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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 308 - (show annotations) (download)
Mon Jan 21 14:57:19 2008 UTC (6 years, 6 months ago) by ph10
File size: 149880 byte(s)
Apply Craig's patch to move no_arg into the RE class.

1 ChangeLog for PCRE
2 ------------------
3
4 Version 7.6 19-Jan-08
5 ---------------------
6
7 1. A character class containing a very large number of characters with
8 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
9 overflow.
10
11 2. Patch to cut out the "long long" test in pcrecpp_unittest when
12 HAVE_LONG_LONG is not defined.
13
14 3. Applied Christian Ehrlicher's patch to update the CMake build files to
15 bring them up to date and include new features. This patch includes:
16
17 - Fixed PH's badly added libz and libbz2 support.
18 - Fixed a problem with static linking.
19 - Added pcredemo.
20 - Fixed dftables problem and added an option.
21 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
22 HAVE_LONG_LONG.
23 - Added readline support for pcretest.
24 - Added an listing of the option settings after cmake has run.
25
26 4. A user submitted a patch to Makefile that makes it easy to created a dll
27 under mingw. I added stuff to Makefile.am that cause it to include this
28 special target, without affecting anything else.
29
30 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
31 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
32 exported in the Windows port". It has not yet been confirmed that the patch
33 solves the problem, but it does no harm.
34
35
36 Version 7.5 10-Jan-08
37 ---------------------
38
39 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
40 values in parens when parsing an RE using the C++ wrapper."
41
42 2. Negative specials like \S did not work in character classes in UTF-8 mode.
43 Characters greater than 255 were excluded from the class instead of being
44 included.
45
46 3. The same bug as (2) above applied to negated POSIX classes such as
47 [:^space:].
48
49 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
50 defined or documented. It seems to have been a typo for PCRE_STATIC, so
51 I have changed it.
52
53 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
54 first named subpattern) and a construct such as (?&a) would reference the
55 first named subpattern whose name started with "a" (in other words, the
56 length check was missing). Both these problems are fixed. "Subpattern name
57 expected" is now given for (?&) (a zero-length name), and this patch also
58 makes it give the same error for \k'' (previously it complained that that
59 was a reference to a non-existent subpattern).
60
61 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
62 this is right because (?- can be followed by option settings as well as by
63 digits. I have, however, made the messages clearer.
64
65 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
66 than the number used in the conditional) now cause a compile-time error.
67 This is actually not compatible with Perl, which accepts such patterns, but
68 treats the conditional as always being FALSE (as PCRE used to), but it
69 seems to me that giving a diagnostic is better.
70
71 8. Change "alphameric" to the more common word "alphanumeric" in comments
72 and messages.
73
74 9. Fix two occurrences of "backslash" in comments that should have been
75 "backspace".
76
77 10. Remove two redundant lines of code that can never be obeyed (their function
78 was moved elsewhere).
79
80 11. The program that makes PCRE's Unicode character property table had a bug
81 which caused it to generate incorrect table entries for sequences of
82 characters that have the same character type, but are in different scripts.
83 It amalgamated them into a single range, with the script of the first of
84 them. In other words, some characters were in the wrong script. There were
85 thirteen such cases, affecting characters in the following ranges:
86
87 U+002b0 - U+002c1
88 U+0060c - U+0060d
89 U+0061e - U+00612
90 U+0064b - U+0065e
91 U+0074d - U+0076d
92 U+01800 - U+01805
93 U+01d00 - U+01d77
94 U+01d9b - U+01dbf
95 U+0200b - U+0200f
96 U+030fc - U+030fe
97 U+03260 - U+0327f
98 U+0fb46 - U+0fbb1
99 U+10450 - U+1049d
100
101 12. The -o option (show only the matching part of a line) for pcregrep was not
102 compatible with GNU grep in that, if there was more than one match in a
103 line, it showed only the first of them. It now behaves in the same way as
104 GNU grep.
105
106 13. If the -o and -v options were combined for pcregrep, it printed a blank
107 line for every non-matching line. GNU grep prints nothing, and pcregrep now
108 does the same. The return code can be used to tell if there were any
109 non-matching lines.
110
111 14. Added --file-offsets and --line-offsets to pcregrep.
112
113 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
114 infinitely looping recursion. The bug was that positive lookaheads were not
115 being skipped when checking for a possible empty match (negative lookaheads
116 and both kinds of lookbehind were skipped).
