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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 233 - (show annotations) (download)
Tue Sep 11 11:26:11 2007 UTC (6 years, 11 months ago) by ph10
File size: 140002 byte(s)
Put back the definition of PCRE_ERROR_NULLWSLIMIT for compatibility, even 
though it is no longer used.

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

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