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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 412 - (show annotations)
Sat Apr 11 10:34:37 2009 UTC (6 years, 2 months ago) by ph10
File size: 165472 byte(s)
Error occurred while calculating annotation data.
Add support for (*UTF8).
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
123
124 Version 7.8 05-Sep-08
125 ---------------------
126
127 1. Replaced UCP searching code with optimized version as implemented for Ad
128 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
129 stage table and inline lookup instead of a function, giving speed ups of 2
130 to 5 times on some simple patterns that I tested. Permission was given to
131 distribute the MultiStage2.py script that generates the tables (it's not in
132 the tarball, but is in the Subversion repository).
133
134 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
135 scripts.
136
137 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
138 a group with a zero qualifier. The result of the study could be incorrect,
139 or the function might crash, depending on the pattern.
140
141 4. Caseless matching was not working for non-ASCII characters in back
142 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
143 It now works when Unicode Property Support is available.
144
145 5. In pcretest, an escape such as \x{de} in the data was always generating
146 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
147 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
148 truncation.
149
150 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
151
152 7. Added two (int) casts to pcregrep when printing the difference of two
153 pointers, in case they are 64-bit values.
154
155 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
156 test 2 if it fails.
157
158 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
159 and a #define of that name to empty if it is not externally set. This is to
160 allow users of MSVC to set it if necessary.
161
162 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
163 the convenience functions in the pcre_get.c source file.
164
165 11. An option change at the start of a pattern that had top-level alternatives
166 could cause overwriting and/or a crash. This command provoked a crash in
167 some environments:
168
169 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
170
171 This potential security problem was recorded as CVE-2008-2371.
172
173 12. For a pattern where the match had to start at the beginning or immediately
174 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
175 pcre_dfa_exec() could read past the end of the passed subject if there was
176 no match. To help with detecting such bugs (e.g. with valgrind), I modified
177 pcretest so that it places the subject at the end of its malloc-ed buffer.
178
179 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
180 exec() might read past the end of the data buffer in UTF-8 mode.
181
182 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
183 the data contained the byte 0x85 as part of a UTF-8 character within its
184 first line. This applied both to normal and DFA matching.
185
186 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
187 /^[^d]*?$/8 failed to match "abc".
188
189 16. Added a missing copyright notice to pcrecpp_internal.h.
190
191 17. Make it more clear in the documentation that values returned from
192 pcre_exec() in ovector are byte offsets, not character counts.
193
194 18. Tidied a few places to stop certain compilers from issuing warnings.
195
196 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
197 supplied by Stefan Weber. I made a further small update for 7.8 because
198 there is a change of source arrangements: the pcre_searchfuncs.c module is
199 replaced by pcre_ucd.c.
200
201
202 Version 7.7 07-May-08
203 ---------------------
204
205 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
206 a string to a long long, pretend we don't even have a long long." This is
207 done by checking for the strtoq, strtoll, and _strtoi64 functions.
208
209 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
210 pre-7.6 versions, which defined a global no_arg variable instead of putting
211 it in the RE class. (See also #8 below.)
212
213 3. Remove a line of dead code, identified by coverity and reported by Nuno
214 Lopes.
215
216 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
217
218 (1) The include/exclude patterns were being applied to the whole pathnames
219 of files, instead of just to the final components.
220
221 (2) If there was more than one level of directory, the subdirectories were
222 skipped unless they satisfied the include/exclude conditions. This is
223 inconsistent with GNU grep (and could even be seen as contrary to the
224 pcregrep specification - which I improved to make it absolutely clear).
225 The action now is always to scan all levels of directory, and just
226 apply the include/exclude patterns to regular files.
227
228 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
229 --exclude_dir in the tests to avoid scanning .svn directories.
230
231 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
232 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
233 doesn't support NULs in patterns.
234
235 7. Added some missing "const"s to declarations of static tables in
236 pcre_compile.c and pcre_dfa_exec.c.
237
238 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
239 caused by fix #2 above. (Subsequently also a second patch to fix the
240 first patch. And a third patch - this was a messy problem.)
241
242 9. Applied Craig's patch to remove the use of push_back().
243
244 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
245 matching function regexec().
246
247 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
248 which, however, unlike Perl's \g{...}, are subroutine calls, not back
249 references. PCRE supports relative numbers with this syntax (I don't think
250 Oniguruma does).
251
252 12. Previously, a group with a zero repeat such as (...){0} was completely
253 omitted from the compiled regex. However, this means that if the group
254 was called as a subroutine from elsewhere in the pattern, things went wrong
255 (an internal error was given). Such groups are now left in the compiled
256 pattern, with a new opcode that causes them to be skipped at execution
257 time.
258
259 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
260 to the way PCRE behaves:
261
262 (a) A lone ] character is dis-allowed (Perl treats it as data).
263
264 (b) A back reference to an unmatched subpattern matches an empty string
265 (Perl fails the current match path).
266
267 (c) A data ] in a character class must be notated as \] because if the
268 first data character in a class is ], it defines an empty class. (In
269 Perl it is not possible to have an empty class.) The empty class []
270 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
271 The negative empty class [^] matches any one character, independently
272 of the DOTALL setting.
273
274 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
275 non-existent subpattern following a character class starting with ']' and
276 containing () gave an internal compiling error instead of "reference to
277 non-existent subpattern". Fortunately, when the pattern did exist, the
278 compiled code was correct. (When scanning forwards to check for the
279 existencd of the subpattern, it was treating the data ']' as terminating
280 the class, so got the count wrong. When actually compiling, the reference
281 was subsequently set up correctly.)
282
283 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
284 it was being rejected as not supported by pcre_dfa_exec(), even though
285 other assertions are supported. I have made pcre_dfa_exec() support
286 (*FAIL).
287
288 16. The implementation of 13c above involved the invention of a new opcode,
289 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
290 cannot be changed at match time, I realized I could make a small
291 improvement to matching performance by compiling OP_ALLANY instead of
292 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
293 on the OP_ANY path.
294
295 17. Compiling pcretest on Windows with readline support failed without the
296 following two fixes: (1) Make the unistd.h include conditional on
297 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
298
299 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
300 ncurses library to be included for pcretest when ReadLine support is
301 requested, but also to allow for it to be overridden. This patch came from
302 Daniel Bergström.
303
304 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
305 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
306 any errors with the current Unicode tables. Thanks to Peter Kankowski for
307 spotting this.
308
309
310 Version 7.6 28-Jan-08
311 ---------------------
312
313 1. A character class containing a very large number of characters with
314 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
315 overflow.
316
317 2. Patch to cut out the "long long" test in pcrecpp_unittest when
318 HAVE_LONG_LONG is not defined.
319
320 3. Applied Christian Ehrlicher's patch to update the CMake build files to
321 bring them up to date and include new features. This patch includes:
322
323 - Fixed PH's badly added libz and libbz2 support.
324 - Fixed a problem with static linking.
325 - Added pcredemo. [But later removed - see 7 below.]
326 - Fixed dftables problem and added an option.
327 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
328 HAVE_LONG_LONG.
329 - Added readline support for pcretest.
330 - Added an listing of the option settings after cmake has run.
331
332 4. A user submitted a patch to Makefile that makes it easy to create
333 "pcre.dll" under mingw when using Configure/Make. I added stuff to
334 Makefile.am that cause it to include this special target, without
335 affecting anything else. Note that the same mingw target plus all
336 the other distribution libraries and programs are now supported
337 when configuring with CMake (see 6 below) instead of with
338 Configure/Make.
339
340 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
341 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
342 exported in the Windows port". It has not yet been confirmed that the patch
343 solves the problem, but it does no harm.
344
345 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
346 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
347 with CMake, and also correct the comment about stack recursion.
348
349 7. Remove the automatic building of pcredemo from the ./configure system and
350 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
351 of a program that users should build themselves after PCRE is installed, so
352 building it automatically is not really right. What is more, it gave
353 trouble in some build environments.
354
355 8. Further tidies to CMakeLists.txt from Sheri and Christian.
356
357
358 Version 7.5 10-Jan-08
359 ---------------------
360
361 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
362 values in parens when parsing an RE using the C++ wrapper."
363
364 2. Negative specials like \S did not work in character classes in UTF-8 mode.
365 Characters greater than 255 were excluded from the class instead of being
366 included.
367
368 3. The same bug as (2) above applied to negated POSIX classes such as
369 [:^space:].
370
371 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
372 defined or documented. It seems to have been a typo for PCRE_STATIC, so
373 I have changed it.
374
375 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
376 first named subpattern) and a construct such as (?&a) would reference the
377 first named subpattern whose name started with "a" (in other words, the
378 length check was missing). Both these problems are fixed. "Subpattern name
379 expected" is now given for (?&) (a zero-length name), and this patch also
380 makes it give the same error for \k'' (previously it complained that that
381 was a reference to a non-existent subpattern).
382
383 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
384 this is right because (?- can be followed by option settings as well as by
385 digits. I have, however, made the messages clearer.
386
387 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
388 than the number used in the conditional) now cause a compile-time error.
389 This is actually not compatible with Perl, which accepts such patterns, but
390 treats the conditional as always being FALSE (as PCRE used to), but it
391 seems to me that giving a diagnostic is better.
392
393 8. Change "alphameric" to the more common word "alphanumeric" in comments
394 and messages.
395
396 9. Fix two occurrences of "backslash" in comments that should have been
397 "backspace".
398
399 10. Remove two redundant lines of code that can never be obeyed (their function
400 was moved elsewhere).
401
402 11. The program that makes PCRE's Unicode character property table had a bug
403 which caused it to generate incorrect table entries for sequences of
404 characters that have the same character type, but are in different scripts.
405 It amalgamated them into a single range, with the script of the first of
406 them. In other words, some characters were in the wrong script. There were
407 thirteen such cases, affecting characters in the following ranges:
408
409 U+002b0 - U+002c1
410 U+0060c - U+0060d
411 U+0061e - U+00612
412 U+0064b - U+0065e
413 U+0074d - U+0076d
414 U+01800 - U+01805
415 U+01d00 - U+01d77
416 U+01d9b - U+01dbf
417 U+0200b - U+0200f
418 U+030fc - U+030fe
419 U+03260 - U+0327f
420 U+0fb46 - U+0fbb1
421 U+10450 - U+1049d
422
423 12. The -o option (show only the matching part of a line) for pcregrep was not
424 compatible with GNU grep in that, if there was more than one match in a
425 line, it showed only the first of them. It now behaves in the same way as
426 GNU grep.
427
428 13. If the -o and -v options were combined for pcregrep, it printed a blank
429 line for every non-matching line. GNU grep prints nothing, and pcregrep now
430 does the same. The return code can be used to tell if there were any
431 non-matching lines.
432
433 14. Added --file-offsets and --line-offsets to pcregrep.
434
435 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
436 infinitely looping recursion. The bug was that positive lookaheads were not
437 being skipped when checking for a possible empty match (negative lookaheads
438 and both kinds of lookbehind were skipped).
439
440 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
441 inclusion of <windows.h> to before rather than after the definition of
442 INVALID_FILE_ATTRIBUTES (patch from David Byron).
443
444 17. Specifying a possessive quantifier with a specific limit for a Unicode
445 character property caused pcre_compile() to compile bad code, which led at
446 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
447 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
448 caused the error; without that there was no problem.
449
450 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
451
452 19. Added --enable-pcretest-libreadline.
453
454 20. In pcrecpp.cc, the variable 'count' was incremented twice in
455 RE::GlobalReplace(). As a result, the number of replacements returned was
456 double what it should be. I removed one of the increments, but Craig sent a
457 later patch that removed the other one (the right fix) and added unit tests
458 that check the return values (which was not done before).
459
460 21. Several CMake things:
461
462 (1) Arranged that, when cmake is used on Unix, the libraries end up with
463 the names libpcre and libpcreposix, not just pcre and pcreposix.
464
465 (2) The above change means that pcretest and pcregrep are now correctly
466 linked with the newly-built libraries, not previously installed ones.
467
468 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
469
470 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
471 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
472 UTF-8 newline character). The key issue is that the pattern starts .*;
473 this means that the match must be either at the beginning, or after a
474 newline. The bug was in the code for advancing after a failed match and
475 checking that the new position followed a newline. It was not taking
476 account of UTF-8 characters correctly.
477
478 23. PCRE was behaving differently from Perl in the way it recognized POSIX
479 character classes. PCRE was not treating the sequence [:...:] as a
480 character class unless the ... were all letters. Perl, however, seems to
481 allow any characters between [: and :], though of course it rejects as
482 unknown any "names" that contain non-letters, because all the known class
483 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
484 for example, whereas PCRE did not - it did not recognize a POSIX character
485 class. This seemed a bit dangerous, so the code has been changed to be
486 closer to Perl. The behaviour is not identical to Perl, because PCRE will
487 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
488 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
489 Perl does, and where it didn't before.
490
491 24. Rewrite so as to remove the single use of %n from pcregrep because in some
492 Windows environments %n is disabled by default.
493
494
495 Version 7.4 21-Sep-07
496 ---------------------
497
498 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
499 means that a class such as [\s] counted as "explicit reference to CR or
500 LF". That isn't really right - the whole point of the change was to try to
501 help when there was an actual mention of one of the two characters. So now
502 the change happens only if \r or \n (or a literal CR or LF) character is
503 encountered.
504
505 2. The 32-bit options word was also used for 6 internal flags, but the numbers
506 of both had grown to the point where there were only 3 bits left.
507 Fortunately, there was spare space in the data structure, and so I have
508 moved the internal flags into a new 16-bit field to free up more option
509 bits.
510
511 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
512 but did not set the internal JCHANGED flag - either of these is enough to
513 control the way the "get" function works - but the PCRE_INFO_JCHANGED
514 facility is supposed to tell if (?J) was ever used, so now (?J) at the
515 start sets both bits.
516
517 4. Added options (at build time, compile time, exec time) to change \R from
518 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
519
520 5. doc/pcresyntax.html was missing from the distribution.
521
522 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
523 compatibility, even though it is no longer used.
524
525 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
526 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
527 windows.h file is present (where different names are used). [This was
528 reversed later after testing - see 16 below.]
529
530 8. Changed all #include <config.h> to #include "config.h". There were also
531 some further <pcre.h> cases that I changed to "pcre.h".
532
533 9. When pcregrep was used with the --colour option, it missed the line ending
534 sequence off the lines that it output.
535
536 10. It was pointed out to me that arrays of string pointers cause lots of
537 relocations when a shared library is dynamically loaded. A technique of
538 using a single long string with a table of offsets can drastically reduce
539 these. I have refactored PCRE in four places to do this. The result is
540 dramatic:
541
542 Originally: 290
543 After changing UCP table: 187
544 After changing error message table: 43
545 After changing table of "verbs" 36
546 After changing table of Posix names 22
547
548 Thanks to the folks working on Gregex for glib for this insight.
549
550 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
551 unicode-properties was also set.
552
553 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
554
555 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
556 checked only for CRLF.
557
558 14. Added casts to pcretest.c to avoid compiler warnings.
559
560 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
561
562 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
563 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
564 entirely. This removes changes made in 7 above.
