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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 409 - (show annotations) (download)
Sat Mar 28 17:10:56 2009 UTC (5 years, 6 months ago) by ph10
File size: 165262 byte(s)
Further USPTR additions.

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

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