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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 296 - (show annotations) (download)
Tue Jan 1 20:09:30 2008 UTC (6 years, 3 months ago) by ph10
File size: 148396 byte(s)
Remove the use of %n from pcregrep.

1 ChangeLog for PCRE
2 ------------------
3
4 Version 7.5 31-Dec-07
5 ---------------------
6
7 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
8 values in parens when parsing an RE using the C++ wrapper."
9
10 2. Negative specials like \S did not work in character classes in UTF-8 mode.
11 Characters greater than 255 were excluded from the class instead of being
12 included.
13
14 3. The same bug as (2) above applied to negated POSIX classes such as
15 [:^space:].
16
17 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
18 defined or documented. It seems to have been a typo for PCRE_STATIC, so
19 I have changed it.
20
21 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
22 first named subpattern) and a construct such as (?&a) would reference the
23 first named subpattern whose name started with "a" (in other words, the
24 length check was missing). Both these problems are fixed. "Subpattern name
25 expected" is now given for (?&) (a zero-length name), and this patch also
26 makes it give the same error for \k'' (previously it complained that that
27 was a reference to a non-existent subpattern).
28
29 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
30 this is right because (?- can be followed by option settings as well as by
31 digits. I have, however, made the messages clearer.
32
33 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
34 than the number used in the conditional) now cause a compile-time error.
35 This is actually not compatible with Perl, which accepts such patterns, but
36 treats the conditional as always being FALSE (as PCRE used to), but it
37 seems to me that giving a diagnostic is better.
38
39 8. Change "alphameric" to the more common word "alphanumeric" in comments
40 and messages.
41
42 9. Fix two occurrences of "backslash" in comments that should have been
43 "backspace".
44
45 10. Remove two redundant lines of code that can never be obeyed (their function
46 was moved elsewhere).
47
48 11. The program that makes PCRE's Unicode character property table had a bug
49 which caused it to generate incorrect table entries for sequences of
50 characters that have the same character type, but are in different scripts.
51 It amalgamated them into a single range, with the script of the first of
52 them. In other words, some characters were in the wrong script. There were
53 thirteen such cases, affecting characters in the following ranges:
54
55 U+002b0 - U+002c1
56 U+0060c - U+0060d
57 U+0061e - U+00612
58 U+0064b - U+0065e
59 U+0074d - U+0076d
60 U+01800 - U+01805
61 U+01d00 - U+01d77
62 U+01d9b - U+01dbf
63 U+0200b - U+0200f
64 U+030fc - U+030fe
65 U+03260 - U+0327f
66 U+0fb46 - U+0fbb1
67 U+10450 - U+1049d
68
69 12. The -o option (show only the matching part of a line) for pcregrep was not
70 compatible with GNU grep in that, if there was more than one match in a
71 line, it showed only the first of them. It now behaves in the same way as
72 GNU grep.
73
74 13. If the -o and -v options were combined for pcregrep, it printed a blank
75 line for every non-matching line. GNU grep prints nothing, and pcregrep now
76 does the same. The return code can be used to tell if there were any
77 non-matching lines.
78
79 14. Added --file-offsets and --line-offsets to pcregrep.
80
81 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
82 infinitely looping recursion. The bug was that positive lookaheads were not
83 being skipped when checking for a possible empty match (negative lookaheads
84 and both kinds of lookbehind were skipped).
85
86 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
87 inclusion of <windows.h> to before rather than after the definition of
88 INVALID_FILE_ATTRIBUTES (patch from David Byron).
89
90 17. Specifying a possessive quantifier with a specific limit for a Unicode
91 character property caused pcre_compile() to compile bad code, which led at
92 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
93 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
94 caused the error; without that there was no problem.
95
96 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
97
98 19. Added --enable-pcretest-libreadline.
99
100 20. In pcrecpp.cc, the variable 'count' was incremented twice in
101 RE::GlobalReplace(). As a result, the number of replacements returned was
102 double what it should be. I have removed one of the increments.
103
104 21. Several CMake things:
105
106 (1) Arranged that, when cmake is used on Unix, the libraries end up with
107 the names libpcre and libpcreposix, not just pcre and pcreposix.
108
109 (2) The above change means that pcretest and pcregrep are now correctly
110 linked with the newly-built libraries, not previously installed ones.
111
112 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
113
114 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
115 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
116 UTF-8 newline character). The key issue is that the pattern starts .*;
117 this means that the match must be either at the beginning, or after a
118 newline. The bug was in the code for advancing after a failed match and
119 checking that the new position followed a newline. It was not taking
120 account of UTF-8 characters correctly.
121
122 23. PCRE was behaving differently from Perl in the way it recognized POSIX
123 character classes. PCRE was not treating the sequence [:...:] as a
124 character class unless the ... were all letters. Perl, however, seems to
125 allow any characters between [: and :], though of course it rejects as
126 unknown any "names" that contain non-letters, because all the known class
127 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
128 for example, whereas PCRE did not - it did not recognize a POSIX character
129 class. This seemed a bit dangerous, so the code has been changed to be
130 closer to Perl. The behaviour is not identical to Perl, because PCRE will
131 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
132 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
133 Perl does, and where it didn't before.
134
135 24. Rewrite so as to remove the single use of %n from pcregrep because in some
136 Windows environments %n is disabled by default.
137
138
139 Version 7.4 21-Sep-07
140 ---------------------
141
142 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
143 means that a class such as [\s] counted as "explicit reference to CR or
144 LF". That isn't really right - the whole point of the change was to try to
145 help when there was an actual mention of one of the two characters. So now
146 the change happens only if \r or \n (or a literal CR or LF) character is
147 encountered.
148
149 2. The 32-bit options word was also used for 6 internal flags, but the numbers
150 of both had grown to the point where there were only 3 bits left.
151 Fortunately, there was spare space in the data structure, and so I have
152 moved the internal flags into a new 16-bit field to free up more option
153 bits.
154
155 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
156 but did not set the internal JCHANGED flag - either of these is enough to
157 control the way the "get" function works - but the PCRE_INFO_JCHANGED
158 facility is supposed to tell if (?J) was ever used, so now (?J) at the
159 start sets both bits.
160
161 4. Added options (at build time, compile time, exec time) to change \R from
162 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
163
164 5. doc/pcresyntax.html was missing from the distribution.
165
166 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
167 compatibility, even though it is no longer used.
168
169 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
170 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
171 windows.h file is present (where different names are used). [This was
172 reversed later after testing - see 16 below.]
173
174 8. Changed all #include <config.h> to #include "config.h". There were also
175 some further <pcre.h> cases that I changed to "pcre.h".
176
177 9. When pcregrep was used with the --colour option, it missed the line ending
178 sequence off the lines that it output.
179
180 10. It was pointed out to me that arrays of string pointers cause lots of
181 relocations when a shared library is dynamically loaded. A technique of
182 using a single long string with a table of offsets can drastically reduce
183 these. I have refactored PCRE in four places to do this. The result is
184 dramatic:
185
186 Originally: 290
187 After changing UCP table: 187
188 After changing error message table: 43
189 After changing table of "verbs" 36
190 After changing table of Posix names 22
191
192 Thanks to the folks working on Gregex for glib for this insight.
193
194 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
195 unicode-properties was also set.
196
197 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
198
199 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
200 checked only for CRLF.
201
202 14. Added casts to pcretest.c to avoid compiler warnings.
203
204 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
205
206 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
207 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
208 entirely. This removes changes made in 7 above.
209
210 17. The CMake files have been updated, and there is now more information about
211 building with CMake in the NON-UNIX-USE document.
212
213
214 Version 7.3 28-Aug-07
215 ---------------------
216
217 1. In the rejigging of the build system that eventually resulted in 7.1, the
218 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
219 brackets there is not right, since it causes compilers to look for an
220 installed pcre.h, not the version that is in the source that is being
221 compiled (which of course may be different). I have changed it back to:
222
223 #include "pcre.h"
224
225 I have a vague recollection that the change was concerned with compiling in
226 different directories, but in the new build system, that is taken care of
227 by the VPATH setting the Makefile.
228
229 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
230 when the subject happened to end in the byte 0x85 (e.g. if the last
231 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
232 characters but of course it shouldn't be taken as a newline when it is part
233 of another character. The bug was that, for an unlimited repeat of . in
234 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
235 characters when looking for a newline.
236
237 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
238
239 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
240 in debug output.
241
242 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
243 long printing in the pcrecpp unittest when running under MinGW.
244
245 6. ESC_K was left out of the EBCDIC table.
246
247 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
248 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
249 limit also applies to "virtual nesting" when a pattern is recursive, and in
250 this case 1000 isn't so big. I have been able to remove this limit at the
251 expense of backing off one optimization in certain circumstances. Normally,
252 when pcre_exec() would call its internal match() function recursively and
253 immediately return the result unconditionally, it uses a "tail recursion"
254 feature to save stack. However, when a subpattern that can match an empty
255 string has an unlimited repetition quantifier, it no longer makes this
256 optimization. That gives it a stack frame in which to save the data for
257 checking that an empty string has been matched. Previously this was taken
258 from the 1000-entry workspace that had been reserved. So now there is no
259 explicit limit, but more stack is used.
260
261 8. Applied Daniel's patches to solve problems with the import/export magic
262 syntax that is required for Windows, and which was going wrong for the
263 pcreposix and pcrecpp parts of the library. These were overlooked when this
264 problem was solved for the main library.
265
266 9. There were some crude static tests to avoid integer overflow when computing
267 the size of patterns that contain repeated groups with explicit upper
268 limits. As the maximum quantifier is 65535, the maximum group length was
269 set at 30,000 so that the product of these two numbers did not overflow a
270 32-bit integer. However, it turns out that people want to use groups that
271 are longer than 30,000 bytes (though not repeat them that many times).
272 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
273 made it possible to implement the integer overflow checks in a much more
274 dynamic way, which I have now done. The artificial limitation on group
275 length has been removed - we now have only the limit on the total length of
276 the compiled pattern, which depends on the LINK_SIZE setting.
277
278 10. Fixed a bug in the documentation for get/copy named substring when
279 duplicate names are permitted. If none of the named substrings are set, the
280 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
281 empty string.
282
283 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
284 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
285 because the ] is interpreted as the first data character and the
286 terminating ] is not found. PCRE has been made compatible with Perl in this
287 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
288 cause memory overwriting.
289
290 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
291 string has been matched (to stop an infinite loop). It was not recognizing
292 a conditional subpattern that could match an empty string if that
293 subpattern was within another subpattern. For example, it looped when
294 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
295 condition was not nested. This bug has been fixed.
296
297 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
298 past the start of the subject in the presence of bytes with the top bit
299 set, for example "\x8aBCD".
300
301 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
302 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
303
304 14. Optimized (?!) to (*FAIL).
305
306 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
307 This restricts code points to be within the range 0 to 0x10FFFF, excluding
308 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
309 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
310 does: it's just the validity check that is more restrictive.
311
312 16. Inserted checks for integer overflows during escape sequence (backslash)
313 processing, and also fixed erroneous offset values for syntax errors during
314 backslash processing.
315
316 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
317 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
318
319 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
320 caused an overrun.
321
322 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
323 something other than just ASCII characters) inside a group that had an
324 unlimited repeat caused a loop at compile time (while checking to see
325 whether the group could match an empty string).
326
327 20. Debugging a pattern containing \p or \P could cause a crash. For example,
328 [\P{Any}] did so. (Error in the code for printing property names.)
329
330 21. An orphan \E inside a character class could cause a crash.
331
332 22. A repeated capturing bracket such as (A)? could cause a wild memory
333 reference during compilation.
334
335 23. There are several functions in pcre_compile() that scan along a compiled
336 expression for various reasons (e.g. to see if it's fixed length for look
337 behind). There were bugs in these functions when a repeated \p or \P was
338 present in the pattern. These operators have additional parameters compared
339 with \d, etc, and these were not being taken into account when moving along
340 the compiled data. Specifically:
341
342 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
343 length.
344
345 (b) An item such as \pL+ within a repeated group could cause crashes or
346 loops.
347
348 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
349 "reference to non-existent subpattern" error.
350
351 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
352
353 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
354 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
355
356 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
357
358 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
359 character were causing crashes (broken optimization).
360
361 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
362 \p or \P) caused a compile-time loop.
363
364 28. More problems have arisen in unanchored patterns when CRLF is a valid line
365 break. For example, the unstudied pattern [\r\n]A does not match the string
366 "\r\nA" because change 7.0/46 below moves the current point on by two
367 characters after failing to match at the start. However, the pattern \nA
368 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
369 the same is true. There doesn't seem any very clean way out of this, but
370 what I have chosen to do makes the common cases work: PCRE now takes note
371 of whether there can be an explicit match for \r or \n anywhere in the
372 pattern, and if so, 7.0/46 no longer applies. As part of this change,
373 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
374 pattern has explicit CR or LF references.
375
376 29. Added (*CR) etc for changing newline setting at start of pattern.
377
378
379 Version 7.2 19-Jun-07
380 ---------------------
381
382 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
383 which is apparently normally available under Windows.
384
385 2. Re-jig the pcregrep tests with different newline settings in an attempt
386 to make them independent of the local environment's newline setting.
387
388 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
389
390 4. Some of the "internals" tests were previously cut out when the link size
391 was not 2, because the output contained actual offsets. The recent new
392 "Z" feature of pcretest means that these can be cut out, making the tests
393 usable with all link sizes.
394
395 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
396 stack recursion. This gives a massive performance boost under BSD, but just
397 a small improvement under Linux. However, it saves one field in the frame
398 in all cases.
399
400 6. Added more features from the forthcoming Perl 5.10:
401
402 (a) (?-n) (where n is a string of digits) is a relative subroutine or
403 recursion call. It refers to the nth most recently opened parentheses.
404
405 (b) (?+n) is also a relative subroutine call; it refers to the nth next
406 to be opened parentheses.
407
408 (c) Conditions that refer to capturing parentheses can be specified
409 relatively, for example, (?(-2)... or (?(+3)...
410
411 (d) \K resets the start of the current match so that everything before
412 is not part of it.