565
566 17. The CMake files have been updated, and there is now more information about
567 building with CMake in the NON-UNIX-USE document.
568
569
570 Version 7.3 28-Aug-07
571 ---------------------
572
573 1. In the rejigging of the build system that eventually resulted in 7.1, the
574 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
575 brackets there is not right, since it causes compilers to look for an
576 installed pcre.h, not the version that is in the source that is being
577 compiled (which of course may be different). I have changed it back to:
578
579 #include "pcre.h"
580
581 I have a vague recollection that the change was concerned with compiling in
582 different directories, but in the new build system, that is taken care of
583 by the VPATH setting the Makefile.
584
585 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
586 when the subject happened to end in the byte 0x85 (e.g. if the last
587 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
588 characters but of course it shouldn't be taken as a newline when it is part
589 of another character. The bug was that, for an unlimited repeat of . in
590 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
591 characters when looking for a newline.
592
593 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
594
595 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
596 in debug output.
597
598 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
599 long printing in the pcrecpp unittest when running under MinGW.
600
601 6. ESC_K was left out of the EBCDIC table.
602
603 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
604 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
605 limit also applies to "virtual nesting" when a pattern is recursive, and in
606 this case 1000 isn't so big. I have been able to remove this limit at the
607 expense of backing off one optimization in certain circumstances. Normally,
608 when pcre_exec() would call its internal match() function recursively and
609 immediately return the result unconditionally, it uses a "tail recursion"
610 feature to save stack. However, when a subpattern that can match an empty
611 string has an unlimited repetition quantifier, it no longer makes this
612 optimization. That gives it a stack frame in which to save the data for
613 checking that an empty string has been matched. Previously this was taken
614 from the 1000-entry workspace that had been reserved. So now there is no
615 explicit limit, but more stack is used.
616
617 8. Applied Daniel's patches to solve problems with the import/export magic
618 syntax that is required for Windows, and which was going wrong for the
619 pcreposix and pcrecpp parts of the library. These were overlooked when this
620 problem was solved for the main library.
621
622 9. There were some crude static tests to avoid integer overflow when computing
623 the size of patterns that contain repeated groups with explicit upper
624 limits. As the maximum quantifier is 65535, the maximum group length was
625 set at 30,000 so that the product of these two numbers did not overflow a
626 32-bit integer. However, it turns out that people want to use groups that
627 are longer than 30,000 bytes (though not repeat them that many times).
628 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
629 made it possible to implement the integer overflow checks in a much more
630 dynamic way, which I have now done. The artificial limitation on group
631 length has been removed - we now have only the limit on the total length of
632 the compiled pattern, which depends on the LINK_SIZE setting.
633
634 10. Fixed a bug in the documentation for get/copy named substring when
635 duplicate names are permitted. If none of the named substrings are set, the
636 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
637 empty string.
638
639 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
640 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
641 because the ] is interpreted as the first data character and the
642 terminating ] is not found. PCRE has been made compatible with Perl in this
643 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
644 cause memory overwriting.
645
646 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
647 string has been matched (to stop an infinite loop). It was not recognizing
648 a conditional subpattern that could match an empty string if that
649 subpattern was within another subpattern. For example, it looped when
650 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
651 condition was not nested. This bug has been fixed.
652
653 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
654 past the start of the subject in the presence of bytes with the top bit
655 set, for example "\x8aBCD".
656
657 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
658 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
659
660 14. Optimized (?!) to (*FAIL).
661
662 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
663 This restricts code points to be within the range 0 to 0x10FFFF, excluding
664 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
665 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
666 does: it's just the validity check that is more restrictive.
667
668 16. Inserted checks for integer overflows during escape sequence (backslash)
669 processing, and also fixed erroneous offset values for syntax errors during
670 backslash processing.
671
672 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
673 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
674
675 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
676 caused an overrun.
677
678 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
679 something other than just ASCII characters) inside a group that had an
680 unlimited repeat caused a loop at compile time (while checking to see
681 whether the group could match an empty string).
682
683 20. Debugging a pattern containing \p or \P could cause a crash. For example,
684 [\P{Any}] did so. (Error in the code for printing property names.)
685
686 21. An orphan \E inside a character class could cause a crash.
687
688 22. A repeated capturing bracket such as (A)? could cause a wild memory
689 reference during compilation.
690
691 23. There are several functions in pcre_compile() that scan along a compiled
692 expression for various reasons (e.g. to see if it's fixed length for look
693 behind). There were bugs in these functions when a repeated \p or \P was
694 present in the pattern. These operators have additional parameters compared
695 with \d, etc, and these were not being taken into account when moving along
696 the compiled data. Specifically:
697
698 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
699 length.
700
701 (b) An item such as \pL+ within a repeated group could cause crashes or
702 loops.
703
704 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
705 "reference to non-existent subpattern" error.
706
707 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
708
709 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
710 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
711
712 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
713
714 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
715 character were causing crashes (broken optimization).
716
717 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
718 \p or \P) caused a compile-time loop.
719
720 28. More problems have arisen in unanchored patterns when CRLF is a valid line
721 break. For example, the unstudied pattern [\r\n]A does not match the string
722 "\r\nA" because change 7.0/46 below moves the current point on by two
723 characters after failing to match at the start. However, the pattern \nA
724 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
725 the same is true. There doesn't seem any very clean way out of this, but
726 what I have chosen to do makes the common cases work: PCRE now takes note
727 of whether there can be an explicit match for \r or \n anywhere in the
728 pattern, and if so, 7.0/46 no longer applies. As part of this change,
729 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
730 pattern has explicit CR or LF references.
731
732 29. Added (*CR) etc for changing newline setting at start of pattern.
733
734
735 Version 7.2 19-Jun-07
736 ---------------------
737
738 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
739 which is apparently normally available under Windows.
740
741 2. Re-jig the pcregrep tests with different newline settings in an attempt
742 to make them independent of the local environment's newline setting.
743
744 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
745
746 4. Some of the "internals" tests were previously cut out when the link size
747 was not 2, because the output contained actual offsets. The recent new
748 "Z" feature of pcretest means that these can be cut out, making the tests
749 usable with all link sizes.
750
751 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
752 stack recursion. This gives a massive performance boost under BSD, but just
753 a small improvement under Linux. However, it saves one field in the frame
754 in all cases.
755
756 6. Added more features from the forthcoming Perl 5.10:
757
758 (a) (?-n) (where n is a string of digits) is a relative subroutine or
759 recursion call. It refers to the nth most recently opened parentheses.
760
761 (b) (?+n) is also a relative subroutine call; it refers to the nth next
762 to be opened parentheses.
763
764 (c) Conditions that refer to capturing parentheses can be specified
765 relatively, for example, (?(-2)... or (?(+3)...
766
767 (d) \K resets the start of the current match so that everything before
768 is not part of it.
769
770 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
771
772 (f) \g{name} is another synonym - part of Perl 5.10's unification of
773 reference syntax.
774
775 (g) (?| introduces a group in which the numbering of parentheses in each
776 alternative starts with the same number.
777
778 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
779
780 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
781 PCRE_INFO_JCHANGED.
782
783 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
784 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
785 for detecting groups that can match an empty string.
786
787 9. A pattern with a very large number of alternatives (more than several
788 hundred) was running out of internal workspace during the pre-compile
789 phase, where pcre_compile() figures out how much memory will be needed. A
790 bit of new cunning has reduced the workspace needed for groups with
791 alternatives. The 1000-alternative test pattern now uses 12 bytes of
792 workspace instead of running out of the 4096 that are available.
793
794 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
795
796 11. Applied patch from Google to remove an optimization that didn't quite work.
797 The report of the bug said:
798
799 pcrecpp::RE("a*").FullMatch("aaa") matches, while
800 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
801 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
802
803 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
804 it matched the wrong number of bytes.
805
806
807 Version 7.1 24-Apr-07
808 ---------------------
809
810 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
811 that is more "standard", making use of automake and other Autotools. There
812 is some re-arrangement of the files and adjustment of comments consequent
813 on this.
814
815 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
816 for recursive directory scanning broke on some systems because the files
817 are not scanned in any specific order and on different systems the order
818 was different. A call to "sort" has been inserted into RunGrepTest for the
819 approprate test as a short-term fix. In the longer term there may be an
820 alternative.
821
822 3. I had an email from Eric Raymond about problems translating some of PCRE's
823 man pages to HTML (despite the fact that I distribute HTML pages, some
824 people do their own conversions for various reasons). The problems
825 concerned the use of low-level troff macros .br and .in. I have therefore
826 removed all such uses from the man pages (some were redundant, some could
827 be replaced by .nf/.fi pairs). The 132html script that I use to generate
828 HTML has been updated to handle .nf/.fi and to complain if it encounters
829 .br or .in.
830
831 4. Updated comments in configure.ac that get placed in config.h.in and also
832 arranged for config.h to be included in the distribution, with the name
833 config.h.generic, for the benefit of those who have to compile without
834 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
835
836 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
837 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
838 makevp.bat for the current PCRE, using the additional files
839 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
840
841 6. A Windows user reported a minor discrepancy with test 2, which turned out
842 to be caused by a trailing space on an input line that had got lost in his
843 copy. The trailing space was an accident, so I've just removed it.
844
845 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
846 that is needed.
847
848 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
849 as "const" (a) because they are and (b) because it helps the PHP
850 maintainers who have recently made a script to detect big data structures
851 in the php code that should be moved to the .rodata section. I remembered
852 to update Builducptable as well, so it won't revert if ucptable.h is ever
853 re-created.
854
855 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
856 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
857 order to be able to cut out the UTF-8 tables in the latter when UTF-8
858 support is not required. This saves 1.5-2K of code, which is important in
859 some applications.
860
861 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
862 so as not to refer to the tables, even though these functions will never be
863 called when UTF-8 support is disabled. Otherwise there are problems with a
864 shared library.
865
866 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
867
868 (a) It was defining its arguments as char * instead of void *.
869
870 (b) It was assuming that all moves were upwards in memory; this was true
871 a long time ago when I wrote it, but is no longer the case.
872
873 The emulated memove() is provided for those environments that have neither
874 memmove() nor bcopy(). I didn't think anyone used it these days, but that
875 is clearly not the case, as these two bugs were recently reported.
876
877 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
878 and Detrail to create the HTML documentation, the .txt form of the man
879 pages, and it removes trailing spaces from listed files. It also creates
880 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
881 case, it wraps all the #defines with #ifndefs. This script should be run
882 before "make dist".
883
884 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
885 with Unicode property support.
886
887 (a) For a maximizing quantifier, if the two different cases of the
888 character were of different lengths in their UTF-8 codings (there are
889 some cases like this - I found 11), and the matching function had to
890 back up over a mixture of the two cases, it incorrectly assumed they
891 were both the same length.
892
893 (b) When PCRE was configured to use the heap rather than the stack for
894 recursion during matching, it was not correctly preserving the data for
895 the other case of a UTF-8 character when checking ahead for a match
896 while processing a minimizing repeat. If the check also involved
897 matching a wide character, but failed, corruption could cause an
898 erroneous result when trying to check for a repeat of the original
899 character.
900
901 13. Some tidying changes to the testing mechanism:
902
903 (a) The RunTest script now detects the internal link size and whether there
904 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
905 values substituted by "configure". (The RunGrepTest script already did
906 this for UTF-8.) The configure.ac script no longer substitutes the
907 relevant variables.
908
909 (b) The debugging options /B and /D in pcretest show the compiled bytecode
910 with length and offset values. This means that the output is different
911 for different internal link sizes. Test 2 is skipped for link sizes
912 other than 2 because of this, bypassing the problem. Unfortunately,
913 there was also a test in test 3 (the locale tests) that used /B and
914 failed for link sizes other than 2. Rather than cut the whole test out,
915 I have added a new /Z option to pcretest that replaces the length and
916 offset values with spaces. This is now used to make test 3 independent
917 of link size. (Test 2 will be tidied up later.)
918
919 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
920 segmentation fault instead of returning the appropriate error message.
921
922 15. In multiline mode when the newline sequence was set to "any", the pattern
923 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
924 This doesn't seem right; it now treats the CRLF combination as the line
925 ending, and so does not match in that case. It's only a pattern such as ^$
926 that would hit this one: something like ^ABC$ would have failed after \r
927 and then tried again after \r\n.
928
929 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
930 in an attempt to make files that differ only in their line terminators
931 compare equal. This works on Linux.
932
933 17. Under certain error circumstances pcregrep might try to free random memory
934 as it exited. This is now fixed, thanks to valgrind.
935
936 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
937 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
938 was because its rules for how to advance for /g after matching an empty
939 string at the end of a line did not allow for this case. They now check for
940 it specially.
941
942 20. pcretest is supposed to handle patterns and data of any length, by
943 extending its buffers when necessary. It was getting this wrong when the
944 buffer for a data line had to be extended.
945
946 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
947 CRLF as a newline sequence.
948
949 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
950 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
951 I have nevertheless tidied it up.
952
953 23. Added some casts to kill warnings from HP-UX ia64 compiler.
954
955 24. Added a man page for pcre-config.
956
957
958 Version 7.0 19-Dec-06
959 ---------------------
960
961 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
962 moving to gcc 4.1.1.
963
964 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
965 sys/time.h, which is documented as needed for this function. It doesn't
966 seem to matter on Linux, but it showed up on some releases of OS X.
967
968 3. It seems that there are systems where bytes whose values are greater than
969 127 match isprint() in the "C" locale. The "C" locale should be the
970 default when a C program starts up. In most systems, only ASCII printing
971 characters match isprint(). This difference caused the output from pcretest
972 to vary, making some of the tests fail. I have changed pcretest so that:
973
974 (a) When it is outputting text in the compiled version of a pattern, bytes
975 other than 32-126 are always shown as hex escapes.
976
977 (b) When it is outputting text that is a matched part of a subject string,
978 it does the same, unless a different locale has been set for the match
979 (using the /L modifier). In this case, it uses isprint() to decide.
980
981 4. Fixed a major bug that caused incorrect computation of the amount of memory
982 required for a compiled pattern when options that changed within the
983 pattern affected the logic of the preliminary scan that determines the
984 length. The relevant options are -x, and -i in UTF-8 mode. The result was
985 that the computed length was too small. The symptoms of this bug were
986 either the PCRE error "internal error: code overflow" from pcre_compile(),
987 or a glibc crash with a message such as "pcretest: free(): invalid next
988 size (fast)". Examples of patterns that provoked this bug (shown in
989 pcretest format) are:
990
991 /(?-x: )/x
992 /(?x)(?-x: \s*#\s*)/
993 /((?i)[\x{c0}])/8
994 /(?i:[\x{c0}])/8
995
996 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
997 is now done differently.
998
999 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
1000 wrapper classes; (b) implement a new function in the C++ scanner that is
1001 more efficient than the old way of doing things because it avoids levels of
1002 recursion in the regex matching; (c) add a paragraph to the documentation
1003 for the FullMatch() function.
1004
1005 6. The escape sequence \n was being treated as whatever was defined as
1006 "newline". Not only was this contrary to the documentation, which states
1007 that \n is character 10 (hex 0A), but it also went horribly wrong when
1008 "newline" was defined as CRLF. This has been fixed.
