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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 300 - (show annotations) (download)
Mon Jan 14 19:43:19 2008 UTC (6 years, 3 months ago) by ph10
File size: 148714 byte(s)
Fix buffer overrun for class with very many characters whose codepoints are 
above 255.

1 ChangeLog for PCRE
2 ------------------
3
4 Version 7.6 14-Jan-08
5 ---------------------
6
7 1. A class with a very large number of characters with codepoints greater
8 than 255 (in UTF-8 mode, of course), caused a buffer overflow.
9
10
11 Version 7.5 10-Jan-08
12 ---------------------
13
14 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
15 values in parens when parsing an RE using the C++ wrapper."
16
17 2. Negative specials like \S did not work in character classes in UTF-8 mode.
18 Characters greater than 255 were excluded from the class instead of being
19 included.
20
21 3. The same bug as (2) above applied to negated POSIX classes such as
22 [:^space:].
23
24 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
25 defined or documented. It seems to have been a typo for PCRE_STATIC, so
26 I have changed it.
27
28 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
29 first named subpattern) and a construct such as (?&a) would reference the
30 first named subpattern whose name started with "a" (in other words, the
31 length check was missing). Both these problems are fixed. "Subpattern name
32 expected" is now given for (?&) (a zero-length name), and this patch also
33 makes it give the same error for \k'' (previously it complained that that
34 was a reference to a non-existent subpattern).
35
36 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
37 this is right because (?- can be followed by option settings as well as by
38 digits. I have, however, made the messages clearer.
39
40 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
41 than the number used in the conditional) now cause a compile-time error.
42 This is actually not compatible with Perl, which accepts such patterns, but
43 treats the conditional as always being FALSE (as PCRE used to), but it
44 seems to me that giving a diagnostic is better.
45
46 8. Change "alphameric" to the more common word "alphanumeric" in comments
47 and messages.
48
49 9. Fix two occurrences of "backslash" in comments that should have been
50 "backspace".
51
52 10. Remove two redundant lines of code that can never be obeyed (their function
53 was moved elsewhere).
54
55 11. The program that makes PCRE's Unicode character property table had a bug
56 which caused it to generate incorrect table entries for sequences of
57 characters that have the same character type, but are in different scripts.
58 It amalgamated them into a single range, with the script of the first of
59 them. In other words, some characters were in the wrong script. There were
60 thirteen such cases, affecting characters in the following ranges:
61
62 U+002b0 - U+002c1
63 U+0060c - U+0060d
64 U+0061e - U+00612
65 U+0064b - U+0065e
66 U+0074d - U+0076d
67 U+01800 - U+01805
68 U+01d00 - U+01d77
69 U+01d9b - U+01dbf
70 U+0200b - U+0200f
71 U+030fc - U+030fe
72 U+03260 - U+0327f
73 U+0fb46 - U+0fbb1
74 U+10450 - U+1049d
75
76 12. The -o option (show only the matching part of a line) for pcregrep was not
77 compatible with GNU grep in that, if there was more than one match in a
78 line, it showed only the first of them. It now behaves in the same way as
79 GNU grep.
80
81 13. If the -o and -v options were combined for pcregrep, it printed a blank
82 line for every non-matching line. GNU grep prints nothing, and pcregrep now
83 does the same. The return code can be used to tell if there were any
84 non-matching lines.
85
86 14. Added --file-offsets and --line-offsets to pcregrep.
87
88 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
89 infinitely looping recursion. The bug was that positive lookaheads were not
90 being skipped when checking for a possible empty match (negative lookaheads
91 and both kinds of lookbehind were skipped).
92
93 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
94 inclusion of <windows.h> to before rather than after the definition of
95 INVALID_FILE_ATTRIBUTES (patch from David Byron).
96
97 17. Specifying a possessive quantifier with a specific limit for a Unicode
98 character property caused pcre_compile() to compile bad code, which led at
99 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
100 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
101 caused the error; without that there was no problem.
102
103 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
104
105 19. Added --enable-pcretest-libreadline.
106
107 20. In pcrecpp.cc, the variable 'count' was incremented twice in
108 RE::GlobalReplace(). As a result, the number of replacements returned was
109 double what it should be. I removed one of the increments, but Craig sent a
110 later patch that removed the other one (the right fix) and added unit tests
111 that check the return values (which was not done before).
112
113 21. Several CMake things:
114
115 (1) Arranged that, when cmake is used on Unix, the libraries end up with
116 the names libpcre and libpcreposix, not just pcre and pcreposix.
117
118 (2) The above change means that pcretest and pcregrep are now correctly
119 linked with the newly-built libraries, not previously installed ones.
120
121 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
122
123 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
124 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
125 UTF-8 newline character). The key issue is that the pattern starts .*;
126 this means that the match must be either at the beginning, or after a
127 newline. The bug was in the code for advancing after a failed match and
128 checking that the new position followed a newline. It was not taking
129 account of UTF-8 characters correctly.
130
131 23. PCRE was behaving differently from Perl in the way it recognized POSIX
132 character classes. PCRE was not treating the sequence [:...:] as a
133 character class unless the ... were all letters. Perl, however, seems to
134 allow any characters between [: and :], though of course it rejects as
135 unknown any "names" that contain non-letters, because all the known class
136 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
137 for example, whereas PCRE did not - it did not recognize a POSIX character
138 class. This seemed a bit dangerous, so the code has been changed to be
139 closer to Perl. The behaviour is not identical to Perl, because PCRE will
140 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
141 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
142 Perl does, and where it didn't before.
143
144 24. Rewrite so as to remove the single use of %n from pcregrep because in some
145 Windows environments %n is disabled by default.
146
147
148 Version 7.4 21-Sep-07
149 ---------------------
150
151 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
152 means that a class such as [\s] counted as "explicit reference to CR or
153 LF". That isn't really right - the whole point of the change was to try to
154 help when there was an actual mention of one of the two characters. So now
155 the change happens only if \r or \n (or a literal CR or LF) character is
156 encountered.
157
158 2. The 32-bit options word was also used for 6 internal flags, but the numbers
159 of both had grown to the point where there were only 3 bits left.
160 Fortunately, there was spare space in the data structure, and so I have
161 moved the internal flags into a new 16-bit field to free up more option
162 bits.
163
164 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
165 but did not set the internal JCHANGED flag - either of these is enough to
166 control the way the "get" function works - but the PCRE_INFO_JCHANGED
167 facility is supposed to tell if (?J) was ever used, so now (?J) at the
168 start sets both bits.
169
170 4. Added options (at build time, compile time, exec time) to change \R from
171 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
172
173 5. doc/pcresyntax.html was missing from the distribution.
174
175 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
176 compatibility, even though it is no longer used.
177
178 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
179 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
180 windows.h file is present (where different names are used). [This was
181 reversed later after testing - see 16 below.]
182
183 8. Changed all #include <config.h> to #include "config.h". There were also
184 some further <pcre.h> cases that I changed to "pcre.h".
185
186 9. When pcregrep was used with the --colour option, it missed the line ending
187 sequence off the lines that it output.
188
189 10. It was pointed out to me that arrays of string pointers cause lots of
190 relocations when a shared library is dynamically loaded. A technique of
191 using a single long string with a table of offsets can drastically reduce
192 these. I have refactored PCRE in four places to do this. The result is
193 dramatic:
194
195 Originally: 290
196 After changing UCP table: 187
197 After changing error message table: 43
198 After changing table of "verbs" 36
199 After changing table of Posix names 22
200
201 Thanks to the folks working on Gregex for glib for this insight.
202
203 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
204 unicode-properties was also set.
205
206 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
207
208 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
209 checked only for CRLF.
210
211 14. Added casts to pcretest.c to avoid compiler warnings.
212
213 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
214
215 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
216 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
217 entirely. This removes changes made in 7 above.
218
219 17. The CMake files have been updated, and there is now more information about
220 building with CMake in the NON-UNIX-USE document.
221
222
223 Version 7.3 28-Aug-07
224 ---------------------
225
226 1. In the rejigging of the build system that eventually resulted in 7.1, the
227 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
228 brackets there is not right, since it causes compilers to look for an
229 installed pcre.h, not the version that is in the source that is being
230 compiled (which of course may be different). I have changed it back to:
231
232 #include "pcre.h"
233
234 I have a vague recollection that the change was concerned with compiling in
235 different directories, but in the new build system, that is taken care of
236 by the VPATH setting the Makefile.
237
238 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
239 when the subject happened to end in the byte 0x85 (e.g. if the last
240 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
241 characters but of course it shouldn't be taken as a newline when it is part
242 of another character. The bug was that, for an unlimited repeat of . in
243 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
244 characters when looking for a newline.
245
246 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
247
248 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
249 in debug output.
250
251 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
252 long printing in the pcrecpp unittest when running under MinGW.
253
254 6. ESC_K was left out of the EBCDIC table.
255
256 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
257 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
258 limit also applies to "virtual nesting" when a pattern is recursive, and in
259 this case 1000 isn't so big. I have been able to remove this limit at the
260 expense of backing off one optimization in certain circumstances. Normally,
261 when pcre_exec() would call its internal match() function recursively and
262 immediately return the result unconditionally, it uses a "tail recursion"
263 feature to save stack. However, when a subpattern that can match an empty
264 string has an unlimited repetition quantifier, it no longer makes this
265 optimization. That gives it a stack frame in which to save the data for
266 checking that an empty string has been matched. Previously this was taken
267 from the 1000-entry workspace that had been reserved. So now there is no
268 explicit limit, but more stack is used.
269
270 8. Applied Daniel's patches to solve problems with the import/export magic
271 syntax that is required for Windows, and which was going wrong for the
272 pcreposix and pcrecpp parts of the library. These were overlooked when this
273 problem was solved for the main library.
274
275 9. There were some crude static tests to avoid integer overflow when computing
276 the size of patterns that contain repeated groups with explicit upper
277 limits. As the maximum quantifier is 65535, the maximum group length was
278 set at 30,000 so that the product of these two numbers did not overflow a
279 32-bit integer. However, it turns out that people want to use groups that
280 are longer than 30,000 bytes (though not repeat them that many times).
281 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
282 made it possible to implement the integer overflow checks in a much more
283 dynamic way, which I have now done. The artificial limitation on group
284 length has been removed - we now have only the limit on the total length of
285 the compiled pattern, which depends on the LINK_SIZE setting.
286
287 10. Fixed a bug in the documentation for get/copy named substring when
288 duplicate names are permitted. If none of the named substrings are set, the
289 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
290 empty string.
291
292 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
293 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
294 because the ] is interpreted as the first data character and the
295 terminating ] is not found. PCRE has been made compatible with Perl in this
296 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
297 cause memory overwriting.
298
299 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
300 string has been matched (to stop an infinite loop). It was not recognizing
301 a conditional subpattern that could match an empty string if that
302 subpattern was within another subpattern. For example, it looped when
303 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
304 condition was not nested. This bug has been fixed.
305
306 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
307 past the start of the subject in the presence of bytes with the top bit
308 set, for example "\x8aBCD".
309
310 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
311 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
312
313 14. Optimized (?!) to (*FAIL).
314
315 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
316 This restricts code points to be within the range 0 to 0x10FFFF, excluding
317 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
318 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
319 does: it's just the validity check that is more restrictive.
320
321 16. Inserted checks for integer overflows during escape sequence (backslash)
322 processing, and also fixed erroneous offset values for syntax errors during
323 backslash processing.
324
325 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
326 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
327
328 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
329 caused an overrun.
330
331 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
332 something other than just ASCII characters) inside a group that had an
333 unlimited repeat caused a loop at compile time (while checking to see
334 whether the group could match an empty string).
335
336 20. Debugging a pattern containing \p or \P could cause a crash. For example,
337 [\P{Any}] did so. (Error in the code for printing property names.)
338
339 21. An orphan \E inside a character class could cause a crash.
340
341 22. A repeated capturing bracket such as (A)? could cause a wild memory
342 reference during compilation.
343
344 23. There are several functions in pcre_compile() that scan along a compiled
345 expression for various reasons (e.g. to see if it's fixed length for look
346 behind). There were bugs in these functions when a repeated \p or \P was
347 present in the pattern. These operators have additional parameters compared
348 with \d, etc, and these were not being taken into account when moving along
349 the compiled data. Specifically:
350
351 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
352 length.
353
354 (b) An item such as \pL+ within a repeated group could cause crashes or
355 loops.
356
357 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
358 "reference to non-existent subpattern" error.
359
360 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
361
362 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
363 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
364
365 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
366
367 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
368 character were causing crashes (broken optimization).
369
370 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
371 \p or \P) caused a compile-time loop.
372
373 28. More problems have arisen in unanchored patterns when CRLF is a valid line
374 break. For example, the unstudied pattern [\r\n]A does not match the string
375 "\r\nA" because change 7.0/46 below moves the current point on by two
376 characters after failing to match at the start. However, the pattern \nA
377 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
378 the same is true. There doesn't seem any very clean way out of this, but
379 what I have chosen to do makes the common cases work: PCRE now takes note
380 of whether there can be an explicit match for \r or \n anywhere in the
381 pattern, and if so, 7.0/46 no longer applies. As part of this change,
382 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
383 pattern has explicit CR or LF references.
384
385 29. Added (*CR) etc for changing newline setting at start of pattern.
386
387
388 Version 7.2 19-Jun-07
389 ---------------------
390
391 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
392 which is apparently normally available under Windows.
393
394 2. Re-jig the pcregrep tests with different newline settings in an attempt
395 to make them independent of the local environment's newline setting.
396
397 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
398
399 4. Some of the "internals" tests were previously cut out when the link size
400 was not 2, because the output contained actual offsets. The recent new
401 "Z" feature of pcretest means that these can be cut out, making the tests
402 usable with all link sizes.
403
404 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
405 stack recursion. This gives a massive performance boost under BSD, but just
406 a small improvement under Linux. However, it saves one field in the frame
407 in all cases.
408
409 6. Added more features from the forthcoming Perl 5.10:
410
411 (a) (?-n) (where n is a string of digits) is a relative subroutine or
412 recursion call. It refers to the nth most recently opened parentheses.
