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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 147 - (show annotations) (download)
Mon Apr 16 13:24:37 2007 UTC (7 years ago) by ph10
File size: 126426 byte(s)
Fix bad handling of data lines longer than 50K by pcretest.

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

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