413
414 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
415
416 (f) \g{name} is another synonym - part of Perl 5.10's unification of
417 reference syntax.
418
419 (g) (?| introduces a group in which the numbering of parentheses in each
420 alternative starts with the same number.
421
422 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
423
424 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
425 PCRE_INFO_JCHANGED.
426
427 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
428 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
429 for detecting groups that can match an empty string.
430
431 9. A pattern with a very large number of alternatives (more than several
432 hundred) was running out of internal workspace during the pre-compile
433 phase, where pcre_compile() figures out how much memory will be needed. A
434 bit of new cunning has reduced the workspace needed for groups with
435 alternatives. The 1000-alternative test pattern now uses 12 bytes of
436 workspace instead of running out of the 4096 that are available.
437
438 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
439
440 11. Applied patch from Google to remove an optimization that didn't quite work.
441 The report of the bug said:
442
443 pcrecpp::RE("a*").FullMatch("aaa") matches, while
444 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
445 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
446
447 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
448 it matched the wrong number of bytes.
449
450
451 Version 7.1 24-Apr-07
452 ---------------------
453
454 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
455 that is more "standard", making use of automake and other Autotools. There
456 is some re-arrangement of the files and adjustment of comments consequent
457 on this.
458
459 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
460 for recursive directory scanning broke on some systems because the files
461 are not scanned in any specific order and on different systems the order
462 was different. A call to "sort" has been inserted into RunGrepTest for the
463 approprate test as a short-term fix. In the longer term there may be an
464 alternative.
465
466 3. I had an email from Eric Raymond about problems translating some of PCRE's
467 man pages to HTML (despite the fact that I distribute HTML pages, some
468 people do their own conversions for various reasons). The problems
469 concerned the use of low-level troff macros .br and .in. I have therefore
470 removed all such uses from the man pages (some were redundant, some could
471 be replaced by .nf/.fi pairs). The 132html script that I use to generate
472 HTML has been updated to handle .nf/.fi and to complain if it encounters
473 .br or .in.
474
475 4. Updated comments in configure.ac that get placed in config.h.in and also
476 arranged for config.h to be included in the distribution, with the name
477 config.h.generic, for the benefit of those who have to compile without
478 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
479
480 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
481 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
482 makevp.bat for the current PCRE, using the additional files
483 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
484
485 6. A Windows user reported a minor discrepancy with test 2, which turned out
486 to be caused by a trailing space on an input line that had got lost in his
487 copy. The trailing space was an accident, so I've just removed it.
488
489 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
490 that is needed.
491
492 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
493 as "const" (a) because they are and (b) because it helps the PHP
494 maintainers who have recently made a script to detect big data structures
495 in the php code that should be moved to the .rodata section. I remembered
496 to update Builducptable as well, so it won't revert if ucptable.h is ever
497 re-created.
498
499 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
500 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
501 order to be able to cut out the UTF-8 tables in the latter when UTF-8
502 support is not required. This saves 1.5-2K of code, which is important in
503 some applications.
504
505 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
506 so as not to refer to the tables, even though these functions will never be
507 called when UTF-8 support is disabled. Otherwise there are problems with a
508 shared library.
509
510 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
511
512 (a) It was defining its arguments as char * instead of void *.
513
514 (b) It was assuming that all moves were upwards in memory; this was true
515 a long time ago when I wrote it, but is no longer the case.
516
517 The emulated memove() is provided for those environments that have neither
518 memmove() nor bcopy(). I didn't think anyone used it these days, but that
519 is clearly not the case, as these two bugs were recently reported.
520
521 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
522 and Detrail to create the HTML documentation, the .txt form of the man
523 pages, and it removes trailing spaces from listed files. It also creates
524 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
525 case, it wraps all the #defines with #ifndefs. This script should be run
526 before "make dist".
527
528 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
529 with Unicode property support.
530
531 (a) For a maximizing quantifier, if the two different cases of the
532 character were of different lengths in their UTF-8 codings (there are
533 some cases like this - I found 11), and the matching function had to
534 back up over a mixture of the two cases, it incorrectly assumed they
535 were both the same length.
536
537 (b) When PCRE was configured to use the heap rather than the stack for
538 recursion during matching, it was not correctly preserving the data for
539 the other case of a UTF-8 character when checking ahead for a match
540 while processing a minimizing repeat. If the check also involved
541 matching a wide character, but failed, corruption could cause an
542 erroneous result when trying to check for a repeat of the original
543 character.
544
545 13. Some tidying changes to the testing mechanism:
546
547 (a) The RunTest script now detects the internal link size and whether there
548 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
549 values substituted by "configure". (The RunGrepTest script already did
550 this for UTF-8.) The configure.ac script no longer substitutes the
551 relevant variables.
552
553 (b) The debugging options /B and /D in pcretest show the compiled bytecode
554 with length and offset values. This means that the output is different
555 for different internal link sizes. Test 2 is skipped for link sizes
556 other than 2 because of this, bypassing the problem. Unfortunately,
557 there was also a test in test 3 (the locale tests) that used /B and
558 failed for link sizes other than 2. Rather than cut the whole test out,
559 I have added a new /Z option to pcretest that replaces the length and
560 offset values with spaces. This is now used to make test 3 independent
561 of link size. (Test 2 will be tidied up later.)
562
563 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
564 segmentation fault instead of returning the appropriate error message.
565
566 15. In multiline mode when the newline sequence was set to "any", the pattern
567 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
568 This doesn't seem right; it now treats the CRLF combination as the line
569 ending, and so does not match in that case. It's only a pattern such as ^$
570 that would hit this one: something like ^ABC$ would have failed after \r
571 and then tried again after \r\n.
572
573 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
574 in an attempt to make files that differ only in their line terminators
575 compare equal. This works on Linux.
576
577 17. Under certain error circumstances pcregrep might try to free random memory
578 as it exited. This is now fixed, thanks to valgrind.
579
580 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
581 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
582 was because its rules for how to advance for /g after matching an empty
583 string at the end of a line did not allow for this case. They now check for
584 it specially.
585
586 20. pcretest is supposed to handle patterns and data of any length, by
587 extending its buffers when necessary. It was getting this wrong when the
588 buffer for a data line had to be extended.
589
590 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
591 CRLF as a newline sequence.
592
593 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
594 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
595 I have nevertheless tidied it up.
596
597 23. Added some casts to kill warnings from HP-UX ia64 compiler.
598
599 24. Added a man page for pcre-config.
600
601
602 Version 7.0 19-Dec-06
603 ---------------------
604
605 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
606 moving to gcc 4.1.1.
607
608 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
609 sys/time.h, which is documented as needed for this function. It doesn't
610 seem to matter on Linux, but it showed up on some releases of OS X.
611
612 3. It seems that there are systems where bytes whose values are greater than
613 127 match isprint() in the "C" locale. The "C" locale should be the
614 default when a C program starts up. In most systems, only ASCII printing
615 characters match isprint(). This difference caused the output from pcretest
616 to vary, making some of the tests fail. I have changed pcretest so that:
617
618 (a) When it is outputting text in the compiled version of a pattern, bytes
619 other than 32-126 are always shown as hex escapes.
620
621 (b) When it is outputting text that is a matched part of a subject string,
622 it does the same, unless a different locale has been set for the match
623 (using the /L modifier). In this case, it uses isprint() to decide.
624
625 4. Fixed a major bug that caused incorrect computation of the amount of memory
626 required for a compiled pattern when options that changed within the
627 pattern affected the logic of the preliminary scan that determines the
628 length. The relevant options are -x, and -i in UTF-8 mode. The result was
629 that the computed length was too small. The symptoms of this bug were
630 either the PCRE error "internal error: code overflow" from pcre_compile(),
631 or a glibc crash with a message such as "pcretest: free(): invalid next
632 size (fast)". Examples of patterns that provoked this bug (shown in
633 pcretest format) are:
634
635 /(?-x: )/x
636 /(?x)(?-x: \s*#\s*)/
637 /((?i)[\x{c0}])/8
638 /(?i:[\x{c0}])/8
639
640 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
641 is now done differently.
642
643 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
644 wrapper classes; (b) implement a new function in the C++ scanner that is
645 more efficient than the old way of doing things because it avoids levels of
646 recursion in the regex matching; (c) add a paragraph to the documentation
647 for the FullMatch() function.
648
649 6. The escape sequence \n was being treated as whatever was defined as
650 "newline". Not only was this contrary to the documentation, which states
651 that \n is character 10 (hex 0A), but it also went horribly wrong when
652 "newline" was defined as CRLF. This has been fixed.
653
654 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
655 was being set to -1 for the "end of line" case (supposedly a value that no
656 character can have). Though this value is never used (the check for end of
657 line is "zero bytes in current character"), it caused compiler complaints.
658 I've changed it to 0xffffffff.
659
660 8. In pcre_version.c, the version string was being built by a sequence of
661 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
662 string (as it is for production releases) called a macro with an empty
663 argument. The C standard says the result of this is undefined. The gcc
664 compiler treats it as an empty string (which was what was wanted) but it is
665 reported that Visual C gives an error. The source has been hacked around to
666 avoid this problem.
667
668 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
669 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
670 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
671 of them did).
672
673 10. Originally, pcretest opened its input and output without "b"; then I was
674 told that "b" was needed in some environments, so it was added for release
675 5.0 to both the input and output. (It makes no difference on Unix-like
676 systems.) Later I was told that it is wrong for the input on Windows. I've
677 now abstracted the modes into two macros, to make it easier to fiddle with
678 them, and removed "b" from the input mode under Windows.
679
680 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
681
682 12. Added -help and --help to pcretest as an official way of being reminded
683 of the options.
684
685 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
686 and pcrecpp.cc because they annoy compilers at high warning levels.
687
688 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
689
690 15. Fixed an occurrence of == in configure.ac that should have been = (shell
691 scripts are not C programs :-) and which was not noticed because it works
692 on Linux.
693
694 16. pcretest is supposed to handle any length of pattern and data line (as one
695 line or as a continued sequence of lines) by extending its input buffer if
696 necessary. This feature was broken for very long pattern lines, leading to
697 a string of junk being passed to pcre_compile() if the pattern was longer
698 than about 50K.
699
700 17. I have done a major re-factoring of the way pcre_compile() computes the
701 amount of memory needed for a compiled pattern. Previously, there was code
702 that made a preliminary scan of the pattern in order to do this. That was
703 OK when PCRE was new, but as the facilities have expanded, it has become
704 harder and harder to keep it in step with the real compile phase, and there
705 have been a number of bugs (see for example, 4 above). I have now found a
706 cunning way of running the real compile function in a "fake" mode that
707 enables it to compute how much memory it would need, while actually only
708 ever using a few hundred bytes of working memory and without too many
709 tests of the mode. This should make future maintenance and development
710 easier. A side effect of this work is that the limit of 200 on the nesting
711 depth of parentheses has been removed (though this was never a serious
712 limitation, I suspect). However, there is a downside: pcre_compile() now
713 runs more slowly than before (30% or more, depending on the pattern). I
714 hope this isn't a big issue. There is no effect on runtime performance.
715
716 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
717 newline (only possible for the last line of a file) and it was a
718 pattern that set a locale (followed by /Lsomething), pcretest crashed.
719
720 19. Added additional timing features to pcretest. (1) The -tm option now times
721 matching only, not compiling. (2) Both -t and -tm can be followed, as a
722 separate command line item, by a number that specifies the number of
723 repeats to use when timing. The default is 50000; this gives better
724 precision, but takes uncomfortably long for very large patterns.
725
726 20. Extended pcre_study() to be more clever in cases where a branch of a
727 subpattern has no definite first character. For example, (a*|b*)[cd] would
728 previously give no result from pcre_study(). Now it recognizes that the
729 first character must be a, b, c, or d.
730
731 21. There was an incorrect error "recursive call could loop indefinitely" if
732 a subpattern (or the entire pattern) that was being tested for matching an
733 empty string contained only one non-empty item after a nested subpattern.
734 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
735 incorrectly, because the \d was being skipped in the check.
736
737 22. The pcretest program now has a new pattern option /B and a command line
738 option -b, which is equivalent to adding /B to every pattern. This causes
739 it to show the compiled bytecode, without the additional information that
740 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
741 is the same as /B/I).
742
743 23. A new optimization is now able automatically to treat some sequences such
744 as a*b as a*+b. More specifically, if something simple (such as a character
745 or a simple class like \d) has an unlimited quantifier, and is followed by
746 something that cannot possibly match the quantified thing, the quantifier
747 is automatically "possessified".
748
749 24. A recursive reference to a subpattern whose number was greater than 39
750 went wrong under certain circumstances in UTF-8 mode. This bug could also
751 have affected the operation of pcre_study().
752
753 25. Realized that a little bit of performance could be had by replacing
754 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
755
756 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
757
758 27. Possessive quantifiers such as a++ were previously implemented by turning
759 them into atomic groups such as ($>a+). Now they have their own opcodes,
760 which improves performance. This includes the automatically created ones
761 from 23 above.
762
763 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
764 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
765 the first matched character to be a colon. This applied both to named and
766 numbered groups.
767
768 29. The ucpinternal.h header file was missing its idempotency #ifdef.
769
770 30. I was sent a "project" file called libpcre.a.dev which I understand makes
771 building PCRE on Windows easier, so I have included it in the distribution.
772
773 31. There is now a check in pcretest against a ridiculously large number being
774 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
775 loop, the loop is abandoned.
776
777 32. Forward references to subpatterns in conditions such as (?(2)...) where
778 subpattern 2 is defined later cause pcre_compile() to search forwards in
779 the pattern for the relevant set of parentheses. This search went wrong
780 when there were unescaped parentheses in a character class, parentheses
781 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
782
783 33. "Subroutine" calls and backreferences were previously restricted to
784 referencing subpatterns earlier in the regex. This restriction has now
785 been removed.
786
787 34. Added a number of extra features that are going to be in Perl 5.10. On the
788 whole, these are just syntactic alternatives for features that PCRE had
789 previously implemented using the Python syntax or my own invention. The
790 other formats are all retained for compatibility.