117
118 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
119 inclusion of <windows.h> to before rather than after the definition of
120 INVALID_FILE_ATTRIBUTES (patch from David Byron).
121
122 17. Specifying a possessive quantifier with a specific limit for a Unicode
123 character property caused pcre_compile() to compile bad code, which led at
124 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
125 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
126 caused the error; without that there was no problem.
127
128 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
129
130 19. Added --enable-pcretest-libreadline.
131
132 20. In pcrecpp.cc, the variable 'count' was incremented twice in
133 RE::GlobalReplace(). As a result, the number of replacements returned was
134 double what it should be. I removed one of the increments, but Craig sent a
135 later patch that removed the other one (the right fix) and added unit tests
136 that check the return values (which was not done before).
137
138 21. Several CMake things:
139
140 (1) Arranged that, when cmake is used on Unix, the libraries end up with
141 the names libpcre and libpcreposix, not just pcre and pcreposix.
142
143 (2) The above change means that pcretest and pcregrep are now correctly
144 linked with the newly-built libraries, not previously installed ones.
145
146 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
147
148 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
149 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
150 UTF-8 newline character). The key issue is that the pattern starts .*;
151 this means that the match must be either at the beginning, or after a
152 newline. The bug was in the code for advancing after a failed match and
153 checking that the new position followed a newline. It was not taking
154 account of UTF-8 characters correctly.
155
156 23. PCRE was behaving differently from Perl in the way it recognized POSIX
157 character classes. PCRE was not treating the sequence [:...:] as a
158 character class unless the ... were all letters. Perl, however, seems to
159 allow any characters between [: and :], though of course it rejects as
160 unknown any "names" that contain non-letters, because all the known class
161 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
162 for example, whereas PCRE did not - it did not recognize a POSIX character
163 class. This seemed a bit dangerous, so the code has been changed to be
164 closer to Perl. The behaviour is not identical to Perl, because PCRE will
165 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
166 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
167 Perl does, and where it didn't before.
168
169 24. Rewrite so as to remove the single use of %n from pcregrep because in some
170 Windows environments %n is disabled by default.
171
172
173 Version 7.4 21-Sep-07
174 ---------------------
175
176 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
177 means that a class such as [\s] counted as "explicit reference to CR or
178 LF". That isn't really right - the whole point of the change was to try to
179 help when there was an actual mention of one of the two characters. So now
180 the change happens only if \r or \n (or a literal CR or LF) character is
181 encountered.
182
183 2. The 32-bit options word was also used for 6 internal flags, but the numbers
184 of both had grown to the point where there were only 3 bits left.
185 Fortunately, there was spare space in the data structure, and so I have
186 moved the internal flags into a new 16-bit field to free up more option
187 bits.
188
189 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
190 but did not set the internal JCHANGED flag - either of these is enough to
191 control the way the "get" function works - but the PCRE_INFO_JCHANGED
192 facility is supposed to tell if (?J) was ever used, so now (?J) at the
193 start sets both bits.
194
195 4. Added options (at build time, compile time, exec time) to change \R from
196 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
197
198 5. doc/pcresyntax.html was missing from the distribution.
199
200 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
201 compatibility, even though it is no longer used.
202
203 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
204 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
205 windows.h file is present (where different names are used). [This was
206 reversed later after testing - see 16 below.]
207
208 8. Changed all #include <config.h> to #include "config.h". There were also
209 some further <pcre.h> cases that I changed to "pcre.h".
210
211 9. When pcregrep was used with the --colour option, it missed the line ending
212 sequence off the lines that it output.
213
214 10. It was pointed out to me that arrays of string pointers cause lots of
215 relocations when a shared library is dynamically loaded. A technique of
216 using a single long string with a table of offsets can drastically reduce
217 these. I have refactored PCRE in four places to do this. The result is
218 dramatic:
219
220 Originally: 290
221 After changing UCP table: 187
222 After changing error message table: 43
223 After changing table of "verbs" 36
224 After changing table of Posix names 22
225
226 Thanks to the folks working on Gregex for glib for this insight.
227
228 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
229 unicode-properties was also set.
230
231 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
232
233 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
234 checked only for CRLF.
235
236 14. Added casts to pcretest.c to avoid compiler warnings.
237
238 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
239
240 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
241 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
242 entirely. This removes changes made in 7 above.
243
244 17. The CMake files have been updated, and there is now more information about
245 building with CMake in the NON-UNIX-USE document.
