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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 420 - (show annotations) (download)
Wed Aug 12 17:32:27 2009 UTC (5 years, 1 month ago) by ph10
File size: 166599 byte(s)
Fix infelicities to do with --count in pcregrep.

1 ChangeLog for PCRE
2 ------------------
3
4 Version 8.00 ??-???-??
5 ----------------------
6
7 1. The table for translating pcre_compile() error codes into POSIX error codes
8 was out-of-date, and there was no check on the pcre_compile() error code
9 being within the table. This could lead to an OK return being given in
10 error.
11
12 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
13 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
14 in a Windows environment.
15
16 3. The pcregrep --count option prints the count for each file even when it is
17 zero, as does GNU grep. However, pcregrep was also printing all files when
18 --files-with-matches was added. Now, when both options are given, it prints
19 counts only for those files that have at least one match. (GNU grep just
20 prints the file name in this circumstance, but including the count seems
21 more useful - otherwise, why use --count?) Also ensured that the
22 combination -clh just lists non-zero counts, with no names.
23
24
25 Version 7.9 11-Apr-09
26 ---------------------
27
28 1. When building with support for bzlib/zlib (pcregrep) and/or readline
29 (pcretest), all targets were linked against these libraries. This included
30 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
31 libraries. This caused unwanted dependencies to be created. This problem
32 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
33 pcretest is linked with readline.
34
35 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
36 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
37 moved outside it again, because FALSE and TRUE are already defined in AIX,
38 but BOOL is not.
39
40 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
41 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
42
43 4. The pcregrep documentation said spaces were inserted as well as colons (or
44 hyphens) following file names and line numbers when outputting matching
45 lines. This is not true; no spaces are inserted. I have also clarified the
46 wording for the --colour (or --color) option.
47
48 5. In pcregrep, when --colour was used with -o, the list of matching strings
49 was not coloured; this is different to GNU grep, so I have changed it to be
50 the same.
51
52 6. When --colo(u)r was used in pcregrep, only the first matching substring in
53 each matching line was coloured. Now it goes on to look for further matches
54 of any of the test patterns, which is the same behaviour as GNU grep.
55
56 7. A pattern that could match an empty string could cause pcregrep to loop; it
57 doesn't make sense to accept an empty string match in pcregrep, so I have
58 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
59 seems to be how GNU grep behaves.
60
61 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
62 start or after a newline", because the conditional assertion was not being
63 correctly handled. The rule now is that both the assertion and what follows
64 in the first alternative must satisfy the test.
65
66 9. If auto-callout was enabled in a pattern with a conditional group whose
67 condition was an assertion, PCRE could crash during matching, both with
68 pcre_exec() and pcre_dfa_exec().
69
70 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
71 used for matching.
72
73 11. Unicode property support in character classes was not working for
74 characters (bytes) greater than 127 when not in UTF-8 mode.
75
76 12. Added the -M command line option to pcretest.
77
78 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
79
80 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
81
82 16. Added comments and documentation about mis-use of no_arg in the C++
83 wrapper.
84
85 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
86 from Martin Jerabek that uses macro names for all relevant character and
87 string constants.
88
89 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
90 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
91 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
92 these, but not everybody uses configure.
93
94 19. A conditional group that had only one branch was not being correctly
95 recognized as an item that could match an empty string. This meant that an
96 enclosing group might also not be so recognized, causing infinite looping
97 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
98 with the subject "ab", where knowledge that the repeated group can match
99 nothing is needed in order to break the loop.
100
101 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
102 exec(), but without supplying a callout function, matching went wrong.
103
104 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
105 leak if the size of the offset vector was greater than 30. When the vector
106 is smaller, the saved offsets during recursion go onto a local stack
107 vector, but for larger vectors malloc() is used. It was failing to free
108 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
109 error, in fact).
110
111 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
112 heapframe that is used only when UTF-8 support is enabled. This caused no
113 problem, but was untidy.
114
115 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
116 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
117 included within another project.
118
119 24. Steven Van Ingelgem's patches to add more options to the CMake support,
120 slightly modified by me:
121
122 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
123 not building pcregrep.
124
125 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
126 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
127
128 25. Forward references, both numeric and by name, in patterns that made use of
129 duplicate group numbers, could behave incorrectly or give incorrect errors,
130 because when scanning forward to find the reference group, PCRE was not
131 taking into account the duplicate group numbers. A pattern such as
132 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
133
134 26. Changed a few more instances of "const unsigned char *" to USPTR, making
135 the feature of a custom pointer more persuasive (as requested by a user).
136
137 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
138 pcretest.c, inside #ifndefs, because it seems they are sometimes already
139 pre-defined.
140
141 28. Added support for (*UTF8) at the start of a pattern.
142
143 29. Arrange for flags added by the "release type" setting in CMake to be shown
144 in the configuration summary.
145
146
147 Version 7.8 05-Sep-08
148 ---------------------
149
150 1. Replaced UCP searching code with optimized version as implemented for Ad
151 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
152 stage table and inline lookup instead of a function, giving speed ups of 2
153 to 5 times on some simple patterns that I tested. Permission was given to
154 distribute the MultiStage2.py script that generates the tables (it's not in
155 the tarball, but is in the Subversion repository).
156
157 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
158 scripts.
159
160 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
161 a group with a zero qualifier. The result of the study could be incorrect,
162 or the function might crash, depending on the pattern.
163
164 4. Caseless matching was not working for non-ASCII characters in back
165 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
166 It now works when Unicode Property Support is available.
167
168 5. In pcretest, an escape such as \x{de} in the data was always generating
169 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
170 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
171 truncation.
172
173 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
174
175 7. Added two (int) casts to pcregrep when printing the difference of two
176 pointers, in case they are 64-bit values.
177
178 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
179 test 2 if it fails.
180
181 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
182 and a #define of that name to empty if it is not externally set. This is to
183 allow users of MSVC to set it if necessary.
184
185 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
186 the convenience functions in the pcre_get.c source file.
187
188 11. An option change at the start of a pattern that had top-level alternatives
189 could cause overwriting and/or a crash. This command provoked a crash in
190 some environments:
191
192 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
193
194 This potential security problem was recorded as CVE-2008-2371.
195
196 12. For a pattern where the match had to start at the beginning or immediately
197 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
198 pcre_dfa_exec() could read past the end of the passed subject if there was
199 no match. To help with detecting such bugs (e.g. with valgrind), I modified
200 pcretest so that it places the subject at the end of its malloc-ed buffer.
201
202 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
203 exec() might read past the end of the data buffer in UTF-8 mode.
204
205 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
206 the data contained the byte 0x85 as part of a UTF-8 character within its
207 first line. This applied both to normal and DFA matching.
208
209 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
210 /^[^d]*?$/8 failed to match "abc".
211
212 16. Added a missing copyright notice to pcrecpp_internal.h.
213
214 17. Make it more clear in the documentation that values returned from
215 pcre_exec() in ovector are byte offsets, not character counts.
216
217 18. Tidied a few places to stop certain compilers from issuing warnings.
218
219 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
220 supplied by Stefan Weber. I made a further small update for 7.8 because
221 there is a change of source arrangements: the pcre_searchfuncs.c module is
222 replaced by pcre_ucd.c.
223
224
225 Version 7.7 07-May-08
226 ---------------------
227
228 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
229 a string to a long long, pretend we don't even have a long long." This is
230 done by checking for the strtoq, strtoll, and _strtoi64 functions.
231
232 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
233 pre-7.6 versions, which defined a global no_arg variable instead of putting
234 it in the RE class. (See also #8 below.)
235
236 3. Remove a line of dead code, identified by coverity and reported by Nuno
237 Lopes.
238
239 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
240
241 (1) The include/exclude patterns were being applied to the whole pathnames
242 of files, instead of just to the final components.
243
244 (2) If there was more than one level of directory, the subdirectories were
245 skipped unless they satisfied the include/exclude conditions. This is
246 inconsistent with GNU grep (and could even be seen as contrary to the
247 pcregrep specification - which I improved to make it absolutely clear).
248 The action now is always to scan all levels of directory, and just
249 apply the include/exclude patterns to regular files.
250
251 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
252 --exclude_dir in the tests to avoid scanning .svn directories.
253
254 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
255 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
256 doesn't support NULs in patterns.
257
258 7. Added some missing "const"s to declarations of static tables in
259 pcre_compile.c and pcre_dfa_exec.c.
260
261 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
262 caused by fix #2 above. (Subsequently also a second patch to fix the
263 first patch. And a third patch - this was a messy problem.)
264
265 9. Applied Craig's patch to remove the use of push_back().
266
267 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
268 matching function regexec().
269
270 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
271 which, however, unlike Perl's \g{...}, are subroutine calls, not back
272 references. PCRE supports relative numbers with this syntax (I don't think
273 Oniguruma does).
274
275 12. Previously, a group with a zero repeat such as (...){0} was completely
276 omitted from the compiled regex. However, this means that if the group
277 was called as a subroutine from elsewhere in the pattern, things went wrong
278 (an internal error was given). Such groups are now left in the compiled
279 pattern, with a new opcode that causes them to be skipped at execution
280 time.
281
282 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
283 to the way PCRE behaves:
284
285 (a) A lone ] character is dis-allowed (Perl treats it as data).
286
287 (b) A back reference to an unmatched subpattern matches an empty string
288 (Perl fails the current match path).
289
290 (c) A data ] in a character class must be notated as \] because if the
291 first data character in a class is ], it defines an empty class. (In
292 Perl it is not possible to have an empty class.) The empty class []
293 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
294 The negative empty class [^] matches any one character, independently
295 of the DOTALL setting.
296
297 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
298 non-existent subpattern following a character class starting with ']' and
299 containing () gave an internal compiling error instead of "reference to
300 non-existent subpattern". Fortunately, when the pattern did exist, the
301 compiled code was correct. (When scanning forwards to check for the
302 existencd of the subpattern, it was treating the data ']' as terminating
303 the class, so got the count wrong. When actually compiling, the reference
304 was subsequently set up correctly.)
305
306 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
307 it was being rejected as not supported by pcre_dfa_exec(), even though
308 other assertions are supported. I have made pcre_dfa_exec() support
309 (*FAIL).
310
311 16. The implementation of 13c above involved the invention of a new opcode,
312 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
313 cannot be changed at match time, I realized I could make a small
314 improvement to matching performance by compiling OP_ALLANY instead of
315 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
316 on the OP_ANY path.
317
318 17. Compiling pcretest on Windows with readline support failed without the
319 following two fixes: (1) Make the unistd.h include conditional on
320 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
321
322 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
323 ncurses library to be included for pcretest when ReadLine support is
324 requested, but also to allow for it to be overridden. This patch came from
325 Daniel Bergström.
326
327 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
328 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
329 any errors with the current Unicode tables. Thanks to Peter Kankowski for
330 spotting this.
331
332
333 Version 7.6 28-Jan-08
334 ---------------------
335
336 1. A character class containing a very large number of characters with
337 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
338 overflow.
339
340 2. Patch to cut out the "long long" test in pcrecpp_unittest when
341 HAVE_LONG_LONG is not defined.
342
343 3. Applied Christian Ehrlicher's patch to update the CMake build files to
344 bring them up to date and include new features. This patch includes:
345
346 - Fixed PH's badly added libz and libbz2 support.
347 - Fixed a problem with static linking.
348 - Added pcredemo. [But later removed - see 7 below.]
349 - Fixed dftables problem and added an option.
350 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
351 HAVE_LONG_LONG.
352 - Added readline support for pcretest.
353 - Added an listing of the option settings after cmake has run.
354
355 4. A user submitted a patch to Makefile that makes it easy to create
356 "pcre.dll" under mingw when using Configure/Make. I added stuff to
357 Makefile.am that cause it to include this special target, without
358 affecting anything else. Note that the same mingw target plus all
359 the other distribution libraries and programs are now supported
360 when configuring with CMake (see 6 below) instead of with
361 Configure/Make.
362
363 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
364 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
365 exported in the Windows port". It has not yet been confirmed that the patch
366 solves the problem, but it does no harm.
367
368 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
369 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
370 with CMake, and also correct the comment about stack recursion.
371
372 7. Remove the automatic building of pcredemo from the ./configure system and
373 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
374 of a program that users should build themselves after PCRE is installed, so
375 building it automatically is not really right. What is more, it gave
376 trouble in some build environments.
377
378 8. Further tidies to CMakeLists.txt from Sheri and Christian.
379
380
381 Version 7.5 10-Jan-08
382 ---------------------
383
384 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
385 values in parens when parsing an RE using the C++ wrapper."
386
387 2. Negative specials like \S did not work in character classes in UTF-8 mode.
388 Characters greater than 255 were excluded from the class instead of being
389 included.
390
391 3. The same bug as (2) above applied to negated POSIX classes such as
392 [:^space:].
393
394 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
395 defined or documented. It seems to have been a typo for PCRE_STATIC, so
396 I have changed it.
397
398 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
399 first named subpattern) and a construct such as (?&a) would reference the
400 first named subpattern whose name started with "a" (in other words, the
401 length check was missing). Both these problems are fixed. "Subpattern name
402 expected" is now given for (?&) (a zero-length name), and this patch also
403 makes it give the same error for \k'' (previously it complained that that
404 was a reference to a non-existent subpattern).
405
406 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
407 this is right because (?- can be followed by option settings as well as by
408 digits. I have, however, made the messages clearer.
409
410 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
411 than the number used in the conditional) now cause a compile-time error.
412 This is actually not compatible with Perl, which accepts such patterns, but
413 treats the conditional as always being FALSE (as PCRE used to), but it
414 seems to me that giving a diagnostic is better.
415
416 8. Change "alphameric" to the more common word "alphanumeric" in comments
417 and messages.
418
419 9. Fix two occurrences of "backslash" in comments that should have been
420 "backspace".
421
422 10. Remove two redundant lines of code that can never be obeyed (their function
423 was moved elsewhere).
424
425 11. The program that makes PCRE's Unicode character property table had a bug
426 which caused it to generate incorrect table entries for sequences of
427 characters that have the same character type, but are in different scripts.
428 It amalgamated them into a single range, with the script of the first of
429 them. In other words, some characters were in the wrong script. There were
430 thirteen such cases, affecting characters in the following ranges:
431
432 U+002b0 - U+002c1
433 U+0060c - U+0060d
434 U+0061e - U+00612
435 U+0064b - U+0065e
436 U+0074d - U+0076d
437 U+01800 - U+01805
438 U+01d00 - U+01d77
439 U+01d9b - U+01dbf
440 U+0200b - U+0200f
441 U+030fc - U+030fe
442 U+03260 - U+0327f
443 U+0fb46 - U+0fbb1
444 U+10450 - U+1049d
445
446 12. The -o option (show only the matching part of a line) for pcregrep was not
447 compatible with GNU grep in that, if there was more than one match in a
448 line, it showed only the first of them. It now behaves in the same way as
449 GNU grep.
450
451 13. If the -o and -v options were combined for pcregrep, it printed a blank
452 line for every non-matching line. GNU grep prints nothing, and pcregrep now
453 does the same. The return code can be used to tell if there were any
454 non-matching lines.
455
456 14. Added --file-offsets and --line-offsets to pcregrep.
457
458 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
459 infinitely looping recursion. The bug was that positive lookaheads were not
460 being skipped when checking for a possible empty match (negative lookaheads
461 and both kinds of lookbehind were skipped).
462
463 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
464 inclusion of <windows.h> to before rather than after the definition of
465 INVALID_FILE_ATTRIBUTES (patch from David Byron).
466
467 17. Specifying a possessive quantifier with a specific limit for a Unicode
468 character property caused pcre_compile() to compile bad code, which led at
469 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
470 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
471 caused the error; without that there was no problem.