791
792 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
793 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
794 also .NET compatible.
795
796 (b) A recursion or subroutine call to a named group can now be defined as
797 (?&name) as well as (?P>name).
798
799 (c) A backreference to a named group can now be defined as \k<name> or
800 \k'name' as well as (?P=name). The new forms, as well as being in Perl
801 5.10, are also .NET compatible.
802
803 (d) A conditional reference to a named group can now use the syntax
804 (?(<name>) or (?('name') as well as (?(name).
805
806 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
807 groups (named and numbered) that are never evaluated inline, but can be
808 called as "subroutines" from elsewhere. In effect, the DEFINE condition
809 is always false. There may be only one alternative in such a group.
810
811 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
812 as the simple (?(R). The condition is true only if the most recent
813 recursion is that of the given number or name. It does not search out
814 through the entire recursion stack.
815
816 (g) The escape \gN or \g{N} has been added, where N is a positive or
817 negative number, specifying an absolute or relative reference.
818
819 35. Tidied to get rid of some further signed/unsigned compiler warnings and
820 some "unreachable code" warnings.
821
822 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
823 things, this adds five new scripts.
824
825 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
826 There were also incompatibilities regarding the handling of \Q..\E inside
827 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
828 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
829
830 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
831 matches an empty string, and forcibly breaks the loop. There were bugs in
832 this code in non-simple cases. For a pattern such as ^(a()*)* matched
833 against aaaa the result was just "a" rather than "aaaa", for example. Two
834 separate and independent bugs (that affected different cases) have been
835 fixed.
836
837 39. Refactored the code to abolish the use of different opcodes for small
838 capturing bracket numbers. This is a tidy that I avoided doing when I
839 removed the limit on the number of capturing brackets for 3.5 back in 2001.
840 The new approach is not only tidier, it makes it possible to reduce the
841 memory needed to fix the previous bug (38).
842
843 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
844 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
845 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
846 mode.
847
848 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
849 report.
850
851 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
852 copy construction and assignment in the C++ wrapper.
853
854 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
855 couple of bugs that could have given wrong results in the "--newline=crlf"
856 case.
857
858 44. Added a number of casts and did some reorganization of signed/unsigned int
859 variables following suggestions from Dair Grant. Also renamed the variable
860 "this" as "item" because it is a C++ keyword.
861
862 45. Arranged for dftables to add
863
864 #include "pcre_internal.h"
865
866 to pcre_chartables.c because without it, gcc 4.x may remove the array
867 definition from the final binary if PCRE is built into a static library and
868 dead code stripping is activated.
869
870 46. For an unanchored pattern, if a match attempt fails at the start of a
871 newline sequence, and the newline setting is CRLF or ANY, and the next two
872 characters are CRLF, advance by two characters instead of one.
873
874
875 Version 6.7 04-Jul-06
876 ---------------------
877
878 1. In order to handle tests when input lines are enormously long, pcretest has
879 been re-factored so that it automatically extends its buffers when
880 necessary. The code is crude, but this _is_ just a test program. The
881 default size has been increased from 32K to 50K.
882
883 2. The code in pcre_study() was using the value of the re argument before
884 testing it for NULL. (Of course, in any sensible call of the function, it
885 won't be NULL.)
886
887 3. The memmove() emulation function in pcre_internal.h, which is used on
888 systems that lack both memmove() and bcopy() - that is, hardly ever -
889 was missing a "static" storage class specifier.
890
891 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
892 containing an extended class (one that cannot be represented by a bitmap
893 because it contains high-valued characters or Unicode property items, e.g.
894 [\pZ]). Almost always one would set UTF-8 mode when processing such a
895 pattern, but PCRE should not loop if you do not (it no longer does).
896 [Detail: two cases were found: (a) a repeated subpattern containing an
897 extended class; (b) a recursive reference to a subpattern that followed a
898 previous extended class. It wasn't skipping over the extended class
899 correctly when UTF-8 mode was not set.]
900
901 5. A negated single-character class was not being recognized as fixed-length
902 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
903 compile error "lookbehind assertion is not fixed length".
904
905 6. The RunPerlTest auxiliary script was showing an unexpected difference
906 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
907 write a Perl script that can interpret lines of an input file either as
908 byte characters or as UTF-8, which is what "perltest" was being required to
909 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
910 can't do is switch easily at run time between having the "use utf8;" pragma
911 or not. In the end, I fudged it by using the RunPerlTest script to insert
912 "use utf8;" explicitly for the UTF-8 tests.
913
914 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
915 the end of the subject string, contrary to the documentation and to what
916 Perl does. This was true of both matching functions. Now it matches only at
917 the start of the subject and immediately after *internal* newlines.
918
919 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
920 a pointer to an int instead of a pointer to an unsigned long int. This
921 caused problems on 64-bit systems.
922
923 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
924 instance of the 'standard' template library not being so standard".
925
926 10. There was no check on the number of named subpatterns nor the maximum
927 length of a subpattern name. The product of these values is used to compute
928 the size of the memory block for a compiled pattern. By supplying a very
929 long subpattern name and a large number of named subpatterns, the size
930 computation could be caused to overflow. This is now prevented by limiting
931 the length of names to 32 characters, and the number of named subpatterns
932 to 10,000.
933
934 11. Subpatterns that are repeated with specific counts have to be replicated in
935 the compiled pattern. The size of memory for this was computed from the
936 length of the subpattern and the repeat count. The latter is limited to
937 65535, but there was no limit on the former, meaning that integer overflow
938 could in principle occur. The compiled length of a repeated subpattern is
939 now limited to 30,000 bytes in order to prevent this.
940
941 12. Added the optional facility to have named substrings with the same name.
942
943 13. Added the ability to use a named substring as a condition, using the
944 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
945 are numbers (not recommended). Forward references are permitted.
946
947 14. Added forward references in named backreferences (if you see what I mean).
948
949 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
950 pattern could run off the end of the subject. For example, the pattern
951 "(?s)(.{1,5})"8 did this with the subject "ab".
952
953 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
954 PCRE_CASELESS was set when matching characters that were quantified with ?
955 or *.
956
957 17. A character class other than a single negated character that had a minimum
958 but no maximum quantifier - for example [ab]{6,} - was not handled
959 correctly by pce_dfa_exec(). It would match only one character.
960
961 18. A valid (though odd) pattern that looked like a POSIX character
962 class but used an invalid character after [ (for example [[,abc,]]) caused
963 pcre_compile() to give the error "Failed: internal error: code overflow" or
964 in some cases to crash with a glibc free() error. This could even happen if
965 the pattern terminated after [[ but there just happened to be a sequence of
966 letters, a binary zero, and a closing ] in the memory that followed.
967
968 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
969 over the years. Originally (before any Unicode support), just the bottom 8
970 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
971 output from "man perlunicode" includes this:
972
973 The regular expression compiler produces polymorphic opcodes. That
974 is, the pattern adapts to the data and automatically switches to
975 the Unicode character scheme when presented with Unicode data--or
976 instead uses a traditional byte scheme when presented with byte
977 data.
978
979 Sadly, a wide octal escape does not cause a switch, and in a string with
980 no other multibyte characters, these octal escapes are treated as before.
981 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
982 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
983 Unicode string.
984
985 I have not perpetrated such confusion in PCRE. Up till now, it took just
986 the bottom 8 bits, as in old Perl. I have now made octal escapes with
987 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
988 translate to the appropriate multibyte character.
989
990 29. Applied some refactoring to reduce the number of warnings from Microsoft
991 and Borland compilers. This has included removing the fudge introduced
992 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
993 a warning about an unused variable.
994
995 21. PCRE has not included VT (character 0x0b) in the set of whitespace
996 characters since release 4.0, because Perl (from release 5.004) does not.
997 [Or at least, is documented not to: some releases seem to be in conflict
998 with the documentation.] However, when a pattern was studied with
999 pcre_study() and all its branches started with \s, PCRE still included VT
1000 as a possible starting character. Of course, this did no harm; it just
1001 caused an unnecessary match attempt.
1002
1003 22. Removed a now-redundant internal flag bit that recorded the fact that case
1004 dependency changed within the pattern. This was once needed for "required
1005 byte" processing, but is no longer used. This recovers a now-scarce options
1006 bit. Also moved the least significant internal flag bit to the most-
1007 significant bit of the word, which was not previously used (hangover from
1008 the days when it was an int rather than a uint) to free up another bit for
1009 the future.
1010
1011 23. Added support for CRLF line endings as well as CR and LF. As well as the
1012 default being selectable at build time, it can now be changed at runtime
1013 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
1014 specify that it is scanning data with non-default line endings.
1015
1016 24. Changed the definition of CXXLINK to make it agree with the definition of
1017 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
1018
1019 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
1020 recursions. This makes a big different to stack usage for some patterns.
1021
1022 26. If a subpattern containing a named recursion or subroutine reference such
1023 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
1024 the space required for the compiled pattern went wrong and gave too small a
1025 value. Depending on the environment, this could lead to "Failed: internal
1026 error: code overflow at offset 49" or "glibc detected double free or
1027 corruption" errors.
1028
1029 27. Applied patches from Google (a) to support the new newline modes and (b) to
1030 advance over multibyte UTF-8 characters in GlobalReplace.
1031
1032 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
1033 difference for some implementation of PCRE in some Windows version.
1034
1035 29. Added some extra testing facilities to pcretest:
1036
1037 \q<number> in a data line sets the "match limit" value
1038 \Q<number> in a data line sets the "match recursion limt" value
1039 -S <number> sets the stack size, where <number> is in megabytes
1040
1041 The -S option isn't available for Windows.
1042
1043
1044 Version 6.6 06-Feb-06
1045 ---------------------
1046
1047 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
1048 in pcreposix.h. I have copied the definition from pcre.h.
1049
1050 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
1051 because pcre.h is no longer a built file.
1052
1053 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
1054 not normally included in the compiled code.
1055
1056
1057 Version 6.5 01-Feb-06
1058 ---------------------
1059
1060 1. When using the partial match feature with pcre_dfa_exec(), it was not
1061 anchoring the second and subsequent partial matches at the new starting
1062 point. This could lead to incorrect results. For example, with the pattern
1063 /1234/, partially matching against "123" and then "a4" gave a match.
1064
1065 2. Changes to pcregrep:
1066
1067 (a) All non-match returns from pcre_exec() were being treated as failures
1068 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
1069 error message is output. Some extra information is given for the
1070 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
1071 probably the only errors that are likely to be caused by users (by
1072 specifying a regex that has nested indefinite repeats, for instance).
1073 If there are more than 20 of these errors, pcregrep is abandoned.
1074
1075 (b) A binary zero was treated as data while matching, but terminated the
1076 output line if it was written out. This has been fixed: binary zeroes
1077 are now no different to any other data bytes.
1078
1079 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
1080 used to set a locale for matching. The --locale=xxxx long option has
1081 been added (no short equivalent) to specify a locale explicitly on the
1082 pcregrep command, overriding the environment variables.
1083
1084 (d) When -B was used with -n, some line numbers in the output were one less
1085 than they should have been.
1086
1087 (e) Added the -o (--only-matching) option.
1088
1089 (f) If -A or -C was used with -c (count only), some lines of context were
1090 accidentally printed for the final match.
1091
1092 (g) Added the -H (--with-filename) option.
1093
1094 (h) The combination of options -rh failed to suppress file names for files
1095 that were found from directory arguments.
1096
1097 (i) Added the -D (--devices) and -d (--directories) options.
1098
1099 (j) Added the -F (--fixed-strings) option.
1100
1101 (k) Allow "-" to be used as a file name for -f as well as for a data file.
1102
1103 (l) Added the --colo(u)r option.
1104
1105 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
1106 is not present by default.
1107
1108 3. A nasty bug was discovered in the handling of recursive patterns, that is,
1109 items such as (?R) or (?1), when the recursion could match a number of
1110 alternatives. If it matched one of the alternatives, but subsequently,
1111 outside the recursion, there was a failure, the code tried to back up into
1112 the recursion. However, because of the way PCRE is implemented, this is not
1113 possible, and the result was an incorrect result from the match.
1114
1115 In order to prevent this happening, the specification of recursion has
1116 been changed so that all such subpatterns are automatically treated as
1117 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
1118
1119 4. I had overlooked the fact that, in some locales, there are characters for
1120 which isalpha() is true but neither isupper() nor islower() are true. In
1121 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
1122 and ordfeminine) are like this. This affected the treatment of \w and \W
1123 when they appeared in character classes, but not when they appeared outside
1124 a character class. The bit map for "word" characters is now created
1125 separately from the results of isalnum() instead of just taking it from the
1126 upper, lower, and digit maps. (Plus the underscore character, of course.)
1127
1128 5. The above bug also affected the handling of POSIX character classes such as
1129 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
1130 permanent tables. Instead, the bit maps for such a class were previously
1131 created as the appropriate unions of the upper, lower, and digit bitmaps.
1132 Now they are created by subtraction from the [[:word:]] class, which has
1133 its own bitmap.
1134
1135 6. The [[:blank:]] character class matches horizontal, but not vertical space.
1136 It is created by subtracting the vertical space characters (\x09, \x0a,
1137 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
1138 subtraction was done in the overall bitmap for a character class, meaning
1139 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
1140 be recognized. This bug has been fixed.
1141
1142 7. Patches from the folks at Google:
1143
1144 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
1145 real life, but is still worth protecting against".
1146
1147 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
1148 regular expressions".
1149
1150 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
1151 have it.
1152
1153 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
1154 "configure" and the latter not, in order to fix a problem somebody had
1155 with compiling the Arg class on HP-UX.
1156
1157 (e) Improve the error-handling of the C++ wrapper a little bit.
1158
1159 (f) New tests for checking recursion limiting.
1160
1161 8. The pcre_memmove() function, which is used only if the environment does not
1162 have a standard memmove() function (and is therefore rarely compiled),
1163 contained two bugs: (a) use of int instead of size_t, and (b) it was not
1164 returning a result (though PCRE never actually uses the result).
1165
1166 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
1167 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
1168 returned instead of calling malloc() with an overflowing number that would
1169 most likely cause subsequent chaos.