413
414 (b) (?+n) is also a relative subroutine call; it refers to the nth next
415 to be opened parentheses.
416
417 (c) Conditions that refer to capturing parentheses can be specified
418 relatively, for example, (?(-2)... or (?(+3)...
419
420 (d) \K resets the start of the current match so that everything before
421 is not part of it.
422
423 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
424
425 (f) \g{name} is another synonym - part of Perl 5.10's unification of
426 reference syntax.
427
428 (g) (?| introduces a group in which the numbering of parentheses in each
429 alternative starts with the same number.
430
431 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
432
433 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
434 PCRE_INFO_JCHANGED.
435
436 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
437 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
438 for detecting groups that can match an empty string.
439
440 9. A pattern with a very large number of alternatives (more than several
441 hundred) was running out of internal workspace during the pre-compile
442 phase, where pcre_compile() figures out how much memory will be needed. A
443 bit of new cunning has reduced the workspace needed for groups with
444 alternatives. The 1000-alternative test pattern now uses 12 bytes of
445 workspace instead of running out of the 4096 that are available.
446
447 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
448
449 11. Applied patch from Google to remove an optimization that didn't quite work.
450 The report of the bug said:
451
452 pcrecpp::RE("a*").FullMatch("aaa") matches, while
453 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
454 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
455
456 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
457 it matched the wrong number of bytes.
458
459
460 Version 7.1 24-Apr-07
461 ---------------------
462
463 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
464 that is more "standard", making use of automake and other Autotools. There
465 is some re-arrangement of the files and adjustment of comments consequent
466 on this.
467
468 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
469 for recursive directory scanning broke on some systems because the files
470 are not scanned in any specific order and on different systems the order
471 was different. A call to "sort" has been inserted into RunGrepTest for the
472 approprate test as a short-term fix. In the longer term there may be an
473 alternative.
474
475 3. I had an email from Eric Raymond about problems translating some of PCRE's
476 man pages to HTML (despite the fact that I distribute HTML pages, some
477 people do their own conversions for various reasons). The problems
478 concerned the use of low-level troff macros .br and .in. I have therefore
479 removed all such uses from the man pages (some were redundant, some could
480 be replaced by .nf/.fi pairs). The 132html script that I use to generate
481 HTML has been updated to handle .nf/.fi and to complain if it encounters
482 .br or .in.
483
484 4. Updated comments in configure.ac that get placed in config.h.in and also
485 arranged for config.h to be included in the distribution, with the name
486 config.h.generic, for the benefit of those who have to compile without
487 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
488
489 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
490 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
491 makevp.bat for the current PCRE, using the additional files
492 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
493
494 6. A Windows user reported a minor discrepancy with test 2, which turned out
495 to be caused by a trailing space on an input line that had got lost in his
496 copy. The trailing space was an accident, so I've just removed it.
497
498 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
499 that is needed.
500
501 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
502 as "const" (a) because they are and (b) because it helps the PHP
503 maintainers who have recently made a script to detect big data structures
504 in the php code that should be moved to the .rodata section. I remembered
505 to update Builducptable as well, so it won't revert if ucptable.h is ever
506 re-created.
507
508 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
509 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
510 order to be able to cut out the UTF-8 tables in the latter when UTF-8
511 support is not required. This saves 1.5-2K of code, which is important in
512 some applications.
513
514 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
515 so as not to refer to the tables, even though these functions will never be
516 called when UTF-8 support is disabled. Otherwise there are problems with a
517 shared library.
518
519 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
520
521 (a) It was defining its arguments as char * instead of void *.
522
523 (b) It was assuming that all moves were upwards in memory; this was true
524 a long time ago when I wrote it, but is no longer the case.
525
526 The emulated memove() is provided for those environments that have neither
527 memmove() nor bcopy(). I didn't think anyone used it these days, but that
528 is clearly not the case, as these two bugs were recently reported.
529
530 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
531 and Detrail to create the HTML documentation, the .txt form of the man
532 pages, and it removes trailing spaces from listed files. It also creates
533 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
534 case, it wraps all the #defines with #ifndefs. This script should be run
535 before "make dist".
536
537 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
538 with Unicode property support.
539
540 (a) For a maximizing quantifier, if the two different cases of the
541 character were of different lengths in their UTF-8 codings (there are
542 some cases like this - I found 11), and the matching function had to
543 back up over a mixture of the two cases, it incorrectly assumed they
544 were both the same length.
545
546 (b) When PCRE was configured to use the heap rather than the stack for
547 recursion during matching, it was not correctly preserving the data for
548 the other case of a UTF-8 character when checking ahead for a match
549 while processing a minimizing repeat. If the check also involved
550 matching a wide character, but failed, corruption could cause an
551 erroneous result when trying to check for a repeat of the original
552 character.
553
554 13. Some tidying changes to the testing mechanism:
555
556 (a) The RunTest script now detects the internal link size and whether there
557 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
558 values substituted by "configure". (The RunGrepTest script already did
559 this for UTF-8.) The configure.ac script no longer substitutes the
560 relevant variables.
561
562 (b) The debugging options /B and /D in pcretest show the compiled bytecode
563 with length and offset values. This means that the output is different
564 for different internal link sizes. Test 2 is skipped for link sizes
565 other than 2 because of this, bypassing the problem. Unfortunately,
566 there was also a test in test 3 (the locale tests) that used /B and
567 failed for link sizes other than 2. Rather than cut the whole test out,
568 I have added a new /Z option to pcretest that replaces the length and
569 offset values with spaces. This is now used to make test 3 independent
570 of link size. (Test 2 will be tidied up later.)
571
572 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
573 segmentation fault instead of returning the appropriate error message.
574
575 15. In multiline mode when the newline sequence was set to "any", the pattern
576 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
577 This doesn't seem right; it now treats the CRLF combination as the line
578 ending, and so does not match in that case. It's only a pattern such as ^$
579 that would hit this one: something like ^ABC$ would have failed after \r
580 and then tried again after \r\n.
581
582 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
583 in an attempt to make files that differ only in their line terminators
584 compare equal. This works on Linux.
585
586 17. Under certain error circumstances pcregrep might try to free random memory
587 as it exited. This is now fixed, thanks to valgrind.
588
589 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
590 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
591 was because its rules for how to advance for /g after matching an empty
592 string at the end of a line did not allow for this case. They now check for
593 it specially.
594
595 20. pcretest is supposed to handle patterns and data of any length, by
596 extending its buffers when necessary. It was getting this wrong when the
597 buffer for a data line had to be extended.
598
599 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
600 CRLF as a newline sequence.
601
602 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
603 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
604 I have nevertheless tidied it up.
605
606 23. Added some casts to kill warnings from HP-UX ia64 compiler.
607
608 24. Added a man page for pcre-config.
609
610
611 Version 7.0 19-Dec-06
612 ---------------------
613
614 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
615 moving to gcc 4.1.1.
616
617 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
618 sys/time.h, which is documented as needed for this function. It doesn't
619 seem to matter on Linux, but it showed up on some releases of OS X.
620
621 3. It seems that there are systems where bytes whose values are greater than
622 127 match isprint() in the "C" locale. The "C" locale should be the
623 default when a C program starts up. In most systems, only ASCII printing
624 characters match isprint(). This difference caused the output from pcretest
625 to vary, making some of the tests fail. I have changed pcretest so that:
626
627 (a) When it is outputting text in the compiled version of a pattern, bytes
628 other than 32-126 are always shown as hex escapes.
629
630 (b) When it is outputting text that is a matched part of a subject string,
631 it does the same, unless a different locale has been set for the match
632 (using the /L modifier). In this case, it uses isprint() to decide.
633
634 4. Fixed a major bug that caused incorrect computation of the amount of memory
635 required for a compiled pattern when options that changed within the
636 pattern affected the logic of the preliminary scan that determines the
637 length. The relevant options are -x, and -i in UTF-8 mode. The result was
638 that the computed length was too small. The symptoms of this bug were
639 either the PCRE error "internal error: code overflow" from pcre_compile(),
640 or a glibc crash with a message such as "pcretest: free(): invalid next
641 size (fast)". Examples of patterns that provoked this bug (shown in
642 pcretest format) are:
643
644 /(?-x: )/x
645 /(?x)(?-x: \s*#\s*)/
646 /((?i)[\x{c0}])/8
647 /(?i:[\x{c0}])/8
648
649 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
650 is now done differently.
651
652 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
653 wrapper classes; (b) implement a new function in the C++ scanner that is
654 more efficient than the old way of doing things because it avoids levels of
655 recursion in the regex matching; (c) add a paragraph to the documentation
656 for the FullMatch() function.
657
658 6. The escape sequence \n was being treated as whatever was defined as
659 "newline". Not only was this contrary to the documentation, which states
660 that \n is character 10 (hex 0A), but it also went horribly wrong when
661 "newline" was defined as CRLF. This has been fixed.
662
663 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
664 was being set to -1 for the "end of line" case (supposedly a value that no
665 character can have). Though this value is never used (the check for end of
666 line is "zero bytes in current character"), it caused compiler complaints.
667 I've changed it to 0xffffffff.
668
669 8. In pcre_version.c, the version string was being built by a sequence of
670 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
671 string (as it is for production releases) called a macro with an empty
672 argument. The C standard says the result of this is undefined. The gcc
673 compiler treats it as an empty string (which was what was wanted) but it is
674 reported that Visual C gives an error. The source has been hacked around to
675 avoid this problem.
676
677 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
678 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
679 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
680 of them did).
681
682 10. Originally, pcretest opened its input and output without "b"; then I was
683 told that "b" was needed in some environments, so it was added for release
684 5.0 to both the input and output. (It makes no difference on Unix-like
685 systems.) Later I was told that it is wrong for the input on Windows. I've
686 now abstracted the modes into two macros, to make it easier to fiddle with
687 them, and removed "b" from the input mode under Windows.
688
689 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
690
691 12. Added -help and --help to pcretest as an official way of being reminded
692 of the options.
693
694 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
695 and pcrecpp.cc because they annoy compilers at high warning levels.
696
697 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
698
699 15. Fixed an occurrence of == in configure.ac that should have been = (shell
700 scripts are not C programs :-) and which was not noticed because it works
701 on Linux.
702
703 16. pcretest is supposed to handle any length of pattern and data line (as one
704 line or as a continued sequence of lines) by extending its input buffer if
705 necessary. This feature was broken for very long pattern lines, leading to
706 a string of junk being passed to pcre_compile() if the pattern was longer
707 than about 50K.
708
709 17. I have done a major re-factoring of the way pcre_compile() computes the
710 amount of memory needed for a compiled pattern. Previously, there was code
711 that made a preliminary scan of the pattern in order to do this. That was
712 OK when PCRE was new, but as the facilities have expanded, it has become
713 harder and harder to keep it in step with the real compile phase, and there
714 have been a number of bugs (see for example, 4 above). I have now found a
715 cunning way of running the real compile function in a "fake" mode that
716 enables it to compute how much memory it would need, while actually only
717 ever using a few hundred bytes of working memory and without too many
718 tests of the mode. This should make future maintenance and development
719 easier. A side effect of this work is that the limit of 200 on the nesting
720 depth of parentheses has been removed (though this was never a serious
721 limitation, I suspect). However, there is a downside: pcre_compile() now
722 runs more slowly than before (30% or more, depending on the pattern). I
723 hope this isn't a big issue. There is no effect on runtime performance.
724
725 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
726 newline (only possible for the last line of a file) and it was a
727 pattern that set a locale (followed by /Lsomething), pcretest crashed.
728
729 19. Added additional timing features to pcretest. (1) The -tm option now times
730 matching only, not compiling. (2) Both -t and -tm can be followed, as a
731 separate command line item, by a number that specifies the number of
732 repeats to use when timing. The default is 50000; this gives better
733 precision, but takes uncomfortably long for very large patterns.
734
735 20. Extended pcre_study() to be more clever in cases where a branch of a
736 subpattern has no definite first character. For example, (a*|b*)[cd] would
737 previously give no result from pcre_study(). Now it recognizes that the
738 first character must be a, b, c, or d.
739
740 21. There was an incorrect error "recursive call could loop indefinitely" if
741 a subpattern (or the entire pattern) that was being tested for matching an
742 empty string contained only one non-empty item after a nested subpattern.
743 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
744 incorrectly, because the \d was being skipped in the check.
745
746 22. The pcretest program now has a new pattern option /B and a command line
747 option -b, which is equivalent to adding /B to every pattern. This causes
748 it to show the compiled bytecode, without the additional information that
749 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
750 is the same as /B/I).
751
752 23. A new optimization is now able automatically to treat some sequences such
753 as a*b as a*+b. More specifically, if something simple (such as a character
754 or a simple class like \d) has an unlimited quantifier, and is followed by
755 something that cannot possibly match the quantified thing, the quantifier
756 is automatically "possessified".
757
758 24. A recursive reference to a subpattern whose number was greater than 39
759 went wrong under certain circumstances in UTF-8 mode. This bug could also
760 have affected the operation of pcre_study().
761
762 25. Realized that a little bit of performance could be had by replacing
763 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
764
765 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
766
767 27. Possessive quantifiers such as a++ were previously implemented by turning
768 them into atomic groups such as ($>a+). Now they have their own opcodes,
769 which improves performance. This includes the automatically created ones
770 from 23 above.
771
772 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
773 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
774 the first matched character to be a colon. This applied both to named and
775 numbered groups.
776
777 29. The ucpinternal.h header file was missing its idempotency #ifdef.
778
779 30. I was sent a "project" file called libpcre.a.dev which I understand makes
780 building PCRE on Windows easier, so I have included it in the distribution.
781
782 31. There is now a check in pcretest against a ridiculously large number being
783 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
784 loop, the loop is abandoned.
785
786 32. Forward references to subpatterns in conditions such as (?(2)...) where
787 subpattern 2 is defined later cause pcre_compile() to search forwards in
788 the pattern for the relevant set of parentheses. This search went wrong
789 when there were unescaped parentheses in a character class, parentheses
790 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
791
792 33. "Subroutine" calls and backreferences were previously restricted to
793 referencing subpatterns earlier in the regex. This restriction has now
794 been removed.
