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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 98 - (show annotations) (download)
Mon Mar 5 15:55:02 2007 UTC (7 years, 1 month ago) by ph10
File size: 119130 byte(s)
This is a second preliminary update, after having done a test build and also a 
check of the maintenance programs. The version is now 7.1-RC1. In autogen.sh, 
autoheader is now before automake, and Tech.Notes (being now in the "maintain" 
directory) are no longer distributed.

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

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