1170
1171 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
1172
1173 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
1174 with this option is matched, the nmatch and pmatch options of regexec() are
1175 ignored.
1176
1177 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
1178 provided in case anyone wants to the the POSIX interface with UTF-8
1179 strings.
1180
1181 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
1182 C++ linking (needed for some HP-UX environments).
1183
1184 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
1185 (unused parameter) and in the pcre_printint() function (omitted "default"
1186 switch label when the default is to do nothing).
1187
1188 15. Added some code to make it possible, when PCRE is compiled as a C++
1189 library, to replace subject pointers for pcre_exec() with a smart pointer
1190 class, thus making it possible to process discontinuous strings.
1191
1192 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
1193 much the same function. They were added by different people who were trying
1194 to make PCRE easy to compile on non-Unix systems. It has been suggested
1195 that PCRE_EXPORT be abolished now that there is more automatic apparatus
1196 for compiling on Windows systems. I have therefore replaced it with
1197 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
1198 defaults to "extern" for C or "extern C" for C++, which works fine on
1199 Unix-like systems. It is now possible to override the value of PCRE_DATA_
1200 SCOPE with something explicit in config.h. In addition:
1201
1202 (a) pcreposix.h still had just "extern" instead of either of these macros;
1203 I have replaced it with PCRE_DATA_SCOPE.
1204
1205 (b) Functions such as _pcre_xclass(), which are internal to the library,
1206 but external in the C sense, all had PCRE_EXPORT in their definitions.
1207 This is apparently wrong for the Windows case, so I have removed it.
1208 (It makes no difference on Unix-like systems.)
1209
1210 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
1211 of recursive calls to match(). This is different to MATCH_LIMIT because
1212 that limits the total number of calls to match(), not all of which increase
1213 the depth of recursion. Limiting the recursion depth limits the amount of
1214 stack (or heap if NO_RECURSE is set) that is used. The default can be set
1215 when PCRE is compiled, and changed at run time. A patch from Google adds
1216 this functionality to the C++ interface.
1217
1218 18. Changes to the handling of Unicode character properties:
1219
1220 (a) Updated the table to Unicode 4.1.0.
1221
1222 (b) Recognize characters that are not in the table as "Cn" (undefined).
1223
1224 (c) I revised the way the table is implemented to a much improved format
1225 which includes recognition of ranges. It now supports the ranges that
1226 are defined in UnicodeData.txt, and it also amalgamates other
1227 characters into ranges. This has reduced the number of entries in the
1228 table from around 16,000 to around 3,000, thus reducing its size
1229 considerably. I realized I did not need to use a tree structure after
1230 all - a binary chop search is just as efficient. Having reduced the
1231 number of entries, I extended their size from 6 bytes to 8 bytes to
1232 allow for more data.
1233
1234 (d) Added support for Unicode script names via properties such as \p{Han}.
1235
1236 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
1237 matching that character.
1238
1239 20. When matching a repeated Unicode property with a minimum greater than zero,
1240 (for example \pL{2,}), PCRE could look past the end of the subject if it
1241 reached it while seeking the minimum number of characters. This could
1242 happen only if some of the characters were more than one byte long, because
1243 there is a check for at least the minimum number of bytes.
1244
1245 21. Refactored the implementation of \p and \P so as to be more general, to
1246 allow for more different types of property in future. This has changed the
1247 compiled form incompatibly. Anybody with saved compiled patterns that use
1248 \p or \P will have to recompile them.
1249
1250 22. Added "Any" and "L&" to the supported property types.
1251
1252 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
1253 but give a compile time error if the value is greater than 0xff.
1254
1255 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
1256 accidentally not being installed or uninstalled.
1257
1258 25. The pcre.h file was built from pcre.h.in, but the only changes that were
1259 made were to insert the current release number. This seemed silly, because
1260 it made things harder for people building PCRE on systems that don't run
1261 "configure". I have turned pcre.h into a distributed file, no longer built
1262 by "configure", with the version identification directly included. There is
1263 no longer a pcre.h.in file.
1264
1265 However, this change necessitated a change to the pcre-config script as
1266 well. It is built from pcre-config.in, and one of the substitutions was the
1267 release number. I have updated configure.ac so that ./configure now finds
1268 the release number by grepping pcre.h.
1269
1270 26. Added the ability to run the tests under valgrind.
1271
1272
1273 Version 6.4 05-Sep-05
1274 ---------------------
1275
1276 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
1277 "--" to be printed when multiple files were scanned, even when none of the
1278 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
1279 consider it to be a bug, and have restored the previous behaviour.
1280
1281 2. A couple of code tidies to get rid of compiler warnings.
1282
1283 3. The pcretest program used to cheat by referring to symbols in the library
1284 whose names begin with _pcre_. These are internal symbols that are not
1285 really supposed to be visible externally, and in some environments it is
1286 possible to suppress them. The cheating is now confined to including
1287 certain files from the library's source, which is a bit cleaner.
1288
1289 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
1290 file's purpose clearer.
1291
1292 5. Reorganized pcre_ucp_findchar().
1293
1294
1295 Version 6.3 15-Aug-05
1296 ---------------------
1297
1298 1. The file libpcre.pc.in did not have general read permission in the tarball.
1299
1300 2. There were some problems when building without C++ support:
1301
1302 (a) If C++ support was not built, "make install" and "make test" still
1303 tried to test it.
1304
1305 (b) There were problems when the value of CXX was explicitly set. Some
1306 changes have been made to try to fix these, and ...
1307
1308 (c) --disable-cpp can now be used to explicitly disable C++ support.
1309
1310 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
1311 backslash in a target when C++ was disabled. This confuses some
1312 versions of "make", apparently. Using an intermediate variable solves
1313 this. (Same for CPP_LOBJ.)
1314
1315 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
1316 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
1317 necessary on certain architectures.
1318
1319 4. Added a setting of -export-symbols-regex to the link command to remove
1320 those symbols that are exported in the C sense, but actually are local
1321 within the library, and not documented. Their names all begin with
1322 "_pcre_". This is not a perfect job, because (a) we have to except some
1323 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
1324 available (and never for static libraries). I have made a note to try to
1325 find a way round (a) in the future.
1326
1327
1328 Version 6.2 01-Aug-05
1329 ---------------------
1330
1331 1. There was no test for integer overflow of quantifier values. A construction
1332 such as {1111111111111111} would give undefined results. What is worse, if
1333 a minimum quantifier for a parenthesized subpattern overflowed and became
1334 negative, the calculation of the memory size went wrong. This could have
1335 led to memory overwriting.
1336
1337 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
1338
1339 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
1340 operating environments where this matters.
1341
1342 4. Applied Giuseppe Maxia's patch to add additional features for controlling
1343 PCRE options from within the C++ wrapper.
1344
1345 5. Named capturing subpatterns were not being correctly counted when a pattern
1346 was compiled. This caused two problems: (a) If there were more than 100
1347 such subpatterns, the calculation of the memory needed for the whole
1348 compiled pattern went wrong, leading to an overflow error. (b) Numerical
1349 back references of the form \12, where the number was greater than 9, were
1350 not recognized as back references, even though there were sufficient
1351 previous subpatterns.
1352
1353 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
1354 versions of gcc, e.g. 2.95.4.
1355
1356
1357 Version 6.1 21-Jun-05
1358 ---------------------
1359
1360 1. There was one reference to the variable "posix" in pcretest.c that was not
1361 surrounded by "#if !defined NOPOSIX".
1362
1363 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
1364 the cross-check on the old pcre_info() function, for the benefit of the
1365 cut-down version of PCRE that is currently imported into Exim.
1366
1367 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
1368 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
1369 patterns that start (?i) but I don't think that matters. The use of (?i) is
1370 just an example; this all applies to the other options as well.
1371
1372 4. Since libtool seems to echo the compile commands it is issuing, the output
1373 from "make" can be reduced a bit by putting "@" in front of each libtool
1374 compile command.
1375
1376 5. Patch from the folks at Google for configure.in to be a bit more thorough
1377 in checking for a suitable C++ installation before trying to compile the
1378 C++ stuff. This should fix a reported problem when a compiler was present,
1379 but no suitable headers.
1380
1381 6. The man pages all had just "PCRE" as their title. I have changed them to
1382 be the relevant file name. I have also arranged that these names are
1383 retained in the file doc/pcre.txt, which is a concatenation in text format
1384 of all the man pages except the little individual ones for each function.
1385
1386 7. The NON-UNIX-USE file had not been updated for the different set of source
1387 files that come with release 6. I also added a few comments about the C++
1388 wrapper.
1389
1390
1391 Version 6.0 07-Jun-05
1392 ---------------------
1393
1394 1. Some minor internal re-organization to help with my DFA experiments.
1395
1396 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
1397 didn't matter for the library itself when fully configured, but did matter
1398 when compiling without UCP support, or within Exim, where the ucp files are
1399 not imported.
1400
1401 3. Refactoring of the library code to split up the various functions into
1402 different source modules. The addition of the new DFA matching code (see
1403 below) to a single monolithic source would have made it really too
1404 unwieldy, quite apart from causing all the code to be include in a
1405 statically linked application, when only some functions are used. This is
1406 relevant even without the DFA addition now that patterns can be compiled in
1407 one application and matched in another.
1408
1409 The downside of splitting up is that there have to be some external
1410 functions and data tables that are used internally in different modules of
1411 the library but which are not part of the API. These have all had their
1412 names changed to start with "_pcre_" so that they are unlikely to clash
1413 with other external names.
1414
1415 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
1416 a different (DFA) algorithm. Although it is slower than the original
1417 function, it does have some advantages for certain types of matching
1418 problem.
1419
1420 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
1421 including restarting after a partial match.
1422
1423 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
1424 defined when compiling for Windows was sent to me. I have put it into the
1425 code, though I have no means of testing or verifying it.
1426
1427 7. Added the pcre_refcount() auxiliary function.
1428
1429 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
1430 match before or at the first newline in the subject string. In pcretest,
1431 the /f option on a pattern can be used to set this.
1432
1433 9. A repeated \w when used in UTF-8 mode with characters greater than 256
1434 would behave wrongly. This has been present in PCRE since release 4.0.
1435
1436 10. A number of changes to the pcregrep command:
1437
1438 (a) Refactored how -x works; insert ^(...)$ instead of setting
1439 PCRE_ANCHORED and checking the length, in preparation for adding
1440 something similar for -w.
1441
1442 (b) Added the -w (match as a word) option.
1443
1444 (c) Refactored the way lines are read and buffered so as to have more
1445 than one at a time available.
1446
1447 (d) Implemented a pcregrep test script.
1448
1449 (e) Added the -M (multiline match) option. This allows patterns to match
1450 over several lines of the subject. The buffering ensures that at least
1451 8K, or the rest of the document (whichever is the shorter) is available
1452 for matching (and similarly the previous 8K for lookbehind assertions).
1453
1454 (f) Changed the --help output so that it now says
1455
1456 -w, --word-regex(p)
1457
1458 instead of two lines, one with "regex" and the other with "regexp"
1459 because that confused at least one person since the short forms are the
1460 same. (This required a bit of code, as the output is generated
1461 automatically from a table. It wasn't just a text change.)
1462
1463 (g) -- can be used to terminate pcregrep options if the next thing isn't an
1464 option but starts with a hyphen. Could be a pattern or a path name
1465 starting with a hyphen, for instance.
1466
1467 (h) "-" can be given as a file name to represent stdin.
1468
1469 (i) When file names are being printed, "(standard input)" is used for
1470 the standard input, for compatibility with GNU grep. Previously
1471 "<stdin>" was used.
1472
1473 (j) The option --label=xxx can be used to supply a name to be used for
1474 stdin when file names are being printed. There is no short form.
1475
1476 (k) Re-factored the options decoding logic because we are going to add
1477 two more options that take data. Such options can now be given in four
1478 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
1479
1480 (l) Added the -A, -B, and -C options for requesting that lines of context
1481 around matches be printed.
1482
1483 (m) Added the -L option to print the names of files that do not contain
1484 any matching lines, that is, the complement of -l.
1485
1486 (n) The return code is 2 if any file cannot be opened, but pcregrep does
1487 continue to scan other files.
1488
1489 (o) The -s option was incorrectly implemented. For compatibility with other
1490 greps, it now suppresses the error message for a non-existent or non-
1491 accessible file (but not the return code). There is a new option called
1492 -q that suppresses the output of matching lines, which was what -s was
1493 previously doing.
1494
1495 (p) Added --include and --exclude options to specify files for inclusion
1496 and exclusion when recursing.
1497
1498 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
1499 Hopefully, it now does.
1500
1501 12. Missing cast in pcre_study().
1502
1503 13. Added an "uninstall" target to the makefile.
1504
1505 14. Replaced "extern" in the function prototypes in Makefile.in with
1506 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
1507 world, but is set differently for Windows.
1508
1509 15. Added a second compiling function called pcre_compile2(). The only
1510 difference is that it has an extra argument, which is a pointer to an
1511 integer error code. When there is a compile-time failure, this is set
1512 non-zero, in addition to the error test pointer being set to point to an
1513 error message. The new argument may be NULL if no error number is required
1514 (but then you may as well call pcre_compile(), which is now just a
1515 wrapper). This facility is provided because some applications need a
1516 numeric error indication, but it has also enabled me to tidy up the way
1517 compile-time errors are handled in the POSIX wrapper.
1518
1519 16. Added VPATH=.libs to the makefile; this should help when building with one
1520 prefix path and installing with another. (Or so I'm told by someone who
1521 knows more about this stuff than I do.)
1522
1523 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
1524 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
1525 match everything, including newlines. This is not POSIX-compatible, but
1526 somebody wanted the feature. From pcretest it can be activated by using
1527 both the P and the s flags.
1528
1529 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
1530
1531 19. libpcre.pc was being incorrectly installed as executable.
1532
1533 20. A couple of places in pcretest check for end-of-line by looking for '\n';
1534 it now also looks for '\r' so that it will work unmodified on Windows.
1535
1536 21. Added Google's contributed C++ wrapper to the distribution.