795
796 34. Added a number of extra features that are going to be in Perl 5.10. On the
797 whole, these are just syntactic alternatives for features that PCRE had
798 previously implemented using the Python syntax or my own invention. The
799 other formats are all retained for compatibility.
800
801 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
802 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
803 also .NET compatible.
804
805 (b) A recursion or subroutine call to a named group can now be defined as
806 (?&name) as well as (?P>name).
807
808 (c) A backreference to a named group can now be defined as \k<name> or
809 \k'name' as well as (?P=name). The new forms, as well as being in Perl
810 5.10, are also .NET compatible.
811
812 (d) A conditional reference to a named group can now use the syntax
813 (?(<name>) or (?('name') as well as (?(name).
814
815 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
816 groups (named and numbered) that are never evaluated inline, but can be
817 called as "subroutines" from elsewhere. In effect, the DEFINE condition
818 is always false. There may be only one alternative in such a group.
819
820 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
821 as the simple (?(R). The condition is true only if the most recent
822 recursion is that of the given number or name. It does not search out
823 through the entire recursion stack.
824
825 (g) The escape \gN or \g{N} has been added, where N is a positive or
826 negative number, specifying an absolute or relative reference.
827
828 35. Tidied to get rid of some further signed/unsigned compiler warnings and
829 some "unreachable code" warnings.
830
831 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
832 things, this adds five new scripts.
833
834 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
835 There were also incompatibilities regarding the handling of \Q..\E inside
836 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
837 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
838
839 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
840 matches an empty string, and forcibly breaks the loop. There were bugs in
841 this code in non-simple cases. For a pattern such as ^(a()*)* matched
842 against aaaa the result was just "a" rather than "aaaa", for example. Two
843 separate and independent bugs (that affected different cases) have been
844 fixed.
845
846 39. Refactored the code to abolish the use of different opcodes for small
847 capturing bracket numbers. This is a tidy that I avoided doing when I
848 removed the limit on the number of capturing brackets for 3.5 back in 2001.
849 The new approach is not only tidier, it makes it possible to reduce the
850 memory needed to fix the previous bug (38).
851
852 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
853 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
854 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
855 mode.
856
857 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
858 report.
859
860 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
861 copy construction and assignment in the C++ wrapper.
862
863 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
864 couple of bugs that could have given wrong results in the "--newline=crlf"
865 case.
866
867 44. Added a number of casts and did some reorganization of signed/unsigned int
868 variables following suggestions from Dair Grant. Also renamed the variable
869 "this" as "item" because it is a C++ keyword.
870
871 45. Arranged for dftables to add
872
873 #include "pcre_internal.h"
874
875 to pcre_chartables.c because without it, gcc 4.x may remove the array
876 definition from the final binary if PCRE is built into a static library and
877 dead code stripping is activated.
878
879 46. For an unanchored pattern, if a match attempt fails at the start of a
880 newline sequence, and the newline setting is CRLF or ANY, and the next two
881 characters are CRLF, advance by two characters instead of one.
882
883
884 Version 6.7 04-Jul-06
885 ---------------------
886
887 1. In order to handle tests when input lines are enormously long, pcretest has
888 been re-factored so that it automatically extends its buffers when
889 necessary. The code is crude, but this _is_ just a test program. The
890 default size has been increased from 32K to 50K.
891
892 2. The code in pcre_study() was using the value of the re argument before
893 testing it for NULL. (Of course, in any sensible call of the function, it
894 won't be NULL.)
895
896 3. The memmove() emulation function in pcre_internal.h, which is used on
897 systems that lack both memmove() and bcopy() - that is, hardly ever -
898 was missing a "static" storage class specifier.
899
900 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
901 containing an extended class (one that cannot be represented by a bitmap
902 because it contains high-valued characters or Unicode property items, e.g.
903 [\pZ]). Almost always one would set UTF-8 mode when processing such a
904 pattern, but PCRE should not loop if you do not (it no longer does).
905 [Detail: two cases were found: (a) a repeated subpattern containing an
906 extended class; (b) a recursive reference to a subpattern that followed a
907 previous extended class. It wasn't skipping over the extended class
908 correctly when UTF-8 mode was not set.]
909
910 5. A negated single-character class was not being recognized as fixed-length
911 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
912 compile error "lookbehind assertion is not fixed length".
913
914 6. The RunPerlTest auxiliary script was showing an unexpected difference
915 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
916 write a Perl script that can interpret lines of an input file either as
917 byte characters or as UTF-8, which is what "perltest" was being required to
918 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
919 can't do is switch easily at run time between having the "use utf8;" pragma
920 or not. In the end, I fudged it by using the RunPerlTest script to insert
921 "use utf8;" explicitly for the UTF-8 tests.
922
923 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
924 the end of the subject string, contrary to the documentation and to what
925 Perl does. This was true of both matching functions. Now it matches only at
926 the start of the subject and immediately after *internal* newlines.
927
928 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
929 a pointer to an int instead of a pointer to an unsigned long int. This
930 caused problems on 64-bit systems.
931
932 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
933 instance of the 'standard' template library not being so standard".
934
935 10. There was no check on the number of named subpatterns nor the maximum
936 length of a subpattern name. The product of these values is used to compute
937 the size of the memory block for a compiled pattern. By supplying a very
938 long subpattern name and a large number of named subpatterns, the size
939 computation could be caused to overflow. This is now prevented by limiting
940 the length of names to 32 characters, and the number of named subpatterns
941 to 10,000.
942
943 11. Subpatterns that are repeated with specific counts have to be replicated in
944 the compiled pattern. The size of memory for this was computed from the
945 length of the subpattern and the repeat count. The latter is limited to
946 65535, but there was no limit on the former, meaning that integer overflow
947 could in principle occur. The compiled length of a repeated subpattern is
948 now limited to 30,000 bytes in order to prevent this.
949
950 12. Added the optional facility to have named substrings with the same name.
951
952 13. Added the ability to use a named substring as a condition, using the
953 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
954 are numbers (not recommended). Forward references are permitted.
955
956 14. Added forward references in named backreferences (if you see what I mean).
957
958 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
959 pattern could run off the end of the subject. For example, the pattern
960 "(?s)(.{1,5})"8 did this with the subject "ab".
961
962 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
963 PCRE_CASELESS was set when matching characters that were quantified with ?
964 or *.
965
966 17. A character class other than a single negated character that had a minimum
967 but no maximum quantifier - for example [ab]{6,} - was not handled
968 correctly by pce_dfa_exec(). It would match only one character.
969
970 18. A valid (though odd) pattern that looked like a POSIX character
971 class but used an invalid character after [ (for example [[,abc,]]) caused
972 pcre_compile() to give the error "Failed: internal error: code overflow" or
973 in some cases to crash with a glibc free() error. This could even happen if
974 the pattern terminated after [[ but there just happened to be a sequence of
975 letters, a binary zero, and a closing ] in the memory that followed.
976
977 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
978 over the years. Originally (before any Unicode support), just the bottom 8
979 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
980 output from "man perlunicode" includes this:
981
982 The regular expression compiler produces polymorphic opcodes. That
983 is, the pattern adapts to the data and automatically switches to
984 the Unicode character scheme when presented with Unicode data--or
985 instead uses a traditional byte scheme when presented with byte
986 data.
987
988 Sadly, a wide octal escape does not cause a switch, and in a string with
989 no other multibyte characters, these octal escapes are treated as before.
990 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
991 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
992 Unicode string.
993
994 I have not perpetrated such confusion in PCRE. Up till now, it took just
995 the bottom 8 bits, as in old Perl. I have now made octal escapes with
996 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
997 translate to the appropriate multibyte character.
998
999 29. Applied some refactoring to reduce the number of warnings from Microsoft
1000 and Borland compilers. This has included removing the fudge introduced
1001 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
1002 a warning about an unused variable.
1003
1004 21. PCRE has not included VT (character 0x0b) in the set of whitespace
1005 characters since release 4.0, because Perl (from release 5.004) does not.
1006 [Or at least, is documented not to: some releases seem to be in conflict
1007 with the documentation.] However, when a pattern was studied with
1008 pcre_study() and all its branches started with \s, PCRE still included VT
1009 as a possible starting character. Of course, this did no harm; it just
1010 caused an unnecessary match attempt.
1011
1012 22. Removed a now-redundant internal flag bit that recorded the fact that case
1013 dependency changed within the pattern. This was once needed for "required
1014 byte" processing, but is no longer used. This recovers a now-scarce options
1015 bit. Also moved the least significant internal flag bit to the most-
1016 significant bit of the word, which was not previously used (hangover from
1017 the days when it was an int rather than a uint) to free up another bit for
1018 the future.
1019
1020 23. Added support for CRLF line endings as well as CR and LF. As well as the
1021 default being selectable at build time, it can now be changed at runtime
1022 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
1023 specify that it is scanning data with non-default line endings.
1024
1025 24. Changed the definition of CXXLINK to make it agree with the definition of
1026 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
1027
1028 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
1029 recursions. This makes a big different to stack usage for some patterns.
1030
1031 26. If a subpattern containing a named recursion or subroutine reference such
1032 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
1033 the space required for the compiled pattern went wrong and gave too small a
1034 value. Depending on the environment, this could lead to "Failed: internal
1035 error: code overflow at offset 49" or "glibc detected double free or
1036 corruption" errors.
1037
1038 27. Applied patches from Google (a) to support the new newline modes and (b) to
1039 advance over multibyte UTF-8 characters in GlobalReplace.
1040
1041 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
1042 difference for some implementation of PCRE in some Windows version.
1043
1044 29. Added some extra testing facilities to pcretest:
1045
1046 \q<number> in a data line sets the "match limit" value
1047 \Q<number> in a data line sets the "match recursion limt" value
1048 -S <number> sets the stack size, where <number> is in megabytes
1049
1050 The -S option isn't available for Windows.
1051
1052
1053 Version 6.6 06-Feb-06
1054 ---------------------
1055
1056 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
1057 in pcreposix.h. I have copied the definition from pcre.h.
1058
1059 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
1060 because pcre.h is no longer a built file.
1061
1062 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
1063 not normally included in the compiled code.
1064
1065
1066 Version 6.5 01-Feb-06
1067 ---------------------
1068
1069 1. When using the partial match feature with pcre_dfa_exec(), it was not
1070 anchoring the second and subsequent partial matches at the new starting
1071 point. This could lead to incorrect results. For example, with the pattern
1072 /1234/, partially matching against "123" and then "a4" gave a match.
1073
1074 2. Changes to pcregrep:
1075
1076 (a) All non-match returns from pcre_exec() were being treated as failures
1077 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
1078 error message is output. Some extra information is given for the
1079 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
1080 probably the only errors that are likely to be caused by users (by
1081 specifying a regex that has nested indefinite repeats, for instance).
1082 If there are more than 20 of these errors, pcregrep is abandoned.
1083
1084 (b) A binary zero was treated as data while matching, but terminated the
1085 output line if it was written out. This has been fixed: binary zeroes
1086 are now no different to any other data bytes.
1087
1088 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
1089 used to set a locale for matching. The --locale=xxxx long option has
1090 been added (no short equivalent) to specify a locale explicitly on the
1091 pcregrep command, overriding the environment variables.
1092
1093 (d) When -B was used with -n, some line numbers in the output were one less
1094 than they should have been.
1095
1096 (e) Added the -o (--only-matching) option.
1097
1098 (f) If -A or -C was used with -c (count only), some lines of context were
1099 accidentally printed for the final match.
1100
1101 (g) Added the -H (--with-filename) option.
1102
1103 (h) The combination of options -rh failed to suppress file names for files
1104 that were found from directory arguments.
1105
1106 (i) Added the -D (--devices) and -d (--directories) options.
1107
1108 (j) Added the -F (--fixed-strings) option.
1109
1110 (k) Allow "-" to be used as a file name for -f as well as for a data file.
1111
1112 (l) Added the --colo(u)r option.
1113
1114 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
1115 is not present by default.
1116
1117 3. A nasty bug was discovered in the handling of recursive patterns, that is,
1118 items such as (?R) or (?1), when the recursion could match a number of
1119 alternatives. If it matched one of the alternatives, but subsequently,
1120 outside the recursion, there was a failure, the code tried to back up into
1121 the recursion. However, because of the way PCRE is implemented, this is not
1122 possible, and the result was an incorrect result from the match.
1123
1124 In order to prevent this happening, the specification of recursion has
1125 been changed so that all such subpatterns are automatically treated as
1126 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
1127
1128 4. I had overlooked the fact that, in some locales, there are characters for
1129 which isalpha() is true but neither isupper() nor islower() are true. In
1130 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
1131 and ordfeminine) are like this. This affected the treatment of \w and \W
1132 when they appeared in character classes, but not when they appeared outside
1133 a character class. The bit map for "word" characters is now created
1134 separately from the results of isalnum() instead of just taking it from the
1135 upper, lower, and digit maps. (Plus the underscore character, of course.)
1136
1137 5. The above bug also affected the handling of POSIX character classes such as
1138 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
1139 permanent tables. Instead, the bit maps for such a class were previously
1140 created as the appropriate unions of the upper, lower, and digit bitmaps.
1141 Now they are created by subtraction from the [[:word:]] class, which has
1142 its own bitmap.
1143
1144 6. The [[:blank:]] character class matches horizontal, but not vertical space.
1145 It is created by subtracting the vertical space characters (\x09, \x0a,
1146 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
1147 subtraction was done in the overall bitmap for a character class, meaning
1148 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
1149 be recognized. This bug has been fixed.
1150
1151 7. Patches from the folks at Google:
1152
1153 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
1154 real life, but is still worth protecting against".
1155
1156 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
1157 regular expressions".
1158
1159 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
1160 have it.
1161
1162 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
1163 "configure" and the latter not, in order to fix a problem somebody had
1164 with compiling the Arg class on HP-UX.
1165
1166 (e) Improve the error-handling of the C++ wrapper a little bit.
1167
1168 (f) New tests for checking recursion limiting.
1169
1170 8. The pcre_memmove() function, which is used only if the environment does not
1171 have a standard memmove() function (and is therefore rarely compiled),
1172 contained two bugs: (a) use of int instead of size_t, and (b) it was not
1173 returning a result (though PCRE never actually uses the result).