246
247
248 Version 7.3 28-Aug-07
249 ---------------------
250
251 1. In the rejigging of the build system that eventually resulted in 7.1, the
252 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
253 brackets there is not right, since it causes compilers to look for an
254 installed pcre.h, not the version that is in the source that is being
255 compiled (which of course may be different). I have changed it back to:
256
257 #include "pcre.h"
258
259 I have a vague recollection that the change was concerned with compiling in
260 different directories, but in the new build system, that is taken care of
261 by the VPATH setting the Makefile.
262
263 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
264 when the subject happened to end in the byte 0x85 (e.g. if the last
265 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
266 characters but of course it shouldn't be taken as a newline when it is part
267 of another character. The bug was that, for an unlimited repeat of . in
268 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
269 characters when looking for a newline.
270
271 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
272
273 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
274 in debug output.
275
276 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
277 long printing in the pcrecpp unittest when running under MinGW.
278
279 6. ESC_K was left out of the EBCDIC table.
280
281 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
282 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
283 limit also applies to "virtual nesting" when a pattern is recursive, and in
284 this case 1000 isn't so big. I have been able to remove this limit at the
285 expense of backing off one optimization in certain circumstances. Normally,
286 when pcre_exec() would call its internal match() function recursively and
287 immediately return the result unconditionally, it uses a "tail recursion"
288 feature to save stack. However, when a subpattern that can match an empty
289 string has an unlimited repetition quantifier, it no longer makes this
290 optimization. That gives it a stack frame in which to save the data for
291 checking that an empty string has been matched. Previously this was taken
292 from the 1000-entry workspace that had been reserved. So now there is no
293 explicit limit, but more stack is used.
294
295 8. Applied Daniel's patches to solve problems with the import/export magic
296 syntax that is required for Windows, and which was going wrong for the
297 pcreposix and pcrecpp parts of the library. These were overlooked when this
298 problem was solved for the main library.
299
300 9. There were some crude static tests to avoid integer overflow when computing
301 the size of patterns that contain repeated groups with explicit upper
302 limits. As the maximum quantifier is 65535, the maximum group length was
303 set at 30,000 so that the product of these two numbers did not overflow a
304 32-bit integer. However, it turns out that people want to use groups that
305 are longer than 30,000 bytes (though not repeat them that many times).
306 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
307 made it possible to implement the integer overflow checks in a much more
308 dynamic way, which I have now done. The artificial limitation on group
309 length has been removed - we now have only the limit on the total length of
310 the compiled pattern, which depends on the LINK_SIZE setting.
311
312 10. Fixed a bug in the documentation for get/copy named substring when
313 duplicate names are permitted. If none of the named substrings are set, the
314 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
315 empty string.
316
317 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
318 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
319 because the ] is interpreted as the first data character and the
320 terminating ] is not found. PCRE has been made compatible with Perl in this
321 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
322 cause memory overwriting.
323
324 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
325 string has been matched (to stop an infinite loop). It was not recognizing
326 a conditional subpattern that could match an empty string if that
327 subpattern was within another subpattern. For example, it looped when
328 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
329 condition was not nested. This bug has been fixed.
330
331 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
332 past the start of the subject in the presence of bytes with the top bit
333 set, for example "\x8aBCD".
334
335 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
336 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
337
338 14. Optimized (?!) to (*FAIL).
339
340 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
341 This restricts code points to be within the range 0 to 0x10FFFF, excluding
342 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
343 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
344 does: it's just the validity check that is more restrictive.
345
346 16. Inserted checks for integer overflows during escape sequence (backslash)
347 processing, and also fixed erroneous offset values for syntax errors during
348 backslash processing.
349
350 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
351 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
352
353 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
354 caused an overrun.
355
356 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
357 something other than just ASCII characters) inside a group that had an
358 unlimited repeat caused a loop at compile time (while checking to see
359 whether the group could match an empty string).
360
361 20. Debugging a pattern containing \p or \P could cause a crash. For example,
362 [\P{Any}] did so. (Error in the code for printing property names.)
363
364 21. An orphan \E inside a character class could cause a crash.
365
366 22. A repeated capturing bracket such as (A)? could cause a wild memory
367 reference during compilation.