1537
1538 22. Added some untidy missing memory free() calls in pcretest, to keep
1539 Electric Fence happy when testing.
1540
1541
1542
1543 Version 5.0 13-Sep-04
1544 ---------------------
1545
1546 1. Internal change: literal characters are no longer packed up into items
1547 containing multiple characters in a single byte-string. Each character
1548 is now matched using a separate opcode. However, there may be more than one
1549 byte in the character in UTF-8 mode.
1550
1551 2. The pcre_callout_block structure has two new fields: pattern_position and
1552 next_item_length. These contain the offset in the pattern to the next match
1553 item, and its length, respectively.
1554
1555 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
1556 insertion of callouts before each pattern item. Added the /C option to
1557 pcretest to make use of this.
1558
1559 4. On the advice of a Windows user, the lines
1560
1561 #if defined(_WIN32) || defined(WIN32)
1562 _setmode( _fileno( stdout ), 0x8000 );
1563 #endif /* defined(_WIN32) || defined(WIN32) */
1564
1565 have been added to the source of pcretest. This apparently does useful
1566 magic in relation to line terminators.
1567
1568 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
1569 for the benefit of those environments where the "b" makes a difference.
1570
1571 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
1572 to know about it. I have put a hack into configure.in that adds in code
1573 to set GCC=yes if CC=icc. This seems to end up at a point in the
1574 generated configure script that is early enough to affect the setting of
1575 compiler options, which is what is needed, but I have no means of testing
1576 whether it really works. (The user who reported this had patched the
1577 generated configure script, which of course I cannot do.)
1578
1579 LATER: After change 22 below (new libtool files), the configure script
1580 seems to know about icc (and also ecc). Therefore, I have commented out
1581 this hack in configure.in.
1582
1583 7. Added support for pkg-config (2 patches were sent in).
1584
1585 8. Negated POSIX character classes that used a combination of internal tables
1586 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
1587 [[:^ascii]]. Typically, they would match almost any characters. The other
1588 POSIX classes were not broken in this way.
1589
1590 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
1591 to find the match, as PCRE was deluded into thinking that the match had to
1592 start at the start point or following a newline. The same bug applied to
1593 patterns with negative forward assertions or any backward assertions
1594 preceding ".*" at the start, unless the pattern required a fixed first
1595 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
1596
1597 10. In UTF-8 mode, when moving forwards in the subject after a failed match
1598 starting at the last subject character, bytes beyond the end of the subject
1599 string were read.
1600
1601 11. Renamed the variable "class" as "classbits" to make life easier for C++
1602 users. (Previously there was a macro definition, but it apparently wasn't
1603 enough.)
1604
1605 12. Added the new field "tables" to the extra data so that tables can be passed
1606 in at exec time, or the internal tables can be re-selected. This allows
1607 a compiled regex to be saved and re-used at a later time by a different
1608 program that might have everything at different addresses.
1609
1610 13. Modified the pcre-config script so that, when run on Solaris, it shows a
1611 -R library as well as a -L library.
1612
1613 14. The debugging options of pcretest (-d on the command line or D on a
1614 pattern) showed incorrect output for anything following an extended class
1615 that contained multibyte characters and which was followed by a quantifier.
1616
1617 15. Added optional support for general category Unicode character properties
1618 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
1619 support. It adds about 90K to the size of the library. The meanings of the
1620 inbuilt class escapes such as \d and \s have NOT been changed.
1621
1622 16. Updated pcredemo.c to include calls to free() to release the memory for the
1623 compiled pattern.
1624
1625 17. The generated file chartables.c was being created in the source directory
1626 instead of in the building directory. This caused the build to fail if the
1627 source directory was different from the building directory, and was
1628 read-only.
1629
1630 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
1631 file. No doubt somebody will tell me if they don't make sense... Also added
1632 Dan Mooney's comments about building on OpenVMS.
1633
1634 19. Added support for partial matching via the PCRE_PARTIAL option for
1635 pcre_exec() and the \P data escape in pcretest.
1636
1637 20. Extended pcretest with 3 new pattern features:
1638
1639 (i) A pattern option of the form ">rest-of-line" causes pcretest to
1640 write the compiled pattern to the file whose name is "rest-of-line".
1641 This is a straight binary dump of the data, with the saved pointer to
1642 the character tables forced to be NULL. The study data, if any, is
1643 written too. After writing, pcretest reads a new pattern.
1644
1645 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
1646 compiled pattern from the given file. There must not be any
1647 occurrences of "<" in the file name (pretty unlikely); if there are,
1648 pcretest will instead treat the initial "<" as a pattern delimiter.
1649 After reading in the pattern, pcretest goes on to read data lines as
1650 usual.
1651
1652 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
1653 and 16-bit fields in a compiled pattern, to simulate a pattern that
1654 was compiled on a host of opposite endianness.
1655
1656 21. The pcre-exec() function can now cope with patterns that were compiled on
1657 hosts of opposite endianness, with this restriction:
1658
1659 As for any compiled expression that is saved and used later, the tables
1660 pointer field cannot be preserved; the extra_data field in the arguments
1661 to pcre_exec() should be used to pass in a tables address if a value
1662 other than the default internal tables were used at compile time.
1663
1664 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
1665 now diagnosed as an error. Previously, most of the time, a negative number
1666 would have been treated as zero, but if in addition "ovector" was passed as
1667 NULL, a crash could occur.
1668
1669 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
1670 new versions from the libtool 1.5 distribution (the last one is a copy of
1671 a file called libtool.m4). This seems to have fixed the need to patch
1672 "configure" to support Darwin 1.3 (which I used to do). However, I still
1673 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
1674 workstation).
1675
1676 24. Changed the PCRE licence to be the more standard "BSD" licence.
1677
1678
1679 Version 4.5 01-Dec-03
1680 ---------------------
1681
1682 1. There has been some re-arrangement of the code for the match() function so
1683 that it can be compiled in a version that does not call itself recursively.
1684 Instead, it keeps those local variables that need separate instances for
1685 each "recursion" in a frame on the heap, and gets/frees frames whenever it
1686 needs to "recurse". Keeping track of where control must go is done by means
1687 of setjmp/longjmp. The whole thing is implemented by a set of macros that
1688 hide most of the details from the main code, and operates only if
1689 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
1690 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
1691 operating.
1692
1693 To make it easier for callers to provide specially tailored get/free
1694 functions for this usage, two new functions, pcre_stack_malloc, and
1695 pcre_stack_free, are used. They are always called in strict stacking order,
1696 and the size of block requested is always the same.
1697
1698 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
1699 PCRE has been compiled to use the stack or the heap for recursion. The
1700 -C option of pcretest uses this to show which version is compiled.
1701
1702 A new data escape \S, is added to pcretest; it causes the amounts of store
1703 obtained and freed by both kinds of malloc/free at match time to be added
1704 to the output.
1705
1706 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
1707 what's available on my current Linux desktop machine.
1708
1709 3. When matching a UTF-8 string, the test for a valid string at the start has
1710 been extended. If start_offset is not zero, PCRE now checks that it points
1711 to a byte that is the start of a UTF-8 character. If not, it returns
1712 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
1713 this is necessary because there may be backward assertions in the pattern.
1714 When matching the same subject several times, it may save resources to use
1715 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
1716
1717 4. The code for checking the validity of UTF-8 strings has been tightened so
1718 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
1719 containing "overlong sequences".
1720
1721 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
1722 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
1723 should have been "&", but it just so happened that all the cases this let
1724 through by mistake were picked up later in the function.
1725
1726 6. I had used a variable called "isblank" - this is a C99 function, causing
1727 some compilers to warn. To avoid this, I renamed it (as "blankclass").
1728
1729 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
1730 prompting; (b) run "./pcretest /dev/null" at the start of the test script
1731 so the version is shown; (c) stop "make test" echoing "./RunTest".
1732
1733 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
1734
1735 9. The prototype for memmove() for systems that don't have it was using
1736 size_t, but the inclusion of the header that defines size_t was later. I've
1737 moved the #includes for the C headers earlier to avoid this.
1738
1739 10. Added some adjustments to the code to make it easier to compiler on certain
1740 special systems:
1741
1742 (a) Some "const" qualifiers were missing.
1743 (b) Added the macro EXPORT before all exported functions; by default this
1744 is defined to be empty.
1745 (c) Changed the dftables auxiliary program (that builds chartables.c) so
1746 that it reads its output file name as an argument instead of writing
1747 to the standard output and assuming this can be redirected.
1748
1749 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
1750 class containing characters with values greater than 255, PCRE compilation
1751 went into a loop.
1752
1753 12. A recursive reference to a subpattern that was within another subpattern
1754 that had a minimum quantifier of zero caused PCRE to crash. For example,
1755 (x(y(?2))z)? provoked this bug with a subject that got as far as the
1756 recursion. If the recursively-called subpattern itself had a zero repeat,
1757 that was OK.
1758
1759 13. In pcretest, the buffer for reading a data line was set at 30K, but the
1760 buffer into which it was copied (for escape processing) was still set at
1761 1024, so long lines caused crashes.
1762
1763 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
1764 "internal error: code overflow...". This applied to any character class
1765 that was followed by a possessive quantifier.
1766
1767 15. Modified the Makefile to add libpcre.la as a prerequisite for
1768 libpcreposix.la because I was told this is needed for a parallel build to
1769 work.
1770
1771 16. If a pattern that contained .* following optional items at the start was
1772 studied, the wrong optimizing data was generated, leading to matching
1773 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
1774 matching string must start with a or b or c. The correct conclusion for
1775 this pattern is that a match can start with any character.
1776
1777
1778 Version 4.4 13-Aug-03
1779 ---------------------
1780
1781 1. In UTF-8 mode, a character class containing characters with values between
1782 127 and 255 was not handled correctly if the compiled pattern was studied.
1783 In fixing this, I have also improved the studying algorithm for such
1784 classes (slightly).
1785
1786 2. Three internal functions had redundant arguments passed to them. Removal
1787 might give a very teeny performance improvement.
1788
1789 3. Documentation bug: the value of the capture_top field in a callout is *one
1790 more than* the number of the hightest numbered captured substring.
1791
1792 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
1793 in incorrectly linking with a previously installed version. They now link
1794 explicitly with libpcre.la.
1795
1796 5. configure.in no longer needs to recognize Cygwin specially.
1797
1798 6. A problem in pcre.in for Windows platforms is fixed.
1799
1800 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
1801 pcretest, it used to include the size of the study block as part of its
1802 output. Unfortunately, the structure contains a field that has a different
1803 size on different hardware architectures. This meant that the tests that
1804 showed this size failed. As the block is currently always of a fixed size,
1805 this information isn't actually particularly useful in pcretest output, so
1806 I have just removed it.
1807
1808 8. Three pre-processor statements accidentally did not start in column 1.
1809 Sadly, there are *still* compilers around that complain, even though
1810 standard C has not required this for well over a decade. Sigh.
1811
1812 9. In pcretest, the code for checking callouts passed small integers in the
1813 callout_data field, which is a void * field. However, some picky compilers
1814 complained about the casts involved for this on 64-bit systems. Now
1815 pcretest passes the address of the small integer instead, which should get
1816 rid of the warnings.
1817
1818 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
1819 both compile and run time, and gives an error if an invalid UTF-8 sequence
1820 is found. There is a option for disabling this check in cases where the
1821 string is known to be correct and/or the maximum performance is wanted.
1822
1823 11. In response to a bug report, I changed one line in Makefile.in from
1824
1825 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
1826 to
1827 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
1828
1829 to look similar to other lines, but I have no way of telling whether this
1830 is the right thing to do, as I do not use Windows. No doubt I'll get told
1831 if it's wrong...
1832
1833
1834 Version 4.3 21-May-03
1835 ---------------------
1836
1837 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
1838 Makefile.
1839
1840 2. Some refactoring to improve the quality of the code:
1841
1842 (i) The utf8_table... variables are now declared "const".
1843
1844 (ii) The code for \cx, which used the "case flipping" table to upper case
1845 lower case letters, now just substracts 32. This is ASCII-specific,
1846 but the whole concept of \cx is ASCII-specific, so it seems
1847 reasonable.
1848
1849 (iii) PCRE was using its character types table to recognize decimal and
1850 hexadecimal digits in the pattern. This is silly, because it handles
1851 only 0-9, a-f, and A-F, but the character types table is locale-
1852 specific, which means strange things might happen. A private
1853 table is now used for this - though it costs 256 bytes, a table is
1854 much faster than multiple explicit tests. Of course, the standard
1855 character types table is still used for matching digits in subject
1856 strings against \d.
1857
1858 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
1859 ending in _t are). So I've renamed it as ESC_tee.
1860
1861 3. The first argument for regexec() in the POSIX wrapper should have been
1862 defined as "const".
1863
1864 4. Changed pcretest to use malloc() for its buffers so that they can be
1865 Electric Fenced for debugging.
1866
1867 5. There were several places in the code where, in UTF-8 mode, PCRE would try
1868 to read one or more bytes before the start of the subject string. Often this
1869 had no effect on PCRE's behaviour, but in some circumstances it could
1870 provoke a segmentation fault.
1871
1872 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
1873 to try to read one or more bytes before the start of the subject string.
1874
1875 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
1876 UTF-8 support could misbehave in various ways if the subject string
1877 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
1878 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
1879 back over UTF-8 characters.)
1880
1881
1882 Version 4.2 14-Apr-03
1883 ---------------------
1884
1885 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
1886
1887 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
1888 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
1889 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
1890 [WIN_PREFIX]: new variable, "cyg" for Cygwin
1891 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
1892 and BUILD_EXEEXT
1893 Note: automatic setting of the BUILD variables is not yet working
1894 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
1895 compile-time but not at link-time
1896 [LINK]: use for linking executables only
1897 make different versions for Windows and non-Windows
1898 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
1899 libraries
1900 [LINK_FOR_BUILD]: new variable
1901 [OBJEXT]: use throughout
1902 [EXEEXT]: use throughout
1903 <winshared>: new target
1904 <wininstall>: new target
1905 <dftables.o>: use native compiler
1906 <dftables>: use native linker
1907 <install>: handle Windows platform correctly
1908 <clean>: ditto
1909 <check>: ditto
1910 copy DLL to top builddir before testing
1911
1912 As part of these changes, -no-undefined was removed again. This was reported
1913 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
1914 in any case.