1174
1175 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
1176 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
1177 returned instead of calling malloc() with an overflowing number that would
1178 most likely cause subsequent chaos.
1179
1180 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
1181
1182 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
1183 with this option is matched, the nmatch and pmatch options of regexec() are
1184 ignored.
1185
1186 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
1187 provided in case anyone wants to the the POSIX interface with UTF-8
1188 strings.
1189
1190 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
1191 C++ linking (needed for some HP-UX environments).
1192
1193 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
1194 (unused parameter) and in the pcre_printint() function (omitted "default"
1195 switch label when the default is to do nothing).
1196
1197 15. Added some code to make it possible, when PCRE is compiled as a C++
1198 library, to replace subject pointers for pcre_exec() with a smart pointer
1199 class, thus making it possible to process discontinuous strings.
1200
1201 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
1202 much the same function. They were added by different people who were trying
1203 to make PCRE easy to compile on non-Unix systems. It has been suggested
1204 that PCRE_EXPORT be abolished now that there is more automatic apparatus
1205 for compiling on Windows systems. I have therefore replaced it with
1206 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
1207 defaults to "extern" for C or "extern C" for C++, which works fine on
1208 Unix-like systems. It is now possible to override the value of PCRE_DATA_
1209 SCOPE with something explicit in config.h. In addition:
1210
1211 (a) pcreposix.h still had just "extern" instead of either of these macros;
1212 I have replaced it with PCRE_DATA_SCOPE.
1213
1214 (b) Functions such as _pcre_xclass(), which are internal to the library,
1215 but external in the C sense, all had PCRE_EXPORT in their definitions.
1216 This is apparently wrong for the Windows case, so I have removed it.
1217 (It makes no difference on Unix-like systems.)
1218
1219 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
1220 of recursive calls to match(). This is different to MATCH_LIMIT because
1221 that limits the total number of calls to match(), not all of which increase
1222 the depth of recursion. Limiting the recursion depth limits the amount of
1223 stack (or heap if NO_RECURSE is set) that is used. The default can be set
1224 when PCRE is compiled, and changed at run time. A patch from Google adds
1225 this functionality to the C++ interface.
1226
1227 18. Changes to the handling of Unicode character properties:
1228
1229 (a) Updated the table to Unicode 4.1.0.
1230
1231 (b) Recognize characters that are not in the table as "Cn" (undefined).
1232
1233 (c) I revised the way the table is implemented to a much improved format
1234 which includes recognition of ranges. It now supports the ranges that
1235 are defined in UnicodeData.txt, and it also amalgamates other
1236 characters into ranges. This has reduced the number of entries in the
1237 table from around 16,000 to around 3,000, thus reducing its size
1238 considerably. I realized I did not need to use a tree structure after
1239 all - a binary chop search is just as efficient. Having reduced the
1240 number of entries, I extended their size from 6 bytes to 8 bytes to
1241 allow for more data.
1242
1243 (d) Added support for Unicode script names via properties such as \p{Han}.
1244
1245 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
1246 matching that character.
1247
1248 20. When matching a repeated Unicode property with a minimum greater than zero,
1249 (for example \pL{2,}), PCRE could look past the end of the subject if it
1250 reached it while seeking the minimum number of characters. This could
1251 happen only if some of the characters were more than one byte long, because
1252 there is a check for at least the minimum number of bytes.
1253
1254 21. Refactored the implementation of \p and \P so as to be more general, to
1255 allow for more different types of property in future. This has changed the
1256 compiled form incompatibly. Anybody with saved compiled patterns that use
1257 \p or \P will have to recompile them.
1258
1259 22. Added "Any" and "L&" to the supported property types.
1260
1261 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
1262 but give a compile time error if the value is greater than 0xff.
1263
1264 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
1265 accidentally not being installed or uninstalled.
1266
1267 25. The pcre.h file was built from pcre.h.in, but the only changes that were
1268 made were to insert the current release number. This seemed silly, because
1269 it made things harder for people building PCRE on systems that don't run
1270 "configure". I have turned pcre.h into a distributed file, no longer built
1271 by "configure", with the version identification directly included. There is
1272 no longer a pcre.h.in file.
1273
1274 However, this change necessitated a change to the pcre-config script as
1275 well. It is built from pcre-config.in, and one of the substitutions was the
1276 release number. I have updated configure.ac so that ./configure now finds
1277 the release number by grepping pcre.h.
1278
1279 26. Added the ability to run the tests under valgrind.
1280
1281
1282 Version 6.4 05-Sep-05
1283 ---------------------
1284
1285 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
1286 "--" to be printed when multiple files were scanned, even when none of the
1287 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
1288 consider it to be a bug, and have restored the previous behaviour.
1289
1290 2. A couple of code tidies to get rid of compiler warnings.
1291
1292 3. The pcretest program used to cheat by referring to symbols in the library
1293 whose names begin with _pcre_. These are internal symbols that are not
1294 really supposed to be visible externally, and in some environments it is
1295 possible to suppress them. The cheating is now confined to including
1296 certain files from the library's source, which is a bit cleaner.
1297
1298 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
1299 file's purpose clearer.
1300
1301 5. Reorganized pcre_ucp_findchar().
1302
1303
1304 Version 6.3 15-Aug-05
1305 ---------------------
1306
1307 1. The file libpcre.pc.in did not have general read permission in the tarball.
1308
1309 2. There were some problems when building without C++ support:
1310
1311 (a) If C++ support was not built, "make install" and "make test" still
1312 tried to test it.
1313
1314 (b) There were problems when the value of CXX was explicitly set. Some
1315 changes have been made to try to fix these, and ...
1316
1317 (c) --disable-cpp can now be used to explicitly disable C++ support.
1318
1319 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
1320 backslash in a target when C++ was disabled. This confuses some
1321 versions of "make", apparently. Using an intermediate variable solves
1322 this. (Same for CPP_LOBJ.)
1323
1324 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
1325 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
1326 necessary on certain architectures.
1327
1328 4. Added a setting of -export-symbols-regex to the link command to remove
1329 those symbols that are exported in the C sense, but actually are local
1330 within the library, and not documented. Their names all begin with
1331 "_pcre_". This is not a perfect job, because (a) we have to except some
1332 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
1333 available (and never for static libraries). I have made a note to try to
1334 find a way round (a) in the future.
1335
1336
1337 Version 6.2 01-Aug-05
1338 ---------------------
1339
1340 1. There was no test for integer overflow of quantifier values. A construction
1341 such as {1111111111111111} would give undefined results. What is worse, if
1342 a minimum quantifier for a parenthesized subpattern overflowed and became
1343 negative, the calculation of the memory size went wrong. This could have
1344 led to memory overwriting.
1345
1346 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
1347
1348 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
1349 operating environments where this matters.
1350
1351 4. Applied Giuseppe Maxia's patch to add additional features for controlling
1352 PCRE options from within the C++ wrapper.
1353
1354 5. Named capturing subpatterns were not being correctly counted when a pattern
1355 was compiled. This caused two problems: (a) If there were more than 100
1356 such subpatterns, the calculation of the memory needed for the whole
1357 compiled pattern went wrong, leading to an overflow error. (b) Numerical
1358 back references of the form \12, where the number was greater than 9, were
1359 not recognized as back references, even though there were sufficient
1360 previous subpatterns.
1361
1362 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
1363 versions of gcc, e.g. 2.95.4.
1364
1365
1366 Version 6.1 21-Jun-05
1367 ---------------------
1368
1369 1. There was one reference to the variable "posix" in pcretest.c that was not
1370 surrounded by "#if !defined NOPOSIX".
1371
1372 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
1373 the cross-check on the old pcre_info() function, for the benefit of the
1374 cut-down version of PCRE that is currently imported into Exim.
1375
1376 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
1377 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
1378 patterns that start (?i) but I don't think that matters. The use of (?i) is
1379 just an example; this all applies to the other options as well.
1380
1381 4. Since libtool seems to echo the compile commands it is issuing, the output
1382 from "make" can be reduced a bit by putting "@" in front of each libtool
1383 compile command.
1384
1385 5. Patch from the folks at Google for configure.in to be a bit more thorough
1386 in checking for a suitable C++ installation before trying to compile the
1387 C++ stuff. This should fix a reported problem when a compiler was present,
1388 but no suitable headers.
1389
1390 6. The man pages all had just "PCRE" as their title. I have changed them to
1391 be the relevant file name. I have also arranged that these names are
1392 retained in the file doc/pcre.txt, which is a concatenation in text format
1393 of all the man pages except the little individual ones for each function.
1394
1395 7. The NON-UNIX-USE file had not been updated for the different set of source
1396 files that come with release 6. I also added a few comments about the C++
1397 wrapper.
1398
1399
1400 Version 6.0 07-Jun-05
1401 ---------------------
1402
1403 1. Some minor internal re-organization to help with my DFA experiments.
1404
1405 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
1406 didn't matter for the library itself when fully configured, but did matter
1407 when compiling without UCP support, or within Exim, where the ucp files are
1408 not imported.
1409
1410 3. Refactoring of the library code to split up the various functions into
1411 different source modules. The addition of the new DFA matching code (see
1412 below) to a single monolithic source would have made it really too
1413 unwieldy, quite apart from causing all the code to be include in a
1414 statically linked application, when only some functions are used. This is
1415 relevant even without the DFA addition now that patterns can be compiled in
1416 one application and matched in another.
1417
1418 The downside of splitting up is that there have to be some external
1419 functions and data tables that are used internally in different modules of
1420 the library but which are not part of the API. These have all had their
1421 names changed to start with "_pcre_" so that they are unlikely to clash
1422 with other external names.
1423
1424 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
1425 a different (DFA) algorithm. Although it is slower than the original
1426 function, it does have some advantages for certain types of matching
1427 problem.
1428
1429 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
1430 including restarting after a partial match.
1431
1432 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
1433 defined when compiling for Windows was sent to me. I have put it into the
1434 code, though I have no means of testing or verifying it.
1435
1436 7. Added the pcre_refcount() auxiliary function.
1437
1438 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
1439 match before or at the first newline in the subject string. In pcretest,
1440 the /f option on a pattern can be used to set this.
1441
1442 9. A repeated \w when used in UTF-8 mode with characters greater than 256
1443 would behave wrongly. This has been present in PCRE since release 4.0.
1444
1445 10. A number of changes to the pcregrep command:
1446
1447 (a) Refactored how -x works; insert ^(...)$ instead of setting
1448 PCRE_ANCHORED and checking the length, in preparation for adding
1449 something similar for -w.
1450
1451 (b) Added the -w (match as a word) option.
1452
1453 (c) Refactored the way lines are read and buffered so as to have more
1454 than one at a time available.
1455
1456 (d) Implemented a pcregrep test script.
1457
1458 (e) Added the -M (multiline match) option. This allows patterns to match
1459 over several lines of the subject. The buffering ensures that at least
1460 8K, or the rest of the document (whichever is the shorter) is available
1461 for matching (and similarly the previous 8K for lookbehind assertions).
1462
1463 (f) Changed the --help output so that it now says
1464
1465 -w, --word-regex(p)
1466
1467 instead of two lines, one with "regex" and the other with "regexp"
1468 because that confused at least one person since the short forms are the
1469 same. (This required a bit of code, as the output is generated
1470 automatically from a table. It wasn't just a text change.)
1471
1472 (g) -- can be used to terminate pcregrep options if the next thing isn't an
1473 option but starts with a hyphen. Could be a pattern or a path name
1474 starting with a hyphen, for instance.
1475
1476 (h) "-" can be given as a file name to represent stdin.
1477
1478 (i) When file names are being printed, "(standard input)" is used for
1479 the standard input, for compatibility with GNU grep. Previously
1480 "<stdin>" was used.
1481
1482 (j) The option --label=xxx can be used to supply a name to be used for
1483 stdin when file names are being printed. There is no short form.
1484
1485 (k) Re-factored the options decoding logic because we are going to add
1486 two more options that take data. Such options can now be given in four
1487 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
1488
1489 (l) Added the -A, -B, and -C options for requesting that lines of context
1490 around matches be printed.
1491
1492 (m) Added the -L option to print the names of files that do not contain
1493 any matching lines, that is, the complement of -l.
1494
1495 (n) The return code is 2 if any file cannot be opened, but pcregrep does
1496 continue to scan other files.
1497
1498 (o) The -s option was incorrectly implemented. For compatibility with other
1499 greps, it now suppresses the error message for a non-existent or non-
1500 accessible file (but not the return code). There is a new option called
1501 -q that suppresses the output of matching lines, which was what -s was
1502 previously doing.
1503
1504 (p) Added --include and --exclude options to specify files for inclusion
1505 and exclusion when recursing.
1506
1507 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
1508 Hopefully, it now does.
1509
1510 12. Missing cast in pcre_study().
1511
1512 13. Added an "uninstall" target to the makefile.
1513
1514 14. Replaced "extern" in the function prototypes in Makefile.in with
1515 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
1516 world, but is set differently for Windows.
1517
1518 15. Added a second compiling function called pcre_compile2(). The only
1519 difference is that it has an extra argument, which is a pointer to an
1520 integer error code. When there is a compile-time failure, this is set
1521 non-zero, in addition to the error test pointer being set to point to an
1522 error message. The new argument may be NULL if no error number is required
1523 (but then you may as well call pcre_compile(), which is now just a
1524 wrapper). This facility is provided because some applications need a
1525 numeric error indication, but it has also enabled me to tidy up the way
1526 compile-time errors are handled in the POSIX wrapper.
1527
1528 16. Added VPATH=.libs to the makefile; this should help when building with one
1529 prefix path and installing with another. (Or so I'm told by someone who
1530 knows more about this stuff than I do.)
1531
1532 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
1533 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
1534 match everything, including newlines. This is not POSIX-compatible, but
1535 somebody wanted the feature. From pcretest it can be activated by using
1536 both the P and the s flags.
1537
1538 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
1539
1540 19. libpcre.pc was being incorrectly installed as executable.