472
473 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
474
475 19. Added --enable-pcretest-libreadline.
476
477 20. In pcrecpp.cc, the variable 'count' was incremented twice in
478 RE::GlobalReplace(). As a result, the number of replacements returned was
479 double what it should be. I removed one of the increments, but Craig sent a
480 later patch that removed the other one (the right fix) and added unit tests
481 that check the return values (which was not done before).
482
483 21. Several CMake things:
484
485 (1) Arranged that, when cmake is used on Unix, the libraries end up with
486 the names libpcre and libpcreposix, not just pcre and pcreposix.
487
488 (2) The above change means that pcretest and pcregrep are now correctly
489 linked with the newly-built libraries, not previously installed ones.
490
491 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
492
493 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
494 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
495 UTF-8 newline character). The key issue is that the pattern starts .*;
496 this means that the match must be either at the beginning, or after a
497 newline. The bug was in the code for advancing after a failed match and
498 checking that the new position followed a newline. It was not taking
499 account of UTF-8 characters correctly.
500
501 23. PCRE was behaving differently from Perl in the way it recognized POSIX
502 character classes. PCRE was not treating the sequence [:...:] as a
503 character class unless the ... were all letters. Perl, however, seems to
504 allow any characters between [: and :], though of course it rejects as
505 unknown any "names" that contain non-letters, because all the known class
506 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
507 for example, whereas PCRE did not - it did not recognize a POSIX character
508 class. This seemed a bit dangerous, so the code has been changed to be
509 closer to Perl. The behaviour is not identical to Perl, because PCRE will
510 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
511 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
512 Perl does, and where it didn't before.
513
514 24. Rewrite so as to remove the single use of %n from pcregrep because in some
515 Windows environments %n is disabled by default.
516
517
518 Version 7.4 21-Sep-07
519 ---------------------
520
521 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
522 means that a class such as [\s] counted as "explicit reference to CR or
523 LF". That isn't really right - the whole point of the change was to try to
524 help when there was an actual mention of one of the two characters. So now
525 the change happens only if \r or \n (or a literal CR or LF) character is
526 encountered.
527
528 2. The 32-bit options word was also used for 6 internal flags, but the numbers
529 of both had grown to the point where there were only 3 bits left.
530 Fortunately, there was spare space in the data structure, and so I have
531 moved the internal flags into a new 16-bit field to free up more option
532 bits.
533
534 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
535 but did not set the internal JCHANGED flag - either of these is enough to
536 control the way the "get" function works - but the PCRE_INFO_JCHANGED
537 facility is supposed to tell if (?J) was ever used, so now (?J) at the
538 start sets both bits.
539
540 4. Added options (at build time, compile time, exec time) to change \R from
541 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
542
543 5. doc/pcresyntax.html was missing from the distribution.
544
545 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
546 compatibility, even though it is no longer used.
547
548 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
549 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
550 windows.h file is present (where different names are used). [This was
551 reversed later after testing - see 16 below.]
552
553 8. Changed all #include <config.h> to #include "config.h". There were also
554 some further <pcre.h> cases that I changed to "pcre.h".
555
556 9. When pcregrep was used with the --colour option, it missed the line ending
557 sequence off the lines that it output.
558
559 10. It was pointed out to me that arrays of string pointers cause lots of
560 relocations when a shared library is dynamically loaded. A technique of
561 using a single long string with a table of offsets can drastically reduce
562 these. I have refactored PCRE in four places to do this. The result is
563 dramatic:
564
565 Originally: 290
566 After changing UCP table: 187
567 After changing error message table: 43
568 After changing table of "verbs" 36
569 After changing table of Posix names 22
570
571 Thanks to the folks working on Gregex for glib for this insight.
572
573 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
574 unicode-properties was also set.
575
576 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
577
578 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
579 checked only for CRLF.
580
581 14. Added casts to pcretest.c to avoid compiler warnings.
582
583 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
584
585 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
586 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
587 entirely. This removes changes made in 7 above.
588
589 17. The CMake files have been updated, and there is now more information about
590 building with CMake in the NON-UNIX-USE document.
591
592
593 Version 7.3 28-Aug-07
594 ---------------------
595
596 1. In the rejigging of the build system that eventually resulted in 7.1, the
597 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
598 brackets there is not right, since it causes compilers to look for an
599 installed pcre.h, not the version that is in the source that is being
600 compiled (which of course may be different). I have changed it back to:
601
602 #include "pcre.h"
603
604 I have a vague recollection that the change was concerned with compiling in
605 different directories, but in the new build system, that is taken care of
606 by the VPATH setting the Makefile.
607
608 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
609 when the subject happened to end in the byte 0x85 (e.g. if the last
610 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
611 characters but of course it shouldn't be taken as a newline when it is part
612 of another character. The bug was that, for an unlimited repeat of . in
613 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
614 characters when looking for a newline.
615
616 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
617
618 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
619 in debug output.
620
621 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
622 long printing in the pcrecpp unittest when running under MinGW.
623
624 6. ESC_K was left out of the EBCDIC table.
625
626 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
627 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
628 limit also applies to "virtual nesting" when a pattern is recursive, and in
629 this case 1000 isn't so big. I have been able to remove this limit at the
630 expense of backing off one optimization in certain circumstances. Normally,
631 when pcre_exec() would call its internal match() function recursively and
632 immediately return the result unconditionally, it uses a "tail recursion"
633 feature to save stack. However, when a subpattern that can match an empty
634 string has an unlimited repetition quantifier, it no longer makes this
635 optimization. That gives it a stack frame in which to save the data for
636 checking that an empty string has been matched. Previously this was taken
637 from the 1000-entry workspace that had been reserved. So now there is no
638 explicit limit, but more stack is used.
639
640 8. Applied Daniel's patches to solve problems with the import/export magic
641 syntax that is required for Windows, and which was going wrong for the
642 pcreposix and pcrecpp parts of the library. These were overlooked when this
643 problem was solved for the main library.
644
645 9. There were some crude static tests to avoid integer overflow when computing
646 the size of patterns that contain repeated groups with explicit upper
647 limits. As the maximum quantifier is 65535, the maximum group length was
648 set at 30,000 so that the product of these two numbers did not overflow a
649 32-bit integer. However, it turns out that people want to use groups that
650 are longer than 30,000 bytes (though not repeat them that many times).
651 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
652 made it possible to implement the integer overflow checks in a much more
653 dynamic way, which I have now done. The artificial limitation on group
654 length has been removed - we now have only the limit on the total length of
655 the compiled pattern, which depends on the LINK_SIZE setting.
656
657 10. Fixed a bug in the documentation for get/copy named substring when
658 duplicate names are permitted. If none of the named substrings are set, the
659 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
660 empty string.
661
662 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
663 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
664 because the ] is interpreted as the first data character and the
665 terminating ] is not found. PCRE has been made compatible with Perl in this
666 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
667 cause memory overwriting.
668
669 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
670 string has been matched (to stop an infinite loop). It was not recognizing
671 a conditional subpattern that could match an empty string if that
672 subpattern was within another subpattern. For example, it looped when
673 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
674 condition was not nested. This bug has been fixed.
675
676 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
677 past the start of the subject in the presence of bytes with the top bit
678 set, for example "\x8aBCD".
679
680 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
681 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
682
683 14. Optimized (?!) to (*FAIL).
684
685 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
686 This restricts code points to be within the range 0 to 0x10FFFF, excluding
687 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
688 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
689 does: it's just the validity check that is more restrictive.
690
691 16. Inserted checks for integer overflows during escape sequence (backslash)
692 processing, and also fixed erroneous offset values for syntax errors during
693 backslash processing.
694
695 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
696 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
697
698 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
699 caused an overrun.
700
701 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
702 something other than just ASCII characters) inside a group that had an
703 unlimited repeat caused a loop at compile time (while checking to see
704 whether the group could match an empty string).
705
706 20. Debugging a pattern containing \p or \P could cause a crash. For example,
707 [\P{Any}] did so. (Error in the code for printing property names.)
708
709 21. An orphan \E inside a character class could cause a crash.
710
711 22. A repeated capturing bracket such as (A)? could cause a wild memory
712 reference during compilation.
713
714 23. There are several functions in pcre_compile() that scan along a compiled
715 expression for various reasons (e.g. to see if it's fixed length for look
716 behind). There were bugs in these functions when a repeated \p or \P was
717 present in the pattern. These operators have additional parameters compared
718 with \d, etc, and these were not being taken into account when moving along
719 the compiled data. Specifically:
720
721 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
722 length.
723
724 (b) An item such as \pL+ within a repeated group could cause crashes or
725 loops.
726
727 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
728 "reference to non-existent subpattern" error.
729
730 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
731
732 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
733 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
734
735 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
736
737 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
738 character were causing crashes (broken optimization).
739
740 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
741 \p or \P) caused a compile-time loop.
742
743 28. More problems have arisen in unanchored patterns when CRLF is a valid line
744 break. For example, the unstudied pattern [\r\n]A does not match the string
745 "\r\nA" because change 7.0/46 below moves the current point on by two
746 characters after failing to match at the start. However, the pattern \nA
747 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
748 the same is true. There doesn't seem any very clean way out of this, but
749 what I have chosen to do makes the common cases work: PCRE now takes note
750 of whether there can be an explicit match for \r or \n anywhere in the
751 pattern, and if so, 7.0/46 no longer applies. As part of this change,
752 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
753 pattern has explicit CR or LF references.
754
755 29. Added (*CR) etc for changing newline setting at start of pattern.
756
757
758 Version 7.2 19-Jun-07
759 ---------------------
760
761 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
762 which is apparently normally available under Windows.
763
764 2. Re-jig the pcregrep tests with different newline settings in an attempt
765 to make them independent of the local environment's newline setting.
766
767 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
768
769 4. Some of the "internals" tests were previously cut out when the link size
770 was not 2, because the output contained actual offsets. The recent new
771 "Z" feature of pcretest means that these can be cut out, making the tests
772 usable with all link sizes.
773
774 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
775 stack recursion. This gives a massive performance boost under BSD, but just
776 a small improvement under Linux. However, it saves one field in the frame
777 in all cases.
778
779 6. Added more features from the forthcoming Perl 5.10:
780
781 (a) (?-n) (where n is a string of digits) is a relative subroutine or
782 recursion call. It refers to the nth most recently opened parentheses.
783
784 (b) (?+n) is also a relative subroutine call; it refers to the nth next
785 to be opened parentheses.
786
787 (c) Conditions that refer to capturing parentheses can be specified
788 relatively, for example, (?(-2)... or (?(+3)...
789
790 (d) \K resets the start of the current match so that everything before
791 is not part of it.
792
793 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
794
795 (f) \g{name} is another synonym - part of Perl 5.10's unification of
796 reference syntax.
797
798 (g) (?| introduces a group in which the numbering of parentheses in each
799 alternative starts with the same number.
800
801 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
802
803 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
804 PCRE_INFO_JCHANGED.
805
806 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
807 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
808 for detecting groups that can match an empty string.
809
810 9. A pattern with a very large number of alternatives (more than several
811 hundred) was running out of internal workspace during the pre-compile
812 phase, where pcre_compile() figures out how much memory will be needed. A
813 bit of new cunning has reduced the workspace needed for groups with
814 alternatives. The 1000-alternative test pattern now uses 12 bytes of
815 workspace instead of running out of the 4096 that are available.
816
817 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
818
819 11. Applied patch from Google to remove an optimization that didn't quite work.
820 The report of the bug said:
821
822 pcrecpp::RE("a*").FullMatch("aaa") matches, while
823 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
824 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
825
826 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
827 it matched the wrong number of bytes.
828
829
830 Version 7.1 24-Apr-07
831 ---------------------
832
833 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
834 that is more "standard", making use of automake and other Autotools. There
835 is some re-arrangement of the files and adjustment of comments consequent
836 on this.
837
838 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
839 for recursive directory scanning broke on some systems because the files
840 are not scanned in any specific order and on different systems the order
841 was different. A call to "sort" has been inserted into RunGrepTest for the
842 approprate test as a short-term fix. In the longer term there may be an
843 alternative.
844
845 3. I had an email from Eric Raymond about problems translating some of PCRE's
846 man pages to HTML (despite the fact that I distribute HTML pages, some
847 people do their own conversions for various reasons). The problems
848 concerned the use of low-level troff macros .br and .in. I have therefore
849 removed all such uses from the man pages (some were redundant, some could
850 be replaced by .nf/.fi pairs). The 132html script that I use to generate
851 HTML has been updated to handle .nf/.fi and to complain if it encounters
852 .br or .in.
853
854 4. Updated comments in configure.ac that get placed in config.h.in and also
855 arranged for config.h to be included in the distribution, with the name
856 config.h.generic, for the benefit of those who have to compile without
857 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
858
859 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
860 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
861 makevp.bat for the current PCRE, using the additional files
862 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
863
864 6. A Windows user reported a minor discrepancy with test 2, which turned out
865 to be caused by a trailing space on an input line that had got lost in his
866 copy. The trailing space was an accident, so I've just removed it.
867
868 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
869 that is needed.
870
871 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
872 as "const" (a) because they are and (b) because it helps the PHP
873 maintainers who have recently made a script to detect big data structures
874 in the php code that should be moved to the .rodata section. I remembered
875 to update Builducptable as well, so it won't revert if ucptable.h is ever
876 re-created.
877
878 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
879 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
880 order to be able to cut out the UTF-8 tables in the latter when UTF-8
881 support is not required. This saves 1.5-2K of code, which is important in
882 some applications.
883
884 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
885 so as not to refer to the tables, even though these functions will never be
886 called when UTF-8 support is disabled. Otherwise there are problems with a
887 shared library.
888
889 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
890
891 (a) It was defining its arguments as char * instead of void *.
892
893 (b) It was assuming that all moves were upwards in memory; this was true
894 a long time ago when I wrote it, but is no longer the case.
895
896 The emulated memove() is provided for those environments that have neither
897 memmove() nor bcopy(). I didn't think anyone used it these days, but that
898 is clearly not the case, as these two bugs were recently reported.
899
900 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
901 and Detrail to create the HTML documentation, the .txt form of the man
902 pages, and it removes trailing spaces from listed files. It also creates
903 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
904 case, it wraps all the #defines with #ifndefs. This script should be run
905 before "make dist".
906
907 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
908 with Unicode property support.
909
910 (a) For a maximizing quantifier, if the two different cases of the
911 character were of different lengths in their UTF-8 codings (there are
912 some cases like this - I found 11), and the matching function had to
913 back up over a mixture of the two cases, it incorrectly assumed they
914 were both the same length.
915
916 (b) When PCRE was configured to use the heap rather than the stack for
917 recursion during matching, it was not correctly preserving the data for
918 the other case of a UTF-8 character when checking ahead for a match
919 while processing a minimizing repeat. If the check also involved
920 matching a wide character, but failed, corruption could cause an
921 erroneous result when trying to check for a repeat of the original
922 character.
923
924 13. Some tidying changes to the testing mechanism:
925
926 (a) The RunTest script now detects the internal link size and whether there
927 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
928 values substituted by "configure". (The RunGrepTest script already did
929 this for UTF-8.) The configure.ac script no longer substitutes the
930 relevant variables.
931
932 (b) The debugging options /B and /D in pcretest show the compiled bytecode
933 with length and offset values. This means that the output is different
934 for different internal link sizes. Test 2 is skipped for link sizes
935 other than 2 because of this, bypassing the problem. Unfortunately,
936 there was also a test in test 3 (the locale tests) that used /B and
937 failed for link sizes other than 2. Rather than cut the whole test out,
938 I have added a new /Z option to pcretest that replaces the length and
939 offset values with spaces. This is now used to make test 3 independent
940 of link size. (Test 2 will be tidied up later.)
