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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 386 - (show annotations) (download)
Tue Mar 10 11:48:33 2009 UTC (5 years, 9 months ago) by ph10
File size: 162312 byte(s)
Add the -M option to pcretest.

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

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