1541
1542 20. A couple of places in pcretest check for end-of-line by looking for '\n';
1543 it now also looks for '\r' so that it will work unmodified on Windows.
1544
1545 21. Added Google's contributed C++ wrapper to the distribution.
1546
1547 22. Added some untidy missing memory free() calls in pcretest, to keep
1548 Electric Fence happy when testing.
1549
1550
1551
1552 Version 5.0 13-Sep-04
1553 ---------------------
1554
1555 1. Internal change: literal characters are no longer packed up into items
1556 containing multiple characters in a single byte-string. Each character
1557 is now matched using a separate opcode. However, there may be more than one
1558 byte in the character in UTF-8 mode.
1559
1560 2. The pcre_callout_block structure has two new fields: pattern_position and
1561 next_item_length. These contain the offset in the pattern to the next match
1562 item, and its length, respectively.
1563
1564 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
1565 insertion of callouts before each pattern item. Added the /C option to
1566 pcretest to make use of this.
1567
1568 4. On the advice of a Windows user, the lines
1569
1570 #if defined(_WIN32) || defined(WIN32)
1571 _setmode( _fileno( stdout ), 0x8000 );
1572 #endif /* defined(_WIN32) || defined(WIN32) */
1573
1574 have been added to the source of pcretest. This apparently does useful
1575 magic in relation to line terminators.
1576
1577 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
1578 for the benefit of those environments where the "b" makes a difference.
1579
1580 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
1581 to know about it. I have put a hack into configure.in that adds in code
1582 to set GCC=yes if CC=icc. This seems to end up at a point in the
1583 generated configure script that is early enough to affect the setting of
1584 compiler options, which is what is needed, but I have no means of testing
1585 whether it really works. (The user who reported this had patched the
1586 generated configure script, which of course I cannot do.)
1587
1588 LATER: After change 22 below (new libtool files), the configure script
1589 seems to know about icc (and also ecc). Therefore, I have commented out
1590 this hack in configure.in.
1591
1592 7. Added support for pkg-config (2 patches were sent in).
1593
1594 8. Negated POSIX character classes that used a combination of internal tables
1595 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
1596 [[:^ascii]]. Typically, they would match almost any characters. The other
1597 POSIX classes were not broken in this way.
1598
1599 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
1600 to find the match, as PCRE was deluded into thinking that the match had to
1601 start at the start point or following a newline. The same bug applied to
1602 patterns with negative forward assertions or any backward assertions
1603 preceding ".*" at the start, unless the pattern required a fixed first
1604 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
1605
1606 10. In UTF-8 mode, when moving forwards in the subject after a failed match
1607 starting at the last subject character, bytes beyond the end of the subject
1608 string were read.
1609
1610 11. Renamed the variable "class" as "classbits" to make life easier for C++
1611 users. (Previously there was a macro definition, but it apparently wasn't
1612 enough.)
1613
1614 12. Added the new field "tables" to the extra data so that tables can be passed
1615 in at exec time, or the internal tables can be re-selected. This allows
1616 a compiled regex to be saved and re-used at a later time by a different
1617 program that might have everything at different addresses.
1618
1619 13. Modified the pcre-config script so that, when run on Solaris, it shows a
1620 -R library as well as a -L library.
1621
1622 14. The debugging options of pcretest (-d on the command line or D on a
1623 pattern) showed incorrect output for anything following an extended class
1624 that contained multibyte characters and which was followed by a quantifier.
1625
1626 15. Added optional support for general category Unicode character properties
1627 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
1628 support. It adds about 90K to the size of the library. The meanings of the
1629 inbuilt class escapes such as \d and \s have NOT been changed.
1630
1631 16. Updated pcredemo.c to include calls to free() to release the memory for the
1632 compiled pattern.
1633
1634 17. The generated file chartables.c was being created in the source directory
1635 instead of in the building directory. This caused the build to fail if the
1636 source directory was different from the building directory, and was
1637 read-only.
1638
1639 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
1640 file. No doubt somebody will tell me if they don't make sense... Also added
1641 Dan Mooney's comments about building on OpenVMS.
1642
1643 19. Added support for partial matching via the PCRE_PARTIAL option for
1644 pcre_exec() and the \P data escape in pcretest.
1645
1646 20. Extended pcretest with 3 new pattern features:
1647
1648 (i) A pattern option of the form ">rest-of-line" causes pcretest to
1649 write the compiled pattern to the file whose name is "rest-of-line".
1650 This is a straight binary dump of the data, with the saved pointer to
1651 the character tables forced to be NULL. The study data, if any, is
1652 written too. After writing, pcretest reads a new pattern.
1653
1654 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
1655 compiled pattern from the given file. There must not be any
1656 occurrences of "<" in the file name (pretty unlikely); if there are,
1657 pcretest will instead treat the initial "<" as a pattern delimiter.
1658 After reading in the pattern, pcretest goes on to read data lines as
1659 usual.
1660
1661 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
1662 and 16-bit fields in a compiled pattern, to simulate a pattern that
1663 was compiled on a host of opposite endianness.
1664
1665 21. The pcre-exec() function can now cope with patterns that were compiled on
1666 hosts of opposite endianness, with this restriction:
1667
1668 As for any compiled expression that is saved and used later, the tables
1669 pointer field cannot be preserved; the extra_data field in the arguments
1670 to pcre_exec() should be used to pass in a tables address if a value
1671 other than the default internal tables were used at compile time.
1672
1673 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
1674 now diagnosed as an error. Previously, most of the time, a negative number
1675 would have been treated as zero, but if in addition "ovector" was passed as
1676 NULL, a crash could occur.
1677
1678 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
1679 new versions from the libtool 1.5 distribution (the last one is a copy of
1680 a file called libtool.m4). This seems to have fixed the need to patch
1681 "configure" to support Darwin 1.3 (which I used to do). However, I still
1682 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
1683 workstation).
1684
1685 24. Changed the PCRE licence to be the more standard "BSD" licence.
1686
1687
1688 Version 4.5 01-Dec-03
1689 ---------------------
1690
1691 1. There has been some re-arrangement of the code for the match() function so
1692 that it can be compiled in a version that does not call itself recursively.
1693 Instead, it keeps those local variables that need separate instances for
1694 each "recursion" in a frame on the heap, and gets/frees frames whenever it
1695 needs to "recurse". Keeping track of where control must go is done by means
1696 of setjmp/longjmp. The whole thing is implemented by a set of macros that
1697 hide most of the details from the main code, and operates only if
1698 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
1699 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
1700 operating.
1701
1702 To make it easier for callers to provide specially tailored get/free
1703 functions for this usage, two new functions, pcre_stack_malloc, and
1704 pcre_stack_free, are used. They are always called in strict stacking order,
1705 and the size of block requested is always the same.
1706
1707 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
1708 PCRE has been compiled to use the stack or the heap for recursion. The
1709 -C option of pcretest uses this to show which version is compiled.
1710
1711 A new data escape \S, is added to pcretest; it causes the amounts of store
1712 obtained and freed by both kinds of malloc/free at match time to be added
1713 to the output.
1714
1715 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
1716 what's available on my current Linux desktop machine.
1717
1718 3. When matching a UTF-8 string, the test for a valid string at the start has
1719 been extended. If start_offset is not zero, PCRE now checks that it points
1720 to a byte that is the start of a UTF-8 character. If not, it returns
1721 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
1722 this is necessary because there may be backward assertions in the pattern.
1723 When matching the same subject several times, it may save resources to use
1724 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
1725
1726 4. The code for checking the validity of UTF-8 strings has been tightened so
1727 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
1728 containing "overlong sequences".
1729
1730 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
1731 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
1732 should have been "&", but it just so happened that all the cases this let
1733 through by mistake were picked up later in the function.
1734
1735 6. I had used a variable called "isblank" - this is a C99 function, causing
1736 some compilers to warn. To avoid this, I renamed it (as "blankclass").
1737
1738 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
1739 prompting; (b) run "./pcretest /dev/null" at the start of the test script
1740 so the version is shown; (c) stop "make test" echoing "./RunTest".
1741
1742 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
1743
1744 9. The prototype for memmove() for systems that don't have it was using
1745 size_t, but the inclusion of the header that defines size_t was later. I've
1746 moved the #includes for the C headers earlier to avoid this.
1747
1748 10. Added some adjustments to the code to make it easier to compiler on certain
1749 special systems:
1750
1751 (a) Some "const" qualifiers were missing.
1752 (b) Added the macro EXPORT before all exported functions; by default this
1753 is defined to be empty.
1754 (c) Changed the dftables auxiliary program (that builds chartables.c) so
1755 that it reads its output file name as an argument instead of writing
1756 to the standard output and assuming this can be redirected.
1757
1758 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
1759 class containing characters with values greater than 255, PCRE compilation
1760 went into a loop.
1761
1762 12. A recursive reference to a subpattern that was within another subpattern
1763 that had a minimum quantifier of zero caused PCRE to crash. For example,
1764 (x(y(?2))z)? provoked this bug with a subject that got as far as the
1765 recursion. If the recursively-called subpattern itself had a zero repeat,
1766 that was OK.
1767
1768 13. In pcretest, the buffer for reading a data line was set at 30K, but the
1769 buffer into which it was copied (for escape processing) was still set at
1770 1024, so long lines caused crashes.
1771
1772 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
1773 "internal error: code overflow...". This applied to any character class
1774 that was followed by a possessive quantifier.
1775
1776 15. Modified the Makefile to add libpcre.la as a prerequisite for
1777 libpcreposix.la because I was told this is needed for a parallel build to
1778 work.
1779
1780 16. If a pattern that contained .* following optional items at the start was
1781 studied, the wrong optimizing data was generated, leading to matching
1782 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
1783 matching string must start with a or b or c. The correct conclusion for
1784 this pattern is that a match can start with any character.
1785
1786
1787 Version 4.4 13-Aug-03
1788 ---------------------
1789
1790 1. In UTF-8 mode, a character class containing characters with values between
1791 127 and 255 was not handled correctly if the compiled pattern was studied.
1792 In fixing this, I have also improved the studying algorithm for such
1793 classes (slightly).
1794
1795 2. Three internal functions had redundant arguments passed to them. Removal
1796 might give a very teeny performance improvement.
1797
1798 3. Documentation bug: the value of the capture_top field in a callout is *one
1799 more than* the number of the hightest numbered captured substring.
1800
1801 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
1802 in incorrectly linking with a previously installed version. They now link
1803 explicitly with libpcre.la.
1804
1805 5. configure.in no longer needs to recognize Cygwin specially.
1806
1807 6. A problem in pcre.in for Windows platforms is fixed.
1808
1809 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
1810 pcretest, it used to include the size of the study block as part of its
1811 output. Unfortunately, the structure contains a field that has a different
1812 size on different hardware architectures. This meant that the tests that
1813 showed this size failed. As the block is currently always of a fixed size,
1814 this information isn't actually particularly useful in pcretest output, so
1815 I have just removed it.
1816
1817 8. Three pre-processor statements accidentally did not start in column 1.
1818 Sadly, there are *still* compilers around that complain, even though
1819 standard C has not required this for well over a decade. Sigh.
1820
1821 9. In pcretest, the code for checking callouts passed small integers in the
1822 callout_data field, which is a void * field. However, some picky compilers
1823 complained about the casts involved for this on 64-bit systems. Now
1824 pcretest passes the address of the small integer instead, which should get
1825 rid of the warnings.
1826
1827 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
1828 both compile and run time, and gives an error if an invalid UTF-8 sequence
1829 is found. There is a option for disabling this check in cases where the
1830 string is known to be correct and/or the maximum performance is wanted.
1831
1832 11. In response to a bug report, I changed one line in Makefile.in from
1833
1834 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
1835 to
1836 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
1837
1838 to look similar to other lines, but I have no way of telling whether this
1839 is the right thing to do, as I do not use Windows. No doubt I'll get told
1840 if it's wrong...
1841
1842
1843 Version 4.3 21-May-03
1844 ---------------------
1845
1846 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
1847 Makefile.
1848
1849 2. Some refactoring to improve the quality of the code:
1850
1851 (i) The utf8_table... variables are now declared "const".
1852
1853 (ii) The code for \cx, which used the "case flipping" table to upper case
1854 lower case letters, now just substracts 32. This is ASCII-specific,
1855 but the whole concept of \cx is ASCII-specific, so it seems
1856 reasonable.
1857
1858 (iii) PCRE was using its character types table to recognize decimal and
1859 hexadecimal digits in the pattern. This is silly, because it handles
1860 only 0-9, a-f, and A-F, but the character types table is locale-
1861 specific, which means strange things might happen. A private
1862 table is now used for this - though it costs 256 bytes, a table is
1863 much faster than multiple explicit tests. Of course, the standard
1864 character types table is still used for matching digits in subject
1865 strings against \d.
1866
1867 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
1868 ending in _t are). So I've renamed it as ESC_tee.
1869
1870 3. The first argument for regexec() in the POSIX wrapper should have been
1871 defined as "const".
1872
1873 4. Changed pcretest to use malloc() for its buffers so that they can be
1874 Electric Fenced for debugging.
1875
1876 5. There were several places in the code where, in UTF-8 mode, PCRE would try
1877 to read one or more bytes before the start of the subject string. Often this
1878 had no effect on PCRE's behaviour, but in some circumstances it could
1879 provoke a segmentation fault.
1880
1881 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
1882 to try to read one or more bytes before the start of the subject string.
1883
1884 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
1885 UTF-8 support could misbehave in various ways if the subject string
1886 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
1887 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
1888 back over UTF-8 characters.)
1889
1890
1891 Version 4.2 14-Apr-03
1892 ---------------------
1893
1894 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
1895
1896 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
1897 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
1898 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
1899 [WIN_PREFIX]: new variable, "cyg" for Cygwin
1900 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
1901 and BUILD_EXEEXT
1902 Note: automatic setting of the BUILD variables is not yet working
1903 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
1904 compile-time but not at link-time
1905 [LINK]: use for linking executables only
1906 make different versions for Windows and non-Windows
1907 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
1908 libraries
1909 [LINK_FOR_BUILD]: new variable
1910 [OBJEXT]: use throughout
1911 [EXEEXT]: use throughout
1912 <winshared>: new target
1913 <wininstall>: new target
1914 <dftables.o>: use native compiler
1915 <dftables>: use native linker
1916 <install>: handle Windows platform correctly
1917 <clean>: ditto
1918 <check>: ditto
1919 copy DLL to top builddir before testing
1920
1921 As part of these changes, -no-undefined was removed again. This was reported
1922 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
1923 in any case.