941
942 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
943 segmentation fault instead of returning the appropriate error message.
944
945 15. In multiline mode when the newline sequence was set to "any", the pattern
946 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
947 This doesn't seem right; it now treats the CRLF combination as the line
948 ending, and so does not match in that case. It's only a pattern such as ^$
949 that would hit this one: something like ^ABC$ would have failed after \r
950 and then tried again after \r\n.
951
952 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
953 in an attempt to make files that differ only in their line terminators
954 compare equal. This works on Linux.
955
956 17. Under certain error circumstances pcregrep might try to free random memory
957 as it exited. This is now fixed, thanks to valgrind.
958
959 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
960 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
961 was because its rules for how to advance for /g after matching an empty
962 string at the end of a line did not allow for this case. They now check for
963 it specially.
964
965 20. pcretest is supposed to handle patterns and data of any length, by
966 extending its buffers when necessary. It was getting this wrong when the
967 buffer for a data line had to be extended.
968
969 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
970 CRLF as a newline sequence.
971
972 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
973 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
974 I have nevertheless tidied it up.
975
976 23. Added some casts to kill warnings from HP-UX ia64 compiler.
977
978 24. Added a man page for pcre-config.
979
980
981 Version 7.0 19-Dec-06
982 ---------------------
983
984 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
985 moving to gcc 4.1.1.
986
987 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
988 sys/time.h, which is documented as needed for this function. It doesn't
989 seem to matter on Linux, but it showed up on some releases of OS X.
990
991 3. It seems that there are systems where bytes whose values are greater than
992 127 match isprint() in the "C" locale. The "C" locale should be the
993 default when a C program starts up. In most systems, only ASCII printing
994 characters match isprint(). This difference caused the output from pcretest
995 to vary, making some of the tests fail. I have changed pcretest so that:
996
997 (a) When it is outputting text in the compiled version of a pattern, bytes
998 other than 32-126 are always shown as hex escapes.
999
1000 (b) When it is outputting text that is a matched part of a subject string,
1001 it does the same, unless a different locale has been set for the match
1002 (using the /L modifier). In this case, it uses isprint() to decide.
1003
1004 4. Fixed a major bug that caused incorrect computation of the amount of memory
1005 required for a compiled pattern when options that changed within the
1006 pattern affected the logic of the preliminary scan that determines the
1007 length. The relevant options are -x, and -i in UTF-8 mode. The result was
1008 that the computed length was too small. The symptoms of this bug were
1009 either the PCRE error "internal error: code overflow" from pcre_compile(),
1010 or a glibc crash with a message such as "pcretest: free(): invalid next
1011 size (fast)". Examples of patterns that provoked this bug (shown in
1012 pcretest format) are:
1013
1014 /(?-x: )/x
1015 /(?x)(?-x: \s*#\s*)/
1016 /((?i)[\x{c0}])/8
1017 /(?i:[\x{c0}])/8
1018
1019 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
1020 is now done differently.
1021
1022 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
1023 wrapper classes; (b) implement a new function in the C++ scanner that is
1024 more efficient than the old way of doing things because it avoids levels of
1025 recursion in the regex matching; (c) add a paragraph to the documentation
1026 for the FullMatch() function.
1027
1028 6. The escape sequence \n was being treated as whatever was defined as
1029 "newline". Not only was this contrary to the documentation, which states
1030 that \n is character 10 (hex 0A), but it also went horribly wrong when
1031 "newline" was defined as CRLF. This has been fixed.
1032
1033 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
1034 was being set to -1 for the "end of line" case (supposedly a value that no
1035 character can have). Though this value is never used (the check for end of
1036 line is "zero bytes in current character"), it caused compiler complaints.
1037 I've changed it to 0xffffffff.
1038
1039 8. In pcre_version.c, the version string was being built by a sequence of
1040 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
1041 string (as it is for production releases) called a macro with an empty
1042 argument. The C standard says the result of this is undefined. The gcc
1043 compiler treats it as an empty string (which was what was wanted) but it is
1044 reported that Visual C gives an error. The source has been hacked around to
1045 avoid this problem.
1046
1047 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
1048 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
1049 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
1050 of them did).
1051
1052 10. Originally, pcretest opened its input and output without "b"; then I was
1053 told that "b" was needed in some environments, so it was added for release
1054 5.0 to both the input and output. (It makes no difference on Unix-like
1055 systems.) Later I was told that it is wrong for the input on Windows. I've
1056 now abstracted the modes into two macros, to make it easier to fiddle with
1057 them, and removed "b" from the input mode under Windows.
1058
1059 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
1060
1061 12. Added -help and --help to pcretest as an official way of being reminded
1062 of the options.
1063
1064 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
1065 and pcrecpp.cc because they annoy compilers at high warning levels.
1066
1067 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
1068
1069 15. Fixed an occurrence of == in configure.ac that should have been = (shell
1070 scripts are not C programs :-) and which was not noticed because it works
1071 on Linux.
1072
1073 16. pcretest is supposed to handle any length of pattern and data line (as one
1074 line or as a continued sequence of lines) by extending its input buffer if
1075 necessary. This feature was broken for very long pattern lines, leading to
1076 a string of junk being passed to pcre_compile() if the pattern was longer
1077 than about 50K.
1078
1079 17. I have done a major re-factoring of the way pcre_compile() computes the
1080 amount of memory needed for a compiled pattern. Previously, there was code
1081 that made a preliminary scan of the pattern in order to do this. That was
1082 OK when PCRE was new, but as the facilities have expanded, it has become
1083 harder and harder to keep it in step with the real compile phase, and there
1084 have been a number of bugs (see for example, 4 above). I have now found a
1085 cunning way of running the real compile function in a "fake" mode that
1086 enables it to compute how much memory it would need, while actually only
1087 ever using a few hundred bytes of working memory and without too many
1088 tests of the mode. This should make future maintenance and development
1089 easier. A side effect of this work is that the limit of 200 on the nesting
1090 depth of parentheses has been removed (though this was never a serious
1091 limitation, I suspect). However, there is a downside: pcre_compile() now
1092 runs more slowly than before (30% or more, depending on the pattern). I
1093 hope this isn't a big issue. There is no effect on runtime performance.
1094
1095 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
1096 newline (only possible for the last line of a file) and it was a
1097 pattern that set a locale (followed by /Lsomething), pcretest crashed.
1098
1099 19. Added additional timing features to pcretest. (1) The -tm option now times
1100 matching only, not compiling. (2) Both -t and -tm can be followed, as a
1101 separate command line item, by a number that specifies the number of
1102 repeats to use when timing. The default is 50000; this gives better
1103 precision, but takes uncomfortably long for very large patterns.
1104
1105 20. Extended pcre_study() to be more clever in cases where a branch of a
1106 subpattern has no definite first character. For example, (a*|b*)[cd] would
1107 previously give no result from pcre_study(). Now it recognizes that the
1108 first character must be a, b, c, or d.
1109
1110 21. There was an incorrect error "recursive call could loop indefinitely" if
1111 a subpattern (or the entire pattern) that was being tested for matching an
1112 empty string contained only one non-empty item after a nested subpattern.
1113 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
1114 incorrectly, because the \d was being skipped in the check.
1115
1116 22. The pcretest program now has a new pattern option /B and a command line
1117 option -b, which is equivalent to adding /B to every pattern. This causes
1118 it to show the compiled bytecode, without the additional information that
1119 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
1120 is the same as /B/I).
1121
1122 23. A new optimization is now able automatically to treat some sequences such
1123 as a*b as a*+b. More specifically, if something simple (such as a character
1124 or a simple class like \d) has an unlimited quantifier, and is followed by
1125 something that cannot possibly match the quantified thing, the quantifier
1126 is automatically "possessified".
1127
1128 24. A recursive reference to a subpattern whose number was greater than 39
1129 went wrong under certain circumstances in UTF-8 mode. This bug could also
1130 have affected the operation of pcre_study().
1131
1132 25. Realized that a little bit of performance could be had by replacing
1133 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
1134
1135 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
1136
1137 27. Possessive quantifiers such as a++ were previously implemented by turning
1138 them into atomic groups such as ($>a+). Now they have their own opcodes,
1139 which improves performance. This includes the automatically created ones
1140 from 23 above.
1141
1142 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
1143 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
1144 the first matched character to be a colon. This applied both to named and
1145 numbered groups.
1146
1147 29. The ucpinternal.h header file was missing its idempotency #ifdef.
1148
1149 30. I was sent a "project" file called libpcre.a.dev which I understand makes
1150 building PCRE on Windows easier, so I have included it in the distribution.
1151
1152 31. There is now a check in pcretest against a ridiculously large number being
1153 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
1154 loop, the loop is abandoned.
1155
1156 32. Forward references to subpatterns in conditions such as (?(2)...) where
1157 subpattern 2 is defined later cause pcre_compile() to search forwards in
1158 the pattern for the relevant set of parentheses. This search went wrong
1159 when there were unescaped parentheses in a character class, parentheses
1160 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
1161
1162 33. "Subroutine" calls and backreferences were previously restricted to
1163 referencing subpatterns earlier in the regex. This restriction has now
1164 been removed.
1165
1166 34. Added a number of extra features that are going to be in Perl 5.10. On the
1167 whole, these are just syntactic alternatives for features that PCRE had
1168 previously implemented using the Python syntax or my own invention. The
1169 other formats are all retained for compatibility.
1170
1171 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
1172 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
1173 also .NET compatible.
1174
1175 (b) A recursion or subroutine call to a named group can now be defined as
1176 (?&name) as well as (?P>name).
1177
1178 (c) A backreference to a named group can now be defined as \k<name> or
1179 \k'name' as well as (?P=name). The new forms, as well as being in Perl
1180 5.10, are also .NET compatible.
1181
1182 (d) A conditional reference to a named group can now use the syntax
1183 (?(<name>) or (?('name') as well as (?(name).
1184
1185 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
1186 groups (named and numbered) that are never evaluated inline, but can be
1187 called as "subroutines" from elsewhere. In effect, the DEFINE condition
1188 is always false. There may be only one alternative in such a group.
1189
1190 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
1191 as the simple (?(R). The condition is true only if the most recent
1192 recursion is that of the given number or name. It does not search out
1193 through the entire recursion stack.
1194
1195 (g) The escape \gN or \g{N} has been added, where N is a positive or
1196 negative number, specifying an absolute or relative reference.
1197
1198 35. Tidied to get rid of some further signed/unsigned compiler warnings and
1199 some "unreachable code" warnings.
1200
1201 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
1202 things, this adds five new scripts.
1203
1204 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
1205 There were also incompatibilities regarding the handling of \Q..\E inside
1206 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
1207 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
1208
1209 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
1210 matches an empty string, and forcibly breaks the loop. There were bugs in
1211 this code in non-simple cases. For a pattern such as ^(a()*)* matched
1212 against aaaa the result was just "a" rather than "aaaa", for example. Two
1213 separate and independent bugs (that affected different cases) have been
1214 fixed.
1215
1216 39. Refactored the code to abolish the use of different opcodes for small
1217 capturing bracket numbers. This is a tidy that I avoided doing when I
1218 removed the limit on the number of capturing brackets for 3.5 back in 2001.
1219 The new approach is not only tidier, it makes it possible to reduce the
1220 memory needed to fix the previous bug (38).
1221
1222 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
1223 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
1224 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
1225 mode.
1226
1227 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
1228 report.
1229
1230 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
1231 copy construction and assignment in the C++ wrapper.
1232
1233 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
1234 couple of bugs that could have given wrong results in the "--newline=crlf"
1235 case.
1236
1237 44. Added a number of casts and did some reorganization of signed/unsigned int
1238 variables following suggestions from Dair Grant. Also renamed the variable
1239 "this" as "item" because it is a C++ keyword.
1240
1241 45. Arranged for dftables to add
1242
1243 #include "pcre_internal.h"
1244
1245 to pcre_chartables.c because without it, gcc 4.x may remove the array
1246 definition from the final binary if PCRE is built into a static library and
1247 dead code stripping is activated.
1248
1249 46. For an unanchored pattern, if a match attempt fails at the start of a
1250 newline sequence, and the newline setting is CRLF or ANY, and the next two
1251 characters are CRLF, advance by two characters instead of one.
1252
1253
1254 Version 6.7 04-Jul-06
1255 ---------------------
1256
1257 1. In order to handle tests when input lines are enormously long, pcretest has
1258 been re-factored so that it automatically extends its buffers when
1259 necessary. The code is crude, but this _is_ just a test program. The
1260 default size has been increased from 32K to 50K.
1261
1262 2. The code in pcre_study() was using the value of the re argument before
1263 testing it for NULL. (Of course, in any sensible call of the function, it
1264 won't be NULL.)
1265
1266 3. The memmove() emulation function in pcre_internal.h, which is used on
1267 systems that lack both memmove() and bcopy() - that is, hardly ever -
1268 was missing a "static" storage class specifier.
1269
1270 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
1271 containing an extended class (one that cannot be represented by a bitmap
1272 because it contains high-valued characters or Unicode property items, e.g.
1273 [\pZ]). Almost always one would set UTF-8 mode when processing such a
1274 pattern, but PCRE should not loop if you do not (it no longer does).
1275 [Detail: two cases were found: (a) a repeated subpattern containing an
1276 extended class; (b) a recursive reference to a subpattern that followed a
1277 previous extended class. It wasn't skipping over the extended class
1278 correctly when UTF-8 mode was not set.]
1279
1280 5. A negated single-character class was not being recognized as fixed-length
1281 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
1282 compile error "lookbehind assertion is not fixed length".
1283
1284 6. The RunPerlTest auxiliary script was showing an unexpected difference
1285 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
1286 write a Perl script that can interpret lines of an input file either as
1287 byte characters or as UTF-8, which is what "perltest" was being required to
1288 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
1289 can't do is switch easily at run time between having the "use utf8;" pragma
1290 or not. In the end, I fudged it by using the RunPerlTest script to insert
1291 "use utf8;" explicitly for the UTF-8 tests.
1292
1293 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
1294 the end of the subject string, contrary to the documentation and to what
1295 Perl does. This was true of both matching functions. Now it matches only at
1296 the start of the subject and immediately after *internal* newlines.
1297
1298 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
1299 a pointer to an int instead of a pointer to an unsigned long int. This
1300 caused problems on 64-bit systems.
1301
1302 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
1303 instance of the 'standard' template library not being so standard".
1304
1305 10. There was no check on the number of named subpatterns nor the maximum
1306 length of a subpattern name. The product of these values is used to compute
1307 the size of the memory block for a compiled pattern. By supplying a very
1308 long subpattern name and a large number of named subpatterns, the size
1309 computation could be caused to overflow. This is now prevented by limiting
1310 the length of names to 32 characters, and the number of named subpatterns
1311 to 10,000.
1312
1313 11. Subpatterns that are repeated with specific counts have to be replicated in
1314 the compiled pattern. The size of memory for this was computed from the
1315 length of the subpattern and the repeat count. The latter is limited to
1316 65535, but there was no limit on the former, meaning that integer overflow
1317 could in principle occur. The compiled length of a repeated subpattern is
1318 now limited to 30,000 bytes in order to prevent this.
1319
1320 12. Added the optional facility to have named substrings with the same name.
1321
1322 13. Added the ability to use a named substring as a condition, using the
1323 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
1324 are numbers (not recommended). Forward references are permitted.
1325
1326 14. Added forward references in named backreferences (if you see what I mean).
1327
1328 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
1329 pattern could run off the end of the subject. For example, the pattern
1330 "(?s)(.{1,5})"8 did this with the subject "ab".
1331
1332 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
1333 PCRE_CASELESS was set when matching characters that were quantified with ?
1334 or *.
