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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 104 - (show annotations) (download)
Tue Mar 6 16:25:15 2007 UTC (7 years, 1 month ago) by ph10
File size: 121060 byte(s)
Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
that is needed.  

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

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