1924
1925 3. Some tidies to get rid of compiler warnings:
1926
1927 . In the match_data structure, match_limit was an unsigned long int, whereas
1928 match_call_count was an int. I've made them both unsigned long ints.
1929
1930 . In pcretest the fact that a const uschar * doesn't automatically cast to
1931 a void * provoked a warning.
1932
1933 . Turning on some more compiler warnings threw up some "shadow" variables
1934 and a few more missing casts.
1935
1936 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
1937 option, a class that contained a single character with a value between 128
1938 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
1939
1940 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
1941 option, a class that contained several characters, but with at least one
1942 whose value was between 128 and 255 caused PCRE to crash.
1943
1944
1945 Version 4.1 12-Mar-03
1946 ---------------------
1947
1948 1. Compiling with gcc -pedantic found a couple of places where casts were
1949 needed, and a string in dftables.c that was longer than standard compilers are
1950 required to support.
1951
1952 2. Compiling with Sun's compiler found a few more places where the code could
1953 be tidied up in order to avoid warnings.
1954
1955 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
1956 first of these names is deprecated in the latest Autoconf in favour of the name
1957 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
1958 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
1959 analogy I have changed it to CFLAGS_FOR_BUILD.
1960
1961 4. Added -no-undefined to the linking command in the Makefile, because this is
1962 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
1963 linking step for the pcreposix library.
1964
1965 5. PCRE was failing to diagnose the case of two named groups with the same
1966 name.
1967
1968 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
1969 literal character that is needed in the subject for a match, and scans along to
1970 ensure that it is present before embarking on the full matching process. This
1971 saves time in cases of nested unlimited repeats that are never going to match.
1972 Problem: the scan can take a lot of time if the subject is very long (e.g.
1973 megabytes), thus penalizing straightforward matches. It is now done only if the
1974 amount of subject to be scanned is less than 1000 bytes.
1975
1976 7. A lesser problem with the same optimization is that it was recording the
1977 first character of an anchored pattern as "needed", thus provoking a search
1978 right along the subject, even when the first match of the pattern was going to
1979 fail. The "needed" character is now not set for anchored patterns, unless it
1980 follows something in the pattern that is of non-fixed length. Thus, it still
1981 fulfils its original purpose of finding quick non-matches in cases of nested
1982 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
1983
1984
1985 Version 4.0 17-Feb-03
1986 ---------------------
1987
1988 1. If a comment in an extended regex that started immediately after a meta-item
1989 extended to the end of string, PCRE compiled incorrect data. This could lead to
1990 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
1991
1992 2. Moved to autoconf 2.53 and libtool 1.4.2.
1993
1994 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
1995 the special perltest8 script is no longer needed - all the tests can be run
1996 from a single perltest script.
1997
1998 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
1999 by \s. It has now been removed in PCRE. This means it isn't recognized as
2000 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
2001 class [:space:] *does* include VT, thereby creating a mess.
2002
2003 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
2004 space and tab.
2005
2006 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
2007 its new features into the main test script, reducing the number of scripts.
2008
2009 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
2010 were backward compatible, and made the (?i) apply to the whole pattern, as if
2011 /i were given. Now it behaves more logically, and applies the option setting
2012 only to what follows. PCRE has been changed to follow suit. However, if it
2013 finds options settings right at the start of the pattern, it extracts them into
2014 the global options, as before. Thus, they show up in the info data.
2015
2016 8. Added support for the \Q...\E escape sequence. Characters in between are
2017 treated as literals. This is slightly different from Perl in that $ and @ are
2018 also handled as literals inside the quotes. In Perl, they will cause variable
2019 interpolation. Note the following examples:
2020
2021 Pattern PCRE matches Perl matches
2022
2023 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
2024 \Qabc\$xyz\E abc\$xyz abc\$xyz
2025 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
2026
2027 For compatibility with Perl, \Q...\E sequences are recognized inside character
2028 classes as well as outside them.
2029
2030 9. Re-organized 3 code statements in pcretest to avoid "overflow in
2031 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
2032 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
2033 signed/unsigned warnings.
2034
2035 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
2036 option for pcretest, so I've replaced it by a simple function that does just
2037 that job.
2038
2039 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
2040 "pcregrep -".
2041
2042 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
2043 Java package. This provides some syntactic sugar for simple cases of what my
2044 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
2045 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
2046 item, you can use this simplified notation. Note that only makes sense with
2047 greedy quantifiers. Consequently, the use of the possessive quantifier forces
2048 greediness, whatever the setting of the PCRE_UNGREEDY option.
2049
2050 13. A change of greediness default within a pattern was not taking effect at
2051 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
2052 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
2053 was abstracted outside.
2054
2055 14. PCRE now supports the \G assertion. It is true when the current matching
2056 position is at the start point of the match. This differs from \A when the
2057 starting offset is non-zero. Used with the /g option of pcretest (or similar
2058 code), it works in the same way as it does for Perl's /g option. If all
2059 alternatives of a regex begin with \G, the expression is anchored to the start
2060 match position, and the "anchored" flag is set in the compiled expression.
2061
2062 15. Some bugs concerning the handling of certain option changes within patterns
2063 have been fixed. These applied to options other than (?ims). For example,
2064 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
2065 been the other way round. Some of this was related to change 7 above.
2066
2067 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
2068 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
2069 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
2070 POSIX classes only within a class (e.g. /[[:alpha:]]/).
2071
2072 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
2073 mode. Unlike ".", it always matches newline, whatever the setting of
2074 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
2075 assertions. Perl allows it, but it doesn't (in general) work because it can't
2076 calculate the length of the lookbehind. At least, that's the case for Perl
2077 5.8.0 - I've been told they are going to document that it doesn't work in
2078 future.
2079
2080 18. Added an error diagnosis for escapes that PCRE does not support: these are
2081 \L, \l, \N, \P, \p, \U, \u, and \X.
2082
2083 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
2084 reading past the end of the pattern in cases such as /[abcd/.
2085
2086 20. PCRE was getting more memory than necessary for patterns with classes that
2087 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
2088
2089 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
2090 compiling PCRE for use with Virtual Pascal.
2091
2092 22. Small fix to the Makefile to make it work properly if the build is done
2093 outside the source tree.
2094
2095 23. Added a new extension: a condition to go with recursion. If a conditional
2096 subpattern starts with (?(R) the "true" branch is used if recursion has
2097 happened, whereas the "false" branch is used only at the top level.
2098
2099 24. When there was a very long string of literal characters (over 255 bytes
2100 without UTF support, over 250 bytes with UTF support), the computation of how
2101 much memory was required could be incorrect, leading to segfaults or other
2102 strange effects.
2103
2104 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
2105 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
2106 there was a subsequent back reference to those brackets. This meant that, for
2107 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
2108 possible to check for precisely this case. All we can do is abandon the
2109 optimization if .* occurs inside capturing brackets when there are any back
2110 references whatsoever. (See below for a better fix that came later.)
2111
2112 26. The handling of the optimization for finding the first character of a
2113 non-anchored pattern, and for finding a character that is required later in the
2114 match were failing in some cases. This didn't break the matching; it just
2115 failed to optimize when it could. The way this is done has been re-implemented.
2116
2117 27. Fixed typo in error message for invalid (?R item (it said "(?p").
2118
2119 28. Added a new feature that provides some of the functionality that Perl
2120 provides with (?{...}). The facility is termed a "callout". The way it is done
2121 in PCRE is for the caller to provide an optional function, by setting
2122 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
2123 global variable. By default it is unset, which disables all calling out. To get
2124 the function called, the regex must include (?C) at appropriate points. This
2125 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
2126 This provides a means of identifying different callout points. When PCRE
2127 reaches such a point in the regex, if pcre_callout has been set, the external
2128 function is called. It is provided with data in a structure called
2129 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
2130 matching continues; if it returns a non-zero value, the match at the current
2131 point fails. However, backtracking will occur if possible. [This was changed
2132 later and other features added - see item 49 below.]
2133
2134 29. pcretest is upgraded to test the callout functionality. It provides a
2135 callout function that displays information. By default, it shows the start of
2136 the match and the current position in the text. There are some new data escapes
2137 to vary what happens:
2138
2139 \C+ in addition, show current contents of captured substrings
2140 \C- do not supply a callout function
2141 \C!n return 1 when callout number n is reached
2142 \C!n!m return 1 when callout number n is reached for the mth time
2143
2144 30. If pcregrep was called with the -l option and just a single file name, it
2145 output "<stdin>" if a match was found, instead of the file name.
2146
2147 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
2148 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
2149 pcre_exec(). This saves a malloc/free per call. The default value of
2150 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
2151 when configuring.
2152
2153 32. The default maximum size of a compiled pattern is 64K. There have been a
2154 few cases of people hitting this limit. The code now uses macros to handle the
2155 storing of links as offsets within the compiled pattern. It defaults to 2-byte
2156 links, but this can be changed to 3 or 4 bytes by --with-link-size when
2157 configuring. Tests 2 and 5 work only with 2-byte links because they output
2158 debugging information about compiled patterns.
2159
2160 33. Internal code re-arrangements:
2161
2162 (a) Moved the debugging function for printing out a compiled regex into
2163 its own source file (printint.c) and used #include to pull it into
2164 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
2165 separate copies.
2166
2167 (b) Defined the list of op-code names for debugging as a macro in
2168 internal.h so that it is next to the definition of the opcodes.
2169
2170 (c) Defined a table of op-code lengths for simpler skipping along compiled
2171 code. This is again a macro in internal.h so that it is next to the
2172 definition of the opcodes.
2173
2174 34. Added support for recursive calls to individual subpatterns, along the
2175 lines of Robin Houston's patch (but implemented somewhat differently).
2176
2177 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
2178 allow it to read and process whole directories in Win32. This code was
2179 contributed by Lionel Fourquaux; it has not been tested by me.
2180
2181 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
2182 used to name a group. Names consist of alphanumerics and underscores, and must
2183 be unique. Back references use the syntax (?P=name) and recursive calls use
2184 (?P>name) which is a PCRE extension to the Python extension. Groups still have
2185 numbers. The function pcre_fullinfo() can be used after compilation to extract
2186 a name/number map. There are three relevant calls:
2187
2188 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
2189 PCRE_INFO_NAMECOUNT yields the number of entries
2190 PCRE_INFO_NAMETABLE yields a pointer to the map.
2191
2192 The map is a vector of fixed-size entries. The size of each entry depends on
2193 the length of the longest name used. The first two bytes of each entry are the
2194 group number, most significant byte first. There follows the corresponding
2195 name, zero terminated. The names are in alphabetical order.
2196
2197 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
2198 case instead of 255. Making it the same both with and without UTF-8 support
2199 means that the same test output works with both.
2200
2201 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
2202 calling malloc() with a zero argument.
2203
2204 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
2205 optimization. I've improved things by keeping a bitmap of backreferences with
2206 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
2207 fact referenced, the optimization can be applied. It is unlikely that a
2208 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
2209 the match to follow \n) will appear inside brackets with a number greater than
2210 31, but if it does, any back reference > 31 suppresses the optimization.
2211
2212 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
2213 of disabling numbered capturing parentheses. Any opening parenthesis that is
2214 not followed by ? behaves as if it were followed by ?: but named parentheses
2215 can still be used for capturing (and they will acquire numbers in the usual
2216 way).
2217
2218 41. Redesigned the return codes from the match() function into yes/no/error so
2219 that errors can be passed back from deep inside the nested calls. A malloc
2220 failure while inside a recursive subpattern call now causes the
2221 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
2222
2223 42. It is now possible to set a limit on the number of times the match()
2224 function is called in a call to pcre_exec(). This facility makes it possible to
2225 limit the amount of recursion and backtracking, though not in a directly
2226 obvious way, because the match() function is used in a number of different
2227 circumstances. The count starts from zero for each position in the subject
2228 string (for non-anchored patterns). The default limit is, for compatibility, a
2229 large number, namely 10 000 000. You can change this in two ways:
2230
2231 (a) When configuring PCRE before making, you can use --with-match-limit=n
2232 to set a default value for the compiled library.
2233
2234 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
2235 a different value is set. See 45 below.
2236
2237 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
2238
2239 43. Added a new function pcre_config(int, void *) to enable run-time extraction
2240 of things that can be changed at compile time. The first argument specifies
2241 what is wanted and the second points to where the information is to be placed.
2242 The current list of available information is:
2243
2244 PCRE_CONFIG_UTF8
2245
2246 The output is an integer that is set to one if UTF-8 support is available;
2247 otherwise it is set to zero.
2248
2249 PCRE_CONFIG_NEWLINE
2250
2251 The output is an integer that it set to the value of the code that is used for
2252 newline. It is either LF (10) or CR (13).
2253
2254 PCRE_CONFIG_LINK_SIZE
2255
2256 The output is an integer that contains the number of bytes used for internal
2257 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
2258
2259 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
2260
2261 The output is an integer that contains the threshold above which the POSIX
2262 interface uses malloc() for output vectors. See item 31 above.
2263
2264 PCRE_CONFIG_MATCH_LIMIT
2265
2266 The output is an unsigned integer that contains the default limit of the number
2267 of match() calls in a pcre_exec() execution. See 42 above.
2268
2269 44. pcretest has been upgraded by the addition of the -C option. This causes it
2270 to extract all the available output from the new pcre_config() function, and to
2271 output it. The program then exits immediately.
2272
2273 45. A need has arisen to pass over additional data with calls to pcre_exec() in
2274 order to support additional features. One way would have been to define
2275 pcre_exec2() (for example) with extra arguments, but this would not have been
2276 extensible, and would also have required all calls to the original function to
2277 be mapped to the new one. Instead, I have chosen to extend the mechanism that
2278 is used for passing in "extra" data from pcre_study().