1335
1336 17. A character class other than a single negated character that had a minimum
1337 but no maximum quantifier - for example [ab]{6,} - was not handled
1338 correctly by pce_dfa_exec(). It would match only one character.
1339
1340 18. A valid (though odd) pattern that looked like a POSIX character
1341 class but used an invalid character after [ (for example [[,abc,]]) caused
1342 pcre_compile() to give the error "Failed: internal error: code overflow" or
1343 in some cases to crash with a glibc free() error. This could even happen if
1344 the pattern terminated after [[ but there just happened to be a sequence of
1345 letters, a binary zero, and a closing ] in the memory that followed.
1346
1347 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
1348 over the years. Originally (before any Unicode support), just the bottom 8
1349 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
1350 output from "man perlunicode" includes this:
1351
1352 The regular expression compiler produces polymorphic opcodes. That
1353 is, the pattern adapts to the data and automatically switches to
1354 the Unicode character scheme when presented with Unicode data--or
1355 instead uses a traditional byte scheme when presented with byte
1356 data.
1357
1358 Sadly, a wide octal escape does not cause a switch, and in a string with
1359 no other multibyte characters, these octal escapes are treated as before.
1360 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
1361 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
1362 Unicode string.
1363
1364 I have not perpetrated such confusion in PCRE. Up till now, it took just
1365 the bottom 8 bits, as in old Perl. I have now made octal escapes with
1366 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
1367 translate to the appropriate multibyte character.
1368
1369 29. Applied some refactoring to reduce the number of warnings from Microsoft
1370 and Borland compilers. This has included removing the fudge introduced
1371 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
1372 a warning about an unused variable.
1373
1374 21. PCRE has not included VT (character 0x0b) in the set of whitespace
1375 characters since release 4.0, because Perl (from release 5.004) does not.
1376 [Or at least, is documented not to: some releases seem to be in conflict
1377 with the documentation.] However, when a pattern was studied with
1378 pcre_study() and all its branches started with \s, PCRE still included VT
1379 as a possible starting character. Of course, this did no harm; it just
1380 caused an unnecessary match attempt.
1381
1382 22. Removed a now-redundant internal flag bit that recorded the fact that case
1383 dependency changed within the pattern. This was once needed for "required
1384 byte" processing, but is no longer used. This recovers a now-scarce options
1385 bit. Also moved the least significant internal flag bit to the most-
1386 significant bit of the word, which was not previously used (hangover from
1387 the days when it was an int rather than a uint) to free up another bit for
1388 the future.
1389
1390 23. Added support for CRLF line endings as well as CR and LF. As well as the
1391 default being selectable at build time, it can now be changed at runtime
1392 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
1393 specify that it is scanning data with non-default line endings.
1394
1395 24. Changed the definition of CXXLINK to make it agree with the definition of
1396 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
1397
1398 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
1399 recursions. This makes a big different to stack usage for some patterns.
1400
1401 26. If a subpattern containing a named recursion or subroutine reference such
1402 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
1403 the space required for the compiled pattern went wrong and gave too small a
1404 value. Depending on the environment, this could lead to "Failed: internal
1405 error: code overflow at offset 49" or "glibc detected double free or
1406 corruption" errors.
1407
1408 27. Applied patches from Google (a) to support the new newline modes and (b) to
1409 advance over multibyte UTF-8 characters in GlobalReplace.
1410
1411 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
1412 difference for some implementation of PCRE in some Windows version.
1413
1414 29. Added some extra testing facilities to pcretest:
1415
1416 \q<number> in a data line sets the "match limit" value
1417 \Q<number> in a data line sets the "match recursion limt" value
1418 -S <number> sets the stack size, where <number> is in megabytes
1419
1420 The -S option isn't available for Windows.
1421
1422
1423 Version 6.6 06-Feb-06
1424 ---------------------
1425
1426 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
1427 in pcreposix.h. I have copied the definition from pcre.h.
1428
1429 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
1430 because pcre.h is no longer a built file.
1431
1432 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
1433 not normally included in the compiled code.
1434
1435
1436 Version 6.5 01-Feb-06
1437 ---------------------
1438
1439 1. When using the partial match feature with pcre_dfa_exec(), it was not
1440 anchoring the second and subsequent partial matches at the new starting
1441 point. This could lead to incorrect results. For example, with the pattern
1442 /1234/, partially matching against "123" and then "a4" gave a match.
1443
1444 2. Changes to pcregrep:
1445
1446 (a) All non-match returns from pcre_exec() were being treated as failures
1447 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
1448 error message is output. Some extra information is given for the
1449 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
1450 probably the only errors that are likely to be caused by users (by
1451 specifying a regex that has nested indefinite repeats, for instance).
1452 If there are more than 20 of these errors, pcregrep is abandoned.
1453
1454 (b) A binary zero was treated as data while matching, but terminated the
1455 output line if it was written out. This has been fixed: binary zeroes
1456 are now no different to any other data bytes.
1457
1458 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
1459 used to set a locale for matching. The --locale=xxxx long option has
1460 been added (no short equivalent) to specify a locale explicitly on the
1461 pcregrep command, overriding the environment variables.
1462
1463 (d) When -B was used with -n, some line numbers in the output were one less
1464 than they should have been.
1465
1466 (e) Added the -o (--only-matching) option.
1467
1468 (f) If -A or -C was used with -c (count only), some lines of context were
1469 accidentally printed for the final match.
1470
1471 (g) Added the -H (--with-filename) option.
1472
1473 (h) The combination of options -rh failed to suppress file names for files
1474 that were found from directory arguments.
1475
1476 (i) Added the -D (--devices) and -d (--directories) options.
1477
1478 (j) Added the -F (--fixed-strings) option.
1479
1480 (k) Allow "-" to be used as a file name for -f as well as for a data file.
1481
1482 (l) Added the --colo(u)r option.
1483
1484 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
1485 is not present by default.
1486
1487 3. A nasty bug was discovered in the handling of recursive patterns, that is,
1488 items such as (?R) or (?1), when the recursion could match a number of
1489 alternatives. If it matched one of the alternatives, but subsequently,
1490 outside the recursion, there was a failure, the code tried to back up into
1491 the recursion. However, because of the way PCRE is implemented, this is not
1492 possible, and the result was an incorrect result from the match.
1493
1494 In order to prevent this happening, the specification of recursion has
1495 been changed so that all such subpatterns are automatically treated as
1496 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
1497
1498 4. I had overlooked the fact that, in some locales, there are characters for
1499 which isalpha() is true but neither isupper() nor islower() are true. In
1500 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
1501 and ordfeminine) are like this. This affected the treatment of \w and \W
1502 when they appeared in character classes, but not when they appeared outside
1503 a character class. The bit map for "word" characters is now created
1504 separately from the results of isalnum() instead of just taking it from the
1505 upper, lower, and digit maps. (Plus the underscore character, of course.)
1506
1507 5. The above bug also affected the handling of POSIX character classes such as
1508 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
1509 permanent tables. Instead, the bit maps for such a class were previously
1510 created as the appropriate unions of the upper, lower, and digit bitmaps.
1511 Now they are created by subtraction from the [[:word:]] class, which has
1512 its own bitmap.
1513
1514 6. The [[:blank:]] character class matches horizontal, but not vertical space.
1515 It is created by subtracting the vertical space characters (\x09, \x0a,
1516 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
1517 subtraction was done in the overall bitmap for a character class, meaning
1518 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
1519 be recognized. This bug has been fixed.
1520
1521 7. Patches from the folks at Google:
1522
1523 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
1524 real life, but is still worth protecting against".
1525
1526 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
1527 regular expressions".
1528
1529 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
1530 have it.
1531
1532 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
1533 "configure" and the latter not, in order to fix a problem somebody had
1534 with compiling the Arg class on HP-UX.
1535
1536 (e) Improve the error-handling of the C++ wrapper a little bit.
1537
1538 (f) New tests for checking recursion limiting.
1539
1540 8. The pcre_memmove() function, which is used only if the environment does not
1541 have a standard memmove() function (and is therefore rarely compiled),
1542 contained two bugs: (a) use of int instead of size_t, and (b) it was not
1543 returning a result (though PCRE never actually uses the result).
1544
1545 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
1546 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
1547 returned instead of calling malloc() with an overflowing number that would
1548 most likely cause subsequent chaos.
1549
1550 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
1551
1552 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
1553 with this option is matched, the nmatch and pmatch options of regexec() are
1554 ignored.
1555
1556 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
1557 provided in case anyone wants to the the POSIX interface with UTF-8
1558 strings.
1559
1560 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
1561 C++ linking (needed for some HP-UX environments).
1562
1563 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
1564 (unused parameter) and in the pcre_printint() function (omitted "default"
1565 switch label when the default is to do nothing).
1566
1567 15. Added some code to make it possible, when PCRE is compiled as a C++
1568 library, to replace subject pointers for pcre_exec() with a smart pointer
1569 class, thus making it possible to process discontinuous strings.
1570
1571 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
1572 much the same function. They were added by different people who were trying
1573 to make PCRE easy to compile on non-Unix systems. It has been suggested
1574 that PCRE_EXPORT be abolished now that there is more automatic apparatus
1575 for compiling on Windows systems. I have therefore replaced it with
1576 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
1577 defaults to "extern" for C or "extern C" for C++, which works fine on
1578 Unix-like systems. It is now possible to override the value of PCRE_DATA_
1579 SCOPE with something explicit in config.h. In addition:
1580
1581 (a) pcreposix.h still had just "extern" instead of either of these macros;
1582 I have replaced it with PCRE_DATA_SCOPE.
1583
1584 (b) Functions such as _pcre_xclass(), which are internal to the library,
1585 but external in the C sense, all had PCRE_EXPORT in their definitions.
1586 This is apparently wrong for the Windows case, so I have removed it.
1587 (It makes no difference on Unix-like systems.)
1588
1589 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
1590 of recursive calls to match(). This is different to MATCH_LIMIT because
1591 that limits the total number of calls to match(), not all of which increase
1592 the depth of recursion. Limiting the recursion depth limits the amount of
1593 stack (or heap if NO_RECURSE is set) that is used. The default can be set
1594 when PCRE is compiled, and changed at run time. A patch from Google adds
1595 this functionality to the C++ interface.
1596
1597 18. Changes to the handling of Unicode character properties:
1598
1599 (a) Updated the table to Unicode 4.1.0.
1600
1601 (b) Recognize characters that are not in the table as "Cn" (undefined).
1602
1603 (c) I revised the way the table is implemented to a much improved format
1604 which includes recognition of ranges. It now supports the ranges that
1605 are defined in UnicodeData.txt, and it also amalgamates other
1606 characters into ranges. This has reduced the number of entries in the
1607 table from around 16,000 to around 3,000, thus reducing its size
1608 considerably. I realized I did not need to use a tree structure after
1609 all - a binary chop search is just as efficient. Having reduced the
1610 number of entries, I extended their size from 6 bytes to 8 bytes to
1611 allow for more data.
1612
1613 (d) Added support for Unicode script names via properties such as \p{Han}.
1614
1615 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
1616 matching that character.
1617
1618 20. When matching a repeated Unicode property with a minimum greater than zero,
1619 (for example \pL{2,}), PCRE could look past the end of the subject if it
1620 reached it while seeking the minimum number of characters. This could
1621 happen only if some of the characters were more than one byte long, because
1622 there is a check for at least the minimum number of bytes.
1623
1624 21. Refactored the implementation of \p and \P so as to be more general, to
1625 allow for more different types of property in future. This has changed the
1626 compiled form incompatibly. Anybody with saved compiled patterns that use
1627 \p or \P will have to recompile them.
1628
1629 22. Added "Any" and "L&" to the supported property types.
1630
1631 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
1632 but give a compile time error if the value is greater than 0xff.
1633
1634 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
1635 accidentally not being installed or uninstalled.
1636
1637 25. The pcre.h file was built from pcre.h.in, but the only changes that were
1638 made were to insert the current release number. This seemed silly, because
1639 it made things harder for people building PCRE on systems that don't run
1640 "configure". I have turned pcre.h into a distributed file, no longer built
1641 by "configure", with the version identification directly included. There is
1642 no longer a pcre.h.in file.
1643
1644 However, this change necessitated a change to the pcre-config script as
1645 well. It is built from pcre-config.in, and one of the substitutions was the
1646 release number. I have updated configure.ac so that ./configure now finds
1647 the release number by grepping pcre.h.
1648
1649 26. Added the ability to run the tests under valgrind.
1650
1651
1652 Version 6.4 05-Sep-05
1653 ---------------------
1654
1655 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
1656 "--" to be printed when multiple files were scanned, even when none of the
1657 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
1658 consider it to be a bug, and have restored the previous behaviour.
1659
1660 2. A couple of code tidies to get rid of compiler warnings.
1661
1662 3. The pcretest program used to cheat by referring to symbols in the library
1663 whose names begin with _pcre_. These are internal symbols that are not
1664 really supposed to be visible externally, and in some environments it is
1665 possible to suppress them. The cheating is now confined to including
1666 certain files from the library's source, which is a bit cleaner.
1667
1668 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
1669 file's purpose clearer.
1670
1671 5. Reorganized pcre_ucp_findchar().
1672
1673
1674 Version 6.3 15-Aug-05
1675 ---------------------
1676
1677 1. The file libpcre.pc.in did not have general read permission in the tarball.
1678
1679 2. There were some problems when building without C++ support:
1680
1681 (a) If C++ support was not built, "make install" and "make test" still
1682 tried to test it.
1683
1684 (b) There were problems when the value of CXX was explicitly set. Some
1685 changes have been made to try to fix these, and ...
1686
1687 (c) --disable-cpp can now be used to explicitly disable C++ support.
1688
1689 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
1690 backslash in a target when C++ was disabled. This confuses some
1691 versions of "make", apparently. Using an intermediate variable solves
1692 this. (Same for CPP_LOBJ.)
1693
1694 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
1695 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
1696 necessary on certain architectures.
1697
1698 4. Added a setting of -export-symbols-regex to the link command to remove
1699 those symbols that are exported in the C sense, but actually are local
1700 within the library, and not documented. Their names all begin with
1701 "_pcre_". This is not a perfect job, because (a) we have to except some
1702 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
1703 available (and never for static libraries). I have made a note to try to
1704 find a way round (a) in the future.
1705
1706
1707 Version 6.2 01-Aug-05
1708 ---------------------
1709
1710 1. There was no test for integer overflow of quantifier values. A construction
1711 such as {1111111111111111} would give undefined results. What is worse, if
1712 a minimum quantifier for a parenthesized subpattern overflowed and became
1713 negative, the calculation of the memory size went wrong. This could have
1714 led to memory overwriting.
1715
1716 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
1717
1718 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
1719 operating environments where this matters.
1720
1721 4. Applied Giuseppe Maxia's patch to add additional features for controlling
1722 PCRE options from within the C++ wrapper.
1723
1724 5. Named capturing subpatterns were not being correctly counted when a pattern
1725 was compiled. This caused two problems: (a) If there were more than 100
1726 such subpatterns, the calculation of the memory needed for the whole
1727 compiled pattern went wrong, leading to an overflow error. (b) Numerical
1728 back references of the form \12, where the number was greater than 9, were
1729 not recognized as back references, even though there were sufficient
1730 previous subpatterns.
1731
1732 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
1733 versions of gcc, e.g. 2.95.4.
1734
1735
1736 Version 6.1 21-Jun-05
1737 ---------------------
1738
1739 1. There was one reference to the variable "posix" in pcretest.c that was not
1740 surrounded by "#if !defined NOPOSIX".
1741
1742 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
1743 the cross-check on the old pcre_info() function, for the benefit of the
1744 cut-down version of PCRE that is currently imported into Exim.
