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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 377 - (show annotations) (download)
Sun Mar 1 12:07:19 2009 UTC (5 years, 1 month ago) by ph10
File size: 161070 byte(s)
Use colour when requested with -o in pcregrep.

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

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