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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 402 - (show annotations) (download)
Sat Mar 21 17:26:03 2009 UTC (5 years, 1 month ago) by ph10
File size: 164206 byte(s)
Add missing #ifdef SUPPORT_UTF8 round heapframe::Xcharptr.

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

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