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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 282 - (show annotations) (download)
Fri Dec 7 19:32:32 2007 UTC (6 years, 11 months ago) by ph10
File size: 145570 byte(s)
Fix non-diagnosis of (?=a)(?R) (positive lookaheads not skipped when checking 
for an empty match).

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

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