368
369 23. There are several functions in pcre_compile() that scan along a compiled
370 expression for various reasons (e.g. to see if it's fixed length for look
371 behind). There were bugs in these functions when a repeated \p or \P was
372 present in the pattern. These operators have additional parameters compared
373 with \d, etc, and these were not being taken into account when moving along
374 the compiled data. Specifically:
375
376 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
377 length.
378
379 (b) An item such as \pL+ within a repeated group could cause crashes or
380 loops.
381
382 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
383 "reference to non-existent subpattern" error.
384
385 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
386
387 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
388 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
389
390 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
391
392 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
393 character were causing crashes (broken optimization).
394
395 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
396 \p or \P) caused a compile-time loop.
397
398 28. More problems have arisen in unanchored patterns when CRLF is a valid line
399 break. For example, the unstudied pattern [\r\n]A does not match the string
400 "\r\nA" because change 7.0/46 below moves the current point on by two
401 characters after failing to match at the start. However, the pattern \nA
402 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
403 the same is true. There doesn't seem any very clean way out of this, but
404 what I have chosen to do makes the common cases work: PCRE now takes note
405 of whether there can be an explicit match for \r or \n anywhere in the
406 pattern, and if so, 7.0/46 no longer applies. As part of this change,
407 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
408 pattern has explicit CR or LF references.
409
410 29. Added (*CR) etc for changing newline setting at start of pattern.
411
412
413 Version 7.2 19-Jun-07
414 ---------------------
415
416 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
417 which is apparently normally available under Windows.
418
419 2. Re-jig the pcregrep tests with different newline settings in an attempt
420 to make them independent of the local environment's newline setting.
421
422 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
423
424 4. Some of the "internals" tests were previously cut out when the link size
425 was not 2, because the output contained actual offsets. The recent new
426 "Z" feature of pcretest means that these can be cut out, making the tests
427 usable with all link sizes.
428
429 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
430 stack recursion. This gives a massive performance boost under BSD, but just
431 a small improvement under Linux. However, it saves one field in the frame
432 in all cases.
433
434 6. Added more features from the forthcoming Perl 5.10:
435
436 (a) (?-n) (where n is a string of digits) is a relative subroutine or
437 recursion call. It refers to the nth most recently opened parentheses.
438
439 (b) (?+n) is also a relative subroutine call; it refers to the nth next
440 to be opened parentheses.
441
442 (c) Conditions that refer to capturing parentheses can be specified
443 relatively, for example, (?(-2)... or (?(+3)...
444
445 (d) \K resets the start of the current match so that everything before
446 is not part of it.
447
448 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
449
450 (f) \g{name} is another synonym - part of Perl 5.10's unification of
451 reference syntax.
452
453 (g) (?| introduces a group in which the numbering of parentheses in each
454 alternative starts with the same number.
455
456 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
457
458 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
459 PCRE_INFO_JCHANGED.
460
461 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
462 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
463 for detecting groups that can match an empty string.
464
465 9. A pattern with a very large number of alternatives (more than several
466 hundred) was running out of internal workspace during the pre-compile
467 phase, where pcre_compile() figures out how much memory will be needed. A
468 bit of new cunning has reduced the workspace needed for groups with
469 alternatives. The 1000-alternative test pattern now uses 12 bytes of
470 workspace instead of running out of the 4096 that are available.
471
472 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
473
474 11. Applied patch from Google to remove an optimization that didn't quite work.
475 The report of the bug said:
476
477 pcrecpp::RE("a*").FullMatch("aaa") matches, while
478 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
479 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
480
481 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
482 it matched the wrong number of bytes.
483
484
485 Version 7.1 24-Apr-07
486 ---------------------
487
488 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
489 that is more "standard", making use of automake and other Autotools. There
490 is some re-arrangement of the files and adjustment of comments consequent
491 on this.
492
493 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
494 for recursive directory scanning broke on some systems because the files
495 are not scanned in any specific order and on different systems the order
496 was different. A call to "sort" has been inserted into RunGrepTest for the
497 approprate test as a short-term fix. In the longer term there may be an
498 alternative.
499
500 3. I had an email from Eric Raymond about problems translating some of PCRE's
501 man pages to HTML (despite the fact that I distribute HTML pages, some
502 people do their own conversions for various reasons). The problems
503 concerned the use of low-level troff macros .br and .in. I have therefore
504 removed all such uses from the man pages (some were redundant, some could
505 be replaced by .nf/.fi pairs). The 132html script that I use to generate
506 HTML has been updated to handle .nf/.fi and to complain if it encounters
507 .br or .in.
