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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 413 - (show annotations) (download)
Sat Apr 11 11:41:52 2009 UTC (5 years, 7 months ago) by ph10
File size: 165585 byte(s)
CMake: arrange for release type compiler flags to be shown in the summary.

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

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