1009
1010 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
1011 was being set to -1 for the "end of line" case (supposedly a value that no
1012 character can have). Though this value is never used (the check for end of
1013 line is "zero bytes in current character"), it caused compiler complaints.
1014 I've changed it to 0xffffffff.
1015
1016 8. In pcre_version.c, the version string was being built by a sequence of
1017 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
1018 string (as it is for production releases) called a macro with an empty
1019 argument. The C standard says the result of this is undefined. The gcc
1020 compiler treats it as an empty string (which was what was wanted) but it is
1021 reported that Visual C gives an error. The source has been hacked around to
1022 avoid this problem.
1023
1024 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
1025 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
1026 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
1027 of them did).
1028
1029 10. Originally, pcretest opened its input and output without "b"; then I was
1030 told that "b" was needed in some environments, so it was added for release
1031 5.0 to both the input and output. (It makes no difference on Unix-like
1032 systems.) Later I was told that it is wrong for the input on Windows. I've
1033 now abstracted the modes into two macros, to make it easier to fiddle with
1034 them, and removed "b" from the input mode under Windows.
1035
1036 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
1037
1038 12. Added -help and --help to pcretest as an official way of being reminded
1039 of the options.
1040
1041 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
1042 and pcrecpp.cc because they annoy compilers at high warning levels.
1043
1044 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
1045
1046 15. Fixed an occurrence of == in configure.ac that should have been = (shell
1047 scripts are not C programs :-) and which was not noticed because it works
1048 on Linux.
1049
1050 16. pcretest is supposed to handle any length of pattern and data line (as one
1051 line or as a continued sequence of lines) by extending its input buffer if
1052 necessary. This feature was broken for very long pattern lines, leading to
1053 a string of junk being passed to pcre_compile() if the pattern was longer
1054 than about 50K.
1055
1056 17. I have done a major re-factoring of the way pcre_compile() computes the
1057 amount of memory needed for a compiled pattern. Previously, there was code
1058 that made a preliminary scan of the pattern in order to do this. That was
1059 OK when PCRE was new, but as the facilities have expanded, it has become
1060 harder and harder to keep it in step with the real compile phase, and there
1061 have been a number of bugs (see for example, 4 above). I have now found a
1062 cunning way of running the real compile function in a "fake" mode that
1063 enables it to compute how much memory it would need, while actually only
1064 ever using a few hundred bytes of working memory and without too many
1065 tests of the mode. This should make future maintenance and development
1066 easier. A side effect of this work is that the limit of 200 on the nesting
1067 depth of parentheses has been removed (though this was never a serious
1068 limitation, I suspect). However, there is a downside: pcre_compile() now
1069 runs more slowly than before (30% or more, depending on the pattern). I
1070 hope this isn't a big issue. There is no effect on runtime performance.
1071
1072 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
1073 newline (only possible for the last line of a file) and it was a
1074 pattern that set a locale (followed by /Lsomething), pcretest crashed.
1075
1076 19. Added additional timing features to pcretest. (1) The -tm option now times
1077 matching only, not compiling. (2) Both -t and -tm can be followed, as a
1078 separate command line item, by a number that specifies the number of
1079 repeats to use when timing. The default is 50000; this gives better
1080 precision, but takes uncomfortably long for very large patterns.
1081
1082 20. Extended pcre_study() to be more clever in cases where a branch of a
1083 subpattern has no definite first character. For example, (a*|b*)[cd] would
1084 previously give no result from pcre_study(). Now it recognizes that the
1085 first character must be a, b, c, or d.
1086
1087 21. There was an incorrect error "recursive call could loop indefinitely" if
1088 a subpattern (or the entire pattern) that was being tested for matching an
1089 empty string contained only one non-empty item after a nested subpattern.
1090 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
1091 incorrectly, because the \d was being skipped in the check.
1092
1093 22. The pcretest program now has a new pattern option /B and a command line
1094 option -b, which is equivalent to adding /B to every pattern. This causes
1095 it to show the compiled bytecode, without the additional information that
1096 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
1097 is the same as /B/I).
1098
1099 23. A new optimization is now able automatically to treat some sequences such
1100 as a*b as a*+b. More specifically, if something simple (such as a character
1101 or a simple class like \d) has an unlimited quantifier, and is followed by
1102 something that cannot possibly match the quantified thing, the quantifier
1103 is automatically "possessified".
1104
1105 24. A recursive reference to a subpattern whose number was greater than 39
1106 went wrong under certain circumstances in UTF-8 mode. This bug could also
1107 have affected the operation of pcre_study().
1108
1109 25. Realized that a little bit of performance could be had by replacing
1110 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
1111
1112 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
1113
1114 27. Possessive quantifiers such as a++ were previously implemented by turning
1115 them into atomic groups such as ($>a+). Now they have their own opcodes,
1116 which improves performance. This includes the automatically created ones
1117 from 23 above.
1118
1119 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
1120 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
1121 the first matched character to be a colon. This applied both to named and
1122 numbered groups.
1123
1124 29. The ucpinternal.h header file was missing its idempotency #ifdef.
1125
1126 30. I was sent a "project" file called libpcre.a.dev which I understand makes
1127 building PCRE on Windows easier, so I have included it in the distribution.
1128
1129 31. There is now a check in pcretest against a ridiculously large number being
1130 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
1131 loop, the loop is abandoned.
1132
1133 32. Forward references to subpatterns in conditions such as (?(2)...) where
1134 subpattern 2 is defined later cause pcre_compile() to search forwards in
1135 the pattern for the relevant set of parentheses. This search went wrong
1136 when there were unescaped parentheses in a character class, parentheses
1137 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
1138
1139 33. "Subroutine" calls and backreferences were previously restricted to
1140 referencing subpatterns earlier in the regex. This restriction has now
1141 been removed.
1142
1143 34. Added a number of extra features that are going to be in Perl 5.10. On the
1144 whole, these are just syntactic alternatives for features that PCRE had
1145 previously implemented using the Python syntax or my own invention. The
1146 other formats are all retained for compatibility.
1147
1148 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
1149 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
1150 also .NET compatible.
1151
1152 (b) A recursion or subroutine call to a named group can now be defined as
1153 (?&name) as well as (?P>name).
1154
1155 (c) A backreference to a named group can now be defined as \k<name> or
1156 \k'name' as well as (?P=name). The new forms, as well as being in Perl
1157 5.10, are also .NET compatible.
1158
1159 (d) A conditional reference to a named group can now use the syntax
1160 (?(<name>) or (?('name') as well as (?(name).
1161
1162 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
1163 groups (named and numbered) that are never evaluated inline, but can be
1164 called as "subroutines" from elsewhere. In effect, the DEFINE condition
1165 is always false. There may be only one alternative in such a group.
1166
1167 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
1168 as the simple (?(R). The condition is true only if the most recent
1169 recursion is that of the given number or name. It does not search out
1170 through the entire recursion stack.
1171
1172 (g) The escape \gN or \g{N} has been added, where N is a positive or
1173 negative number, specifying an absolute or relative reference.
1174
1175 35. Tidied to get rid of some further signed/unsigned compiler warnings and
1176 some "unreachable code" warnings.
1177
1178 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
1179 things, this adds five new scripts.
1180
1181 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
1182 There were also incompatibilities regarding the handling of \Q..\E inside
1183 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
1184 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
1185
1186 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
1187 matches an empty string, and forcibly breaks the loop. There were bugs in
1188 this code in non-simple cases. For a pattern such as ^(a()*)* matched
1189 against aaaa the result was just "a" rather than "aaaa", for example. Two
1190 separate and independent bugs (that affected different cases) have been
1191 fixed.
1192
1193 39. Refactored the code to abolish the use of different opcodes for small
1194 capturing bracket numbers. This is a tidy that I avoided doing when I
1195 removed the limit on the number of capturing brackets for 3.5 back in 2001.
1196 The new approach is not only tidier, it makes it possible to reduce the
1197 memory needed to fix the previous bug (38).
1198
1199 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
1200 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
1201 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
1202 mode.
1203
1204 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
1205 report.
1206
1207 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
1208 copy construction and assignment in the C++ wrapper.
1209
1210 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
1211 couple of bugs that could have given wrong results in the "--newline=crlf"
1212 case.
1213
1214 44. Added a number of casts and did some reorganization of signed/unsigned int
1215 variables following suggestions from Dair Grant. Also renamed the variable
1216 "this" as "item" because it is a C++ keyword.
1217
1218 45. Arranged for dftables to add
1219
1220 #include "pcre_internal.h"
1221
1222 to pcre_chartables.c because without it, gcc 4.x may remove the array
1223 definition from the final binary if PCRE is built into a static library and
1224 dead code stripping is activated.
1225
1226 46. For an unanchored pattern, if a match attempt fails at the start of a
1227 newline sequence, and the newline setting is CRLF or ANY, and the next two
1228 characters are CRLF, advance by two characters instead of one.
1229
1230
1231 Version 6.7 04-Jul-06
1232 ---------------------
1233
1234 1. In order to handle tests when input lines are enormously long, pcretest has
1235 been re-factored so that it automatically extends its buffers when
1236 necessary. The code is crude, but this _is_ just a test program. The
1237 default size has been increased from 32K to 50K.
1238
1239 2. The code in pcre_study() was using the value of the re argument before
1240 testing it for NULL. (Of course, in any sensible call of the function, it
1241 won't be NULL.)
1242
1243 3. The memmove() emulation function in pcre_internal.h, which is used on
1244 systems that lack both memmove() and bcopy() - that is, hardly ever -
1245 was missing a "static" storage class specifier.
1246
1247 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
1248 containing an extended class (one that cannot be represented by a bitmap
1249 because it contains high-valued characters or Unicode property items, e.g.
1250 [\pZ]). Almost always one would set UTF-8 mode when processing such a
1251 pattern, but PCRE should not loop if you do not (it no longer does).
1252 [Detail: two cases were found: (a) a repeated subpattern containing an
1253 extended class; (b) a recursive reference to a subpattern that followed a
1254 previous extended class. It wasn't skipping over the extended class
1255 correctly when UTF-8 mode was not set.]
1256
1257 5. A negated single-character class was not being recognized as fixed-length
1258 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
1259 compile error "lookbehind assertion is not fixed length".
1260
1261 6. The RunPerlTest auxiliary script was showing an unexpected difference
1262 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
1263 write a Perl script that can interpret lines of an input file either as
1264 byte characters or as UTF-8, which is what "perltest" was being required to
1265 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
1266 can't do is switch easily at run time between having the "use utf8;" pragma
1267 or not. In the end, I fudged it by using the RunPerlTest script to insert
1268 "use utf8;" explicitly for the UTF-8 tests.
1269
1270 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
1271 the end of the subject string, contrary to the documentation and to what
1272 Perl does. This was true of both matching functions. Now it matches only at
1273 the start of the subject and immediately after *internal* newlines.
1274
1275 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
1276 a pointer to an int instead of a pointer to an unsigned long int. This
1277 caused problems on 64-bit systems.
1278
1279 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
1280 instance of the 'standard' template library not being so standard".
1281
1282 10. There was no check on the number of named subpatterns nor the maximum
1283 length of a subpattern name. The product of these values is used to compute
1284 the size of the memory block for a compiled pattern. By supplying a very
1285 long subpattern name and a large number of named subpatterns, the size
1286 computation could be caused to overflow. This is now prevented by limiting
1287 the length of names to 32 characters, and the number of named subpatterns
1288 to 10,000.
1289
1290 11. Subpatterns that are repeated with specific counts have to be replicated in
1291 the compiled pattern. The size of memory for this was computed from the
1292 length of the subpattern and the repeat count. The latter is limited to
1293 65535, but there was no limit on the former, meaning that integer overflow
1294 could in principle occur. The compiled length of a repeated subpattern is
1295 now limited to 30,000 bytes in order to prevent this.
1296
1297 12. Added the optional facility to have named substrings with the same name.
1298
1299 13. Added the ability to use a named substring as a condition, using the
1300 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
1301 are numbers (not recommended). Forward references are permitted.
1302
1303 14. Added forward references in named backreferences (if you see what I mean).
1304
1305 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
1306 pattern could run off the end of the subject. For example, the pattern
1307 "(?s)(.{1,5})"8 did this with the subject "ab".
1308
1309 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
1310 PCRE_CASELESS was set when matching characters that were quantified with ?
1311 or *.
1312
1313 17. A character class other than a single negated character that had a minimum
1314 but no maximum quantifier - for example [ab]{6,} - was not handled
1315 correctly by pce_dfa_exec(). It would match only one character.
1316
1317 18. A valid (though odd) pattern that looked like a POSIX character
1318 class but used an invalid character after [ (for example [[,abc,]]) caused
1319 pcre_compile() to give the error "Failed: internal error: code overflow" or
1320 in some cases to crash with a glibc free() error. This could even happen if
1321 the pattern terminated after [[ but there just happened to be a sequence of
1322 letters, a binary zero, and a closing ] in the memory that followed.
1323
1324 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
1325 over the years. Originally (before any Unicode support), just the bottom 8
1326 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
1327 output from "man perlunicode" includes this:
1328
1329 The regular expression compiler produces polymorphic opcodes. That
1330 is, the pattern adapts to the data and automatically switches to
1331 the Unicode character scheme when presented with Unicode data--or
1332 instead uses a traditional byte scheme when presented with byte
1333 data.
1334
1335 Sadly, a wide octal escape does not cause a switch, and in a string with
1336 no other multibyte characters, these octal escapes are treated as before.
1337 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
1338 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
1339 Unicode string.
1340
1341 I have not perpetrated such confusion in PCRE. Up till now, it took just
1342 the bottom 8 bits, as in old Perl. I have now made octal escapes with
1343 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
1344 translate to the appropriate multibyte character.
1345
1346 29. Applied some refactoring to reduce the number of warnings from Microsoft
1347 and Borland compilers. This has included removing the fudge introduced
1348 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
1349 a warning about an unused variable.
1350
1351 21. PCRE has not included VT (character 0x0b) in the set of whitespace
1352 characters since release 4.0, because Perl (from release 5.004) does not.
1353 [Or at least, is documented not to: some releases seem to be in conflict
1354 with the documentation.] However, when a pattern was studied with
1355 pcre_study() and all its branches started with \s, PCRE still included VT
1356 as a possible starting character. Of course, this did no harm; it just
1357 caused an unnecessary match attempt.
1358
1359 22. Removed a now-redundant internal flag bit that recorded the fact that case
1360 dependency changed within the pattern. This was once needed for "required
1361 byte" processing, but is no longer used. This recovers a now-scarce options
1362 bit. Also moved the least significant internal flag bit to the most-
1363 significant bit of the word, which was not previously used (hangover from
1364 the days when it was an int rather than a uint) to free up another bit for
1365 the future.
1366
1367 23. Added support for CRLF line endings as well as CR and LF. As well as the
1368 default being selectable at build time, it can now be changed at runtime
1369 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
1370 specify that it is scanning data with non-default line endings.
1371
1372 24. Changed the definition of CXXLINK to make it agree with the definition of
1373 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
1374
1375 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
1376 recursions. This makes a big different to stack usage for some patterns.