2279
2280 The pcre_extra structure is now exposed and defined in pcre.h. It currently
2281 contains the following fields:
2282
2283 flags a bitmap indicating which of the following fields are set
2284 study_data opaque data from pcre_study()
2285 match_limit a way of specifying a limit on match() calls for a specific
2286 call to pcre_exec()
2287 callout_data data for callouts (see 49 below)
2288
2289 The flag bits are also defined in pcre.h, and are
2290
2291 PCRE_EXTRA_STUDY_DATA
2292 PCRE_EXTRA_MATCH_LIMIT
2293 PCRE_EXTRA_CALLOUT_DATA
2294
2295 The pcre_study() function now returns one of these new pcre_extra blocks, with
2296 the actual study data pointed to by the study_data field, and the
2297 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
2298 before. That is, this change is entirely upwards-compatible and requires no
2299 change to existing code.
2300
2301 If you want to pass in additional data to pcre_exec(), you can either place it
2302 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
2303 block.
2304
2305 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
2306 data string contains the escape sequence \M, pcretest calls pcre_exec() several
2307 times with different match limits, until it finds the minimum value needed for
2308 pcre_exec() to complete. The value is then output. This can be instructive; for
2309 most simple matches the number is quite small, but for pathological cases it
2310 gets very large very quickly.
2311
2312 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
2313 returns the size of the data block pointed to by the study_data field in a
2314 pcre_extra block, that is, the value that was passed as the argument to
2315 pcre_malloc() when PCRE was getting memory in which to place the information
2316 created by pcre_study(). The fourth argument should point to a size_t variable.
2317 pcretest has been extended so that this information is shown after a successful
2318 pcre_study() call when information about the compiled regex is being displayed.
2319
2320 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
2321 because what follows is always an absolute path. (Later: it turns out that this
2322 is more than cosmetic for MinGW, because it doesn't like empty path
2323 components.)
2324
2325 49. Some changes have been made to the callout feature (see 28 above):
2326
2327 (i) A callout function now has three choices for what it returns:
2328
2329 0 => success, carry on matching
2330 > 0 => failure at this point, but backtrack if possible
2331 < 0 => serious error, return this value from pcre_exec()
2332
2333 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
2334 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
2335 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
2336 use by callout functions. It will never be used by PCRE itself.
2337
2338 (ii) The pcre_extra structure (see 45 above) has a void * field called
2339 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
2340 pcre_callout_block structure has a field of the same name. The contents of
2341 the field passed in the pcre_extra structure are passed to the callout
2342 function in the corresponding field in the callout block. This makes it
2343 easier to use the same callout-containing regex from multiple threads. For
2344 testing, the pcretest program has a new data escape
2345
2346 \C*n pass the number n (may be negative) as callout_data
2347
2348 If the callout function in pcretest receives a non-zero value as
2349 callout_data, it returns that value.
2350
2351 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
2352 there were some redundant $(CFLAGS) in commands that are now specified as
2353 $(LINK), which already includes $(CFLAGS).
2354
2355 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
2356 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
2357 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
2358 one-byte characters throughout. Note that case-insensitive matching applies
2359 only to characters whose values are less than 256. PCRE doesn't support the
2360 notion of cases for higher-valued characters.
2361
2362 (i) A character class whose characters are all within 0-255 is handled as
2363 a bit map, and the map is inverted for negative classes. Previously, a
2364 character > 255 always failed to match such a class; however it should
2365 match if the class was a negative one (e.g. [^ab]). This has been fixed.
2366
2367 (ii) A negated character class with a single character < 255 is coded as
2368 "not this character" (OP_NOT). This wasn't working properly when the test
2369 character was multibyte, either singly or repeated.
2370
2371 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
2372 mode, for example: \x{100}{2,3}.
2373
2374 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
2375 singly or repeated) now correctly test multibyte characters. However,
2376 PCRE doesn't recognize any characters with values greater than 255 as
2377 digits, spaces, or word characters. Such characters always match \D, \S,
2378 and \W, and never match \d, \s, or \w.
2379
2380 (v) Classes may now contain characters and character ranges with values
2381 greater than 255. For example: [ab\x{100}-\x{400}].
2382
2383 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
2384 PCRE in UTF-8 mode.
2385
2386 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
2387 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
2388 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
2389 value.)
2390
2391 53. The single man page has become too large. I have therefore split it up into
2392 a number of separate man pages. These also give rise to individual HTML pages;
2393 these are now put in a separate directory, and there is an index.html page that
2394 lists them all. Some hyperlinking between the pages has been installed.
2395
2396 54. Added convenience functions for handling named capturing parentheses.
2397
2398 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
2399 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
2400 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
2401 are faulted.
2402
2403 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
2404 calling configure. These values are used when compiling the dftables.c program
2405 which is run to generate the source of the default character tables. They
2406 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
2407 you will need to set these values.
2408
2409 57. Updated the building process for Windows DLL, as provided by Fred Cox.
2410
2411
2412 Version 3.9 02-Jan-02
2413 ---------------------
2414
2415 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
2416
2417 2. If --disable-static was given, the building process failed when trying to
2418 build pcretest and pcregrep. (For some reason it was using libtool to compile
2419 them, which is not right, as they aren't part of the library.)
2420
2421
2422 Version 3.8 18-Dec-01
2423 ---------------------
2424
2425 1. The experimental UTF-8 code was completely screwed up. It was packing the
2426 bytes in the wrong order. How dumb can you get?
2427
2428
2429 Version 3.7 29-Oct-01
2430 ---------------------
2431
2432 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
2433 This caused pcretest, when used on the test data, to segfault. Unfortunately,
2434 this didn't happen under Solaris 8, where I normally test things.
2435
2436 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
2437 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
2438 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
2439 this fix an hour or so after the initial 3.7 release.)
2440
2441
2442 Version 3.6 23-Oct-01
2443 ---------------------
2444
2445 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
2446 offsets passed as NULL with zero offset count.
2447
2448 2. The config.guess and config.sub files had not been updated when I moved to
2449 the latest autoconf.
2450
2451
2452 Version 3.5 15-Aug-01
2453 ---------------------
2454
2455 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
2456 had been forgotten.
2457
2458 2. By using declared but undefined structures, we can avoid using "void"
2459 definitions in pcre.h while keeping the internal definitions of the structures
2460 private.
2461
2462 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
2463 user point of view, this means that both static and shared libraries are built
2464 by default, but this can be individually controlled. More of the work of
2465 handling this static/shared cases is now inside libtool instead of PCRE's make
2466 file.
2467
2468 4. The pcretest utility is now installed along with pcregrep because it is
2469 useful for users (to test regexs) and by doing this, it automatically gets
2470 relinked by libtool. The documentation has been turned into a man page, so
2471 there are now .1, .txt, and .html versions in /doc.
2472
2473 5. Upgrades to pcregrep:
2474 (i) Added long-form option names like gnu grep.
2475 (ii) Added --help to list all options with an explanatory phrase.
2476 (iii) Added -r, --recursive to recurse into sub-directories.
2477 (iv) Added -f, --file to read patterns from a file.
2478
2479 6. pcre_exec() was referring to its "code" argument before testing that
2480 argument for NULL (and giving an error if it was NULL).
2481
2482 7. Upgraded Makefile.in to allow for compiling in a different directory from
2483 the source directory.
2484
2485 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
2486 options bits, the pointer it was passed was to an int instead of to an unsigned
2487 long int. This mattered only on 64-bit systems.
2488
2489 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
2490 generated) instead of pcre.in, which it its source. Also made the same change
2491 in several of the .c files.
2492
2493 10. A new release of gcc defines printf() as a macro, which broke pcretest
2494 because it had an ifdef in the middle of a string argument for printf(). Fixed
2495 by using separate calls to printf().
2496
2497 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
2498 script, to force use of CR or LF instead of \n in the source. On non-Unix
2499 systems, the value can be set in config.h.
2500
2501 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
2502 absolute limit. Changed the text of the error message to make this clear, and
2503 likewise updated the man page.
2504
2505 13. The limit of 99 on the number of capturing subpatterns has been removed.
2506 The new limit is 65535, which I hope will not be a "real" limit.
2507
2508
2509 Version 3.4 22-Aug-00
2510 ---------------------
2511
2512 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
2513
2514 2. Diagnose condition (?(0) as an error instead of crashing on matching.
2515
2516
2517 Version 3.3 01-Aug-00
2518 ---------------------
2519
2520 1. If an octal character was given, but the value was greater than \377, it
2521 was not getting masked to the least significant bits, as documented. This could
2522 lead to crashes in some systems.
2523
2524 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
2525 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
2526
2527 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
2528 These just pass their arguments on to (pcre_free)(), but they are provided
2529 because some uses of PCRE bind it to non-C systems that can call its functions,
2530 but cannot call free() or pcre_free() directly.
2531
2532 4. Add "make test" as a synonym for "make check". Corrected some comments in
2533 the Makefile.
2534
2535 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
2536 Makefile.
2537
2538 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
2539 command called pgrep for grepping around the active processes.
2540
2541 7. Added the beginnings of support for UTF-8 character strings.
2542
2543 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
2544 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
2545 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
2546 out for the ar command.)
2547
2548
2549 Version 3.2 12-May-00
2550 ---------------------
2551
2552 This is purely a bug fixing release.
2553
2554 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
2555 of ZA. This was just one example of several cases that could provoke this bug,
2556 which was introduced by change 9 of version 2.00. The code for breaking
2557 infinite loops after an iteration that matches an empty string was't working
2558 correctly.
2559
2560 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
2561 when matched against abbab (for example). After matching an empty string, it
2562 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
2563 caused it to match further down the string than it should.
2564
2565 3. The code contained an inclusion of sys/types.h. It isn't clear why this
2566 was there because it doesn't seem to be needed, and it causes trouble on some
2567 systems, as it is not a Standard C header. It has been removed.
2568
2569 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
2570 were reported on the Macintosh. The changes were from
2571
2572 while ((c = *(++ptr)) != 0 && c != '\n');
2573 to
2574 while ((c = *(++ptr)) != 0 && c != '\n') ;
2575
2576 Totally extraordinary, but if that's what it takes...
2577
2578 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
2579 available. Added HAVE_BCOPY and an autoconf test for it; if neither
2580 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
2581 assumes the way PCRE uses memmove() (always moving upwards).
2582
2583 6. PCRE is being used in one environment where strchr() is not available. There
2584 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
2585 faster code anyway.
2586
2587
2588 Version 3.1 09-Feb-00
2589 ---------------------
2590
2591 The only change in this release is the fixing of some bugs in Makefile.in for
2592 the "install" target:
2593
2594 (1) It was failing to install pcreposix.h.
2595
2596 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
2597
2598
2599 Version 3.0 01-Feb-00
2600 ---------------------
2601
2602 1. Add support for the /+ modifier to perltest (to output $` like it does in
2603 pcretest).
2604
2605 2. Add support for the /g modifier to perltest.
2606
2607 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
2608 matches null strings.
2609
2610 4. Fix perltest so that it doesn't do unwanted things when fed an empty
2611 pattern. Perl treats empty patterns specially - it reuses the most recent
2612 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
2613 effect.
2614
2615 5. The POSIX interface was broken in that it was just handing over the POSIX
2616 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
2617 required a bigger vector, with some working space on the end. This means that
2618 the POSIX wrapper now has to get and free some memory, and copy the results.
2619
2620 6. Added some simple autoconf support, placing the test data and the
2621 documentation in separate directories, re-organizing some of the
2622 information files, and making it build pcre-config (a GNU standard). Also added
2623 libtool support for building PCRE as a shared library, which is now the
2624 default.
2625
2626 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
2627 09 are not valid octal constants. Single digits will be used for minor values
2628 less than 10.
2629
2630 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
2631 existing programs that set these in the POSIX interface can use PCRE without
2632 modification.
2633
2634 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
2635 return all that pcre_info() returns, plus additional data. The pcre_info()
2636 function is retained for compatibility, but is considered to be obsolete.
2637
2638 10. Added experimental recursion feature (?R) to handle one common case that
2639 Perl 5.6 will be able to do with (?p{...}).
2640
2641 11. Added support for POSIX character classes like [:alpha:], which Perl is
2642 adopting.
2643
2644
2645 Version 2.08 31-Aug-99
2646 ----------------------
2647
2648 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
2649 trying to match at the startoffset position, but instead was moving forward to
2650 the next newline as if a previous match had failed.
2651
2652 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
2653 and could get into a loop if a null string was matched other than at the start
2654 of the subject.
2655
2656 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
2657 be distinguished at compile time, and for completeness also added PCRE_DATE.
2658
2659 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
2660 in GnuWin32 environments.
2661
2662
2663 Version 2.07 29-Jul-99
2664 ----------------------
2665
2666 1. The documentation is now supplied in plain text form and HTML as well as in
2667 the form of man page sources.
2668
2669 2. C++ compilers don't like assigning (void *) values to other pointer types.
2670 In particular this affects malloc(). Although there is no problem in Standard
2671 C, I've put in casts to keep C++ compilers happy.
2672
2673 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
2674 should be (const char *).
2675
2676 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
2677 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
2678 However, I haven't made this a standard facility. The documentation doesn't
2679 mention it, and the Makefile doesn't support it.
2680
2681 5. The Makefile now contains an "install" target, with editable destinations at
2682 the top of the file. The pcretest program is not installed.
2683
2684 6. pgrep -V now gives the PCRE version number and date.
2685
2686 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
2687 causing the entire string to be ignored, instead of just the last character.
2688
2689 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
2690 non-matching string, it can take a very, very long time, even for strings of
2691 quite modest length, because of the nested recursion. PCRE now does better in
2692 some of these cases. It does this by remembering the last required literal
2693 character in the pattern, and pre-searching the subject to ensure it is present
2694 before running the real match. In other words, it applies a heuristic to detect
2695 some types of certain failure quickly, and in the above example, if presented
2696 with a string that has no trailing " it gives "no match" very quickly.
2697
2698 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
2699 other alternatives are tried instead.
2700
2701
2702 Version 2.06 09-Jun-99
2703 ----------------------
2704
2705 1. Change pcretest's output for amount of store used to show just the code
2706 space, because the remainder (the data block) varies in size between 32-bit and
2707 64-bit systems.