1745
1746 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
1747 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
1748 patterns that start (?i) but I don't think that matters. The use of (?i) is
1749 just an example; this all applies to the other options as well.
1750
1751 4. Since libtool seems to echo the compile commands it is issuing, the output
1752 from "make" can be reduced a bit by putting "@" in front of each libtool
1753 compile command.
1754
1755 5. Patch from the folks at Google for configure.in to be a bit more thorough
1756 in checking for a suitable C++ installation before trying to compile the
1757 C++ stuff. This should fix a reported problem when a compiler was present,
1758 but no suitable headers.
1759
1760 6. The man pages all had just "PCRE" as their title. I have changed them to
1761 be the relevant file name. I have also arranged that these names are
1762 retained in the file doc/pcre.txt, which is a concatenation in text format
1763 of all the man pages except the little individual ones for each function.
1764
1765 7. The NON-UNIX-USE file had not been updated for the different set of source
1766 files that come with release 6. I also added a few comments about the C++
1767 wrapper.
1768
1769
1770 Version 6.0 07-Jun-05
1771 ---------------------
1772
1773 1. Some minor internal re-organization to help with my DFA experiments.
1774
1775 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
1776 didn't matter for the library itself when fully configured, but did matter
1777 when compiling without UCP support, or within Exim, where the ucp files are
1778 not imported.
1779
1780 3. Refactoring of the library code to split up the various functions into
1781 different source modules. The addition of the new DFA matching code (see
1782 below) to a single monolithic source would have made it really too
1783 unwieldy, quite apart from causing all the code to be include in a
1784 statically linked application, when only some functions are used. This is
1785 relevant even without the DFA addition now that patterns can be compiled in
1786 one application and matched in another.
1787
1788 The downside of splitting up is that there have to be some external
1789 functions and data tables that are used internally in different modules of
1790 the library but which are not part of the API. These have all had their
1791 names changed to start with "_pcre_" so that they are unlikely to clash
1792 with other external names.
1793
1794 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
1795 a different (DFA) algorithm. Although it is slower than the original
1796 function, it does have some advantages for certain types of matching
1797 problem.
1798
1799 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
1800 including restarting after a partial match.
1801
1802 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
1803 defined when compiling for Windows was sent to me. I have put it into the
1804 code, though I have no means of testing or verifying it.
1805
1806 7. Added the pcre_refcount() auxiliary function.
1807
1808 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
1809 match before or at the first newline in the subject string. In pcretest,
1810 the /f option on a pattern can be used to set this.
1811
1812 9. A repeated \w when used in UTF-8 mode with characters greater than 256
1813 would behave wrongly. This has been present in PCRE since release 4.0.
1814
1815 10. A number of changes to the pcregrep command:
1816
1817 (a) Refactored how -x works; insert ^(...)$ instead of setting
1818 PCRE_ANCHORED and checking the length, in preparation for adding
1819 something similar for -w.
1820
1821 (b) Added the -w (match as a word) option.
1822
1823 (c) Refactored the way lines are read and buffered so as to have more
1824 than one at a time available.
1825
1826 (d) Implemented a pcregrep test script.
1827
1828 (e) Added the -M (multiline match) option. This allows patterns to match
1829 over several lines of the subject. The buffering ensures that at least
1830 8K, or the rest of the document (whichever is the shorter) is available
1831 for matching (and similarly the previous 8K for lookbehind assertions).
1832
1833 (f) Changed the --help output so that it now says
1834
1835 -w, --word-regex(p)
1836
1837 instead of two lines, one with "regex" and the other with "regexp"
1838 because that confused at least one person since the short forms are the
1839 same. (This required a bit of code, as the output is generated
1840 automatically from a table. It wasn't just a text change.)
1841
1842 (g) -- can be used to terminate pcregrep options if the next thing isn't an
1843 option but starts with a hyphen. Could be a pattern or a path name
1844 starting with a hyphen, for instance.
1845
1846 (h) "-" can be given as a file name to represent stdin.
1847
1848 (i) When file names are being printed, "(standard input)" is used for
1849 the standard input, for compatibility with GNU grep. Previously
1850 "<stdin>" was used.
1851
1852 (j) The option --label=xxx can be used to supply a name to be used for
1853 stdin when file names are being printed. There is no short form.
1854
1855 (k) Re-factored the options decoding logic because we are going to add
1856 two more options that take data. Such options can now be given in four
1857 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
1858
1859 (l) Added the -A, -B, and -C options for requesting that lines of context
1860 around matches be printed.
1861
1862 (m) Added the -L option to print the names of files that do not contain
1863 any matching lines, that is, the complement of -l.
1864
1865 (n) The return code is 2 if any file cannot be opened, but pcregrep does
1866 continue to scan other files.
1867
1868 (o) The -s option was incorrectly implemented. For compatibility with other
1869 greps, it now suppresses the error message for a non-existent or non-
1870 accessible file (but not the return code). There is a new option called
1871 -q that suppresses the output of matching lines, which was what -s was
1872 previously doing.
1873
1874 (p) Added --include and --exclude options to specify files for inclusion
1875 and exclusion when recursing.
1876
1877 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
1878 Hopefully, it now does.
1879
1880 12. Missing cast in pcre_study().
1881
1882 13. Added an "uninstall" target to the makefile.
1883
1884 14. Replaced "extern" in the function prototypes in Makefile.in with
1885 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
1886 world, but is set differently for Windows.
1887
1888 15. Added a second compiling function called pcre_compile2(). The only
1889 difference is that it has an extra argument, which is a pointer to an
1890 integer error code. When there is a compile-time failure, this is set
1891 non-zero, in addition to the error test pointer being set to point to an
1892 error message. The new argument may be NULL if no error number is required
1893 (but then you may as well call pcre_compile(), which is now just a
1894 wrapper). This facility is provided because some applications need a
1895 numeric error indication, but it has also enabled me to tidy up the way
1896 compile-time errors are handled in the POSIX wrapper.
1897
1898 16. Added VPATH=.libs to the makefile; this should help when building with one
1899 prefix path and installing with another. (Or so I'm told by someone who
1900 knows more about this stuff than I do.)
1901
1902 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
1903 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
1904 match everything, including newlines. This is not POSIX-compatible, but
1905 somebody wanted the feature. From pcretest it can be activated by using
1906 both the P and the s flags.
1907
1908 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
1909
1910 19. libpcre.pc was being incorrectly installed as executable.
1911
1912 20. A couple of places in pcretest check for end-of-line by looking for '\n';
1913 it now also looks for '\r' so that it will work unmodified on Windows.
1914
1915 21. Added Google's contributed C++ wrapper to the distribution.
1916
1917 22. Added some untidy missing memory free() calls in pcretest, to keep
1918 Electric Fence happy when testing.
1919
1920
1921
1922 Version 5.0 13-Sep-04
1923 ---------------------
1924
1925 1. Internal change: literal characters are no longer packed up into items
1926 containing multiple characters in a single byte-string. Each character
1927 is now matched using a separate opcode. However, there may be more than one
1928 byte in the character in UTF-8 mode.
1929
1930 2. The pcre_callout_block structure has two new fields: pattern_position and
1931 next_item_length. These contain the offset in the pattern to the next match
1932 item, and its length, respectively.
1933
1934 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
1935 insertion of callouts before each pattern item. Added the /C option to
1936 pcretest to make use of this.
1937
1938 4. On the advice of a Windows user, the lines
1939
1940 #if defined(_WIN32) || defined(WIN32)
1941 _setmode( _fileno( stdout ), 0x8000 );
1942 #endif /* defined(_WIN32) || defined(WIN32) */
1943
1944 have been added to the source of pcretest. This apparently does useful
1945 magic in relation to line terminators.
1946
1947 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
1948 for the benefit of those environments where the "b" makes a difference.
1949
1950 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
1951 to know about it. I have put a hack into configure.in that adds in code
1952 to set GCC=yes if CC=icc. This seems to end up at a point in the
1953 generated configure script that is early enough to affect the setting of
1954 compiler options, which is what is needed, but I have no means of testing
1955 whether it really works. (The user who reported this had patched the
1956 generated configure script, which of course I cannot do.)
1957
1958 LATER: After change 22 below (new libtool files), the configure script
1959 seems to know about icc (and also ecc). Therefore, I have commented out
1960 this hack in configure.in.
1961
1962 7. Added support for pkg-config (2 patches were sent in).
1963
1964 8. Negated POSIX character classes that used a combination of internal tables
1965 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
1966 [[:^ascii]]. Typically, they would match almost any characters. The other
1967 POSIX classes were not broken in this way.
1968
1969 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
1970 to find the match, as PCRE was deluded into thinking that the match had to
1971 start at the start point or following a newline. The same bug applied to
1972 patterns with negative forward assertions or any backward assertions
1973 preceding ".*" at the start, unless the pattern required a fixed first
1974 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
1975
1976 10. In UTF-8 mode, when moving forwards in the subject after a failed match
1977 starting at the last subject character, bytes beyond the end of the subject
1978 string were read.
1979
1980 11. Renamed the variable "class" as "classbits" to make life easier for C++
1981 users. (Previously there was a macro definition, but it apparently wasn't
1982 enough.)
1983
1984 12. Added the new field "tables" to the extra data so that tables can be passed
1985 in at exec time, or the internal tables can be re-selected. This allows
1986 a compiled regex to be saved and re-used at a later time by a different
1987 program that might have everything at different addresses.
1988
1989 13. Modified the pcre-config script so that, when run on Solaris, it shows a
1990 -R library as well as a -L library.
1991
1992 14. The debugging options of pcretest (-d on the command line or D on a
1993 pattern) showed incorrect output for anything following an extended class
1994 that contained multibyte characters and which was followed by a quantifier.
1995
1996 15. Added optional support for general category Unicode character properties
1997 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
1998 support. It adds about 90K to the size of the library. The meanings of the
1999 inbuilt class escapes such as \d and \s have NOT been changed.
2000
2001 16. Updated pcredemo.c to include calls to free() to release the memory for the
2002 compiled pattern.
2003
2004 17. The generated file chartables.c was being created in the source directory
2005 instead of in the building directory. This caused the build to fail if the
2006 source directory was different from the building directory, and was
2007 read-only.
2008
2009 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
2010 file. No doubt somebody will tell me if they don't make sense... Also added
2011 Dan Mooney's comments about building on OpenVMS.
2012
2013 19. Added support for partial matching via the PCRE_PARTIAL option for
2014 pcre_exec() and the \P data escape in pcretest.
2015
2016 20. Extended pcretest with 3 new pattern features:
2017
2018 (i) A pattern option of the form ">rest-of-line" causes pcretest to
2019 write the compiled pattern to the file whose name is "rest-of-line".
2020 This is a straight binary dump of the data, with the saved pointer to
2021 the character tables forced to be NULL. The study data, if any, is
2022 written too. After writing, pcretest reads a new pattern.
2023
2024 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
2025 compiled pattern from the given file. There must not be any
2026 occurrences of "<" in the file name (pretty unlikely); if there are,
2027 pcretest will instead treat the initial "<" as a pattern delimiter.
2028 After reading in the pattern, pcretest goes on to read data lines as
2029 usual.
2030
2031 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
2032 and 16-bit fields in a compiled pattern, to simulate a pattern that
2033 was compiled on a host of opposite endianness.
2034
2035 21. The pcre-exec() function can now cope with patterns that were compiled on
2036 hosts of opposite endianness, with this restriction:
2037
2038 As for any compiled expression that is saved and used later, the tables
2039 pointer field cannot be preserved; the extra_data field in the arguments
2040 to pcre_exec() should be used to pass in a tables address if a value
2041 other than the default internal tables were used at compile time.
2042
2043 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
2044 now diagnosed as an error. Previously, most of the time, a negative number
2045 would have been treated as zero, but if in addition "ovector" was passed as
2046 NULL, a crash could occur.
2047
2048 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
2049 new versions from the libtool 1.5 distribution (the last one is a copy of
2050 a file called libtool.m4). This seems to have fixed the need to patch
2051 "configure" to support Darwin 1.3 (which I used to do). However, I still
2052 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
2053 workstation).
2054
2055 24. Changed the PCRE licence to be the more standard "BSD" licence.
2056
2057
2058 Version 4.5 01-Dec-03
2059 ---------------------
2060
2061 1. There has been some re-arrangement of the code for the match() function so
2062 that it can be compiled in a version that does not call itself recursively.
2063 Instead, it keeps those local variables that need separate instances for
2064 each "recursion" in a frame on the heap, and gets/frees frames whenever it
2065 needs to "recurse". Keeping track of where control must go is done by means
2066 of setjmp/longjmp. The whole thing is implemented by a set of macros that
2067 hide most of the details from the main code, and operates only if
2068 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
2069 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
2070 operating.
2071
2072 To make it easier for callers to provide specially tailored get/free
2073 functions for this usage, two new functions, pcre_stack_malloc, and
2074 pcre_stack_free, are used. They are always called in strict stacking order,
2075 and the size of block requested is always the same.
2076
2077 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
2078 PCRE has been compiled to use the stack or the heap for recursion. The
2079 -C option of pcretest uses this to show which version is compiled.
2080
2081 A new data escape \S, is added to pcretest; it causes the amounts of store
2082 obtained and freed by both kinds of malloc/free at match time to be added
2083 to the output.
2084
2085 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
2086 what's available on my current Linux desktop machine.
2087
2088 3. When matching a UTF-8 string, the test for a valid string at the start has
2089 been extended. If start_offset is not zero, PCRE now checks that it points
2090 to a byte that is the start of a UTF-8 character. If not, it returns
2091 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
2092 this is necessary because there may be backward assertions in the pattern.
2093 When matching the same subject several times, it may save resources to use
2094 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
2095
2096 4. The code for checking the validity of UTF-8 strings has been tightened so
2097 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
2098 containing "overlong sequences".
2099
2100 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
2101 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
2102 should have been "&", but it just so happened that all the cases this let
2103 through by mistake were picked up later in the function.
2104
2105 6. I had used a variable called "isblank" - this is a C99 function, causing
2106 some compilers to warn. To avoid this, I renamed it (as "blankclass").
2107
2108 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
2109 prompting; (b) run "./pcretest /dev/null" at the start of the test script
2110 so the version is shown; (c) stop "make test" echoing "./RunTest".
2111
2112 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
2113
2114 9. The prototype for memmove() for systems that don't have it was using
2115 size_t, but the inclusion of the header that defines size_t was later. I've
2116 moved the #includes for the C headers earlier to avoid this.
2117
2118 10. Added some adjustments to the code to make it easier to compiler on certain
2119 special systems:
2120
2121 (a) Some "const" qualifiers were missing.
2122 (b) Added the macro EXPORT before all exported functions; by default this
2123 is defined to be empty.
2124 (c) Changed the dftables auxiliary program (that builds chartables.c) so
2125 that it reads its output file name as an argument instead of writing
2126 to the standard output and assuming this can be redirected.
2127
2128 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
2129 class containing characters with values greater than 255, PCRE compilation
2130 went into a loop.
2131
2132 12. A recursive reference to a subpattern that was within another subpattern
2133 that had a minimum quantifier of zero caused PCRE to crash. For example,
2134 (x(y(?2))z)? provoked this bug with a subject that got as far as the
2135 recursion. If the recursively-called subpattern itself had a zero repeat,
2136 that was OK.
2137
2138 13. In pcretest, the buffer for reading a data line was set at 30K, but the
2139 buffer into which it was copied (for escape processing) was still set at
2140 1024, so long lines caused crashes.
2141
2142 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
2143 "internal error: code overflow...". This applied to any character class
2144 that was followed by a possessive quantifier.
2145
2146 15. Modified the Makefile to add libpcre.la as a prerequisite for
2147 libpcreposix.la because I was told this is needed for a parallel build to
2148 work.