1377
1378 26. If a subpattern containing a named recursion or subroutine reference such
1379 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
1380 the space required for the compiled pattern went wrong and gave too small a
1381 value. Depending on the environment, this could lead to "Failed: internal
1382 error: code overflow at offset 49" or "glibc detected double free or
1383 corruption" errors.
1384
1385 27. Applied patches from Google (a) to support the new newline modes and (b) to
1386 advance over multibyte UTF-8 characters in GlobalReplace.
1387
1388 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
1389 difference for some implementation of PCRE in some Windows version.
1390
1391 29. Added some extra testing facilities to pcretest:
1392
1393 \q<number> in a data line sets the "match limit" value
1394 \Q<number> in a data line sets the "match recursion limt" value
1395 -S <number> sets the stack size, where <number> is in megabytes
1396
1397 The -S option isn't available for Windows.
1398
1399
1400 Version 6.6 06-Feb-06
1401 ---------------------
1402
1403 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
1404 in pcreposix.h. I have copied the definition from pcre.h.
1405
1406 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
1407 because pcre.h is no longer a built file.
1408
1409 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
1410 not normally included in the compiled code.
1411
1412
1413 Version 6.5 01-Feb-06
1414 ---------------------
1415
1416 1. When using the partial match feature with pcre_dfa_exec(), it was not
1417 anchoring the second and subsequent partial matches at the new starting
1418 point. This could lead to incorrect results. For example, with the pattern
1419 /1234/, partially matching against "123" and then "a4" gave a match.
1420
1421 2. Changes to pcregrep:
1422
1423 (a) All non-match returns from pcre_exec() were being treated as failures
1424 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
1425 error message is output. Some extra information is given for the
1426 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
1427 probably the only errors that are likely to be caused by users (by
1428 specifying a regex that has nested indefinite repeats, for instance).
1429 If there are more than 20 of these errors, pcregrep is abandoned.
1430
1431 (b) A binary zero was treated as data while matching, but terminated the
1432 output line if it was written out. This has been fixed: binary zeroes
1433 are now no different to any other data bytes.
1434
1435 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
1436 used to set a locale for matching. The --locale=xxxx long option has
1437 been added (no short equivalent) to specify a locale explicitly on the
1438 pcregrep command, overriding the environment variables.
1439
1440 (d) When -B was used with -n, some line numbers in the output were one less
1441 than they should have been.
1442
1443 (e) Added the -o (--only-matching) option.
1444
1445 (f) If -A or -C was used with -c (count only), some lines of context were
1446 accidentally printed for the final match.
1447
1448 (g) Added the -H (--with-filename) option.
1449
1450 (h) The combination of options -rh failed to suppress file names for files
1451 that were found from directory arguments.
1452
1453 (i) Added the -D (--devices) and -d (--directories) options.
1454
1455 (j) Added the -F (--fixed-strings) option.
1456
1457 (k) Allow "-" to be used as a file name for -f as well as for a data file.
1458
1459 (l) Added the --colo(u)r option.
1460
1461 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
1462 is not present by default.
1463
1464 3. A nasty bug was discovered in the handling of recursive patterns, that is,
1465 items such as (?R) or (?1), when the recursion could match a number of
1466 alternatives. If it matched one of the alternatives, but subsequently,
1467 outside the recursion, there was a failure, the code tried to back up into
1468 the recursion. However, because of the way PCRE is implemented, this is not
1469 possible, and the result was an incorrect result from the match.
1470
1471 In order to prevent this happening, the specification of recursion has
1472 been changed so that all such subpatterns are automatically treated as
1473 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
1474
1475 4. I had overlooked the fact that, in some locales, there are characters for
1476 which isalpha() is true but neither isupper() nor islower() are true. In
1477 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
1478 and ordfeminine) are like this. This affected the treatment of \w and \W
1479 when they appeared in character classes, but not when they appeared outside
1480 a character class. The bit map for "word" characters is now created
1481 separately from the results of isalnum() instead of just taking it from the
1482 upper, lower, and digit maps. (Plus the underscore character, of course.)
1483
1484 5. The above bug also affected the handling of POSIX character classes such as
1485 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
1486 permanent tables. Instead, the bit maps for such a class were previously
1487 created as the appropriate unions of the upper, lower, and digit bitmaps.
1488 Now they are created by subtraction from the [[:word:]] class, which has
1489 its own bitmap.
1490
1491 6. The [[:blank:]] character class matches horizontal, but not vertical space.
1492 It is created by subtracting the vertical space characters (\x09, \x0a,
1493 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
1494 subtraction was done in the overall bitmap for a character class, meaning
1495 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
1496 be recognized. This bug has been fixed.
1497
1498 7. Patches from the folks at Google:
1499
1500 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
1501 real life, but is still worth protecting against".
1502
1503 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
1504 regular expressions".
1505
1506 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
1507 have it.
1508
1509 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
1510 "configure" and the latter not, in order to fix a problem somebody had
1511 with compiling the Arg class on HP-UX.
1512
1513 (e) Improve the error-handling of the C++ wrapper a little bit.
1514
1515 (f) New tests for checking recursion limiting.
1516
1517 8. The pcre_memmove() function, which is used only if the environment does not
1518 have a standard memmove() function (and is therefore rarely compiled),
1519 contained two bugs: (a) use of int instead of size_t, and (b) it was not
1520 returning a result (though PCRE never actually uses the result).
1521
1522 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
1523 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
1524 returned instead of calling malloc() with an overflowing number that would
1525 most likely cause subsequent chaos.
1526
1527 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
1528
1529 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
1530 with this option is matched, the nmatch and pmatch options of regexec() are
1531 ignored.
1532
1533 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
1534 provided in case anyone wants to the the POSIX interface with UTF-8
1535 strings.
1536
1537 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
1538 C++ linking (needed for some HP-UX environments).
1539
1540 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
1541 (unused parameter) and in the pcre_printint() function (omitted "default"
1542 switch label when the default is to do nothing).
1543
1544 15. Added some code to make it possible, when PCRE is compiled as a C++
1545 library, to replace subject pointers for pcre_exec() with a smart pointer
1546 class, thus making it possible to process discontinuous strings.
1547
1548 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
1549 much the same function. They were added by different people who were trying
1550 to make PCRE easy to compile on non-Unix systems. It has been suggested
1551 that PCRE_EXPORT be abolished now that there is more automatic apparatus
1552 for compiling on Windows systems. I have therefore replaced it with
1553 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
1554 defaults to "extern" for C or "extern C" for C++, which works fine on
1555 Unix-like systems. It is now possible to override the value of PCRE_DATA_
1556 SCOPE with something explicit in config.h. In addition:
1557
1558 (a) pcreposix.h still had just "extern" instead of either of these macros;
1559 I have replaced it with PCRE_DATA_SCOPE.
1560
1561 (b) Functions such as _pcre_xclass(), which are internal to the library,
1562 but external in the C sense, all had PCRE_EXPORT in their definitions.
1563 This is apparently wrong for the Windows case, so I have removed it.
1564 (It makes no difference on Unix-like systems.)
1565
1566 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
1567 of recursive calls to match(). This is different to MATCH_LIMIT because
1568 that limits the total number of calls to match(), not all of which increase
1569 the depth of recursion. Limiting the recursion depth limits the amount of
1570 stack (or heap if NO_RECURSE is set) that is used. The default can be set
1571 when PCRE is compiled, and changed at run time. A patch from Google adds
1572 this functionality to the C++ interface.
1573
1574 18. Changes to the handling of Unicode character properties:
1575
1576 (a) Updated the table to Unicode 4.1.0.
1577
1578 (b) Recognize characters that are not in the table as "Cn" (undefined).
1579
1580 (c) I revised the way the table is implemented to a much improved format
1581 which includes recognition of ranges. It now supports the ranges that
1582 are defined in UnicodeData.txt, and it also amalgamates other
1583 characters into ranges. This has reduced the number of entries in the
1584 table from around 16,000 to around 3,000, thus reducing its size
1585 considerably. I realized I did not need to use a tree structure after
1586 all - a binary chop search is just as efficient. Having reduced the
1587 number of entries, I extended their size from 6 bytes to 8 bytes to
1588 allow for more data.
1589
1590 (d) Added support for Unicode script names via properties such as \p{Han}.
1591
1592 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
1593 matching that character.
1594
1595 20. When matching a repeated Unicode property with a minimum greater than zero,
1596 (for example \pL{2,}), PCRE could look past the end of the subject if it
1597 reached it while seeking the minimum number of characters. This could
1598 happen only if some of the characters were more than one byte long, because
1599 there is a check for at least the minimum number of bytes.
1600
1601 21. Refactored the implementation of \p and \P so as to be more general, to
1602 allow for more different types of property in future. This has changed the
1603 compiled form incompatibly. Anybody with saved compiled patterns that use
1604 \p or \P will have to recompile them.
1605
1606 22. Added "Any" and "L&" to the supported property types.
1607
1608 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
1609 but give a compile time error if the value is greater than 0xff.
1610
1611 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
1612 accidentally not being installed or uninstalled.
1613
1614 25. The pcre.h file was built from pcre.h.in, but the only changes that were
1615 made were to insert the current release number. This seemed silly, because
1616 it made things harder for people building PCRE on systems that don't run
1617 "configure". I have turned pcre.h into a distributed file, no longer built
1618 by "configure", with the version identification directly included. There is
1619 no longer a pcre.h.in file.
1620
1621 However, this change necessitated a change to the pcre-config script as
1622 well. It is built from pcre-config.in, and one of the substitutions was the
1623 release number. I have updated configure.ac so that ./configure now finds
1624 the release number by grepping pcre.h.
1625
1626 26. Added the ability to run the tests under valgrind.
1627
1628
1629 Version 6.4 05-Sep-05
1630 ---------------------
1631
1632 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
1633 "--" to be printed when multiple files were scanned, even when none of the
1634 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
1635 consider it to be a bug, and have restored the previous behaviour.
1636
1637 2. A couple of code tidies to get rid of compiler warnings.
1638
1639 3. The pcretest program used to cheat by referring to symbols in the library
1640 whose names begin with _pcre_. These are internal symbols that are not
1641 really supposed to be visible externally, and in some environments it is
1642 possible to suppress them. The cheating is now confined to including
1643 certain files from the library's source, which is a bit cleaner.
1644
1645 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
1646 file's purpose clearer.
1647
1648 5. Reorganized pcre_ucp_findchar().
1649
1650
1651 Version 6.3 15-Aug-05
1652 ---------------------
1653
1654 1. The file libpcre.pc.in did not have general read permission in the tarball.
1655
1656 2. There were some problems when building without C++ support:
1657
1658 (a) If C++ support was not built, "make install" and "make test" still
1659 tried to test it.
1660
1661 (b) There were problems when the value of CXX was explicitly set. Some
1662 changes have been made to try to fix these, and ...
1663
1664 (c) --disable-cpp can now be used to explicitly disable C++ support.
1665
1666 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
1667 backslash in a target when C++ was disabled. This confuses some
1668 versions of "make", apparently. Using an intermediate variable solves
1669 this. (Same for CPP_LOBJ.)
1670
1671 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
1672 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
1673 necessary on certain architectures.
1674
1675 4. Added a setting of -export-symbols-regex to the link command to remove
1676 those symbols that are exported in the C sense, but actually are local
1677 within the library, and not documented. Their names all begin with
1678 "_pcre_". This is not a perfect job, because (a) we have to except some
1679 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
1680 available (and never for static libraries). I have made a note to try to
1681 find a way round (a) in the future.
1682
1683
1684 Version 6.2 01-Aug-05
1685 ---------------------
1686
1687 1. There was no test for integer overflow of quantifier values. A construction
1688 such as {1111111111111111} would give undefined results. What is worse, if
1689 a minimum quantifier for a parenthesized subpattern overflowed and became
1690 negative, the calculation of the memory size went wrong. This could have
1691 led to memory overwriting.
1692
1693 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
1694
1695 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
1696 operating environments where this matters.
1697
1698 4. Applied Giuseppe Maxia's patch to add additional features for controlling
1699 PCRE options from within the C++ wrapper.
1700
1701 5. Named capturing subpatterns were not being correctly counted when a pattern
1702 was compiled. This caused two problems: (a) If there were more than 100
1703 such subpatterns, the calculation of the memory needed for the whole
1704 compiled pattern went wrong, leading to an overflow error. (b) Numerical
1705 back references of the form \12, where the number was greater than 9, were
1706 not recognized as back references, even though there were sufficient
1707 previous subpatterns.
1708
1709 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
1710 versions of gcc, e.g. 2.95.4.
1711
1712
1713 Version 6.1 21-Jun-05
1714 ---------------------
1715
1716 1. There was one reference to the variable "posix" in pcretest.c that was not
1717 surrounded by "#if !defined NOPOSIX".
1718
1719 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
1720 the cross-check on the old pcre_info() function, for the benefit of the
1721 cut-down version of PCRE that is currently imported into Exim.
1722
1723 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
1724 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
1725 patterns that start (?i) but I don't think that matters. The use of (?i) is
1726 just an example; this all applies to the other options as well.
1727
1728 4. Since libtool seems to echo the compile commands it is issuing, the output
1729 from "make" can be reduced a bit by putting "@" in front of each libtool
1730 compile command.
1731
1732 5. Patch from the folks at Google for configure.in to be a bit more thorough
1733 in checking for a suitable C++ installation before trying to compile the
1734 C++ stuff. This should fix a reported problem when a compiler was present,
1735 but no suitable headers.
1736
1737 6. The man pages all had just "PCRE" as their title. I have changed them to
1738 be the relevant file name. I have also arranged that these names are
1739 retained in the file doc/pcre.txt, which is a concatenation in text format
1740 of all the man pages except the little individual ones for each function.
1741
1742 7. The NON-UNIX-USE file had not been updated for the different set of source
1743 files that come with release 6. I also added a few comments about the C++
1744 wrapper.
1745
1746
1747 Version 6.0 07-Jun-05
1748 ---------------------
1749
1750 1. Some minor internal re-organization to help with my DFA experiments.
1751
1752 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
1753 didn't matter for the library itself when fully configured, but did matter
1754 when compiling without UCP support, or within Exim, where the ucp files are
1755 not imported.
1756
1757 3. Refactoring of the library code to split up the various functions into
1758 different source modules. The addition of the new DFA matching code (see
1759 below) to a single monolithic source would have made it really too
1760 unwieldy, quite apart from causing all the code to be include in a
1761 statically linked application, when only some functions are used. This is
1762 relevant even without the DFA addition now that patterns can be compiled in
1763 one application and matched in another.
1764
1765 The downside of splitting up is that there have to be some external
1766 functions and data tables that are used internally in different modules of
1767 the library but which are not part of the API. These have all had their
1768 names changed to start with "_pcre_" so that they are unlikely to clash
1769 with other external names.
1770
1771 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
1772 a different (DFA) algorithm. Although it is slower than the original
1773 function, it does have some advantages for certain types of matching
1774 problem.
1775
1776 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
1777 including restarting after a partial match.
1778
1779 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
1780 defined when compiling for Windows was sent to me. I have put it into the
1781 code, though I have no means of testing or verifying it.
1782
1783 7. Added the pcre_refcount() auxiliary function.