2708
2709 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
2710 start matching at. This allows lookbehinds to work when searching for multiple
2711 occurrences in a string.
2712
2713 3. Added additional options to pcretest for testing multiple occurrences:
2714
2715 /+ outputs the rest of the string that follows a match
2716 /g loops for multiple occurrences, using the new startoffset argument
2717 /G loops for multiple occurrences by passing an incremented pointer
2718
2719 4. PCRE wasn't doing the "first character" optimization for patterns starting
2720 with \b or \B, though it was doing it for other lookbehind assertions. That is,
2721 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
2722 the letter 'x'. On long subject strings, this gives a significant speed-up.
2723
2724
2725 Version 2.05 21-Apr-99
2726 ----------------------
2727
2728 1. Changed the type of magic_number from int to long int so that it works
2729 properly on 16-bit systems.
2730
2731 2. Fixed a bug which caused patterns starting with .* not to work correctly
2732 when the subject string contained newline characters. PCRE was assuming
2733 anchoring for such patterns in all cases, which is not correct because .* will
2734 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
2735 DOTALL is set at top level; otherwise it knows that patterns starting with .*
2736 must be retried after every newline in the subject.
2737
2738
2739 Version 2.04 18-Feb-99
2740 ----------------------
2741
2742 1. For parenthesized subpatterns with repeats whose minimum was zero, the
2743 computation of the store needed to hold the pattern was incorrect (too large).
2744 If such patterns were nested a few deep, this could multiply and become a real
2745 problem.
2746
2747 2. Added /M option to pcretest to show the memory requirement of a specific
2748 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
2749
2750 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
2751 compiled in such a way that the backtracking after subsequent failure was
2752 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
2753 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
2754
2755
2756 Version 2.03 02-Feb-99
2757 ----------------------
2758
2759 1. Fixed typo and small mistake in man page.
2760
2761 2. Added 4th condition (GPL supersedes if conflict) and created separate
2762 LICENCE file containing the conditions.
2763
2764 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
2765 Perl, that is the internal \ allows the delimiter to be included in the
2766 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
2767 the final delimiter, add \ to the end of the pattern (to test the error).
2768
2769 4. Added the convenience functions for extracting substrings after a successful
2770 match. Updated pcretest to make it able to test these functions.
2771
2772
2773 Version 2.02 14-Jan-99
2774 ----------------------
2775
2776 1. Initialized the working variables associated with each extraction so that
2777 their saving and restoring doesn't refer to uninitialized store.
2778
2779 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
2780 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
2781 fix the problem.
2782
2783 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
2784 calls, and wasn't printing the correct value for compiling calls. Increased the
2785 default value of LOOPREPEAT, and the number of significant figures in the
2786 times.
2787
2788 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
2789
2790 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
2791 a building problem on Windows NT with a FAT file system.
2792
2793
2794 Version 2.01 21-Oct-98
2795 ----------------------
2796
2797 1. Changed the API for pcre_compile() to allow for the provision of a pointer
2798 to character tables built by pcre_maketables() in the current locale. If NULL
2799 is passed, the default tables are used.
2800
2801
2802 Version 2.00 24-Sep-98
2803 ----------------------
2804
2805 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
2806 it any more.
2807
2808 2. Allow quantification of (?>) groups, and make it work correctly.
2809
2810 3. The first character computation wasn't working for (?>) groups.
2811
2812 4. Correct the implementation of \Z (it is permitted to match on the \n at the
2813 end of the subject) and add 5.005's \z, which really does match only at the
2814 very end of the subject.
2815
2816 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
2817
2818 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
2819 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
2820 localized options. All options to pcre_study() were also removed.
2821
2822 7. Add other new features from 5.005:
2823
2824 $(?<= positive lookbehind
2825 $(?<! negative lookbehind
2826 (?imsx-imsx) added the unsetting capability
2827 such a setting is global if at outer level; local otherwise
2828 (?imsx-imsx:) non-capturing groups with option setting
2829 (?(cond)re|re) conditional pattern matching
2830
2831 A backreference to itself in a repeated group matches the previous
2832 captured string.
2833
2834 8. General tidying up of studying (both automatic and via "study")
2835 consequential on the addition of new assertions.
2836
2837 9. As in 5.005, unlimited repeated groups that could match an empty substring
2838 are no longer faulted at compile time. Instead, the loop is forcibly broken at
2839 runtime if any iteration does actually match an empty substring.
2840
2841 10. Include the RunTest script in the distribution.
2842
2843 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
2844 discrepancies, some of which were old and were also with respect to 5.004. They
2845 have now been fixed.
2846
2847
2848 Version 1.09 28-Apr-98
2849 ----------------------
2850
2851 1. A negated single character class followed by a quantifier with a minimum
2852 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
2853 program crashes, or just wrong answers. This did not apply to negated classes
2854 containing more than one character, or to minima other than one.
2855
2856
2857 Version 1.08 27-Mar-98
2858 ----------------------
2859
2860 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
2861
2862 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
2863 latter must appear before anything that relies on it in the pattern.
2864
2865
2866 Version 1.07 16-Feb-98
2867 ----------------------
2868
2869 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
2870 repeat of a potentially empty string).
2871
2872
2873 Version 1.06 23-Jan-98
2874 ----------------------
2875
2876 1. Added Markus Oberhumer's little patches for C++.
2877
2878 2. Literal strings longer than 255 characters were broken.
2879
2880
2881 Version 1.05 23-Dec-97
2882 ----------------------
2883
2884 1. Negated character classes containing more than one character were failing if
2885 PCRE_CASELESS was set at run time.
2886
2887
2888 Version 1.04 19-Dec-97
2889 ----------------------
2890
2891 1. Corrected the man page, where some "const" qualifiers had been omitted.
2892
2893 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
2894 input syntax.
2895
2896 3. Fixed memory leak which occurred when a regex with back references was
2897 matched with an offsets vector that wasn't big enough. The temporary memory
2898 that is used in this case wasn't being freed if the match failed.
2899
2900 4. Tidied pcretest to ensure it frees memory that it gets.
2901
2902 5. Temporary memory was being obtained in the case where the passed offsets
2903 vector was exactly big enough.
2904
2905 6. Corrected definition of offsetof() from change 5 below.
2906
2907 7. I had screwed up change 6 below and broken the rules for the use of
2908 setjmp(). Now fixed.
2909
2910
2911 Version 1.03 18-Dec-97
2912 ----------------------
2913
2914 1. A erroneous regex with a missing opening parenthesis was correctly
2915 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
2916 on some systems.
2917
2918 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
2919 it was reported that one broken compiler failed on the former because "code" is
2920 also an independent variable.
2921
2922 3. The erroneous regex a[]b caused an array overrun reference.
2923
2924 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
2925 fail on data ending with that character. (It was going on too far, and checking
2926 the next character, typically a binary zero.) This was specific to the
2927 optimized code for single-character negative classes.
2928
2929 5. Added a contributed patch from the TIN world which does the following:
2930
2931 + Add an undef for memmove, in case the the system defines a macro for it.
2932
2933 + Add a definition of offsetof(), in case there isn't one. (I don't know
2934 the reason behind this - offsetof() is part of the ANSI standard - but
2935 it does no harm).
2936
2937 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
2938 most of the places where whitespace preceded '#'. I have given up and
2939 allowed the remaining 2 cases to be at the margin.
2940
2941 + Rename some variables in pcre to eliminate shadowing. This seems very
2942 pedantic, but does no harm, of course.
2943
2944 6. Moved the call to setjmp() into its own function, to get rid of warnings
2945 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
2946
2947 7. Constructs such as \d{8,} were compiling into the equivalent of
2948 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
2949 outcome, but in this particular case used more store than had been allocated,
2950 which caused the bug to be discovered because it threw up an internal error.
2951
2952 8. The debugging code in both pcre and pcretest for outputting the compiled
2953 form of a regex was going wrong in the case of back references followed by
2954 curly-bracketed repeats.
2955
2956
2957 Version 1.02 12-Dec-97
2958 ----------------------
2959
2960 1. Typos in pcre.3 and comments in the source fixed.
2961
2962 2. Applied a contributed patch to get rid of places where it used to remove
2963 'const' from variables, and fixed some signed/unsigned and uninitialized
2964 variable warnings.
2965
2966 3. Added the "runtest" target to Makefile.
2967
2968 4. Set default compiler flag to -O2 rather than just -O.
2969
2970
2971 Version 1.01 19-Nov-97
2972 ----------------------
2973
2974 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
2975 like /([ab]*)*/, that is, for classes with more than one character in them.
2976
2977 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
2978 as /((?>a*))*/ (a PCRE_EXTRA facility).
2979
2980
2981 Version 1.00 18-Nov-97
2982 ----------------------
2983
2984 1. Added compile-time macros to support systems such as SunOS4 which don't have
2985 memmove() or strerror() but have other things that can be used instead.
2986
2987 2. Arranged that "make clean" removes the executables.
2988
2989
2990 Version 0.99 27-Oct-97
2991 ----------------------
2992
2993 1. Fixed bug in code for optimizing classes with only one character. It was
2994 initializing a 32-byte map regardless, which could cause it to run off the end
2995 of the memory it had got.
2996
2997 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
2998
2999
3000 Version 0.98 22-Oct-97
3001 ----------------------
3002
3003 1. Fixed bug in code for handling temporary memory usage when there are more
3004 back references than supplied space in the ovector. This could cause segfaults.
3005
3006
3007 Version 0.97 21-Oct-97
3008 ----------------------
3009
3010 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
3011
3012 2. Optimized negated single characters not to use a bit map.
3013
3014 3. Brought error texts together as macro definitions; clarified some of them;
3015 fixed one that was wrong - it said "range out of order" when it meant "invalid
3016 escape sequence".
3017
3018 4. Changed some char * arguments to const char *.
3019
3020 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
3021
3022 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
3023 pcretest.
3024
3025
3026 Version 0.96 16-Oct-97
3027 ----------------------
3028
3029 1. Added a simple "pgrep" utility to the distribution.
3030
3031 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
3032 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
3033 where "ddd" means "one or more decimal digits".
3034
3035 3. Fixed serious bug. If a pattern had a back reference, but the call to
3036 pcre_exec() didn't supply a large enough ovector to record the related
3037 identifying subpattern, the match always failed. PCRE now remembers the number
3038 of the largest back reference, and gets some temporary memory in which to save
3039 the offsets during matching if necessary, in order to ensure that
3040 backreferences always work.
3041
3042 4. Increased the compatibility with Perl in a number of ways:
3043
3044 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
3045 to request this handling. The option can be set at compile or exec time.
3046
3047 (b) $ matches before a terminating newline by default; an option
3048 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
3049 mode). The option can be set at compile or exec time.
3050
3051 (c) The handling of \ followed by a digit other than 0 is now supposed to be
3052 the same as Perl's. If the decimal number it represents is less than 10
3053 or there aren't that many previous left capturing parentheses, an octal
3054 escape is read. Inside a character class, it's always an octal escape,
3055 even if it is a single digit.
3056
3057 (d) An escaped but undefined alphabetic character is taken as a literal,
3058 unless PCRE_EXTRA is set. Currently this just reserves the remaining
3059 escapes.
3060
3061 (e) {0} is now permitted. (The previous item is removed from the compiled
3062 pattern).
3063
3064 5. Changed all the names of code files so that the basic parts are no longer
3065 than 10 characters, and abolished the teeny "globals.c" file.
3066
3067 6. Changed the handling of character classes; they are now done with a 32-byte
3068 bit map always.
3069
3070 7. Added the -d and /D options to pcretest to make it possible to look at the
3071 internals of compilation without having to recompile pcre.
3072
3073
3074 Version 0.95 23-Sep-97
3075 ----------------------
3076
3077 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
3078 \x20 at the start of a run of normal characters. These were being treated as
3079 real characters, instead of the source characters being re-checked.
3080
3081
3082 Version 0.94 18-Sep-97
3083 ----------------------
3084
3085 1. The functions are now thread-safe, with the caveat that the global variables
3086 containing pointers to malloc() and free() or alternative functions are the
3087 same for all threads.
3088
3089 2. Get pcre_study() to generate a bitmap of initial characters for non-
3090 anchored patterns when this is possible, and use it if passed to pcre_exec().
3091
3092
3093 Version 0.93 15-Sep-97
3094 ----------------------
3095
3096 1. /(b)|(:+)/ was computing an incorrect first character.
3097
3098 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
3099 but not actually doing anything yet.
3100
3101 3. Treat "-" characters in classes that cannot be part of ranges as literals,
3102 as Perl does (e.g. [-az] or [az-]).
3103
3104 4. Set the anchored flag if a branch starts with .* or .*? because that tests
3105 all possible positions.
3106
3107 5. Split up into different modules to avoid including unneeded functions in a
3108 compiled binary. However, compile and exec are still in one module. The "study"
3109 function is split off.
3110
3111 6. The character tables are now in a separate module whose source is generated
3112 by an auxiliary program - but can then be edited by hand if required. There are
3113 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
3114 toupper() in the code.
3115
3116 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
3117 make them global. Abolish the function for setting them, as the caller can now
3118 set them directly.
3119
3120
3121 Version 0.92 11-Sep-97
3122 ----------------------
3123
3124 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
3125 (e.g. /a{1,3}/) was broken (I mis-optimized it).
3126
3127 2. Caseless matching was not working in character classes if the characters in
3128 the pattern were in upper case.
3129
3130 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
3131
3132 4. Make PCRE_ANCHORED public and accept as a compile option.
3133
3134 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
3135 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
3136 pass them.
3137
3138 6. Give an error if bad option bits passed at compile or run time.
3139
3140 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
3141 pcretest to cause it to pass that flag.
3142
3143 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
3144 options, and the first character, if set.
3145
3146 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
3147
3148
3149 Version 0.91 10-Sep-97
3150 ----------------------
3151
3152 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
3153 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
3154
3155 2. PCRE was looping on encountering an indefinitely repeated back reference to
3156 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
3157 Perl does - treats the match as successful.
3158
3159 ****

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