2149
2150 16. If a pattern that contained .* following optional items at the start was
2151 studied, the wrong optimizing data was generated, leading to matching
2152 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
2153 matching string must start with a or b or c. The correct conclusion for
2154 this pattern is that a match can start with any character.
2155
2156
2157 Version 4.4 13-Aug-03
2158 ---------------------
2159
2160 1. In UTF-8 mode, a character class containing characters with values between
2161 127 and 255 was not handled correctly if the compiled pattern was studied.
2162 In fixing this, I have also improved the studying algorithm for such
2163 classes (slightly).
2164
2165 2. Three internal functions had redundant arguments passed to them. Removal
2166 might give a very teeny performance improvement.
2167
2168 3. Documentation bug: the value of the capture_top field in a callout is *one
2169 more than* the number of the hightest numbered captured substring.
2170
2171 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
2172 in incorrectly linking with a previously installed version. They now link
2173 explicitly with libpcre.la.
2174
2175 5. configure.in no longer needs to recognize Cygwin specially.
2176
2177 6. A problem in pcre.in for Windows platforms is fixed.
2178
2179 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
2180 pcretest, it used to include the size of the study block as part of its
2181 output. Unfortunately, the structure contains a field that has a different
2182 size on different hardware architectures. This meant that the tests that
2183 showed this size failed. As the block is currently always of a fixed size,
2184 this information isn't actually particularly useful in pcretest output, so
2185 I have just removed it.
2186
2187 8. Three pre-processor statements accidentally did not start in column 1.
2188 Sadly, there are *still* compilers around that complain, even though
2189 standard C has not required this for well over a decade. Sigh.
2190
2191 9. In pcretest, the code for checking callouts passed small integers in the
2192 callout_data field, which is a void * field. However, some picky compilers
2193 complained about the casts involved for this on 64-bit systems. Now
2194 pcretest passes the address of the small integer instead, which should get
2195 rid of the warnings.
2196
2197 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
2198 both compile and run time, and gives an error if an invalid UTF-8 sequence
2199 is found. There is a option for disabling this check in cases where the
2200 string is known to be correct and/or the maximum performance is wanted.
2201
2202 11. In response to a bug report, I changed one line in Makefile.in from
2203
2204 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
2205 to
2206 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
2207
2208 to look similar to other lines, but I have no way of telling whether this
2209 is the right thing to do, as I do not use Windows. No doubt I'll get told
2210 if it's wrong...
2211
2212
2213 Version 4.3 21-May-03
2214 ---------------------
2215
2216 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
2217 Makefile.
2218
2219 2. Some refactoring to improve the quality of the code:
2220
2221 (i) The utf8_table... variables are now declared "const".
2222
2223 (ii) The code for \cx, which used the "case flipping" table to upper case
2224 lower case letters, now just substracts 32. This is ASCII-specific,
2225 but the whole concept of \cx is ASCII-specific, so it seems
2226 reasonable.
2227
2228 (iii) PCRE was using its character types table to recognize decimal and
2229 hexadecimal digits in the pattern. This is silly, because it handles
2230 only 0-9, a-f, and A-F, but the character types table is locale-
2231 specific, which means strange things might happen. A private
2232 table is now used for this - though it costs 256 bytes, a table is
2233 much faster than multiple explicit tests. Of course, the standard
2234 character types table is still used for matching digits in subject
2235 strings against \d.
2236
2237 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
2238 ending in _t are). So I've renamed it as ESC_tee.
2239
2240 3. The first argument for regexec() in the POSIX wrapper should have been
2241 defined as "const".
2242
2243 4. Changed pcretest to use malloc() for its buffers so that they can be
2244 Electric Fenced for debugging.
2245
2246 5. There were several places in the code where, in UTF-8 mode, PCRE would try
2247 to read one or more bytes before the start of the subject string. Often this
2248 had no effect on PCRE's behaviour, but in some circumstances it could
2249 provoke a segmentation fault.
2250
2251 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
2252 to try to read one or more bytes before the start of the subject string.
2253
2254 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
2255 UTF-8 support could misbehave in various ways if the subject string
2256 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
2257 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
2258 back over UTF-8 characters.)
2259
2260
2261 Version 4.2 14-Apr-03
2262 ---------------------
2263
2264 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
2265
2266 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
2267 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
2268 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
2269 [WIN_PREFIX]: new variable, "cyg" for Cygwin
2270 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
2271 and BUILD_EXEEXT
2272 Note: automatic setting of the BUILD variables is not yet working
2273 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
2274 compile-time but not at link-time
2275 [LINK]: use for linking executables only
2276 make different versions for Windows and non-Windows
2277 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
2278 libraries
2279 [LINK_FOR_BUILD]: new variable
2280 [OBJEXT]: use throughout
2281 [EXEEXT]: use throughout
2282 <winshared>: new target
2283 <wininstall>: new target
2284 <dftables.o>: use native compiler
2285 <dftables>: use native linker
2286 <install>: handle Windows platform correctly
2287 <clean>: ditto
2288 <check>: ditto
2289 copy DLL to top builddir before testing
2290
2291 As part of these changes, -no-undefined was removed again. This was reported
2292 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
2293 in any case.
2294
2295 3. Some tidies to get rid of compiler warnings:
2296
2297 . In the match_data structure, match_limit was an unsigned long int, whereas
2298 match_call_count was an int. I've made them both unsigned long ints.
2299
2300 . In pcretest the fact that a const uschar * doesn't automatically cast to
2301 a void * provoked a warning.
2302
2303 . Turning on some more compiler warnings threw up some "shadow" variables
2304 and a few more missing casts.
2305
2306 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
2307 option, a class that contained a single character with a value between 128
2308 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
2309
2310 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
2311 option, a class that contained several characters, but with at least one
2312 whose value was between 128 and 255 caused PCRE to crash.
2313
2314
2315 Version 4.1 12-Mar-03
2316 ---------------------
2317
2318 1. Compiling with gcc -pedantic found a couple of places where casts were
2319 needed, and a string in dftables.c that was longer than standard compilers are
2320 required to support.
2321
2322 2. Compiling with Sun's compiler found a few more places where the code could
2323 be tidied up in order to avoid warnings.
2324
2325 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
2326 first of these names is deprecated in the latest Autoconf in favour of the name
2327 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
2328 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
2329 analogy I have changed it to CFLAGS_FOR_BUILD.
2330
2331 4. Added -no-undefined to the linking command in the Makefile, because this is
2332 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
2333 linking step for the pcreposix library.
2334
2335 5. PCRE was failing to diagnose the case of two named groups with the same
2336 name.
2337
2338 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
2339 literal character that is needed in the subject for a match, and scans along to
2340 ensure that it is present before embarking on the full matching process. This
2341 saves time in cases of nested unlimited repeats that are never going to match.
2342 Problem: the scan can take a lot of time if the subject is very long (e.g.
2343 megabytes), thus penalizing straightforward matches. It is now done only if the
2344 amount of subject to be scanned is less than 1000 bytes.
2345
2346 7. A lesser problem with the same optimization is that it was recording the
2347 first character of an anchored pattern as "needed", thus provoking a search
2348 right along the subject, even when the first match of the pattern was going to
2349 fail. The "needed" character is now not set for anchored patterns, unless it
2350 follows something in the pattern that is of non-fixed length. Thus, it still
2351 fulfils its original purpose of finding quick non-matches in cases of nested
2352 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
2353
2354
2355 Version 4.0 17-Feb-03
2356 ---------------------
2357
2358 1. If a comment in an extended regex that started immediately after a meta-item
2359 extended to the end of string, PCRE compiled incorrect data. This could lead to
2360 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
2361
2362 2. Moved to autoconf 2.53 and libtool 1.4.2.
2363
2364 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
2365 the special perltest8 script is no longer needed - all the tests can be run
2366 from a single perltest script.
2367
2368 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
2369 by \s. It has now been removed in PCRE. This means it isn't recognized as
2370 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
2371 class [:space:] *does* include VT, thereby creating a mess.
2372
2373 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
2374 space and tab.
2375
2376 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
2377 its new features into the main test script, reducing the number of scripts.
2378
2379 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
2380 were backward compatible, and made the (?i) apply to the whole pattern, as if
2381 /i were given. Now it behaves more logically, and applies the option setting
2382 only to what follows. PCRE has been changed to follow suit. However, if it
2383 finds options settings right at the start of the pattern, it extracts them into
2384 the global options, as before. Thus, they show up in the info data.
2385
2386 8. Added support for the \Q...\E escape sequence. Characters in between are
2387 treated as literals. This is slightly different from Perl in that $ and @ are
2388 also handled as literals inside the quotes. In Perl, they will cause variable
2389 interpolation. Note the following examples:
2390
2391 Pattern PCRE matches Perl matches
2392
2393 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
2394 \Qabc\$xyz\E abc\$xyz abc\$xyz
2395 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
2396
2397 For compatibility with Perl, \Q...\E sequences are recognized inside character
2398 classes as well as outside them.
2399
2400 9. Re-organized 3 code statements in pcretest to avoid "overflow in
2401 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
2402 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
2403 signed/unsigned warnings.
2404
2405 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
2406 option for pcretest, so I've replaced it by a simple function that does just
2407 that job.
2408
2409 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
2410 "pcregrep -".
2411
2412 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
2413 Java package. This provides some syntactic sugar for simple cases of what my
2414 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
2415 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
2416 item, you can use this simplified notation. Note that only makes sense with
2417 greedy quantifiers. Consequently, the use of the possessive quantifier forces
2418 greediness, whatever the setting of the PCRE_UNGREEDY option.
2419
2420 13. A change of greediness default within a pattern was not taking effect at
2421 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
2422 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
2423 was abstracted outside.
2424
2425 14. PCRE now supports the \G assertion. It is true when the current matching
2426 position is at the start point of the match. This differs from \A when the
2427 starting offset is non-zero. Used with the /g option of pcretest (or similar
2428 code), it works in the same way as it does for Perl's /g option. If all
2429 alternatives of a regex begin with \G, the expression is anchored to the start
2430 match position, and the "anchored" flag is set in the compiled expression.
2431
2432 15. Some bugs concerning the handling of certain option changes within patterns
2433 have been fixed. These applied to options other than (?ims). For example,
2434 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
2435 been the other way round. Some of this was related to change 7 above.
2436
2437 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
2438 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
2439 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
2440 POSIX classes only within a class (e.g. /[[:alpha:]]/).
2441
2442 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
2443 mode. Unlike ".", it always matches newline, whatever the setting of
2444 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
2445 assertions. Perl allows it, but it doesn't (in general) work because it can't
2446 calculate the length of the lookbehind. At least, that's the case for Perl
2447 5.8.0 - I've been told they are going to document that it doesn't work in
2448 future.
2449
2450 18. Added an error diagnosis for escapes that PCRE does not support: these are
2451 \L, \l, \N, \P, \p, \U, \u, and \X.
2452
2453 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
2454 reading past the end of the pattern in cases such as /[abcd/.
2455
2456 20. PCRE was getting more memory than necessary for patterns with classes that
2457 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
2458
2459 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
2460 compiling PCRE for use with Virtual Pascal.
2461
2462 22. Small fix to the Makefile to make it work properly if the build is done
2463 outside the source tree.
2464
2465 23. Added a new extension: a condition to go with recursion. If a conditional
2466 subpattern starts with (?(R) the "true" branch is used if recursion has
2467 happened, whereas the "false" branch is used only at the top level.
2468
2469 24. When there was a very long string of literal characters (over 255 bytes
2470 without UTF support, over 250 bytes with UTF support), the computation of how
2471 much memory was required could be incorrect, leading to segfaults or other
2472 strange effects.
2473
2474 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
2475 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
2476 there was a subsequent back reference to those brackets. This meant that, for
2477 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
2478 possible to check for precisely this case. All we can do is abandon the
2479 optimization if .* occurs inside capturing brackets when there are any back
2480 references whatsoever. (See below for a better fix that came later.)
2481
2482 26. The handling of the optimization for finding the first character of a
2483 non-anchored pattern, and for finding a character that is required later in the
2484 match were failing in some cases. This didn't break the matching; it just
2485 failed to optimize when it could. The way this is done has been re-implemented.
2486
2487 27. Fixed typo in error message for invalid (?R item (it said "(?p").
2488
2489 28. Added a new feature that provides some of the functionality that Perl
2490 provides with (?{...}). The facility is termed a "callout". The way it is done
2491 in PCRE is for the caller to provide an optional function, by setting
2492 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
2493 global variable. By default it is unset, which disables all calling out. To get
2494 the function called, the regex must include (?C) at appropriate points. This
2495 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
2496 This provides a means of identifying different callout points. When PCRE
2497 reaches such a point in the regex, if pcre_callout has been set, the external
2498 function is called. It is provided with data in a structure called
2499 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
2500 matching continues; if it returns a non-zero value, the match at the current
2501 point fails. However, backtracking will occur if possible. [This was changed
2502 later and other features added - see item 49 below.]
2503
2504 29. pcretest is upgraded to test the callout functionality. It provides a
2505 callout function that displays information. By default, it shows the start of
2506 the match and the current position in the text. There are some new data escapes
2507 to vary what happens:
2508
2509 \C+ in addition, show current contents of captured substrings
2510 \C- do not supply a callout function
2511 \C!n return 1 when callout number n is reached
2512 \C!n!m return 1 when callout number n is reached for the mth time
2513
2514 30. If pcregrep was called with the -l option and just a single file name, it
2515 output "<stdin>" if a match was found, instead of the file name.
2516
2517 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
2518 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
2519 pcre_exec(). This saves a malloc/free per call. The default value of
2520 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
2521 when configuring.
2522
2523 32. The default maximum size of a compiled pattern is 64K. There have been a
2524 few cases of people hitting this limit. The code now uses macros to handle the
2525 storing of links as offsets within the compiled pattern. It defaults to 2-byte
2526 links, but this can be changed to 3 or 4 bytes by --with-link-size when
2527 configuring. Tests 2 and 5 work only with 2-byte links because they output
2528 debugging information about compiled patterns.
2529
2530 33. Internal code re-arrangements:
2531
2532 (a) Moved the debugging function for printing out a compiled regex into
2533 its own source file (printint.c) and used #include to pull it into
2534 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
2535 separate copies.
2536
2537 (b) Defined the list of op-code names for debugging as a macro in
2538 internal.h so that it is next to the definition of the opcodes.
2539
2540 (c) Defined a table of op-code lengths for simpler skipping along compiled
2541 code. This is again a macro in internal.h so that it is next to the
2542 definition of the opcodes.
2543
2544 34. Added support for recursive calls to individual subpatterns, along the
2545 lines of Robin Houston's patch (but implemented somewhat differently).
2546
2547 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
2548 allow it to read and process whole directories in Win32. This code was
2549 contributed by Lionel Fourquaux; it has not been tested by me.
2550
2551 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
2552 used to name a group. Names consist of alphanumerics and underscores, and must
2553 be unique. Back references use the syntax (?P=name) and recursive calls use
2554 (?P>name) which is a PCRE extension to the Python extension. Groups still have
2555 numbers. The function pcre_fullinfo() can be used after compilation to extract
2556 a name/number map. There are three relevant calls:
2557
2558 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
2559 PCRE_INFO_NAMECOUNT yields the number of entries
2560 PCRE_INFO_NAMETABLE yields a pointer to the map.
2561
2562 The map is a vector of fixed-size entries. The size of each entry depends on
2563 the length of the longest name used. The first two bytes of each entry are the
2564 group number, most significant byte first. There follows the corresponding
2565 name, zero terminated. The names are in alphabetical order.
2566
2567 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
2568 case instead of 255. Making it the same both with and without UTF-8 support
2569 means that the same test output works with both.