1784
1785 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
1786 match before or at the first newline in the subject string. In pcretest,
1787 the /f option on a pattern can be used to set this.
1788
1789 9. A repeated \w when used in UTF-8 mode with characters greater than 256
1790 would behave wrongly. This has been present in PCRE since release 4.0.
1791
1792 10. A number of changes to the pcregrep command:
1793
1794 (a) Refactored how -x works; insert ^(...)$ instead of setting
1795 PCRE_ANCHORED and checking the length, in preparation for adding
1796 something similar for -w.
1797
1798 (b) Added the -w (match as a word) option.
1799
1800 (c) Refactored the way lines are read and buffered so as to have more
1801 than one at a time available.
1802
1803 (d) Implemented a pcregrep test script.
1804
1805 (e) Added the -M (multiline match) option. This allows patterns to match
1806 over several lines of the subject. The buffering ensures that at least
1807 8K, or the rest of the document (whichever is the shorter) is available
1808 for matching (and similarly the previous 8K for lookbehind assertions).
1809
1810 (f) Changed the --help output so that it now says
1811
1812 -w, --word-regex(p)
1813
1814 instead of two lines, one with "regex" and the other with "regexp"
1815 because that confused at least one person since the short forms are the
1816 same. (This required a bit of code, as the output is generated
1817 automatically from a table. It wasn't just a text change.)
1818
1819 (g) -- can be used to terminate pcregrep options if the next thing isn't an
1820 option but starts with a hyphen. Could be a pattern or a path name
1821 starting with a hyphen, for instance.
1822
1823 (h) "-" can be given as a file name to represent stdin.
1824
1825 (i) When file names are being printed, "(standard input)" is used for
1826 the standard input, for compatibility with GNU grep. Previously
1827 "<stdin>" was used.
1828
1829 (j) The option --label=xxx can be used to supply a name to be used for
1830 stdin when file names are being printed. There is no short form.
1831
1832 (k) Re-factored the options decoding logic because we are going to add
1833 two more options that take data. Such options can now be given in four
1834 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
1835
1836 (l) Added the -A, -B, and -C options for requesting that lines of context
1837 around matches be printed.
1838
1839 (m) Added the -L option to print the names of files that do not contain
1840 any matching lines, that is, the complement of -l.
1841
1842 (n) The return code is 2 if any file cannot be opened, but pcregrep does
1843 continue to scan other files.
1844
1845 (o) The -s option was incorrectly implemented. For compatibility with other
1846 greps, it now suppresses the error message for a non-existent or non-
1847 accessible file (but not the return code). There is a new option called
1848 -q that suppresses the output of matching lines, which was what -s was
1849 previously doing.
1850
1851 (p) Added --include and --exclude options to specify files for inclusion
1852 and exclusion when recursing.
1853
1854 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
1855 Hopefully, it now does.
1856
1857 12. Missing cast in pcre_study().
1858
1859 13. Added an "uninstall" target to the makefile.
1860
1861 14. Replaced "extern" in the function prototypes in Makefile.in with
1862 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
1863 world, but is set differently for Windows.
1864
1865 15. Added a second compiling function called pcre_compile2(). The only
1866 difference is that it has an extra argument, which is a pointer to an
1867 integer error code. When there is a compile-time failure, this is set
1868 non-zero, in addition to the error test pointer being set to point to an
1869 error message. The new argument may be NULL if no error number is required
1870 (but then you may as well call pcre_compile(), which is now just a
1871 wrapper). This facility is provided because some applications need a
1872 numeric error indication, but it has also enabled me to tidy up the way
1873 compile-time errors are handled in the POSIX wrapper.
1874
1875 16. Added VPATH=.libs to the makefile; this should help when building with one
1876 prefix path and installing with another. (Or so I'm told by someone who
1877 knows more about this stuff than I do.)
1878
1879 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
1880 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
1881 match everything, including newlines. This is not POSIX-compatible, but
1882 somebody wanted the feature. From pcretest it can be activated by using
1883 both the P and the s flags.
1884
1885 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
1886
1887 19. libpcre.pc was being incorrectly installed as executable.
1888
1889 20. A couple of places in pcretest check for end-of-line by looking for '\n';
1890 it now also looks for '\r' so that it will work unmodified on Windows.
1891
1892 21. Added Google's contributed C++ wrapper to the distribution.
1893
1894 22. Added some untidy missing memory free() calls in pcretest, to keep
1895 Electric Fence happy when testing.
1896
1897
1898
1899 Version 5.0 13-Sep-04
1900 ---------------------
1901
1902 1. Internal change: literal characters are no longer packed up into items
1903 containing multiple characters in a single byte-string. Each character
1904 is now matched using a separate opcode. However, there may be more than one
1905 byte in the character in UTF-8 mode.
1906
1907 2. The pcre_callout_block structure has two new fields: pattern_position and
1908 next_item_length. These contain the offset in the pattern to the next match
1909 item, and its length, respectively.
1910
1911 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
1912 insertion of callouts before each pattern item. Added the /C option to
1913 pcretest to make use of this.
1914
1915 4. On the advice of a Windows user, the lines
1916
1917 #if defined(_WIN32) || defined(WIN32)
1918 _setmode( _fileno( stdout ), 0x8000 );
1919 #endif /* defined(_WIN32) || defined(WIN32) */
1920
1921 have been added to the source of pcretest. This apparently does useful
1922 magic in relation to line terminators.
1923
1924 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
1925 for the benefit of those environments where the "b" makes a difference.
1926
1927 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
1928 to know about it. I have put a hack into configure.in that adds in code
1929 to set GCC=yes if CC=icc. This seems to end up at a point in the
1930 generated configure script that is early enough to affect the setting of
1931 compiler options, which is what is needed, but I have no means of testing
1932 whether it really works. (The user who reported this had patched the
1933 generated configure script, which of course I cannot do.)
1934
1935 LATER: After change 22 below (new libtool files), the configure script
1936 seems to know about icc (and also ecc). Therefore, I have commented out
1937 this hack in configure.in.
1938
1939 7. Added support for pkg-config (2 patches were sent in).
1940
1941 8. Negated POSIX character classes that used a combination of internal tables
1942 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
1943 [[:^ascii]]. Typically, they would match almost any characters. The other
1944 POSIX classes were not broken in this way.
1945
1946 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
1947 to find the match, as PCRE was deluded into thinking that the match had to
1948 start at the start point or following a newline. The same bug applied to
1949 patterns with negative forward assertions or any backward assertions
1950 preceding ".*" at the start, unless the pattern required a fixed first
1951 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
1952
1953 10. In UTF-8 mode, when moving forwards in the subject after a failed match
1954 starting at the last subject character, bytes beyond the end of the subject
1955 string were read.
1956
1957 11. Renamed the variable "class" as "classbits" to make life easier for C++
1958 users. (Previously there was a macro definition, but it apparently wasn't
1959 enough.)
1960
1961 12. Added the new field "tables" to the extra data so that tables can be passed
1962 in at exec time, or the internal tables can be re-selected. This allows
1963 a compiled regex to be saved and re-used at a later time by a different
1964 program that might have everything at different addresses.
1965
1966 13. Modified the pcre-config script so that, when run on Solaris, it shows a
1967 -R library as well as a -L library.
1968
1969 14. The debugging options of pcretest (-d on the command line or D on a
1970 pattern) showed incorrect output for anything following an extended class
1971 that contained multibyte characters and which was followed by a quantifier.
1972
1973 15. Added optional support for general category Unicode character properties
1974 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
1975 support. It adds about 90K to the size of the library. The meanings of the
1976 inbuilt class escapes such as \d and \s have NOT been changed.
1977
1978 16. Updated pcredemo.c to include calls to free() to release the memory for the
1979 compiled pattern.
1980
1981 17. The generated file chartables.c was being created in the source directory
1982 instead of in the building directory. This caused the build to fail if the
1983 source directory was different from the building directory, and was
1984 read-only.
1985
1986 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
1987 file. No doubt somebody will tell me if they don't make sense... Also added
1988 Dan Mooney's comments about building on OpenVMS.
1989
1990 19. Added support for partial matching via the PCRE_PARTIAL option for
1991 pcre_exec() and the \P data escape in pcretest.
1992
1993 20. Extended pcretest with 3 new pattern features:
1994
1995 (i) A pattern option of the form ">rest-of-line" causes pcretest to
1996 write the compiled pattern to the file whose name is "rest-of-line".
1997 This is a straight binary dump of the data, with the saved pointer to
1998 the character tables forced to be NULL. The study data, if any, is
1999 written too. After writing, pcretest reads a new pattern.
2000
2001 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
2002 compiled pattern from the given file. There must not be any
2003 occurrences of "<" in the file name (pretty unlikely); if there are,
2004 pcretest will instead treat the initial "<" as a pattern delimiter.
2005 After reading in the pattern, pcretest goes on to read data lines as
2006 usual.
2007
2008 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
2009 and 16-bit fields in a compiled pattern, to simulate a pattern that
2010 was compiled on a host of opposite endianness.
2011
2012 21. The pcre-exec() function can now cope with patterns that were compiled on
2013 hosts of opposite endianness, with this restriction:
2014
2015 As for any compiled expression that is saved and used later, the tables
2016 pointer field cannot be preserved; the extra_data field in the arguments
2017 to pcre_exec() should be used to pass in a tables address if a value
2018 other than the default internal tables were used at compile time.
2019
2020 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
2021 now diagnosed as an error. Previously, most of the time, a negative number
2022 would have been treated as zero, but if in addition "ovector" was passed as
2023 NULL, a crash could occur.
2024
2025 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
2026 new versions from the libtool 1.5 distribution (the last one is a copy of
2027 a file called libtool.m4). This seems to have fixed the need to patch
2028 "configure" to support Darwin 1.3 (which I used to do). However, I still
2029 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
2030 workstation).
2031
2032 24. Changed the PCRE licence to be the more standard "BSD" licence.
2033
2034
2035 Version 4.5 01-Dec-03
2036 ---------------------
2037
2038 1. There has been some re-arrangement of the code for the match() function so
2039 that it can be compiled in a version that does not call itself recursively.
2040 Instead, it keeps those local variables that need separate instances for
2041 each "recursion" in a frame on the heap, and gets/frees frames whenever it
2042 needs to "recurse". Keeping track of where control must go is done by means
2043 of setjmp/longjmp. The whole thing is implemented by a set of macros that
2044 hide most of the details from the main code, and operates only if
2045 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
2046 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
2047 operating.
2048
2049 To make it easier for callers to provide specially tailored get/free
2050 functions for this usage, two new functions, pcre_stack_malloc, and
2051 pcre_stack_free, are used. They are always called in strict stacking order,
2052 and the size of block requested is always the same.
2053
2054 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
2055 PCRE has been compiled to use the stack or the heap for recursion. The
2056 -C option of pcretest uses this to show which version is compiled.
2057
2058 A new data escape \S, is added to pcretest; it causes the amounts of store
2059 obtained and freed by both kinds of malloc/free at match time to be added
2060 to the output.
2061
2062 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
2063 what's available on my current Linux desktop machine.
2064
2065 3. When matching a UTF-8 string, the test for a valid string at the start has
2066 been extended. If start_offset is not zero, PCRE now checks that it points
2067 to a byte that is the start of a UTF-8 character. If not, it returns
2068 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
2069 this is necessary because there may be backward assertions in the pattern.
2070 When matching the same subject several times, it may save resources to use
2071 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
2072
2073 4. The code for checking the validity of UTF-8 strings has been tightened so
2074 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
2075 containing "overlong sequences".
2076
2077 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
2078 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
2079 should have been "&", but it just so happened that all the cases this let
2080 through by mistake were picked up later in the function.
2081
2082 6. I had used a variable called "isblank" - this is a C99 function, causing
2083 some compilers to warn. To avoid this, I renamed it (as "blankclass").
2084
2085 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
2086 prompting; (b) run "./pcretest /dev/null" at the start of the test script
2087 so the version is shown; (c) stop "make test" echoing "./RunTest".
2088
2089 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
2090
2091 9. The prototype for memmove() for systems that don't have it was using
2092 size_t, but the inclusion of the header that defines size_t was later. I've
2093 moved the #includes for the C headers earlier to avoid this.
2094
2095 10. Added some adjustments to the code to make it easier to compiler on certain
2096 special systems:
2097
2098 (a) Some "const" qualifiers were missing.
2099 (b) Added the macro EXPORT before all exported functions; by default this
2100 is defined to be empty.
2101 (c) Changed the dftables auxiliary program (that builds chartables.c) so
2102 that it reads its output file name as an argument instead of writing
2103 to the standard output and assuming this can be redirected.
2104
2105 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
2106 class containing characters with values greater than 255, PCRE compilation
2107 went into a loop.
2108
2109 12. A recursive reference to a subpattern that was within another subpattern
2110 that had a minimum quantifier of zero caused PCRE to crash. For example,
2111 (x(y(?2))z)? provoked this bug with a subject that got as far as the
2112 recursion. If the recursively-called subpattern itself had a zero repeat,
2113 that was OK.
2114
2115 13. In pcretest, the buffer for reading a data line was set at 30K, but the
2116 buffer into which it was copied (for escape processing) was still set at
2117 1024, so long lines caused crashes.
2118
2119 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
2120 "internal error: code overflow...". This applied to any character class
2121 that was followed by a possessive quantifier.
2122
2123 15. Modified the Makefile to add libpcre.la as a prerequisite for
2124 libpcreposix.la because I was told this is needed for a parallel build to
2125 work.
2126
2127 16. If a pattern that contained .* following optional items at the start was
2128 studied, the wrong optimizing data was generated, leading to matching
2129 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
2130 matching string must start with a or b or c. The correct conclusion for
2131 this pattern is that a match can start with any character.
2132
2133
2134 Version 4.4 13-Aug-03
2135 ---------------------
2136
2137 1. In UTF-8 mode, a character class containing characters with values between
2138 127 and 255 was not handled correctly if the compiled pattern was studied.
2139 In fixing this, I have also improved the studying algorithm for such
2140 classes (slightly).
2141
2142 2. Three internal functions had redundant arguments passed to them. Removal
2143 might give a very teeny performance improvement.
2144
2145 3. Documentation bug: the value of the capture_top field in a callout is *one
2146 more than* the number of the hightest numbered captured substring.
2147
2148 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
2149 in incorrectly linking with a previously installed version. They now link
2150 explicitly with libpcre.la.
2151
2152 5. configure.in no longer needs to recognize Cygwin specially.
2153
2154 6. A problem in pcre.in for Windows platforms is fixed.
2155
2156 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
2157 pcretest, it used to include the size of the study block as part of its
2158 output. Unfortunately, the structure contains a field that has a different
2159 size on different hardware architectures. This meant that the tests that
2160 showed this size failed. As the block is currently always of a fixed size,
2161 this information isn't actually particularly useful in pcretest output, so
2162 I have just removed it.
2163
2164 8. Three pre-processor statements accidentally did not start in column 1.
2165 Sadly, there are *still* compilers around that complain, even though
2166 standard C has not required this for well over a decade. Sigh.