1915
1916 3. Some tidies to get rid of compiler warnings:
1917
1918 . In the match_data structure, match_limit was an unsigned long int, whereas
1919 match_call_count was an int. I've made them both unsigned long ints.
1920
1921 . In pcretest the fact that a const uschar * doesn't automatically cast to
1922 a void * provoked a warning.
1923
1924 . Turning on some more compiler warnings threw up some "shadow" variables
1925 and a few more missing casts.
1926
1927 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
1928 option, a class that contained a single character with a value between 128
1929 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
1930
1931 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
1932 option, a class that contained several characters, but with at least one
1933 whose value was between 128 and 255 caused PCRE to crash.
1934
1935
1936 Version 4.1 12-Mar-03
1937 ---------------------
1938
1939 1. Compiling with gcc -pedantic found a couple of places where casts were
1940 needed, and a string in dftables.c that was longer than standard compilers are
1941 required to support.
1942
1943 2. Compiling with Sun's compiler found a few more places where the code could
1944 be tidied up in order to avoid warnings.
1945
1946 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
1947 first of these names is deprecated in the latest Autoconf in favour of the name
1948 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
1949 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
1950 analogy I have changed it to CFLAGS_FOR_BUILD.
1951
1952 4. Added -no-undefined to the linking command in the Makefile, because this is
1953 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
1954 linking step for the pcreposix library.
1955
1956 5. PCRE was failing to diagnose the case of two named groups with the same
1957 name.
1958
1959 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
1960 literal character that is needed in the subject for a match, and scans along to
1961 ensure that it is present before embarking on the full matching process. This
1962 saves time in cases of nested unlimited repeats that are never going to match.
1963 Problem: the scan can take a lot of time if the subject is very long (e.g.
1964 megabytes), thus penalizing straightforward matches. It is now done only if the
1965 amount of subject to be scanned is less than 1000 bytes.
1966
1967 7. A lesser problem with the same optimization is that it was recording the
1968 first character of an anchored pattern as "needed", thus provoking a search
1969 right along the subject, even when the first match of the pattern was going to
1970 fail. The "needed" character is now not set for anchored patterns, unless it
1971 follows something in the pattern that is of non-fixed length. Thus, it still
1972 fulfils its original purpose of finding quick non-matches in cases of nested
1973 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
1974
1975
1976 Version 4.0 17-Feb-03
1977 ---------------------
1978
1979 1. If a comment in an extended regex that started immediately after a meta-item
1980 extended to the end of string, PCRE compiled incorrect data. This could lead to
1981 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
1982
1983 2. Moved to autoconf 2.53 and libtool 1.4.2.
1984
1985 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
1986 the special perltest8 script is no longer needed - all the tests can be run
1987 from a single perltest script.
1988
1989 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
1990 by \s. It has now been removed in PCRE. This means it isn't recognized as
1991 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
1992 class [:space:] *does* include VT, thereby creating a mess.
1993
1994 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
1995 space and tab.
1996
1997 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
1998 its new features into the main test script, reducing the number of scripts.
1999
2000 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
2001 were backward compatible, and made the (?i) apply to the whole pattern, as if
2002 /i were given. Now it behaves more logically, and applies the option setting
2003 only to what follows. PCRE has been changed to follow suit. However, if it
2004 finds options settings right at the start of the pattern, it extracts them into
2005 the global options, as before. Thus, they show up in the info data.
2006
2007 8. Added support for the \Q...\E escape sequence. Characters in between are
2008 treated as literals. This is slightly different from Perl in that $ and @ are
2009 also handled as literals inside the quotes. In Perl, they will cause variable
2010 interpolation. Note the following examples:
2011
2012 Pattern PCRE matches Perl matches
2013
2014 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
2015 \Qabc\$xyz\E abc\$xyz abc\$xyz
2016 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
2017
2018 For compatibility with Perl, \Q...\E sequences are recognized inside character
2019 classes as well as outside them.
2020
2021 9. Re-organized 3 code statements in pcretest to avoid "overflow in
2022 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
2023 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
2024 signed/unsigned warnings.
2025
2026 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
2027 option for pcretest, so I've replaced it by a simple function that does just
2028 that job.
2029
2030 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
2031 "pcregrep -".
2032
2033 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
2034 Java package. This provides some syntactic sugar for simple cases of what my
2035 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
2036 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
2037 item, you can use this simplified notation. Note that only makes sense with
2038 greedy quantifiers. Consequently, the use of the possessive quantifier forces
2039 greediness, whatever the setting of the PCRE_UNGREEDY option.
2040
2041 13. A change of greediness default within a pattern was not taking effect at
2042 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
2043 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
2044 was abstracted outside.
2045
2046 14. PCRE now supports the \G assertion. It is true when the current matching
2047 position is at the start point of the match. This differs from \A when the
2048 starting offset is non-zero. Used with the /g option of pcretest (or similar
2049 code), it works in the same way as it does for Perl's /g option. If all
2050 alternatives of a regex begin with \G, the expression is anchored to the start
2051 match position, and the "anchored" flag is set in the compiled expression.
2052
2053 15. Some bugs concerning the handling of certain option changes within patterns
2054 have been fixed. These applied to options other than (?ims). For example,
2055 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
2056 been the other way round. Some of this was related to change 7 above.
2057
2058 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
2059 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
2060 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
2061 POSIX classes only within a class (e.g. /[[:alpha:]]/).
2062
2063 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
2064 mode. Unlike ".", it always matches newline, whatever the setting of
2065 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
2066 assertions. Perl allows it, but it doesn't (in general) work because it can't
2067 calculate the length of the lookbehind. At least, that's the case for Perl
2068 5.8.0 - I've been told they are going to document that it doesn't work in
2069 future.
2070
2071 18. Added an error diagnosis for escapes that PCRE does not support: these are
2072 \L, \l, \N, \P, \p, \U, \u, and \X.
2073
2074 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
2075 reading past the end of the pattern in cases such as /[abcd/.
2076
2077 20. PCRE was getting more memory than necessary for patterns with classes that
2078 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
2079
2080 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
2081 compiling PCRE for use with Virtual Pascal.
2082
2083 22. Small fix to the Makefile to make it work properly if the build is done
2084 outside the source tree.
2085
2086 23. Added a new extension: a condition to go with recursion. If a conditional
2087 subpattern starts with (?(R) the "true" branch is used if recursion has
2088 happened, whereas the "false" branch is used only at the top level.
2089
2090 24. When there was a very long string of literal characters (over 255 bytes
2091 without UTF support, over 250 bytes with UTF support), the computation of how
2092 much memory was required could be incorrect, leading to segfaults or other
2093 strange effects.
2094
2095 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
2096 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
2097 there was a subsequent back reference to those brackets. This meant that, for
2098 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
2099 possible to check for precisely this case. All we can do is abandon the
2100 optimization if .* occurs inside capturing brackets when there are any back
2101 references whatsoever. (See below for a better fix that came later.)
2102
2103 26. The handling of the optimization for finding the first character of a
2104 non-anchored pattern, and for finding a character that is required later in the
2105 match were failing in some cases. This didn't break the matching; it just
2106 failed to optimize when it could. The way this is done has been re-implemented.
2107
2108 27. Fixed typo in error message for invalid (?R item (it said "(?p").
2109
2110 28. Added a new feature that provides some of the functionality that Perl
2111 provides with (?{...}). The facility is termed a "callout". The way it is done
2112 in PCRE is for the caller to provide an optional function, by setting
2113 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
2114 global variable. By default it is unset, which disables all calling out. To get
2115 the function called, the regex must include (?C) at appropriate points. This
2116 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
2117 This provides a means of identifying different callout points. When PCRE
2118 reaches such a point in the regex, if pcre_callout has been set, the external
2119 function is called. It is provided with data in a structure called
2120 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
2121 matching continues; if it returns a non-zero value, the match at the current
2122 point fails. However, backtracking will occur if possible. [This was changed
2123 later and other features added - see item 49 below.]
2124
2125 29. pcretest is upgraded to test the callout functionality. It provides a
2126 callout function that displays information. By default, it shows the start of
2127 the match and the current position in the text. There are some new data escapes
2128 to vary what happens:
2129
2130 \C+ in addition, show current contents of captured substrings
2131 \C- do not supply a callout function
2132 \C!n return 1 when callout number n is reached
2133 \C!n!m return 1 when callout number n is reached for the mth time
2134
2135 30. If pcregrep was called with the -l option and just a single file name, it
2136 output "<stdin>" if a match was found, instead of the file name.
2137
2138 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
2139 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
2140 pcre_exec(). This saves a malloc/free per call. The default value of
2141 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
2142 when configuring.
2143
2144 32. The default maximum size of a compiled pattern is 64K. There have been a
2145 few cases of people hitting this limit. The code now uses macros to handle the
2146 storing of links as offsets within the compiled pattern. It defaults to 2-byte
2147 links, but this can be changed to 3 or 4 bytes by --with-link-size when
2148 configuring. Tests 2 and 5 work only with 2-byte links because they output
2149 debugging information about compiled patterns.
2150
2151 33. Internal code re-arrangements:
2152
2153 (a) Moved the debugging function for printing out a compiled regex into
2154 its own source file (printint.c) and used #include to pull it into
2155 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
2156 separate copies.
2157
2158 (b) Defined the list of op-code names for debugging as a macro in
2159 internal.h so that it is next to the definition of the opcodes.
2160
2161 (c) Defined a table of op-code lengths for simpler skipping along compiled
2162 code. This is again a macro in internal.h so that it is next to the
2163 definition of the opcodes.
2164
2165 34. Added support for recursive calls to individual subpatterns, along the
2166 lines of Robin Houston's patch (but implemented somewhat differently).
2167
2168 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
2169 allow it to read and process whole directories in Win32. This code was
2170 contributed by Lionel Fourquaux; it has not been tested by me.
2171
2172 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
2173 used to name a group. Names consist of alphanumerics and underscores, and must
2174 be unique. Back references use the syntax (?P=name) and recursive calls use
2175 (?P>name) which is a PCRE extension to the Python extension. Groups still have
2176 numbers. The function pcre_fullinfo() can be used after compilation to extract
2177 a name/number map. There are three relevant calls:
2178
2179 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
2180 PCRE_INFO_NAMECOUNT yields the number of entries
2181 PCRE_INFO_NAMETABLE yields a pointer to the map.
2182
2183 The map is a vector of fixed-size entries. The size of each entry depends on
2184 the length of the longest name used. The first two bytes of each entry are the
2185 group number, most significant byte first. There follows the corresponding
2186 name, zero terminated. The names are in alphabetical order.
2187
2188 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
2189 case instead of 255. Making it the same both with and without UTF-8 support
2190 means that the same test output works with both.
2191
2192 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
2193 calling malloc() with a zero argument.
2194
2195 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
2196 optimization. I've improved things by keeping a bitmap of backreferences with
2197 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
2198 fact referenced, the optimization can be applied. It is unlikely that a
2199 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
2200 the match to follow \n) will appear inside brackets with a number greater than
2201 31, but if it does, any back reference > 31 suppresses the optimization.
2202
2203 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
2204 of disabling numbered capturing parentheses. Any opening parenthesis that is
2205 not followed by ? behaves as if it were followed by ?: but named parentheses
2206 can still be used for capturing (and they will acquire numbers in the usual
2207 way).
2208
2209 41. Redesigned the return codes from the match() function into yes/no/error so
2210 that errors can be passed back from deep inside the nested calls. A malloc
2211 failure while inside a recursive subpattern call now causes the
2212 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
2213
2214 42. It is now possible to set a limit on the number of times the match()
2215 function is called in a call to pcre_exec(). This facility makes it possible to
2216 limit the amount of recursion and backtracking, though not in a directly
2217 obvious way, because the match() function is used in a number of different
2218 circumstances. The count starts from zero for each position in the subject
2219 string (for non-anchored patterns). The default limit is, for compatibility, a
2220 large number, namely 10 000 000. You can change this in two ways:
2221
2222 (a) When configuring PCRE before making, you can use --with-match-limit=n
2223 to set a default value for the compiled library.
2224
2225 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
2226 a different value is set. See 45 below.
2227
2228 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
2229
2230 43. Added a new function pcre_config(int, void *) to enable run-time extraction
2231 of things that can be changed at compile time. The first argument specifies
2232 what is wanted and the second points to where the information is to be placed.
2233 The current list of available information is:
2234
2235 PCRE_CONFIG_UTF8
2236
2237 The output is an integer that is set to one if UTF-8 support is available;
2238 otherwise it is set to zero.
2239
2240 PCRE_CONFIG_NEWLINE
2241
2242 The output is an integer that it set to the value of the code that is used for
2243 newline. It is either LF (10) or CR (13).
2244
2245 PCRE_CONFIG_LINK_SIZE
2246
2247 The output is an integer that contains the number of bytes used for internal
2248 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
2249
2250 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
2251
2252 The output is an integer that contains the threshold above which the POSIX
2253 interface uses malloc() for output vectors. See item 31 above.
2254
2255 PCRE_CONFIG_MATCH_LIMIT
2256
2257 The output is an unsigned integer that contains the default limit of the number
2258 of match() calls in a pcre_exec() execution. See 42 above.
2259
2260 44. pcretest has been upgraded by the addition of the -C option. This causes it
2261 to extract all the available output from the new pcre_config() function, and to
2262 output it. The program then exits immediately.
2263
2264 45. A need has arisen to pass over additional data with calls to pcre_exec() in
2265 order to support additional features. One way would have been to define
2266 pcre_exec2() (for example) with extra arguments, but this would not have been
2267 extensible, and would also have required all calls to the original function to
2268 be mapped to the new one. Instead, I have chosen to extend the mechanism that
2269 is used for passing in "extra" data from pcre_study().
