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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 298 - (show annotations) (download)
Thu Jan 10 17:09:12 2008 UTC (6 years, 11 months ago) by ph10
File size: 148520 byte(s)
Tidies for the 7.5 release.

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

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