2167
2168 9. In pcretest, the code for checking callouts passed small integers in the
2169 callout_data field, which is a void * field. However, some picky compilers
2170 complained about the casts involved for this on 64-bit systems. Now
2171 pcretest passes the address of the small integer instead, which should get
2172 rid of the warnings.
2173
2174 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
2175 both compile and run time, and gives an error if an invalid UTF-8 sequence
2176 is found. There is a option for disabling this check in cases where the
2177 string is known to be correct and/or the maximum performance is wanted.
2178
2179 11. In response to a bug report, I changed one line in Makefile.in from
2180
2181 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
2182 to
2183 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
2184
2185 to look similar to other lines, but I have no way of telling whether this
2186 is the right thing to do, as I do not use Windows. No doubt I'll get told
2187 if it's wrong...
2188
2189
2190 Version 4.3 21-May-03
2191 ---------------------
2192
2193 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
2194 Makefile.
2195
2196 2. Some refactoring to improve the quality of the code:
2197
2198 (i) The utf8_table... variables are now declared "const".
2199
2200 (ii) The code for \cx, which used the "case flipping" table to upper case
2201 lower case letters, now just substracts 32. This is ASCII-specific,
2202 but the whole concept of \cx is ASCII-specific, so it seems
2203 reasonable.
2204
2205 (iii) PCRE was using its character types table to recognize decimal and
2206 hexadecimal digits in the pattern. This is silly, because it handles
2207 only 0-9, a-f, and A-F, but the character types table is locale-
2208 specific, which means strange things might happen. A private
2209 table is now used for this - though it costs 256 bytes, a table is
2210 much faster than multiple explicit tests. Of course, the standard
2211 character types table is still used for matching digits in subject
2212 strings against \d.
2213
2214 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
2215 ending in _t are). So I've renamed it as ESC_tee.
2216
2217 3. The first argument for regexec() in the POSIX wrapper should have been
2218 defined as "const".
2219
2220 4. Changed pcretest to use malloc() for its buffers so that they can be
2221 Electric Fenced for debugging.
2222
2223 5. There were several places in the code where, in UTF-8 mode, PCRE would try
2224 to read one or more bytes before the start of the subject string. Often this
2225 had no effect on PCRE's behaviour, but in some circumstances it could
2226 provoke a segmentation fault.
2227
2228 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
2229 to try to read one or more bytes before the start of the subject string.
2230
2231 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
2232 UTF-8 support could misbehave in various ways if the subject string
2233 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
2234 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
2235 back over UTF-8 characters.)
2236
2237
2238 Version 4.2 14-Apr-03
2239 ---------------------
2240
2241 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
2242
2243 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
2244 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
2245 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
2246 [WIN_PREFIX]: new variable, "cyg" for Cygwin
2247 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
2248 and BUILD_EXEEXT
2249 Note: automatic setting of the BUILD variables is not yet working
2250 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
2251 compile-time but not at link-time
2252 [LINK]: use for linking executables only
2253 make different versions for Windows and non-Windows
2254 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
2255 libraries
2256 [LINK_FOR_BUILD]: new variable
2257 [OBJEXT]: use throughout
2258 [EXEEXT]: use throughout
2259 <winshared>: new target
2260 <wininstall>: new target
2261 <dftables.o>: use native compiler
2262 <dftables>: use native linker
2263 <install>: handle Windows platform correctly
2264 <clean>: ditto
2265 <check>: ditto
2266 copy DLL to top builddir before testing
2267
2268 As part of these changes, -no-undefined was removed again. This was reported
2269 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
2270 in any case.
2271
2272 3. Some tidies to get rid of compiler warnings:
2273
2274 . In the match_data structure, match_limit was an unsigned long int, whereas
2275 match_call_count was an int. I've made them both unsigned long ints.
2276
2277 . In pcretest the fact that a const uschar * doesn't automatically cast to
2278 a void * provoked a warning.
2279
2280 . Turning on some more compiler warnings threw up some "shadow" variables
2281 and a few more missing casts.
2282
2283 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
2284 option, a class that contained a single character with a value between 128
2285 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
2286
2287 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
2288 option, a class that contained several characters, but with at least one
2289 whose value was between 128 and 255 caused PCRE to crash.
2290
2291
2292 Version 4.1 12-Mar-03
2293 ---------------------
2294
2295 1. Compiling with gcc -pedantic found a couple of places where casts were
2296 needed, and a string in dftables.c that was longer than standard compilers are
2297 required to support.
2298
2299 2. Compiling with Sun's compiler found a few more places where the code could
2300 be tidied up in order to avoid warnings.
2301
2302 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
2303 first of these names is deprecated in the latest Autoconf in favour of the name
2304 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
2305 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
2306 analogy I have changed it to CFLAGS_FOR_BUILD.
2307
2308 4. Added -no-undefined to the linking command in the Makefile, because this is
2309 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
2310 linking step for the pcreposix library.
2311
2312 5. PCRE was failing to diagnose the case of two named groups with the same
2313 name.
2314
2315 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
2316 literal character that is needed in the subject for a match, and scans along to
2317 ensure that it is present before embarking on the full matching process. This
2318 saves time in cases of nested unlimited repeats that are never going to match.
2319 Problem: the scan can take a lot of time if the subject is very long (e.g.
2320 megabytes), thus penalizing straightforward matches. It is now done only if the
2321 amount of subject to be scanned is less than 1000 bytes.
2322
2323 7. A lesser problem with the same optimization is that it was recording the
2324 first character of an anchored pattern as "needed", thus provoking a search
2325 right along the subject, even when the first match of the pattern was going to
2326 fail. The "needed" character is now not set for anchored patterns, unless it
2327 follows something in the pattern that is of non-fixed length. Thus, it still
2328 fulfils its original purpose of finding quick non-matches in cases of nested
2329 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
2330
2331
2332 Version 4.0 17-Feb-03
2333 ---------------------
2334
2335 1. If a comment in an extended regex that started immediately after a meta-item
2336 extended to the end of string, PCRE compiled incorrect data. This could lead to
2337 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
2338
2339 2. Moved to autoconf 2.53 and libtool 1.4.2.
2340
2341 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
2342 the special perltest8 script is no longer needed - all the tests can be run
2343 from a single perltest script.
2344
2345 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
2346 by \s. It has now been removed in PCRE. This means it isn't recognized as
2347 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
2348 class [:space:] *does* include VT, thereby creating a mess.
2349
2350 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
2351 space and tab.
2352
2353 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
2354 its new features into the main test script, reducing the number of scripts.
2355
2356 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
2357 were backward compatible, and made the (?i) apply to the whole pattern, as if
2358 /i were given. Now it behaves more logically, and applies the option setting
2359 only to what follows. PCRE has been changed to follow suit. However, if it
2360 finds options settings right at the start of the pattern, it extracts them into
2361 the global options, as before. Thus, they show up in the info data.
2362
2363 8. Added support for the \Q...\E escape sequence. Characters in between are
2364 treated as literals. This is slightly different from Perl in that $ and @ are
2365 also handled as literals inside the quotes. In Perl, they will cause variable
2366 interpolation. Note the following examples:
2367
2368 Pattern PCRE matches Perl matches
2369
2370 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
2371 \Qabc\$xyz\E abc\$xyz abc\$xyz
2372 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
2373
2374 For compatibility with Perl, \Q...\E sequences are recognized inside character
2375 classes as well as outside them.
2376
2377 9. Re-organized 3 code statements in pcretest to avoid "overflow in
2378 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
2379 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
2380 signed/unsigned warnings.
2381
2382 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
2383 option for pcretest, so I've replaced it by a simple function that does just
2384 that job.
2385
2386 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
2387 "pcregrep -".
2388
2389 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
2390 Java package. This provides some syntactic sugar for simple cases of what my
2391 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
2392 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
2393 item, you can use this simplified notation. Note that only makes sense with
2394 greedy quantifiers. Consequently, the use of the possessive quantifier forces
2395 greediness, whatever the setting of the PCRE_UNGREEDY option.
2396
2397 13. A change of greediness default within a pattern was not taking effect at
2398 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
2399 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
2400 was abstracted outside.
2401
2402 14. PCRE now supports the \G assertion. It is true when the current matching
2403 position is at the start point of the match. This differs from \A when the
2404 starting offset is non-zero. Used with the /g option of pcretest (or similar
2405 code), it works in the same way as it does for Perl's /g option. If all
2406 alternatives of a regex begin with \G, the expression is anchored to the start
2407 match position, and the "anchored" flag is set in the compiled expression.
2408
2409 15. Some bugs concerning the handling of certain option changes within patterns
2410 have been fixed. These applied to options other than (?ims). For example,
2411 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
2412 been the other way round. Some of this was related to change 7 above.
2413
2414 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
2415 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
2416 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
2417 POSIX classes only within a class (e.g. /[[:alpha:]]/).
2418
2419 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
2420 mode. Unlike ".", it always matches newline, whatever the setting of
2421 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
2422 assertions. Perl allows it, but it doesn't (in general) work because it can't
2423 calculate the length of the lookbehind. At least, that's the case for Perl
2424 5.8.0 - I've been told they are going to document that it doesn't work in
2425 future.
2426
2427 18. Added an error diagnosis for escapes that PCRE does not support: these are
2428 \L, \l, \N, \P, \p, \U, \u, and \X.
2429
2430 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
2431 reading past the end of the pattern in cases such as /[abcd/.
2432
2433 20. PCRE was getting more memory than necessary for patterns with classes that
2434 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
2435
2436 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
2437 compiling PCRE for use with Virtual Pascal.
2438
2439 22. Small fix to the Makefile to make it work properly if the build is done
2440 outside the source tree.
2441
2442 23. Added a new extension: a condition to go with recursion. If a conditional
2443 subpattern starts with (?(R) the "true" branch is used if recursion has
2444 happened, whereas the "false" branch is used only at the top level.
2445
2446 24. When there was a very long string of literal characters (over 255 bytes
2447 without UTF support, over 250 bytes with UTF support), the computation of how
2448 much memory was required could be incorrect, leading to segfaults or other
2449 strange effects.
2450
2451 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
2452 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
2453 there was a subsequent back reference to those brackets. This meant that, for
2454 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
2455 possible to check for precisely this case. All we can do is abandon the
2456 optimization if .* occurs inside capturing brackets when there are any back
2457 references whatsoever. (See below for a better fix that came later.)
2458
2459 26. The handling of the optimization for finding the first character of a
2460 non-anchored pattern, and for finding a character that is required later in the
2461 match were failing in some cases. This didn't break the matching; it just
2462 failed to optimize when it could. The way this is done has been re-implemented.
2463
2464 27. Fixed typo in error message for invalid (?R item (it said "(?p").
2465
2466 28. Added a new feature that provides some of the functionality that Perl
2467 provides with (?{...}). The facility is termed a "callout". The way it is done
2468 in PCRE is for the caller to provide an optional function, by setting
2469 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
2470 global variable. By default it is unset, which disables all calling out. To get
2471 the function called, the regex must include (?C) at appropriate points. This
2472 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
2473 This provides a means of identifying different callout points. When PCRE
2474 reaches such a point in the regex, if pcre_callout has been set, the external
2475 function is called. It is provided with data in a structure called
2476 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
2477 matching continues; if it returns a non-zero value, the match at the current
2478 point fails. However, backtracking will occur if possible. [This was changed
2479 later and other features added - see item 49 below.]
2480
2481 29. pcretest is upgraded to test the callout functionality. It provides a
2482 callout function that displays information. By default, it shows the start of
2483 the match and the current position in the text. There are some new data escapes
2484 to vary what happens:
2485
2486 \C+ in addition, show current contents of captured substrings
2487 \C- do not supply a callout function
2488 \C!n return 1 when callout number n is reached
2489 \C!n!m return 1 when callout number n is reached for the mth time
2490
2491 30. If pcregrep was called with the -l option and just a single file name, it
2492 output "<stdin>" if a match was found, instead of the file name.
2493
2494 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
2495 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
2496 pcre_exec(). This saves a malloc/free per call. The default value of
2497 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
2498 when configuring.
2499
2500 32. The default maximum size of a compiled pattern is 64K. There have been a
2501 few cases of people hitting this limit. The code now uses macros to handle the
2502 storing of links as offsets within the compiled pattern. It defaults to 2-byte
2503 links, but this can be changed to 3 or 4 bytes by --with-link-size when
2504 configuring. Tests 2 and 5 work only with 2-byte links because they output
2505 debugging information about compiled patterns.
2506
2507 33. Internal code re-arrangements:
2508
2509 (a) Moved the debugging function for printing out a compiled regex into
2510 its own source file (printint.c) and used #include to pull it into
2511 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
2512 separate copies.
2513
2514 (b) Defined the list of op-code names for debugging as a macro in
2515 internal.h so that it is next to the definition of the opcodes.
2516
2517 (c) Defined a table of op-code lengths for simpler skipping along compiled
2518 code. This is again a macro in internal.h so that it is next to the
2519 definition of the opcodes.
2520
2521 34. Added support for recursive calls to individual subpatterns, along the
2522 lines of Robin Houston's patch (but implemented somewhat differently).
2523
2524 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
2525 allow it to read and process whole directories in Win32. This code was
2526 contributed by Lionel Fourquaux; it has not been tested by me.
2527
2528 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
2529 used to name a group. Names consist of alphanumerics and underscores, and must
2530 be unique. Back references use the syntax (?P=name) and recursive calls use
2531 (?P>name) which is a PCRE extension to the Python extension. Groups still have
2532 numbers. The function pcre_fullinfo() can be used after compilation to extract
2533 a name/number map. There are three relevant calls:
2534
2535 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
2536 PCRE_INFO_NAMECOUNT yields the number of entries
2537 PCRE_INFO_NAMETABLE yields a pointer to the map.
2538
2539 The map is a vector of fixed-size entries. The size of each entry depends on
2540 the length of the longest name used. The first two bytes of each entry are the
2541 group number, most significant byte first. There follows the corresponding
2542 name, zero terminated. The names are in alphabetical order.
2543
2544 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
2545 case instead of 255. Making it the same both with and without UTF-8 support
2546 means that the same test output works with both.
2547
2548 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
2549 calling malloc() with a zero argument.
2550
2551 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
2552 optimization. I've improved things by keeping a bitmap of backreferences with
2553 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
2554 fact referenced, the optimization can be applied. It is unlikely that a
2555 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
2556 the match to follow \n) will appear inside brackets with a number greater than
2557 31, but if it does, any back reference > 31 suppresses the optimization.
2558
2559 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
2560 of disabling numbered capturing parentheses. Any opening parenthesis that is
2561 not followed by ? behaves as if it were followed by ?: but named parentheses
2562 can still be used for capturing (and they will acquire numbers in the usual
2563 way).
2564
2565 41. Redesigned the return codes from the match() function into yes/no/error so
2566 that errors can be passed back from deep inside the nested calls. A malloc
2567 failure while inside a recursive subpattern call now causes the
2568 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
2569
2570 42. It is now possible to set a limit on the number of times the match()
2571 function is called in a call to pcre_exec(). This facility makes it possible to
2572 limit the amount of recursion and backtracking, though not in a directly
2573 obvious way, because the match() function is used in a number of different
2574 circumstances. The count starts from zero for each position in the subject
2575 string (for non-anchored patterns). The default limit is, for compatibility, a
2576 large number, namely 10 000 000. You can change this in two ways:
2577
2578 (a) When configuring PCRE before making, you can use --with-match-limit=n
2579 to set a default value for the compiled library.