508
509 4. Updated comments in configure.ac that get placed in config.h.in and also
510 arranged for config.h to be included in the distribution, with the name
511 config.h.generic, for the benefit of those who have to compile without
512 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
513
514 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
515 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
516 makevp.bat for the current PCRE, using the additional files
517 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
518
519 6. A Windows user reported a minor discrepancy with test 2, which turned out
520 to be caused by a trailing space on an input line that had got lost in his
521 copy. The trailing space was an accident, so I've just removed it.
522
523 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
524 that is needed.
525
526 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
527 as "const" (a) because they are and (b) because it helps the PHP
528 maintainers who have recently made a script to detect big data structures
529 in the php code that should be moved to the .rodata section. I remembered
530 to update Builducptable as well, so it won't revert if ucptable.h is ever
531 re-created.
532
533 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
534 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
535 order to be able to cut out the UTF-8 tables in the latter when UTF-8
536 support is not required. This saves 1.5-2K of code, which is important in
537 some applications.
538
539 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
540 so as not to refer to the tables, even though these functions will never be
541 called when UTF-8 support is disabled. Otherwise there are problems with a
542 shared library.
543
544 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
545
546 (a) It was defining its arguments as char * instead of void *.
547
548 (b) It was assuming that all moves were upwards in memory; this was true
549 a long time ago when I wrote it, but is no longer the case.
550
551 The emulated memove() is provided for those environments that have neither
552 memmove() nor bcopy(). I didn't think anyone used it these days, but that
553 is clearly not the case, as these two bugs were recently reported.
554
555 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
556 and Detrail to create the HTML documentation, the .txt form of the man
557 pages, and it removes trailing spaces from listed files. It also creates
558 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
559 case, it wraps all the #defines with #ifndefs. This script should be run
560 before "make dist".
561
562 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
563 with Unicode property support.
564
565 (a) For a maximizing quantifier, if the two different cases of the
566 character were of different lengths in their UTF-8 codings (there are
567 some cases like this - I found 11), and the matching function had to
568 back up over a mixture of the two cases, it incorrectly assumed they
569 were both the same length.
570
571 (b) When PCRE was configured to use the heap rather than the stack for
572 recursion during matching, it was not correctly preserving the data for
573 the other case of a UTF-8 character when checking ahead for a match
574 while processing a minimizing repeat. If the check also involved
575 matching a wide character, but failed, corruption could cause an
576 erroneous result when trying to check for a repeat of the original
577 character.
578
579 13. Some tidying changes to the testing mechanism:
580
581 (a) The RunTest script now detects the internal link size and whether there
582 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
583 values substituted by "configure". (The RunGrepTest script already did
584 this for UTF-8.) The configure.ac script no longer substitutes the
585 relevant variables.
586
587 (b) The debugging options /B and /D in pcretest show the compiled bytecode
588 with length and offset values. This means that the output is different
589 for different internal link sizes. Test 2 is skipped for link sizes
590 other than 2 because of this, bypassing the problem. Unfortunately,
591 there was also a test in test 3 (the locale tests) that used /B and
592 failed for link sizes other than 2. Rather than cut the whole test out,
593 I have added a new /Z option to pcretest that replaces the length and
594 offset values with spaces. This is now used to make test 3 independent
595 of link size. (Test 2 will be tidied up later.)
596
597 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
598 segmentation fault instead of returning the appropriate error message.
599
600 15. In multiline mode when the newline sequence was set to "any", the pattern
601 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
602 This doesn't seem right; it now treats the CRLF combination as the line
603 ending, and so does not match in that case. It's only a pattern such as ^$
604 that would hit this one: something like ^ABC$ would have failed after \r
605 and then tried again after \r\n.
606
607 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
608 in an attempt to make files that differ only in their line terminators
609 compare equal. This works on Linux.
610
611 17. Under certain error circumstances pcregrep might try to free random memory
612 as it exited. This is now fixed, thanks to valgrind.
613
614 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
615 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
616 was because its rules for how to advance for /g after matching an empty
617 string at the end of a line did not allow for this case. They now check for
618 it specially.
619
620 20. pcretest is supposed to handle patterns and data of any length, by
621 extending its buffers when necessary. It was getting this wrong when the
622 buffer for a data line had to be extended.
623
624 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
625 CRLF as a newline sequence.
626
627 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
628 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
629 I have nevertheless tidied it up.
630
631 23. Added some casts to kill warnings from HP-UX ia64 compiler.
