necessary, in order to ensure that
3415 backreferences always work.
3416
3417 4. Increased the compatibility with Perl in a number of ways:
3418
3419 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
3420 to request this handling. The option can be set at compile or exec time.
3421
3422 (b) $ matches before a terminating newline by default; an option
3423 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
3424 mode). The option can be set at compile or exec time.
3425
3426 (c) The handling of \ followed by a digit other than 0 is now supposed to be
3427 the same as Perl's. If the decimal number it represents is less than 10
3428 or there aren't that many previous left capturing parentheses, an octal
3429 escape is read. Inside a character class, it's always an octal escape,
3430 even if it is a single digit.
3431
3432 (d) An escaped but undefined alphabetic character is taken as a literal,
3433 unless PCRE_EXTRA is set. Currently this just reserves the remaining
3434 escapes.
3435
3436 (e) {0} is now permitted. (The previous item is removed from the compiled
3437 pattern).
3438
3439 5. Changed all the names of code files so that the basic parts are no longer
3440 than 10 characters, and abolished the teeny "globals.c" file.
3441
3442 6. Changed the handling of character classes; they are now done with a 32-byte
3443 bit map always.
3444
3445 7. Added the -d and /D options to pcretest to make it possible to look at the
3446 internals of compilation without having to recompile pcre.
3447
3448
3449 Version 0.95 23-Sep-97
3450 ----------------------
3451
3452 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
3453 \x20 at the start of a run of normal characters. These were being treated as
3454 real characters, instead of the source characters being re-checked.
3455
3456
3457 Version 0.94 18-Sep-97
3458 ----------------------
3459
3460 1. The functions are now thread-safe, with the caveat that the global variables
3461 containing pointers to malloc() and free() or alternative functions are the
3462 same for all threads.
3463
3464 2. Get pcre_study() to generate a bitmap of initial characters for non-
3465 anchored patterns when this is possible, and use it if passed to pcre_exec().
3466
3467
3468 Version 0.93 15-Sep-97
3469 ----------------------
3470
3471 1. /(b)|(:+)/ was computing an incorrect first character.
3472
3473 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
3474 but not actually doing anything yet.
3475
3476 3. Treat "-" characters in classes that cannot be part of ranges as literals,
3477 as Perl does (e.g. [-az] or [az-]).
3478
3479 4. Set the anchored flag if a branch starts with .* or .*? because that tests
3480 all possible positions.
3481
3482 5. Split up into different modules to avoid including unneeded functions in a
3483 compiled binary. However, compile and exec are still in one module. The "study"
3484 function is split off.
3485
3486 6. The character tables are now in a separate module whose source is generated
3487 by an auxiliary program - but can then be edited by hand if required. There are
3488 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
3489 toupper() in the code.
3490
3491 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
3492 make them global. Abolish the function for setting them, as the caller can now
3493 set them directly.
3494
3495
3496 Version 0.92 11-Sep-97
3497 ----------------------
3498
3499 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
3500 (e.g. /a{1,3}/) was broken (I mis-optimized it).
3501
3502 2. Caseless matching was not working in character classes if the characters in
3503 the pattern were in upper case.
3504
3505 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
3506
3507 4. Make PCRE_ANCHORED public and accept as a compile option.
3508
3509 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
3510 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
3511 pass them.
3512
3513 6. Give an error if bad option bits passed at compile or run time.
3514
3515 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
3516 pcretest to cause it to pass that flag.
3517
3518 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
3519 options, and the first character, if set.
3520
3521 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
3522
3523
3524 Version 0.91 10-Sep-97
3525 ----------------------
3526
3527 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
3528 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
3529
3530 2. PCRE was looping on encountering an indefinitely repeated back reference to
3531 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
3532 Perl does - treats the match as successful.
3533
3534 ****

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