2580
2581 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
2582 a different value is set. See 45 below.
2583
2584 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
2585
2586 43. Added a new function pcre_config(int, void *) to enable run-time extraction
2587 of things that can be changed at compile time. The first argument specifies
2588 what is wanted and the second points to where the information is to be placed.
2589 The current list of available information is:
2590
2591 PCRE_CONFIG_UTF8
2592
2593 The output is an integer that is set to one if UTF-8 support is available;
2594 otherwise it is set to zero.
2595
2596 PCRE_CONFIG_NEWLINE
2597
2598 The output is an integer that it set to the value of the code that is used for
2599 newline. It is either LF (10) or CR (13).
2600
2601 PCRE_CONFIG_LINK_SIZE
2602
2603 The output is an integer that contains the number of bytes used for internal
2604 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
2605
2606 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
2607
2608 The output is an integer that contains the threshold above which the POSIX
2609 interface uses malloc() for output vectors. See item 31 above.
2610
2611 PCRE_CONFIG_MATCH_LIMIT
2612
2613 The output is an unsigned integer that contains the default limit of the number
2614 of match() calls in a pcre_exec() execution. See 42 above.
2615
2616 44. pcretest has been upgraded by the addition of the -C option. This causes it
2617 to extract all the available output from the new pcre_config() function, and to
2618 output it. The program then exits immediately.
2619
2620 45. A need has arisen to pass over additional data with calls to pcre_exec() in
2621 order to support additional features. One way would have been to define
2622 pcre_exec2() (for example) with extra arguments, but this would not have been
2623 extensible, and would also have required all calls to the original function to
2624 be mapped to the new one. Instead, I have chosen to extend the mechanism that
2625 is used for passing in "extra" data from pcre_study().
2626
2627 The pcre_extra structure is now exposed and defined in pcre.h. It currently
2628 contains the following fields:
2629
2630 flags a bitmap indicating which of the following fields are set
2631 study_data opaque data from pcre_study()
2632 match_limit a way of specifying a limit on match() calls for a specific
2633 call to pcre_exec()
2634 callout_data data for callouts (see 49 below)
2635
2636 The flag bits are also defined in pcre.h, and are
2637
2638 PCRE_EXTRA_STUDY_DATA
2639 PCRE_EXTRA_MATCH_LIMIT
2640 PCRE_EXTRA_CALLOUT_DATA
2641
2642 The pcre_study() function now returns one of these new pcre_extra blocks, with
2643 the actual study data pointed to by the study_data field, and the
2644 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
2645 before. That is, this change is entirely upwards-compatible and requires no
2646 change to existing code.
2647
2648 If you want to pass in additional data to pcre_exec(), you can either place it
2649 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
2650 block.
2651
2652 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
2653 data string contains the escape sequence \M, pcretest calls pcre_exec() several
2654 times with different match limits, until it finds the minimum value needed for
2655 pcre_exec() to complete. The value is then output. This can be instructive; for
2656 most simple matches the number is quite small, but for pathological cases it
2657 gets very large very quickly.
2658
2659 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
2660 returns the size of the data block pointed to by the study_data field in a
2661 pcre_extra block, that is, the value that was passed as the argument to
2662 pcre_malloc() when PCRE was getting memory in which to place the information
2663 created by pcre_study(). The fourth argument should point to a size_t variable.
2664 pcretest has been extended so that this information is shown after a successful
2665 pcre_study() call when information about the compiled regex is being displayed.
2666
2667 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
2668 because what follows is always an absolute path. (Later: it turns out that this
2669 is more than cosmetic for MinGW, because it doesn't like empty path
2670 components.)
2671
2672 49. Some changes have been made to the callout feature (see 28 above):
2673
2674 (i) A callout function now has three choices for what it returns:
2675
2676 0 => success, carry on matching
2677 > 0 => failure at this point, but backtrack if possible
2678 < 0 => serious error, return this value from pcre_exec()
2679
2680 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
2681 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
2682 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
2683 use by callout functions. It will never be used by PCRE itself.
2684
2685 (ii) The pcre_extra structure (see 45 above) has a void * field called
2686 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
2687 pcre_callout_block structure has a field of the same name. The contents of
2688 the field passed in the pcre_extra structure are passed to the callout
2689 function in the corresponding field in the callout block. This makes it
2690 easier to use the same callout-containing regex from multiple threads. For
2691 testing, the pcretest program has a new data escape
2692
2693 \C*n pass the number n (may be negative) as callout_data
2694
2695 If the callout function in pcretest receives a non-zero value as
2696 callout_data, it returns that value.
2697
2698 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
2699 there were some redundant $(CFLAGS) in commands that are now specified as
2700 $(LINK), which already includes $(CFLAGS).
2701
2702 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
2703 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
2704 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
2705 one-byte characters throughout. Note that case-insensitive matching applies
2706 only to characters whose values are less than 256. PCRE doesn't support the
2707 notion of cases for higher-valued characters.
2708
2709 (i) A character class whose characters are all within 0-255 is handled as
2710 a bit map, and the map is inverted for negative classes. Previously, a
2711 character > 255 always failed to match such a class; however it should
2712 match if the class was a negative one (e.g. [^ab]). This has been fixed.
2713
2714 (ii) A negated character class with a single character < 255 is coded as
2715 "not this character" (OP_NOT). This wasn't working properly when the test
2716 character was multibyte, either singly or repeated.
2717
2718 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
2719 mode, for example: \x{100}{2,3}.
2720
2721 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
2722 singly or repeated) now correctly test multibyte characters. However,
2723 PCRE doesn't recognize any characters with values greater than 255 as
2724 digits, spaces, or word characters. Such characters always match \D, \S,
2725 and \W, and never match \d, \s, or \w.
2726
2727 (v) Classes may now contain characters and character ranges with values
2728 greater than 255. For example: [ab\x{100}-\x{400}].
2729
2730 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
2731 PCRE in UTF-8 mode.
2732
2733 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
2734 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
2735 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
2736 value.)
2737
2738 53. The single man page has become too large. I have therefore split it up into
2739 a number of separate man pages. These also give rise to individual HTML pages;
2740 these are now put in a separate directory, and there is an index.html page that
2741 lists them all. Some hyperlinking between the pages has been installed.
2742
2743 54. Added convenience functions for handling named capturing parentheses.
2744
2745 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
2746 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
2747 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
2748 are faulted.
2749
2750 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
2751 calling configure. These values are used when compiling the dftables.c program
2752 which is run to generate the source of the default character tables. They
2753 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
2754 you will need to set these values.
2755
2756 57. Updated the building process for Windows DLL, as provided by Fred Cox.
2757
2758
2759 Version 3.9 02-Jan-02
2760 ---------------------
2761
2762 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
2763
2764 2. If --disable-static was given, the building process failed when trying to
2765 build pcretest and pcregrep. (For some reason it was using libtool to compile
2766 them, which is not right, as they aren't part of the library.)
2767
2768
2769 Version 3.8 18-Dec-01
2770 ---------------------
2771
2772 1. The experimental UTF-8 code was completely screwed up. It was packing the
2773 bytes in the wrong order. How dumb can you get?
2774
2775
2776 Version 3.7 29-Oct-01
2777 ---------------------
2778
2779 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
2780 This caused pcretest, when used on the test data, to segfault. Unfortunately,
2781 this didn't happen under Solaris 8, where I normally test things.
2782
2783 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
2784 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
2785 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
2786 this fix an hour or so after the initial 3.7 release.)
2787
2788
2789 Version 3.6 23-Oct-01
2790 ---------------------
2791
2792 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
2793 offsets passed as NULL with zero offset count.
2794
2795 2. The config.guess and config.sub files had not been updated when I moved to
2796 the latest autoconf.
2797
2798
2799 Version 3.5 15-Aug-01
2800 ---------------------
2801
2802 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
2803 had been forgotten.
2804
2805 2. By using declared but undefined structures, we can avoid using "void"
2806 definitions in pcre.h while keeping the internal definitions of the structures
2807 private.
2808
2809 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
2810 user point of view, this means that both static and shared libraries are built
2811 by default, but this can be individually controlled. More of the work of
2812 handling this static/shared cases is now inside libtool instead of PCRE's make
2813 file.
2814
2815 4. The pcretest utility is now installed along with pcregrep because it is
2816 useful for users (to test regexs) and by doing this, it automatically gets
2817 relinked by libtool. The documentation has been turned into a man page, so
2818 there are now .1, .txt, and .html versions in /doc.
2819
2820 5. Upgrades to pcregrep:
2821 (i) Added long-form option names like gnu grep.
2822 (ii) Added --help to list all options with an explanatory phrase.
2823 (iii) Added -r, --recursive to recurse into sub-directories.
2824 (iv) Added -f, --file to read patterns from a file.
2825
2826 6. pcre_exec() was referring to its "code" argument before testing that
2827 argument for NULL (and giving an error if it was NULL).
2828
2829 7. Upgraded Makefile.in to allow for compiling in a different directory from
2830 the source directory.
2831
2832 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
2833 options bits, the pointer it was passed was to an int instead of to an unsigned
2834 long int. This mattered only on 64-bit systems.
2835
2836 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
2837 generated) instead of pcre.in, which it its source. Also made the same change
2838 in several of the .c files.
2839
2840 10. A new release of gcc defines printf() as a macro, which broke pcretest
2841 because it had an ifdef in the middle of a string argument for printf(). Fixed
2842 by using separate calls to printf().
2843
2844 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
2845 script, to force use of CR or LF instead of \n in the source. On non-Unix
2846 systems, the value can be set in config.h.
2847
2848 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
2849 absolute limit. Changed the text of the error message to make this clear, and
2850 likewise updated the man page.
2851
2852 13. The limit of 99 on the number of capturing subpatterns has been removed.
2853 The new limit is 65535, which I hope will not be a "real" limit.
2854
2855
2856 Version 3.4 22-Aug-00
2857 ---------------------
2858
2859 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
2860
2861 2. Diagnose condition (?(0) as an error instead of crashing on matching.
2862
2863
2864 Version 3.3 01-Aug-00
2865 ---------------------
2866
2867 1. If an octal character was given, but the value was greater than \377, it
2868 was not getting masked to the least significant bits, as documented. This could
2869 lead to crashes in some systems.
2870
2871 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
2872 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
2873
2874 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
2875 These just pass their arguments on to (pcre_free)(), but they are provided
2876 because some uses of PCRE bind it to non-C systems that can call its functions,
2877 but cannot call free() or pcre_free() directly.
2878
2879 4. Add "make test" as a synonym for "make check". Corrected some comments in
2880 the Makefile.
2881
2882 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
2883 Makefile.
2884
2885 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
2886 command called pgrep for grepping around the active processes.
2887
2888 7. Added the beginnings of support for UTF-8 character strings.
2889
2890 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
2891 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
2892 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
2893 out for the ar command.)
2894
2895
2896 Version 3.2 12-May-00
2897 ---------------------
2898
2899 This is purely a bug fixing release.
2900
2901 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
2902 of ZA. This was just one example of several cases that could provoke this bug,
2903 which was introduced by change 9 of version 2.00. The code for breaking
2904 infinite loops after an iteration that matches an empty string was't working
2905 correctly.
2906
2907 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
2908 when matched against abbab (for example). After matching an empty string, it
2909 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
2910 caused it to match further down the string than it should.
2911
2912 3. The code contained an inclusion of sys/types.h. It isn't clear why this
2913 was there because it doesn't seem to be needed, and it causes trouble on some
2914 systems, as it is not a Standard C header. It has been removed.
2915
2916 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
2917 were reported on the Macintosh. The changes were from
2918
2919 while ((c = *(++ptr)) != 0 && c != '\n');
2920 to
2921 while ((c = *(++ptr)) != 0 && c != '\n') ;
2922
2923 Totally extraordinary, but if that's what it takes...
2924
2925 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
2926 available. Added HAVE_BCOPY and an autoconf test for it; if neither
2927 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
2928 assumes the way PCRE uses memmove() (always moving upwards).
2929
2930 6. PCRE is being used in one environment where strchr() is not available. There
2931 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
2932 faster code anyway.
2933
2934
2935 Version 3.1 09-Feb-00
2936 ---------------------
2937
2938 The only change in this release is the fixing of some bugs in Makefile.in for
2939 the "install" target:
2940
2941 (1) It was failing to install pcreposix.h.
2942
2943 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
2944
2945
2946 Version 3.0 01-Feb-00
2947 ---------------------
2948
2949 1. Add support for the /+ modifier to perltest (to output $` like it does in
2950 pcretest).
2951
2952 2. Add support for the /g modifier to perltest.
2953
2954 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
2955 matches null strings.
2956
2957 4. Fix perltest so that it doesn't do unwanted things when fed an empty
2958 pattern. Perl treats empty patterns specially - it reuses the most recent
2959 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
2960 effect.
2961
2962 5. The POSIX interface was broken in that it was just handing over the POSIX
2963 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
2964 required a bigger vector, with some working space on the end. This means that
2965 the POSIX wrapper now has to get and free some memory, and copy the results.
2966
2967 6. Added some simple autoconf support, placing the test data and the
2968 documentation in separate directories, re-organizing some of the
2969 information files, and making it build pcre-config (a GNU standard). Also added
2970 libtool support for building PCRE as a shared library, which is now the
2971 default.
2972
2973 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
2974 09 are not valid octal constants. Single digits will be used for minor values
2975 less than 10.
2976
2977 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
2978 existing programs that set these in the POSIX interface can use PCRE without
2979 modification.
2980
2981 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
2982 return all that pcre_info() returns, plus additional data. The pcre_info()
2983 function is retained for compatibility, but is considered to be obsolete.
2984
2985 10. Added experimental recursion feature (?R) to handle one common case that
2986 Perl 5.6 will be able to do with (?p{...}).
2987
2988 11. Added support for POSIX character classes like [:alpha:], which Perl is
2989 adopting.
2990
2991
2992 Version 2.08 31-Aug-99
2993 ----------------------
2994
2995 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
2996 trying to match at the startoffset position, but instead was moving forward to
2997 the next newline as if a previous match had failed.
2998
2999 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
3000 and could get into a loop if a null string was matched other than at the start
3001 of the subject.
3002
3003 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
3004 be distinguished at compile time, and for completeness also added PCRE_DATE.
3005
3006 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
3007 in GnuWin32 environments.
3008
3009
3010 Version 2.07 29-Jul-99
3011 ----------------------
3012
3013 1. The documentation is now supplied in plain text form and HTML as well as in
3014 the form of man page sources.
3015
3016 2. C++ compilers don't like assigning (void *) values to other pointer types.
3017 In particular this affects malloc(). Although there is no problem in Standard
3018 C, I've put in casts to keep C++ compilers happy.
3019
3020 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
3021 should be (const char *).
3022
3023 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
3024 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
3025 However, I haven't made this a standard facility. The documentation doesn't
3026 mention it, and the Makefile doesn't support it.