2270
2271 The pcre_extra structure is now exposed and defined in pcre.h. It currently
2272 contains the following fields:
2273
2274 flags a bitmap indicating which of the following fields are set
2275 study_data opaque data from pcre_study()
2276 match_limit a way of specifying a limit on match() calls for a specific
2277 call to pcre_exec()
2278 callout_data data for callouts (see 49 below)
2279
2280 The flag bits are also defined in pcre.h, and are
2281
2282 PCRE_EXTRA_STUDY_DATA
2283 PCRE_EXTRA_MATCH_LIMIT
2284 PCRE_EXTRA_CALLOUT_DATA
2285
2286 The pcre_study() function now returns one of these new pcre_extra blocks, with
2287 the actual study data pointed to by the study_data field, and the
2288 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
2289 before. That is, this change is entirely upwards-compatible and requires no
2290 change to existing code.
2291
2292 If you want to pass in additional data to pcre_exec(), you can either place it
2293 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
2294 block.
2295
2296 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
2297 data string contains the escape sequence \M, pcretest calls pcre_exec() several
2298 times with different match limits, until it finds the minimum value needed for
2299 pcre_exec() to complete. The value is then output. This can be instructive; for
2300 most simple matches the number is quite small, but for pathological cases it
2301 gets very large very quickly.
2302
2303 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
2304 returns the size of the data block pointed to by the study_data field in a
2305 pcre_extra block, that is, the value that was passed as the argument to
2306 pcre_malloc() when PCRE was getting memory in which to place the information
2307 created by pcre_study(). The fourth argument should point to a size_t variable.
2308 pcretest has been extended so that this information is shown after a successful
2309 pcre_study() call when information about the compiled regex is being displayed.
2310
2311 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
2312 because what follows is always an absolute path. (Later: it turns out that this
2313 is more than cosmetic for MinGW, because it doesn't like empty path
2314 components.)
2315
2316 49. Some changes have been made to the callout feature (see 28 above):
2317
2318 (i) A callout function now has three choices for what it returns:
2319
2320 0 => success, carry on matching
2321 > 0 => failure at this point, but backtrack if possible
2322 < 0 => serious error, return this value from pcre_exec()
2323
2324 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
2325 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
2326 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
2327 use by callout functions. It will never be used by PCRE itself.
2328
2329 (ii) The pcre_extra structure (see 45 above) has a void * field called
2330 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
2331 pcre_callout_block structure has a field of the same name. The contents of
2332 the field passed in the pcre_extra structure are passed to the callout
2333 function in the corresponding field in the callout block. This makes it
2334 easier to use the same callout-containing regex from multiple threads. For
2335 testing, the pcretest program has a new data escape
2336
2337 \C*n pass the number n (may be negative) as callout_data
2338
2339 If the callout function in pcretest receives a non-zero value as
2340 callout_data, it returns that value.
2341
2342 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
2343 there were some redundant $(CFLAGS) in commands that are now specified as
2344 $(LINK), which already includes $(CFLAGS).
2345
2346 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
2347 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
2348 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
2349 one-byte characters throughout. Note that case-insensitive matching applies
2350 only to characters whose values are less than 256. PCRE doesn't support the
2351 notion of cases for higher-valued characters.
2352
2353 (i) A character class whose characters are all within 0-255 is handled as
2354 a bit map, and the map is inverted for negative classes. Previously, a
2355 character > 255 always failed to match such a class; however it should
2356 match if the class was a negative one (e.g. [^ab]). This has been fixed.
2357
2358 (ii) A negated character class with a single character < 255 is coded as
2359 "not this character" (OP_NOT). This wasn't working properly when the test
2360 character was multibyte, either singly or repeated.
2361
2362 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
2363 mode, for example: \x{100}{2,3}.
2364
2365 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
2366 singly or repeated) now correctly test multibyte characters. However,
2367 PCRE doesn't recognize any characters with values greater than 255 as
2368 digits, spaces, or word characters. Such characters always match \D, \S,
2369 and \W, and never match \d, \s, or \w.
2370
2371 (v) Classes may now contain characters and character ranges with values
2372 greater than 255. For example: [ab\x{100}-\x{400}].
2373
2374 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
2375 PCRE in UTF-8 mode.
2376
2377 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
2378 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
2379 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
2380 value.)
2381
2382 53. The single man page has become too large. I have therefore split it up into
2383 a number of separate man pages. These also give rise to individual HTML pages;
2384 these are now put in a separate directory, and there is an index.html page that
2385 lists them all. Some hyperlinking between the pages has been installed.
2386
2387 54. Added convenience functions for handling named capturing parentheses.
2388
2389 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
2390 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
2391 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
2392 are faulted.
2393
2394 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
2395 calling configure. These values are used when compiling the dftables.c program
2396 which is run to generate the source of the default character tables. They
2397 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
2398 you will need to set these values.
2399
2400 57. Updated the building process for Windows DLL, as provided by Fred Cox.
2401
2402
2403 Version 3.9 02-Jan-02
2404 ---------------------
2405
2406 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
2407
2408 2. If --disable-static was given, the building process failed when trying to
2409 build pcretest and pcregrep. (For some reason it was using libtool to compile
2410 them, which is not right, as they aren't part of the library.)
2411
2412
2413 Version 3.8 18-Dec-01
2414 ---------------------
2415
2416 1. The experimental UTF-8 code was completely screwed up. It was packing the
2417 bytes in the wrong order. How dumb can you get?
2418
2419
2420 Version 3.7 29-Oct-01
2421 ---------------------
2422
2423 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
2424 This caused pcretest, when used on the test data, to segfault. Unfortunately,
2425 this didn't happen under Solaris 8, where I normally test things.
2426
2427 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
2428 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
2429 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
2430 this fix an hour or so after the initial 3.7 release.)
2431
2432
2433 Version 3.6 23-Oct-01
2434 ---------------------
2435
2436 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
2437 offsets passed as NULL with zero offset count.
2438
2439 2. The config.guess and config.sub files had not been updated when I moved to
2440 the latest autoconf.
2441
2442
2443 Version 3.5 15-Aug-01
2444 ---------------------
2445
2446 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
2447 had been forgotten.
2448
2449 2. By using declared but undefined structures, we can avoid using "void"
2450 definitions in pcre.h while keeping the internal definitions of the structures
2451 private.
2452
2453 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
2454 user point of view, this means that both static and shared libraries are built
2455 by default, but this can be individually controlled. More of the work of
2456 handling this static/shared cases is now inside libtool instead of PCRE's make
2457 file.
2458
2459 4. The pcretest utility is now installed along with pcregrep because it is
2460 useful for users (to test regexs) and by doing this, it automatically gets
2461 relinked by libtool. The documentation has been turned into a man page, so
2462 there are now .1, .txt, and .html versions in /doc.
2463
2464 5. Upgrades to pcregrep:
2465 (i) Added long-form option names like gnu grep.
2466 (ii) Added --help to list all options with an explanatory phrase.
2467 (iii) Added -r, --recursive to recurse into sub-directories.
2468 (iv) Added -f, --file to read patterns from a file.
2469
2470 6. pcre_exec() was referring to its "code" argument before testing that
2471 argument for NULL (and giving an error if it was NULL).
2472
2473 7. Upgraded Makefile.in to allow for compiling in a different directory from
2474 the source directory.
2475
2476 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
2477 options bits, the pointer it was passed was to an int instead of to an unsigned
2478 long int. This mattered only on 64-bit systems.
2479
2480 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
2481 generated) instead of pcre.in, which it its source. Also made the same change
2482 in several of the .c files.
2483
2484 10. A new release of gcc defines printf() as a macro, which broke pcretest
2485 because it had an ifdef in the middle of a string argument for printf(). Fixed
2486 by using separate calls to printf().
2487
2488 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
2489 script, to force use of CR or LF instead of \n in the source. On non-Unix
2490 systems, the value can be set in config.h.
2491
2492 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
2493 absolute limit. Changed the text of the error message to make this clear, and
2494 likewise updated the man page.
2495
2496 13. The limit of 99 on the number of capturing subpatterns has been removed.
2497 The new limit is 65535, which I hope will not be a "real" limit.
2498
2499
2500 Version 3.4 22-Aug-00
2501 ---------------------
2502
2503 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
2504
2505 2. Diagnose condition (?(0) as an error instead of crashing on matching.
2506
2507
2508 Version 3.3 01-Aug-00
2509 ---------------------
2510
2511 1. If an octal character was given, but the value was greater than \377, it
2512 was not getting masked to the least significant bits, as documented. This could
2513 lead to crashes in some systems.
2514
2515 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
2516 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
2517
2518 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
2519 These just pass their arguments on to (pcre_free)(), but they are provided
2520 because some uses of PCRE bind it to non-C systems that can call its functions,
2521 but cannot call free() or pcre_free() directly.
2522
2523 4. Add "make test" as a synonym for "make check". Corrected some comments in
2524 the Makefile.
2525
2526 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
2527 Makefile.
2528
2529 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
2530 command called pgrep for grepping around the active processes.
2531
2532 7. Added the beginnings of support for UTF-8 character strings.
2533
2534 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
2535 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
2536 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
2537 out for the ar command.)
2538
2539
2540 Version 3.2 12-May-00
2541 ---------------------
2542
2543 This is purely a bug fixing release.
2544
2545 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
2546 of ZA. This was just one example of several cases that could provoke this bug,
2547 which was introduced by change 9 of version 2.00. The code for breaking
2548 infinite loops after an iteration that matches an empty string was't working
2549 correctly.
2550
2551 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
2552 when matched against abbab (for example). After matching an empty string, it
2553 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
2554 caused it to match further down the string than it should.
2555
2556 3. The code contained an inclusion of sys/types.h. It isn't clear why this
2557 was there because it doesn't seem to be needed, and it causes trouble on some
2558 systems, as it is not a Standard C header. It has been removed.
2559
2560 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
2561 were reported on the Macintosh. The changes were from
2562
2563 while ((c = *(++ptr)) != 0 && c != '\n');
2564 to
2565 while ((c = *(++ptr)) != 0 && c != '\n') ;
2566
2567 Totally extraordinary, but if that's what it takes...
2568
2569 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
2570 available. Added HAVE_BCOPY and an autoconf test for it; if neither
2571 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
2572 assumes the way PCRE uses memmove() (always moving upwards).
2573
2574 6. PCRE is being used in one environment where strchr() is not available. There
2575 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
2576 faster code anyway.
2577
2578
2579 Version 3.1 09-Feb-00
2580 ---------------------
2581
2582 The only change in this release is the fixing of some bugs in Makefile.in for
2583 the "install" target:
2584
2585 (1) It was failing to install pcreposix.h.
2586
2587 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
2588
2589
2590 Version 3.0 01-Feb-00
2591 ---------------------
2592
2593 1. Add support for the /+ modifier to perltest (to output $` like it does in
2594 pcretest).
2595
2596 2. Add support for the /g modifier to perltest.
2597
2598 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
2599 matches null strings.
2600
2601 4. Fix perltest so that it doesn't do unwanted things when fed an empty
2602 pattern. Perl treats empty patterns specially - it reuses the most recent
2603 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
2604 effect.
2605
2606 5. The POSIX interface was broken in that it was just handing over the POSIX
2607 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
2608 required a bigger vector, with some working space on the end. This means that
2609 the POSIX wrapper now has to get and free some memory, and copy the results.
2610
2611 6. Added some simple autoconf support, placing the test data and the
2612 documentation in separate directories, re-organizing some of the
2613 information files, and making it build pcre-config (a GNU standard). Also added
2614 libtool support for building PCRE as a shared library, which is now the
2615 default.
2616
2617 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
2618 09 are not valid octal constants. Single digits will be used for minor values
2619 less than 10.
2620
2621 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
2622 existing programs that set these in the POSIX interface can use PCRE without
2623 modification.
2624
2625 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
2626 return all that pcre_info() returns, plus additional data. The pcre_info()
2627 function is retained for compatibility, but is considered to be obsolete.
2628
2629 10. Added experimental recursion feature (?R) to handle one common case that
2630 Perl 5.6 will be able to do with (?p{...}).
2631
2632 11. Added support for POSIX character classes like [:alpha:], which Perl is
2633 adopting.
2634
2635
2636 Version 2.08 31-Aug-99
2637 ----------------------
2638
2639 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
2640 trying to match at the startoffset position, but instead was moving forward to
2641 the next newline as if a previous match had failed.
2642
2643 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
2644 and could get into a loop if a null string was matched other than at the start
2645 of the subject.
2646
2647 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
2648 be distinguished at compile time, and for completeness also added PCRE_DATE.
2649
2650 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
2651 in GnuWin32 environments.
2652
2653
2654 Version 2.07 29-Jul-99
2655 ----------------------
2656
2657 1. The documentation is now supplied in plain text form and HTML as well as in
2658 the form of man page sources.
2659
2660 2. C++ compilers don't like assigning (void *) values to other pointer types.
2661 In particular this affects malloc(). Although there is no problem in Standard
2662 C, I've put in casts to keep C++ compilers happy.
2663
2664 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
2665 should be (const char *).
2666
2667 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
2668 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
2669 However, I haven't made this a standard facility. The documentation doesn't
2670 mention it, and the Makefile doesn't support it.
2671
2672 5. The Makefile now contains an "install" target, with editable destinations at
2673 the top of the file. The pcretest program is not installed.
2674
2675 6. pgrep -V now gives the PCRE version number and date.
2676
2677 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
2678 causing the entire string to be ignored, instead of just the last character.
2679
2680 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
2681 non-matching string, it can take a very, very long time, even for strings of
2682 quite modest length, because of the nested recursion. PCRE now does better in
2683 some of these cases. It does this by remembering the last required literal
2684 character in the pattern, and pre-searching the subject to ensure it is present
2685 before running the real match. In other words, it applies a heuristic to detect
2686 some types of certain failure quickly, and in the above example, if presented
2687 with a string that has no trailing " it gives "no match" very quickly.
2688
2689 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
2690 other alternatives are tried instead.
2691
2692
2693 Version 2.06 09-Jun-99
2694 ----------------------
2695
2696 1. Change pcretest's output for amount of store used to show just the code
2697 space, because the remainder (the data block) varies in size between 32-bit and
2698 64-bit systems.