2570
2571 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
2572 calling malloc() with a zero argument.
2573
2574 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
2575 optimization. I've improved things by keeping a bitmap of backreferences with
2576 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
2577 fact referenced, the optimization can be applied. It is unlikely that a
2578 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
2579 the match to follow \n) will appear inside brackets with a number greater than
2580 31, but if it does, any back reference > 31 suppresses the optimization.
2581
2582 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
2583 of disabling numbered capturing parentheses. Any opening parenthesis that is
2584 not followed by ? behaves as if it were followed by ?: but named parentheses
2585 can still be used for capturing (and they will acquire numbers in the usual
2586 way).
2587
2588 41. Redesigned the return codes from the match() function into yes/no/error so
2589 that errors can be passed back from deep inside the nested calls. A malloc
2590 failure while inside a recursive subpattern call now causes the
2591 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
2592
2593 42. It is now possible to set a limit on the number of times the match()
2594 function is called in a call to pcre_exec(). This facility makes it possible to
2595 limit the amount of recursion and backtracking, though not in a directly
2596 obvious way, because the match() function is used in a number of different
2597 circumstances. The count starts from zero for each position in the subject
2598 string (for non-anchored patterns). The default limit is, for compatibility, a
2599 large number, namely 10 000 000. You can change this in two ways:
2600
2601 (a) When configuring PCRE before making, you can use --with-match-limit=n
2602 to set a default value for the compiled library.
2603
2604 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
2605 a different value is set. See 45 below.
2606
2607 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
2608
2609 43. Added a new function pcre_config(int, void *) to enable run-time extraction
2610 of things that can be changed at compile time. The first argument specifies
2611 what is wanted and the second points to where the information is to be placed.
2612 The current list of available information is:
2613
2614 PCRE_CONFIG_UTF8
2615
2616 The output is an integer that is set to one if UTF-8 support is available;
2617 otherwise it is set to zero.
2618
2619 PCRE_CONFIG_NEWLINE
2620
2621 The output is an integer that it set to the value of the code that is used for
2622 newline. It is either LF (10) or CR (13).
2623
2624 PCRE_CONFIG_LINK_SIZE
2625
2626 The output is an integer that contains the number of bytes used for internal
2627 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
2628
2629 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
2630
2631 The output is an integer that contains the threshold above which the POSIX
2632 interface uses malloc() for output vectors. See item 31 above.
2633
2634 PCRE_CONFIG_MATCH_LIMIT
2635
2636 The output is an unsigned integer that contains the default limit of the number
2637 of match() calls in a pcre_exec() execution. See 42 above.
2638
2639 44. pcretest has been upgraded by the addition of the -C option. This causes it
2640 to extract all the available output from the new pcre_config() function, and to
2641 output it. The program then exits immediately.
2642
2643 45. A need has arisen to pass over additional data with calls to pcre_exec() in
2644 order to support additional features. One way would have been to define
2645 pcre_exec2() (for example) with extra arguments, but this would not have been
2646 extensible, and would also have required all calls to the original function to
2647 be mapped to the new one. Instead, I have chosen to extend the mechanism that
2648 is used for passing in "extra" data from pcre_study().
2649
2650 The pcre_extra structure is now exposed and defined in pcre.h. It currently
2651 contains the following fields:
2652
2653 flags a bitmap indicating which of the following fields are set
2654 study_data opaque data from pcre_study()
2655 match_limit a way of specifying a limit on match() calls for a specific
2656 call to pcre_exec()
2657 callout_data data for callouts (see 49 below)
2658
2659 The flag bits are also defined in pcre.h, and are
2660
2661 PCRE_EXTRA_STUDY_DATA
2662 PCRE_EXTRA_MATCH_LIMIT
2663 PCRE_EXTRA_CALLOUT_DATA
2664
2665 The pcre_study() function now returns one of these new pcre_extra blocks, with
2666 the actual study data pointed to by the study_data field, and the
2667 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
2668 before. That is, this change is entirely upwards-compatible and requires no
2669 change to existing code.
2670
2671 If you want to pass in additional data to pcre_exec(), you can either place it
2672 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
2673 block.
2674
2675 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
2676 data string contains the escape sequence \M, pcretest calls pcre_exec() several
2677 times with different match limits, until it finds the minimum value needed for
2678 pcre_exec() to complete. The value is then output. This can be instructive; for
2679 most simple matches the number is quite small, but for pathological cases it
2680 gets very large very quickly.
2681
2682 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
2683 returns the size of the data block pointed to by the study_data field in a
2684 pcre_extra block, that is, the value that was passed as the argument to
2685 pcre_malloc() when PCRE was getting memory in which to place the information
2686 created by pcre_study(). The fourth argument should point to a size_t variable.
2687 pcretest has been extended so that this information is shown after a successful
2688 pcre_study() call when information about the compiled regex is being displayed.
2689
2690 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
2691 because what follows is always an absolute path. (Later: it turns out that this
2692 is more than cosmetic for MinGW, because it doesn't like empty path
2693 components.)
2694
2695 49. Some changes have been made to the callout feature (see 28 above):
2696
2697 (i) A callout function now has three choices for what it returns:
2698
2699 0 => success, carry on matching
2700 > 0 => failure at this point, but backtrack if possible
2701 < 0 => serious error, return this value from pcre_exec()
2702
2703 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
2704 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
2705 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
2706 use by callout functions. It will never be used by PCRE itself.
2707
2708 (ii) The pcre_extra structure (see 45 above) has a void * field called
2709 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
2710 pcre_callout_block structure has a field of the same name. The contents of
2711 the field passed in the pcre_extra structure are passed to the callout
2712 function in the corresponding field in the callout block. This makes it
2713 easier to use the same callout-containing regex from multiple threads. For
2714 testing, the pcretest program has a new data escape
2715
2716 \C*n pass the number n (may be negative) as callout_data
2717
2718 If the callout function in pcretest receives a non-zero value as
2719 callout_data, it returns that value.
2720
2721 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
2722 there were some redundant $(CFLAGS) in commands that are now specified as
2723 $(LINK), which already includes $(CFLAGS).
2724
2725 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
2726 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
2727 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
2728 one-byte characters throughout. Note that case-insensitive matching applies
2729 only to characters whose values are less than 256. PCRE doesn't support the
2730 notion of cases for higher-valued characters.
2731
2732 (i) A character class whose characters are all within 0-255 is handled as
2733 a bit map, and the map is inverted for negative classes. Previously, a
2734 character > 255 always failed to match such a class; however it should
2735 match if the class was a negative one (e.g. [^ab]). This has been fixed.
2736
2737 (ii) A negated character class with a single character < 255 is coded as
2738 "not this character" (OP_NOT). This wasn't working properly when the test
2739 character was multibyte, either singly or repeated.
2740
2741 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
2742 mode, for example: \x{100}{2,3}.
2743
2744 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
2745 singly or repeated) now correctly test multibyte characters. However,
2746 PCRE doesn't recognize any characters with values greater than 255 as
2747 digits, spaces, or word characters. Such characters always match \D, \S,
2748 and \W, and never match \d, \s, or \w.
2749
2750 (v) Classes may now contain characters and character ranges with values
2751 greater than 255. For example: [ab\x{100}-\x{400}].
2752
2753 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
2754 PCRE in UTF-8 mode.
2755
2756 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
2757 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
2758 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
2759 value.)
2760
2761 53. The single man page has become too large. I have therefore split it up into
2762 a number of separate man pages. These also give rise to individual HTML pages;
2763 these are now put in a separate directory, and there is an index.html page that
2764 lists them all. Some hyperlinking between the pages has been installed.
2765
2766 54. Added convenience functions for handling named capturing parentheses.
2767
2768 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
2769 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
2770 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
2771 are faulted.
2772
2773 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
2774 calling configure. These values are used when compiling the dftables.c program
2775 which is run to generate the source of the default character tables. They
2776 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
2777 you will need to set these values.
2778
2779 57. Updated the building process for Windows DLL, as provided by Fred Cox.
2780
2781
2782 Version 3.9 02-Jan-02
2783 ---------------------
2784
2785 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
2786
2787 2. If --disable-static was given, the building process failed when trying to
2788 build pcretest and pcregrep. (For some reason it was using libtool to compile
2789 them, which is not right, as they aren't part of the library.)
2790
2791
2792 Version 3.8 18-Dec-01
2793 ---------------------
2794
2795 1. The experimental UTF-8 code was completely screwed up. It was packing the
2796 bytes in the wrong order. How dumb can you get?
2797
2798
2799 Version 3.7 29-Oct-01
2800 ---------------------
2801
2802 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
2803 This caused pcretest, when used on the test data, to segfault. Unfortunately,
2804 this didn't happen under Solaris 8, where I normally test things.
2805
2806 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
2807 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
2808 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
2809 this fix an hour or so after the initial 3.7 release.)
2810
2811
2812 Version 3.6 23-Oct-01
2813 ---------------------
2814
2815 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
2816 offsets passed as NULL with zero offset count.
2817
2818 2. The config.guess and config.sub files had not been updated when I moved to
2819 the latest autoconf.
2820
2821
2822 Version 3.5 15-Aug-01
2823 ---------------------
2824
2825 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
2826 had been forgotten.
2827
2828 2. By using declared but undefined structures, we can avoid using "void"
2829 definitions in pcre.h while keeping the internal definitions of the structures
2830 private.
2831
2832 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
2833 user point of view, this means that both static and shared libraries are built
2834 by default, but this can be individually controlled. More of the work of
2835 handling this static/shared cases is now inside libtool instead of PCRE's make
2836 file.
2837
2838 4. The pcretest utility is now installed along with pcregrep because it is
2839 useful for users (to test regexs) and by doing this, it automatically gets
2840 relinked by libtool. The documentation has been turned into a man page, so
2841 there are now .1, .txt, and .html versions in /doc.
2842
2843 5. Upgrades to pcregrep:
2844 (i) Added long-form option names like gnu grep.
2845 (ii) Added --help to list all options with an explanatory phrase.
2846 (iii) Added -r, --recursive to recurse into sub-directories.
2847 (iv) Added -f, --file to read patterns from a file.
2848
2849 6. pcre_exec() was referring to its "code" argument before testing that
2850 argument for NULL (and giving an error if it was NULL).
2851
2852 7. Upgraded Makefile.in to allow for compiling in a different directory from
2853 the source directory.
2854
2855 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
2856 options bits, the pointer it was passed was to an int instead of to an unsigned
2857 long int. This mattered only on 64-bit systems.
2858
2859 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
2860 generated) instead of pcre.in, which it its source. Also made the same change
2861 in several of the .c files.
2862
2863 10. A new release of gcc defines printf() as a macro, which broke pcretest
2864 because it had an ifdef in the middle of a string argument for printf(). Fixed
2865 by using separate calls to printf().
2866
2867 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
2868 script, to force use of CR or LF instead of \n in the source. On non-Unix
2869 systems, the value can be set in config.h.
2870
2871 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
2872 absolute limit. Changed the text of the error message to make this clear, and
2873 likewise updated the man page.
2874
2875 13. The limit of 99 on the number of capturing subpatterns has been removed.
2876 The new limit is 65535, which I hope will not be a "real" limit.
2877
2878
2879 Version 3.4 22-Aug-00
2880 ---------------------
2881
2882 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
2883
2884 2. Diagnose condition (?(0) as an error instead of crashing on matching.
2885
2886
2887 Version 3.3 01-Aug-00
2888 ---------------------
2889
2890 1. If an octal character was given, but the value was greater than \377, it
2891 was not getting masked to the least significant bits, as documented. This could
2892 lead to crashes in some systems.
2893
2894 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
2895 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
2896
2897 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
2898 These just pass their arguments on to (pcre_free)(), but they are provided
2899 because some uses of PCRE bind it to non-C systems that can call its functions,
2900 but cannot call free() or pcre_free() directly.
2901
2902 4. Add "make test" as a synonym for "make check". Corrected some comments in
2903 the Makefile.
2904
2905 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
2906 Makefile.
2907
2908 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
2909 command called pgrep for grepping around the active processes.
2910
2911 7. Added the beginnings of support for UTF-8 character strings.
2912
2913 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
2914 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
2915 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
2916 out for the ar command.)
2917
2918
2919 Version 3.2 12-May-00
2920 ---------------------
2921
2922 This is purely a bug fixing release.
2923
2924 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
2925 of ZA. This was just one example of several cases that could provoke this bug,
2926 which was introduced by change 9 of version 2.00. The code for breaking
2927 infinite loops after an iteration that matches an empty string was't working
2928 correctly.
2929
2930 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
2931 when matched against abbab (for example). After matching an empty string, it
2932 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
2933 caused it to match further down the string than it should.
2934
2935 3. The code contained an inclusion of sys/types.h. It isn't clear why this
2936 was there because it doesn't seem to be needed, and it causes trouble on some
2937 systems, as it is not a Standard C header. It has been removed.
2938
2939 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
2940 were reported on the Macintosh. The changes were from
2941
2942 while ((c = *(++ptr)) != 0 && c != '\n');
2943 to
2944 while ((c = *(++ptr)) != 0 && c != '\n') ;
2945
2946 Totally extraordinary, but if that's what it takes...
2947
2948 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
2949 available. Added HAVE_BCOPY and an autoconf test for it; if neither
2950 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
2951 assumes the way PCRE uses memmove() (always moving upwards).
2952
2953 6. PCRE is being used in one environment where strchr() is not available. There
2954 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
2955 faster code anyway.
2956
2957
2958 Version 3.1 09-Feb-00
2959 ---------------------
2960
2961 The only change in this release is the fixing of some bugs in Makefile.in for
2962 the "install" target:
2963
2964 (1) It was failing to install pcreposix.h.
2965
2966 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
2967
2968
2969 Version 3.0 01-Feb-00
2970 ---------------------
2971
2972 1. Add support for the /+ modifier to perltest (to output $` like it does in
2973 pcretest).
2974
2975 2. Add support for the /g modifier to perltest.
2976
2977 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
2978 matches null strings.
2979
2980 4. Fix perltest so that it doesn't do unwanted things when fed an empty
2981 pattern. Perl treats empty patterns specially - it reuses the most recent
2982 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
2983 effect.
2984
2985 5. The POSIX interface was broken in that it was just handing over the POSIX
2986 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
2987 required a bigger vector, with some working space on the end. This means that
2988 the POSIX wrapper now has to get and free some memory, and copy the results.
2989
2990 6. Added some simple autoconf support, placing the test data and the
2991 documentation in separate directories, re-organizing some of the
2992 information files, and making it build pcre-config (a GNU standard). Also added
2993 libtool support for building PCRE as a shared library, which is now the
2994 default.
2995
2996 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
2997 09 are not valid octal constants. Single digits will be used for minor values
2998 less than 10.
2999
3000 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
3001 existing programs that set these in the POSIX interface can use PCRE without
3002 modification.
3003
3004 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
3005 return all that pcre_info() returns, plus additional data. The pcre_info()
3006 function is retained for compatibility, but is considered to be obsolete.
3007
3008 10. Added experimental recursion feature (?R) to handle one common case that
3009 Perl 5.6 will be able to do with (?p{...}).
3010
3011 11. Added support for POSIX character classes like [:alpha:], which Perl is
3012 adopting.
3013
3014
3015 Version 2.08 31-Aug-99
3016 ----------------------
3017
3018 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
3019 trying to match at the startoffset position, but instead was moving forward to
3020 the next newline as if a previous match had failed.
3021
3022 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
3023 and could get into a loop if a null string was matched other than at the start
3024 of the subject.
3025
3026 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
3027 be distinguished at compile time, and for completeness also added PCRE_DATE.
3028
3029 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
3030 in GnuWin32 environments.
