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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 276 - (show annotations) (download)
Wed Nov 21 15:39:20 2007 UTC (6 years, 8 months ago) by ph10
File size: 143980 byte(s)
Remove two redunant, never-reachable lines of code whose function had been 
moved.

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

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