2699
2700 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
2701 start matching at. This allows lookbehinds to work when searching for multiple
2702 occurrences in a string.
2703
2704 3. Added additional options to pcretest for testing multiple occurrences:
2705
2706 /+ outputs the rest of the string that follows a match
2707 /g loops for multiple occurrences, using the new startoffset argument
2708 /G loops for multiple occurrences by passing an incremented pointer
2709
2710 4. PCRE wasn't doing the "first character" optimization for patterns starting
2711 with \b or \B, though it was doing it for other lookbehind assertions. That is,
2712 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
2713 the letter 'x'. On long subject strings, this gives a significant speed-up.
2714
2715
2716 Version 2.05 21-Apr-99
2717 ----------------------
2718
2719 1. Changed the type of magic_number from int to long int so that it works
2720 properly on 16-bit systems.
2721
2722 2. Fixed a bug which caused patterns starting with .* not to work correctly
2723 when the subject string contained newline characters. PCRE was assuming
2724 anchoring for such patterns in all cases, which is not correct because .* will
2725 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
2726 DOTALL is set at top level; otherwise it knows that patterns starting with .*
2727 must be retried after every newline in the subject.
2728
2729
2730 Version 2.04 18-Feb-99
2731 ----------------------
2732
2733 1. For parenthesized subpatterns with repeats whose minimum was zero, the
2734 computation of the store needed to hold the pattern was incorrect (too large).
2735 If such patterns were nested a few deep, this could multiply and become a real
2736 problem.
2737
2738 2. Added /M option to pcretest to show the memory requirement of a specific
2739 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
2740
2741 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
2742 compiled in such a way that the backtracking after subsequent failure was
2743 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
2744 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
2745
2746
2747 Version 2.03 02-Feb-99
2748 ----------------------
2749
2750 1. Fixed typo and small mistake in man page.
2751
2752 2. Added 4th condition (GPL supersedes if conflict) and created separate
2753 LICENCE file containing the conditions.
2754
2755 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
2756 Perl, that is the internal \ allows the delimiter to be included in the
2757 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
2758 the final delimiter, add \ to the end of the pattern (to test the error).
2759
2760 4. Added the convenience functions for extracting substrings after a successful
2761 match. Updated pcretest to make it able to test these functions.
2762
2763
2764 Version 2.02 14-Jan-99
2765 ----------------------
2766
2767 1. Initialized the working variables associated with each extraction so that
2768 their saving and restoring doesn't refer to uninitialized store.
2769
2770 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
2771 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
2772 fix the problem.
2773
2774 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
2775 calls, and wasn't printing the correct value for compiling calls. Increased the
2776 default value of LOOPREPEAT, and the number of significant figures in the
2777 times.
2778
2779 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
2780
2781 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
2782 a building problem on Windows NT with a FAT file system.
2783
2784
2785 Version 2.01 21-Oct-98
2786 ----------------------
2787
2788 1. Changed the API for pcre_compile() to allow for the provision of a pointer
2789 to character tables built by pcre_maketables() in the current locale. If NULL
2790 is passed, the default tables are used.
2791
2792
2793 Version 2.00 24-Sep-98
2794 ----------------------
2795
2796 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
2797 it any more.
2798
2799 2. Allow quantification of (?>) groups, and make it work correctly.
2800
2801 3. The first character computation wasn't working for (?>) groups.
2802
2803 4. Correct the implementation of \Z (it is permitted to match on the \n at the
2804 end of the subject) and add 5.005's \z, which really does match only at the
2805 very end of the subject.
2806
2807 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
2808
2809 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
2810 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
2811 localized options. All options to pcre_study() were also removed.
2812
2813 7. Add other new features from 5.005:
2814
2815 $(?<= positive lookbehind
2816 $(?<! negative lookbehind
2817 (?imsx-imsx) added the unsetting capability
2818 such a setting is global if at outer level; local otherwise
2819 (?imsx-imsx:) non-capturing groups with option setting
2820 (?(cond)re|re) conditional pattern matching
2821
2822 A backreference to itself in a repeated group matches the previous
2823 captured string.
2824
2825 8. General tidying up of studying (both automatic and via "study")
2826 consequential on the addition of new assertions.
2827
2828 9. As in 5.005, unlimited repeated groups that could match an empty substring
2829 are no longer faulted at compile time. Instead, the loop is forcibly broken at
2830 runtime if any iteration does actually match an empty substring.
2831
2832 10. Include the RunTest script in the distribution.
2833
2834 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
2835 discrepancies, some of which were old and were also with respect to 5.004. They
2836 have now been fixed.
2837
2838
2839 Version 1.09 28-Apr-98
2840 ----------------------
2841
2842 1. A negated single character class followed by a quantifier with a minimum
2843 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
2844 program crashes, or just wrong answers. This did not apply to negated classes
2845 containing more than one character, or to minima other than one.
2846
2847
2848 Version 1.08 27-Mar-98
2849 ----------------------
2850
2851 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
2852
2853 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
2854 latter must appear before anything that relies on it in the pattern.
2855
2856
2857 Version 1.07 16-Feb-98
2858 ----------------------
2859
2860 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
2861 repeat of a potentially empty string).
2862
2863
2864 Version 1.06 23-Jan-98
2865 ----------------------
2866
2867 1. Added Markus Oberhumer's little patches for C++.
2868
2869 2. Literal strings longer than 255 characters were broken.
2870
2871
2872 Version 1.05 23-Dec-97
2873 ----------------------
2874
2875 1. Negated character classes containing more than one character were failing if
2876 PCRE_CASELESS was set at run time.
2877
2878
2879 Version 1.04 19-Dec-97
2880 ----------------------
2881
2882 1. Corrected the man page, where some "const" qualifiers had been omitted.
2883
2884 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
2885 input syntax.
2886
2887 3. Fixed memory leak which occurred when a regex with back references was
2888 matched with an offsets vector that wasn't big enough. The temporary memory
2889 that is used in this case wasn't being freed if the match failed.
2890
2891 4. Tidied pcretest to ensure it frees memory that it gets.
2892
2893 5. Temporary memory was being obtained in the case where the passed offsets
2894 vector was exactly big enough.
2895
2896 6. Corrected definition of offsetof() from change 5 below.
2897
2898 7. I had screwed up change 6 below and broken the rules for the use of
2899 setjmp(). Now fixed.
2900
2901
2902 Version 1.03 18-Dec-97
2903 ----------------------
2904
2905 1. A erroneous regex with a missing opening parenthesis was correctly
2906 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
2907 on some systems.
2908
2909 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
2910 it was reported that one broken compiler failed on the former because "code" is
2911 also an independent variable.
2912
2913 3. The erroneous regex a[]b caused an array overrun reference.
2914
2915 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
2916 fail on data ending with that character. (It was going on too far, and checking
2917 the next character, typically a binary zero.) This was specific to the
2918 optimized code for single-character negative classes.
2919
2920 5. Added a contributed patch from the TIN world which does the following:
2921
2922 + Add an undef for memmove, in case the the system defines a macro for it.
2923
2924 + Add a definition of offsetof(), in case there isn't one. (I don't know
2925 the reason behind this - offsetof() is part of the ANSI standard - but
2926 it does no harm).
2927
2928 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
2929 most of the places where whitespace preceded '#'. I have given up and
2930 allowed the remaining 2 cases to be at the margin.
2931
2932 + Rename some variables in pcre to eliminate shadowing. This seems very
2933 pedantic, but does no harm, of course.
2934
2935 6. Moved the call to setjmp() into its own function, to get rid of warnings
2936 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
2937
2938 7. Constructs such as \d{8,} were compiling into the equivalent of
2939 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
2940 outcome, but in this particular case used more store than had been allocated,
2941 which caused the bug to be discovered because it threw up an internal error.
2942
2943 8. The debugging code in both pcre and pcretest for outputting the compiled
2944 form of a regex was going wrong in the case of back references followed by
2945 curly-bracketed repeats.
2946
2947
2948 Version 1.02 12-Dec-97
2949 ----------------------
2950
2951 1. Typos in pcre.3 and comments in the source fixed.
2952
2953 2. Applied a contributed patch to get rid of places where it used to remove
2954 'const' from variables, and fixed some signed/unsigned and uninitialized
2955 variable warnings.
2956
2957 3. Added the "runtest" target to Makefile.
2958
2959 4. Set default compiler flag to -O2 rather than just -O.
2960
2961
2962 Version 1.01 19-Nov-97
2963 ----------------------
2964
2965 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
2966 like /([ab]*)*/, that is, for classes with more than one character in them.
2967
2968 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
2969 as /((?>a*))*/ (a PCRE_EXTRA facility).
2970
2971
2972 Version 1.00 18-Nov-97
2973 ----------------------
2974
2975 1. Added compile-time macros to support systems such as SunOS4 which don't have
2976 memmove() or strerror() but have other things that can be used instead.
2977
2978 2. Arranged that "make clean" removes the executables.
2979
2980
2981 Version 0.99 27-Oct-97
2982 ----------------------
2983
2984 1. Fixed bug in code for optimizing classes with only one character. It was
2985 initializing a 32-byte map regardless, which could cause it to run off the end
2986 of the memory it had got.
2987
2988 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
2989
2990
2991 Version 0.98 22-Oct-97
2992 ----------------------
2993
2994 1. Fixed bug in code for handling temporary memory usage when there are more
2995 back references than supplied space in the ovector. This could cause segfaults.
2996
2997
2998 Version 0.97 21-Oct-97
2999 ----------------------
3000
3001 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
3002
3003 2. Optimized negated single characters not to use a bit map.
3004
3005 3. Brought error texts together as macro definitions; clarified some of them;
3006 fixed one that was wrong - it said "range out of order" when it meant "invalid
3007 escape sequence".
3008
3009 4. Changed some char * arguments to const char *.
3010
3011 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
3012
3013 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
3014 pcretest.
3015
3016
3017 Version 0.96 16-Oct-97
3018 ----------------------
3019
3020 1. Added a simple "pgrep" utility to the distribution.
3021
3022 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
3023 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
3024 where "ddd" means "one or more decimal digits".
3025
3026 3. Fixed serious bug. If a pattern had a back reference, but the call to
3027 pcre_exec() didn't supply a large enough ovector to record the related
3028 identifying subpattern, the match always failed. PCRE now remembers the number
3029 of the largest back reference, and gets some temporary memory in which to save
3030 the offsets during matching if necessary, in order to ensure that
3031 backreferences always work.
3032
3033 4. Increased the compatibility with Perl in a number of ways:
3034
3035 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
3036 to request this handling. The option can be set at compile or exec time.
3037
3038 (b) $ matches before a terminating newline by default; an option
3039 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
3040 mode). The option can be set at compile or exec time.
3041
3042 (c) The handling of \ followed by a digit other than 0 is now supposed to be
3043 the same as Perl's. If the decimal number it represents is less than 10
3044 or there aren't that many previous left capturing parentheses, an octal
3045 escape is read. Inside a character class, it's always an octal escape,
3046 even if it is a single digit.
3047
3048 (d) An escaped but undefined alphabetic character is taken as a literal,
3049 unless PCRE_EXTRA is set. Currently this just reserves the remaining
3050 escapes.
3051
3052 (e) {0} is now permitted. (The previous item is removed from the compiled
3053 pattern).
3054
3055 5. Changed all the names of code files so that the basic parts are no longer
3056 than 10 characters, and abolished the teeny "globals.c" file.
3057
3058 6. Changed the handling of character classes; they are now done with a 32-byte
3059 bit map always.
3060
3061 7. Added the -d and /D options to pcretest to make it possible to look at the
3062 internals of compilation without having to recompile pcre.
3063
3064
3065 Version 0.95 23-Sep-97
3066 ----------------------
3067
3068 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
3069 \x20 at the start of a run of normal characters. These were being treated as
3070 real characters, instead of the source characters being re-checked.
3071
3072
3073 Version 0.94 18-Sep-97
3074 ----------------------
3075
3076 1. The functions are now thread-safe, with the caveat that the global variables
3077 containing pointers to malloc() and free() or alternative functions are the
3078 same for all threads.
3079
3080 2. Get pcre_study() to generate a bitmap of initial characters for non-
3081 anchored patterns when this is possible, and use it if passed to pcre_exec().
3082
3083
3084 Version 0.93 15-Sep-97
3085 ----------------------
3086
3087 1. /(b)|(:+)/ was computing an incorrect first character.
3088
3089 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
3090 but not actually doing anything yet.
3091
3092 3. Treat "-" characters in classes that cannot be part of ranges as literals,
3093 as Perl does (e.g. [-az] or [az-]).
3094
3095 4. Set the anchored flag if a branch starts with .* or .*? because that tests
3096 all possible positions.
3097
3098 5. Split up into different modules to avoid including unneeded functions in a
3099 compiled binary. However, compile and exec are still in one module. The "study"
3100 function is split off.
3101
3102 6. The character tables are now in a separate module whose source is generated
3103 by an auxiliary program - but can then be edited by hand if required. There are
3104 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
3105 toupper() in the code.
3106
3107 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
3108 make them global. Abolish the function for setting them, as the caller can now
3109 set them directly.
3110
3111
3112 Version 0.92 11-Sep-97
3113 ----------------------
3114
3115 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
3116 (e.g. /a{1,3}/) was broken (I mis-optimized it).
3117
3118 2. Caseless matching was not working in character classes if the characters in
3119 the pattern were in upper case.
3120
3121 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
3122
3123 4. Make PCRE_ANCHORED public and accept as a compile option.
3124
3125 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
3126 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
3127 pass them.
3128
3129 6. Give an error if bad option bits passed at compile or run time.
3130
3131 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
3132 pcretest to cause it to pass that flag.
3133
3134 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
3135 options, and the first character, if set.
3136
3137 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
3138
3139
3140 Version 0.91 10-Sep-97
3141 ----------------------
3142
3143 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
3144 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
3145
3146 2. PCRE was looping on encountering an indefinitely repeated back reference to
3147 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
3148 Perl does - treats the match as successful.
3149
3150 ****

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