3031
3032
3033 Version 2.07 29-Jul-99
3034 ----------------------
3035
3036 1. The documentation is now supplied in plain text form and HTML as well as in
3037 the form of man page sources.
3038
3039 2. C++ compilers don't like assigning (void *) values to other pointer types.
3040 In particular this affects malloc(). Although there is no problem in Standard
3041 C, I've put in casts to keep C++ compilers happy.
3042
3043 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
3044 should be (const char *).
3045
3046 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
3047 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
3048 However, I haven't made this a standard facility. The documentation doesn't
3049 mention it, and the Makefile doesn't support it.
3050
3051 5. The Makefile now contains an "install" target, with editable destinations at
3052 the top of the file. The pcretest program is not installed.
3053
3054 6. pgrep -V now gives the PCRE version number and date.
3055
3056 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
3057 causing the entire string to be ignored, instead of just the last character.
3058
3059 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
3060 non-matching string, it can take a very, very long time, even for strings of
3061 quite modest length, because of the nested recursion. PCRE now does better in
3062 some of these cases. It does this by remembering the last required literal
3063 character in the pattern, and pre-searching the subject to ensure it is present
3064 before running the real match. In other words, it applies a heuristic to detect
3065 some types of certain failure quickly, and in the above example, if presented
3066 with a string that has no trailing " it gives "no match" very quickly.
3067
3068 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
3069 other alternatives are tried instead.
3070
3071
3072 Version 2.06 09-Jun-99
3073 ----------------------
3074
3075 1. Change pcretest's output for amount of store used to show just the code
3076 space, because the remainder (the data block) varies in size between 32-bit and
3077 64-bit systems.
3078
3079 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
3080 start matching at. This allows lookbehinds to work when searching for multiple
3081 occurrences in a string.
3082
3083 3. Added additional options to pcretest for testing multiple occurrences:
3084
3085 /+ outputs the rest of the string that follows a match
3086 /g loops for multiple occurrences, using the new startoffset argument
3087 /G loops for multiple occurrences by passing an incremented pointer
3088
3089 4. PCRE wasn't doing the "first character" optimization for patterns starting
3090 with \b or \B, though it was doing it for other lookbehind assertions. That is,
3091 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
3092 the letter 'x'. On long subject strings, this gives a significant speed-up.
3093
3094
3095 Version 2.05 21-Apr-99
3096 ----------------------
3097
3098 1. Changed the type of magic_number from int to long int so that it works
3099 properly on 16-bit systems.
3100
3101 2. Fixed a bug which caused patterns starting with .* not to work correctly
3102 when the subject string contained newline characters. PCRE was assuming
3103 anchoring for such patterns in all cases, which is not correct because .* will
3104 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
3105 DOTALL is set at top level; otherwise it knows that patterns starting with .*
3106 must be retried after every newline in the subject.
3107
3108
3109 Version 2.04 18-Feb-99
3110 ----------------------
3111
3112 1. For parenthesized subpatterns with repeats whose minimum was zero, the
3113 computation of the store needed to hold the pattern was incorrect (too large).
3114 If such patterns were nested a few deep, this could multiply and become a real
3115 problem.
3116
3117 2. Added /M option to pcretest to show the memory requirement of a specific
3118 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
3119
3120 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
3121 compiled in such a way that the backtracking after subsequent failure was
3122 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
3123 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
3124
3125
3126 Version 2.03 02-Feb-99
3127 ----------------------
3128
3129 1. Fixed typo and small mistake in man page.
3130
3131 2. Added 4th condition (GPL supersedes if conflict) and created separate
3132 LICENCE file containing the conditions.
3133
3134 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
3135 Perl, that is the internal \ allows the delimiter to be included in the
3136 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
3137 the final delimiter, add \ to the end of the pattern (to test the error).
3138
3139 4. Added the convenience functions for extracting substrings after a successful
3140 match. Updated pcretest to make it able to test these functions.
3141
3142
3143 Version 2.02 14-Jan-99
3144 ----------------------
3145
3146 1. Initialized the working variables associated with each extraction so that
3147 their saving and restoring doesn't refer to uninitialized store.
3148
3149 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
3150 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
3151 fix the problem.
3152
3153 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
3154 calls, and wasn't printing the correct value for compiling calls. Increased the
3155 default value of LOOPREPEAT, and the number of significant figures in the
3156 times.
3157
3158 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
3159
3160 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
3161 a building problem on Windows NT with a FAT file system.
3162
3163
3164 Version 2.01 21-Oct-98
3165 ----------------------
3166
3167 1. Changed the API for pcre_compile() to allow for the provision of a pointer
3168 to character tables built by pcre_maketables() in the current locale. If NULL
3169 is passed, the default tables are used.
3170
3171
3172 Version 2.00 24-Sep-98
3173 ----------------------
3174
3175 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
3176 it any more.
3177
3178 2. Allow quantification of (?>) groups, and make it work correctly.
3179
3180 3. The first character computation wasn't working for (?>) groups.
3181
3182 4. Correct the implementation of \Z (it is permitted to match on the \n at the
3183 end of the subject) and add 5.005's \z, which really does match only at the
3184 very end of the subject.
3185
3186 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
3187
3188 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
3189 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
3190 localized options. All options to pcre_study() were also removed.
3191
3192 7. Add other new features from 5.005:
3193
3194 $(?<= positive lookbehind
3195 $(?<! negative lookbehind
3196 (?imsx-imsx) added the unsetting capability
3197 such a setting is global if at outer level; local otherwise
3198 (?imsx-imsx:) non-capturing groups with option setting
3199 (?(cond)re|re) conditional pattern matching
3200
3201 A backreference to itself in a repeated group matches the previous
3202 captured string.
3203
3204 8. General tidying up of studying (both automatic and via "study")
3205 consequential on the addition of new assertions.
3206
3207 9. As in 5.005, unlimited repeated groups that could match an empty substring
3208 are no longer faulted at compile time. Instead, the loop is forcibly broken at
3209 runtime if any iteration does actually match an empty substring.
3210
3211 10. Include the RunTest script in the distribution.
3212
3213 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
3214 discrepancies, some of which were old and were also with respect to 5.004. They
3215 have now been fixed.
3216
3217
3218 Version 1.09 28-Apr-98
3219 ----------------------
3220
3221 1. A negated single character class followed by a quantifier with a minimum
3222 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
3223 program crashes, or just wrong answers. This did not apply to negated classes
3224 containing more than one character, or to minima other than one.
3225
3226
3227 Version 1.08 27-Mar-98
3228 ----------------------
3229
3230 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
3231
3232 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
3233 latter must appear before anything that relies on it in the pattern.
3234
3235
3236 Version 1.07 16-Feb-98
3237 ----------------------
3238
3239 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
3240 repeat of a potentially empty string).
3241
3242
3243 Version 1.06 23-Jan-98
3244 ----------------------
3245
3246 1. Added Markus Oberhumer's little patches for C++.
3247
3248 2. Literal strings longer than 255 characters were broken.
3249
3250
3251 Version 1.05 23-Dec-97
3252 ----------------------
3253
3254 1. Negated character classes containing more than one character were failing if
3255 PCRE_CASELESS was set at run time.
3256
3257
3258 Version 1.04 19-Dec-97
3259 ----------------------
3260
3261 1. Corrected the man page, where some "const" qualifiers had been omitted.
3262
3263 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
3264 input syntax.
3265
3266 3. Fixed memory leak which occurred when a regex with back references was
3267 matched with an offsets vector that wasn't big enough. The temporary memory
3268 that is used in this case wasn't being freed if the match failed.
3269
3270 4. Tidied pcretest to ensure it frees memory that it gets.
3271
3272 5. Temporary memory was being obtained in the case where the passed offsets
3273 vector was exactly big enough.
3274
3275 6. Corrected definition of offsetof() from change 5 below.
3276
3277 7. I had screwed up change 6 below and broken the rules for the use of
3278 setjmp(). Now fixed.
3279
3280
3281 Version 1.03 18-Dec-97
3282 ----------------------
3283
3284 1. A erroneous regex with a missing opening parenthesis was correctly
3285 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
3286 on some systems.
3287
3288 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
3289 it was reported that one broken compiler failed on the former because "code" is
3290 also an independent variable.
3291
3292 3. The erroneous regex a[]b caused an array overrun reference.
3293
3294 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
3295 fail on data ending with that character. (It was going on too far, and checking
3296 the next character, typically a binary zero.) This was specific to the
3297 optimized code for single-character negative classes.
3298
3299 5. Added a contributed patch from the TIN world which does the following:
3300
3301 + Add an undef for memmove, in case the the system defines a macro for it.
3302
3303 + Add a definition of offsetof(), in case there isn't one. (I don't know
3304 the reason behind this - offsetof() is part of the ANSI standard - but
3305 it does no harm).
3306
3307 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
3308 most of the places where whitespace preceded '#'. I have given up and
3309 allowed the remaining 2 cases to be at the margin.
3310
3311 + Rename some variables in pcre to eliminate shadowing. This seems very
3312 pedantic, but does no harm, of course.
3313
3314 6. Moved the call to setjmp() into its own function, to get rid of warnings
3315 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
3316
3317 7. Constructs such as \d{8,} were compiling into the equivalent of
3318 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
3319 outcome, but in this particular case used more store than had been allocated,
3320 which caused the bug to be discovered because it threw up an internal error.
3321
3322 8. The debugging code in both pcre and pcretest for outputting the compiled
3323 form of a regex was going wrong in the case of back references followed by
3324 curly-bracketed repeats.
3325
3326
3327 Version 1.02 12-Dec-97
3328 ----------------------
3329
3330 1. Typos in pcre.3 and comments in the source fixed.
3331
3332 2. Applied a contributed patch to get rid of places where it used to remove
3333 'const' from variables, and fixed some signed/unsigned and uninitialized
3334 variable warnings.
3335
3336 3. Added the "runtest" target to Makefile.
3337
3338 4. Set default compiler flag to -O2 rather than just -O.
3339
3340
3341 Version 1.01 19-Nov-97
3342 ----------------------
3343
3344 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
3345 like /([ab]*)*/, that is, for classes with more than one character in them.
3346
3347 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
3348 as /((?>a*))*/ (a PCRE_EXTRA facility).
3349
3350
3351 Version 1.00 18-Nov-97
3352 ----------------------
3353
3354 1. Added compile-time macros to support systems such as SunOS4 which don't have
3355 memmove() or strerror() but have other things that can be used instead.
3356
3357 2. Arranged that "make clean" removes the executables.
3358
3359
3360 Version 0.99 27-Oct-97
3361 ----------------------
3362
3363 1. Fixed bug in code for optimizing classes with only one character. It was
3364 initializing a 32-byte map regardless, which could cause it to run off the end
3365 of the memory it had got.
3366
3367 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
3368
3369
3370 Version 0.98 22-Oct-97
3371 ----------------------
3372
3373 1. Fixed bug in code for handling temporary memory usage when there are more
3374 back references than supplied space in the ovector. This could cause segfaults.
3375
3376
3377 Version 0.97 21-Oct-97
3378 ----------------------
3379
3380 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
3381
3382 2. Optimized negated single characters not to use a bit map.
3383
3384 3. Brought error texts together as macro definitions; clarified some of them;
3385 fixed one that was wrong - it said "range out of order" when it meant "invalid
3386 escape sequence".
3387
3388 4. Changed some char * arguments to const char *.
3389
3390 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
3391
3392 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
3393 pcretest.
3394
3395
3396 Version 0.96 16-Oct-97
3397 ----------------------
3398
3399 1. Added a simple "pgrep" utility to the distribution.
3400
3401 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
3402 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
3403 where "ddd" means "one or more decimal digits".
3404
3405 3. Fixed serious bug. If a pattern had a back reference, but the call to
3406 pcre_exec() didn't supply a large enough ovector to record the related
3407 identifying subpattern, the match always failed. PCRE now remembers the number
3408 of the largest back reference, and gets some temporary memory in which to save
3409 the offsets during matching if necessary, in order to ensure that
3410 backreferences always work.
3411
3412 4. Increased the compatibility with Perl in a number of ways:
3413
3414 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
3415 to request this handling. The option can be set at compile or exec time.
3416
3417 (b) $ matches before a terminating newline by default; an option
3418 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
3419 mode). The option can be set at compile or exec time.
3420
3421 (c) The handling of \ followed by a digit other than 0 is now supposed to be
3422 the same as Perl's. If the decimal number it represents is less than 10
3423 or there aren't that many previous left capturing parentheses, an octal
3424 escape is read. Inside a character class, it's always an octal escape,
3425 even if it is a single digit.
3426
3427 (d) An escaped but undefined alphabetic character is taken as a literal,
3428 unless PCRE_EXTRA is set. Currently this just reserves the remaining
3429 escapes.
3430
3431 (e) {0} is now permitted. (The previous item is removed from the compiled
3432 pattern).
3433
3434 5. Changed all the names of code files so that the basic parts are no longer
3435 than 10 characters, and abolished the teeny "globals.c" file.
3436
3437 6. Changed the handling of character classes; they are now done with a 32-byte
3438 bit map always.
3439
3440 7. Added the -d and /D options to pcretest to make it possible to look at the
3441 internals of compilation without having to recompile pcre.
3442
3443
3444 Version 0.95 23-Sep-97
3445 ----------------------
3446
3447 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
3448 \x20 at the start of a run of normal characters. These were being treated as
3449 real characters, instead of the source characters being re-checked.
3450
3451
3452 Version 0.94 18-Sep-97
3453 ----------------------
3454
3455 1. The functions are now thread-safe, with the caveat that the global variables
3456 containing pointers to malloc() and free() or alternative functions are the
3457 same for all threads.
3458
3459 2. Get pcre_study() to generate a bitmap of initial characters for non-
3460 anchored patterns when this is possible, and use it if passed to pcre_exec().
3461
3462
3463 Version 0.93 15-Sep-97
3464 ----------------------
3465
3466 1. /(b)|(:+)/ was computing an incorrect first character.
3467
3468 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
3469 but not actually doing anything yet.
3470
3471 3. Treat "-" characters in classes that cannot be part of ranges as literals,
3472 as Perl does (e.g. [-az] or [az-]).
3473
3474 4. Set the anchored flag if a branch starts with .* or .*? because that tests
3475 all possible positions.
3476
3477 5. Split up into different modules to avoid including unneeded functions in a
3478 compiled binary. However, compile and exec are still in one module. The "study"
3479 function is split off.
3480
3481 6. The character tables are now in a separate module whose source is generated
3482 by an auxiliary program - but can then be edited by hand if required. There are
3483 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
3484 toupper() in the code.
3485
3486 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
3487 make them global. Abolish the function for setting them, as the caller can now
3488 set them directly.
3489
3490
3491 Version 0.92 11-Sep-97
3492 ----------------------
3493
3494 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
3495 (e.g. /a{1,3}/) was broken (I mis-optimized it).
3496
3497 2. Caseless matching was not working in character classes if the characters in
3498 the pattern were in upper case.
3499
3500 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
3501
3502 4. Make PCRE_ANCHORED public and accept as a compile option.
3503
3504 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
3505 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
3506 pass them.
3507
3508 6. Give an error if bad option bits passed at compile or run time.
3509
3510 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
3511 pcretest to cause it to pass that flag.
3512
3513 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
3514 options, and the first character, if set.
3515
3516 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
3517
3518
3519 Version 0.91 10-Sep-97
3520 ----------------------
3521
3522 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
3523 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
3524
3525 2. PCRE was looping on encountering an indefinitely repeated back reference to
3526 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
3527 Perl does - treats the match as successful.
3528
3529 ****

Properties

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

webmaster@exim.org
ViewVC Help
Powered by ViewVC 1.1.12