3027
3028 5. The Makefile now contains an "install" target, with editable destinations at
3029 the top of the file. The pcretest program is not installed.
3030
3031 6. pgrep -V now gives the PCRE version number and date.
3032
3033 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
3034 causing the entire string to be ignored, instead of just the last character.
3035
3036 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
3037 non-matching string, it can take a very, very long time, even for strings of
3038 quite modest length, because of the nested recursion. PCRE now does better in
3039 some of these cases. It does this by remembering the last required literal
3040 character in the pattern, and pre-searching the subject to ensure it is present
3041 before running the real match. In other words, it applies a heuristic to detect
3042 some types of certain failure quickly, and in the above example, if presented
3043 with a string that has no trailing " it gives "no match" very quickly.
3044
3045 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
3046 other alternatives are tried instead.
3047
3048
3049 Version 2.06 09-Jun-99
3050 ----------------------
3051
3052 1. Change pcretest's output for amount of store used to show just the code
3053 space, because the remainder (the data block) varies in size between 32-bit and
3054 64-bit systems.
3055
3056 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
3057 start matching at. This allows lookbehinds to work when searching for multiple
3058 occurrences in a string.
3059
3060 3. Added additional options to pcretest for testing multiple occurrences:
3061
3062 /+ outputs the rest of the string that follows a match
3063 /g loops for multiple occurrences, using the new startoffset argument
3064 /G loops for multiple occurrences by passing an incremented pointer
3065
3066 4. PCRE wasn't doing the "first character" optimization for patterns starting
3067 with \b or \B, though it was doing it for other lookbehind assertions. That is,
3068 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
3069 the letter 'x'. On long subject strings, this gives a significant speed-up.
3070
3071
3072 Version 2.05 21-Apr-99
3073 ----------------------
3074
3075 1. Changed the type of magic_number from int to long int so that it works
3076 properly on 16-bit systems.
3077
3078 2. Fixed a bug which caused patterns starting with .* not to work correctly
3079 when the subject string contained newline characters. PCRE was assuming
3080 anchoring for such patterns in all cases, which is not correct because .* will
3081 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
3082 DOTALL is set at top level; otherwise it knows that patterns starting with .*
3083 must be retried after every newline in the subject.
3084
3085
3086 Version 2.04 18-Feb-99
3087 ----------------------
3088
3089 1. For parenthesized subpatterns with repeats whose minimum was zero, the
3090 computation of the store needed to hold the pattern was incorrect (too large).
3091 If such patterns were nested a few deep, this could multiply and become a real
3092 problem.
3093
3094 2. Added /M option to pcretest to show the memory requirement of a specific
3095 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
3096
3097 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
3098 compiled in such a way that the backtracking after subsequent failure was
3099 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
3100 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
3101
3102
3103 Version 2.03 02-Feb-99
3104 ----------------------
3105
3106 1. Fixed typo and small mistake in man page.
3107
3108 2. Added 4th condition (GPL supersedes if conflict) and created separate
3109 LICENCE file containing the conditions.
3110
3111 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
3112 Perl, that is the internal \ allows the delimiter to be included in the
3113 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
3114 the final delimiter, add \ to the end of the pattern (to test the error).
3115
3116 4. Added the convenience functions for extracting substrings after a successful
3117 match. Updated pcretest to make it able to test these functions.
3118
3119
3120 Version 2.02 14-Jan-99
3121 ----------------------
3122
3123 1. Initialized the working variables associated with each extraction so that
3124 their saving and restoring doesn't refer to uninitialized store.
3125
3126 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
3127 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
3128 fix the problem.
3129
3130 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
3131 calls, and wasn't printing the correct value for compiling calls. Increased the
3132 default value of LOOPREPEAT, and the number of significant figures in the
3133 times.
3134
3135 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
3136
3137 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
3138 a building problem on Windows NT with a FAT file system.
3139
3140
3141 Version 2.01 21-Oct-98
3142 ----------------------
3143
3144 1. Changed the API for pcre_compile() to allow for the provision of a pointer
3145 to character tables built by pcre_maketables() in the current locale. If NULL
3146 is passed, the default tables are used.
3147
3148
3149 Version 2.00 24-Sep-98
3150 ----------------------
3151
3152 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
3153 it any more.
3154
3155 2. Allow quantification of (?>) groups, and make it work correctly.
3156
3157 3. The first character computation wasn't working for (?>) groups.
3158
3159 4. Correct the implementation of \Z (it is permitted to match on the \n at the
3160 end of the subject) and add 5.005's \z, which really does match only at the
3161 very end of the subject.
3162
3163 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
3164
3165 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
3166 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
3167 localized options. All options to pcre_study() were also removed.
3168
3169 7. Add other new features from 5.005:
3170
3171 $(?<= positive lookbehind
3172 $(?<! negative lookbehind
3173 (?imsx-imsx) added the unsetting capability
3174 such a setting is global if at outer level; local otherwise
3175 (?imsx-imsx:) non-capturing groups with option setting
3176 (?(cond)re|re) conditional pattern matching
3177
3178 A backreference to itself in a repeated group matches the previous
3179 captured string.
3180
3181 8. General tidying up of studying (both automatic and via "study")
3182 consequential on the addition of new assertions.
3183
3184 9. As in 5.005, unlimited repeated groups that could match an empty substring
3185 are no longer faulted at compile time. Instead, the loop is forcibly broken at
3186 runtime if any iteration does actually match an empty substring.
3187
3188 10. Include the RunTest script in the distribution.
3189
3190 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
3191 discrepancies, some of which were old and were also with respect to 5.004. They
3192 have now been fixed.
3193
3194
3195 Version 1.09 28-Apr-98
3196 ----------------------
3197
3198 1. A negated single character class followed by a quantifier with a minimum
3199 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
3200 program crashes, or just wrong answers. This did not apply to negated classes
3201 containing more than one character, or to minima other than one.
3202
3203
3204 Version 1.08 27-Mar-98
3205 ----------------------
3206
3207 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
3208
3209 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
3210 latter must appear before anything that relies on it in the pattern.
3211
3212
3213 Version 1.07 16-Feb-98
3214 ----------------------
3215
3216 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
3217 repeat of a potentially empty string).
3218
3219
3220 Version 1.06 23-Jan-98
3221 ----------------------
3222
3223 1. Added Markus Oberhumer's little patches for C++.
3224
3225 2. Literal strings longer than 255 characters were broken.
3226
3227
3228 Version 1.05 23-Dec-97
3229 ----------------------
3230
3231 1. Negated character classes containing more than one character were failing if
3232 PCRE_CASELESS was set at run time.
3233
3234
3235 Version 1.04 19-Dec-97
3236 ----------------------
3237
3238 1. Corrected the man page, where some "const" qualifiers had been omitted.
3239
3240 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
3241 input syntax.
3242
3243 3. Fixed memory leak which occurred when a regex with back references was
3244 matched with an offsets vector that wasn't big enough. The temporary memory
3245 that is used in this case wasn't being freed if the match failed.
3246
3247 4. Tidied pcretest to ensure it frees memory that it gets.
3248
3249 5. Temporary memory was being obtained in the case where the passed offsets
3250 vector was exactly big enough.
3251
3252 6. Corrected definition of offsetof() from change 5 below.
3253
3254 7. I had screwed up change 6 below and broken the rules for the use of
3255 setjmp(). Now fixed.
3256
3257
3258 Version 1.03 18-Dec-97
3259 ----------------------
3260
3261 1. A erroneous regex with a missing opening parenthesis was correctly
3262 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
3263 on some systems.
3264
3265 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
3266 it was reported that one broken compiler failed on the former because "code" is
3267 also an independent variable.
3268
3269 3. The erroneous regex a[]b caused an array overrun reference.
3270
3271 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
3272 fail on data ending with that character. (It was going on too far, and checking
3273 the next character, typically a binary zero.) This was specific to the
3274 optimized code for single-character negative classes.
3275
3276 5. Added a contributed patch from the TIN world which does the following:
3277
3278 + Add an undef for memmove, in case the the system defines a macro for it.
3279
3280 + Add a definition of offsetof(), in case there isn't one. (I don't know
3281 the reason behind this - offsetof() is part of the ANSI standard - but
3282 it does no harm).
3283
3284 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
3285 most of the places where whitespace preceded '#'. I have given up and
3286 allowed the remaining 2 cases to be at the margin.
3287
3288 + Rename some variables in pcre to eliminate shadowing. This seems very
3289 pedantic, but does no harm, of course.
3290
3291 6. Moved the call to setjmp() into its own function, to get rid of warnings
3292 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
3293
3294 7. Constructs such as \d{8,} were compiling into the equivalent of
3295 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
3296 outcome, but in this particular case used more store than had been allocated,
3297 which caused the bug to be discovered because it threw up an internal error.
3298
3299 8. The debugging code in both pcre and pcretest for outputting the compiled
3300 form of a regex was going wrong in the case of back references followed by
3301 curly-bracketed repeats.
3302
3303
3304 Version 1.02 12-Dec-97
3305 ----------------------
3306
3307 1. Typos in pcre.3 and comments in the source fixed.
3308
3309 2. Applied a contributed patch to get rid of places where it used to remove
3310 'const' from variables, and fixed some signed/unsigned and uninitialized
3311 variable warnings.
3312
3313 3. Added the "runtest" target to Makefile.
3314
3315 4. Set default compiler flag to -O2 rather than just -O.
3316
3317
3318 Version 1.01 19-Nov-97
3319 ----------------------
3320
3321 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
3322 like /([ab]*)*/, that is, for classes with more than one character in them.
3323
3324 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
3325 as /((?>a*))*/ (a PCRE_EXTRA facility).
3326
3327
3328 Version 1.00 18-Nov-97
3329 ----------------------
3330
3331 1. Added compile-time macros to support systems such as SunOS4 which don't have
3332 memmove() or strerror() but have other things that can be used instead.
3333
3334 2. Arranged that "make clean" removes the executables.
3335
3336
3337 Version 0.99 27-Oct-97
3338 ----------------------
3339
3340 1. Fixed bug in code for optimizing classes with only one character. It was
3341 initializing a 32-byte map regardless, which could cause it to run off the end
3342 of the memory it had got.
3343
3344 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
3345
3346
3347 Version 0.98 22-Oct-97
3348 ----------------------
3349
3350 1. Fixed bug in code for handling temporary memory usage when there are more
3351 back references than supplied space in the ovector. This could cause segfaults.
3352
3353
3354 Version 0.97 21-Oct-97
3355 ----------------------
3356
3357 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
3358
3359 2. Optimized negated single characters not to use a bit map.
3360
3361 3. Brought error texts together as macro definitions; clarified some of them;
3362 fixed one that was wrong - it said "range out of order" when it meant "invalid
3363 escape sequence".
3364
3365 4. Changed some char * arguments to const char *.
3366
3367 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
3368
3369 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
3370 pcretest.
3371
3372
3373 Version 0.96 16-Oct-97
3374 ----------------------
3375
3376 1. Added a simple "pgrep" utility to the distribution.
3377
3378 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
3379 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
3380 where "ddd" means "one or more decimal digits".
3381
3382 3. Fixed serious bug. If a pattern had a back reference, but the call to
3383 pcre_exec() didn't supply a large enough ovector to record the related
3384 identifying subpattern, the match always failed. PCRE now remembers the number
3385 of the largest back reference, and gets some temporary memory in which to save
3386 the offsets during matching if necessary, in order to ensure that
3387 backreferences always work.
3388
3389 4. Increased the compatibility with Perl in a number of ways:
3390
3391 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
3392 to request this handling. The option can be set at compile or exec time.
3393
3394 (b) $ matches before a terminating newline by default; an option
3395 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
3396 mode). The option can be set at compile or exec time.
3397
3398 (c) The handling of \ followed by a digit other than 0 is now supposed to be
3399 the same as Perl's. If the decimal number it represents is less than 10
3400 or there aren't that many previous left capturing parentheses, an octal
3401 escape is read. Inside a character class, it's always an octal escape,
3402 even if it is a single digit.
3403
3404 (d) An escaped but undefined alphabetic character is taken as a literal,
3405 unless PCRE_EXTRA is set. Currently this just reserves the remaining
3406 escapes.
3407
3408 (e) {0} is now permitted. (The previous item is removed from the compiled
3409 pattern).
3410
3411 5. Changed all the names of code files so that the basic parts are no longer
3412 than 10 characters, and abolished the teeny "globals.c" file.
3413
3414 6. Changed the handling of character classes; they are now done with a 32-byte
3415 bit map always.
3416
3417 7. Added the -d and /D options to pcretest to make it possible to look at the
3418 internals of compilation without having to recompile pcre.
3419
3420
3421 Version 0.95 23-Sep-97
3422 ----------------------
3423
3424 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
3425 \x20 at the start of a run of normal characters. These were being treated as
3426 real characters, instead of the source characters being re-checked.
3427
3428
3429 Version 0.94 18-Sep-97
3430 ----------------------
3431
3432 1. The functions are now thread-safe, with the caveat that the global variables
3433 containing pointers to malloc() and free() or alternative functions are the
3434 same for all threads.
3435
3436 2. Get pcre_study() to generate a bitmap of initial characters for non-
3437 anchored patterns when this is possible, and use it if passed to pcre_exec().
3438
3439
3440 Version 0.93 15-Sep-97
3441 ----------------------
3442
3443 1. /(b)|(:+)/ was computing an incorrect first character.
3444
3445 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
3446 but not actually doing anything yet.
3447
3448 3. Treat "-" characters in classes that cannot be part of ranges as literals,
3449 as Perl does (e.g. [-az] or [az-]).
3450
3451 4. Set the anchored flag if a branch starts with .* or .*? because that tests
3452 all possible positions.
3453
3454 5. Split up into different modules to avoid including unneeded functions in a
3455 compiled binary. However, compile and exec are still in one module. The "study"
3456 function is split off.
3457
3458 6. The character tables are now in a separate module whose source is generated
3459 by an auxiliary program - but can then be edited by hand if required. There are
3460 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
3461 toupper() in the code.
3462
3463 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
3464 make them global. Abolish the function for setting them, as the caller can now
3465 set them directly.
3466
3467
3468 Version 0.92 11-Sep-97
3469 ----------------------
3470
3471 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
3472 (e.g. /a{1,3}/) was broken (I mis-optimized it).
3473
3474 2. Caseless matching was not working in character classes if the characters in
3475 the pattern were in upper case.
3476
3477 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
3478
3479 4. Make PCRE_ANCHORED public and accept as a compile option.
3480
3481 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
3482 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
3483 pass them.
3484
3485 6. Give an error if bad option bits passed at compile or run time.
3486
3487 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
3488 pcretest to cause it to pass that flag.
3489
3490 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
3491 options, and the first character, if set.
3492
3493 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
3494
3495
3496 Version 0.91 10-Sep-97
3497 ----------------------
3498
3499 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
3500 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
3501
3502 2. PCRE was looping on encountering an indefinitely repeated back reference to
3503 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
3504 Perl does - treats the match as successful.
3505
3506 ****

Properties

Name Value
svn:eol-style native
svn:keywords "Author Date Id Revision Url"

  ViewVC Help
Powered by ViewVC 1.1.5