632
633 24. Added a man page for pcre-config.
634
635
636 Version 7.0 19-Dec-06
637 ---------------------
638
639 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
640 moving to gcc 4.1.1.
641
642 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
643 sys/time.h, which is documented as needed for this function. It doesn't
644 seem to matter on Linux, but it showed up on some releases of OS X.
645
646 3. It seems that there are systems where bytes whose values are greater than
647 127 match isprint() in the "C" locale. The "C" locale should be the
648 default when a C program starts up. In most systems, only ASCII printing
649 characters match isprint(). This difference caused the output from pcretest
650 to vary, making some of the tests fail. I have changed pcretest so that:
651
652 (a) When it is outputting text in the compiled version of a pattern, bytes
653 other than 32-126 are always shown as hex escapes.
654
655 (b) When it is outputting text that is a matched part of a subject string,
656 it does the same, unless a different locale has been set for the match
657 (using the /L modifier). In this case, it uses isprint() to decide.
658
659 4. Fixed a major bug that caused incorrect computation of the amount of memory
660 required for a compiled pattern when options that changed within the
661 pattern affected the logic of the preliminary scan that determines the
662 length. The relevant options are -x, and -i in UTF-8 mode. The result was
663 that the computed length was too small. The symptoms of this bug were
664 either the PCRE error "internal error: code overflow" from pcre_compile(),
665 or a glibc crash with a message such as "pcretest: free(): invalid next
666 size (fast)". Examples of patterns that provoked this bug (shown in
667 pcretest format) are:
668
669 /(?-x: )/x
670 /(?x)(?-x: \s*#\s*)/
671 /((?i)[\x{c0}])/8
672 /(?i:[\x{c0}])/8
673
674 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
675 is now done differently.
676
677 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
678 wrapper classes; (b) implement a new function in the C++ scanner that is
679 more efficient than the old way of doing things because it avoids levels of
680 recursion in the regex matching; (c) add a paragraph to the documentation
681 for the FullMatch() function.
682
683 6. The escape sequence \n was being treated as whatever was defined as
684 "newline". Not only was this contrary to the documentation, which states
685 that \n is character 10 (hex 0A), but it also went horribly wrong when
686 "newline" was defined as CRLF. This has been fixed.
687
688 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
689 was being set to -1 for the "end of line" case (supposedly a value that no
690 character can have). Though this value is never used (the check for end of
691 line is "zero bytes in current character"), it caused compiler complaints.
692 I've changed it to 0xffffffff.
693
694 8. In pcre_version.c, the version string was being built by a sequence of
695 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
696 string (as it is for production releases) called a macro with an empty
697 argument. The C standard says the result of this is undefined. The gcc
698 compiler treats it as an empty string (which was what was wanted) but it is
699 reported that Visual C gives an error. The source has been hacked around to
700 avoid this problem.
701
702 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
703 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
704 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
705 of them did).
706
707 10. Originally, pcretest opened its input and output without "b"; then I was
708 told that "b" was needed in some environments, so it was added for release
709 5.0 to both the input and output. (It makes no difference on Unix-like
710 systems.) Later I was told that it is wrong for the input on Windows. I've
711 now abstracted the modes into two macros, to make it easier to fiddle with
712 them, and removed "b" from the input mode under Windows.
713
714 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
715
716 12. Added -help and --help to pcretest as an official way of being reminded
717 of the options.
718
719 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
720 and pcrecpp.cc because they annoy compilers at high warning levels.
721
722 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
723
724 15. Fixed an occurrence of == in configure.ac that should have been = (shell
725 scripts are not C programs :-) and which was not noticed because it works
726 on Linux.
727
728 16. pcretest is supposed to handle any length of pattern and data line (as one
729 line or as a continued sequence of lines) by extending its input buffer if
730 necessary. This feature was broken for very long pattern lines, leading to
731 a string of junk being passed to pcre_compile() if the pattern was longer
732 than about 50K.
733
734 17. I have done a major re-factoring of the way pcre_compile() computes the
735 amount of memory needed for a compiled pattern. Previously, there was code
736 that made a preliminary scan of the pattern in order to do this. That was
737 OK when PCRE was new, but as the facilities have expanded, it has become
738 harder and harder to keep it in step with the real compile phase, and there
739 have been a number of bugs (see for example, 4 above). I have now found a
740 cunning way of running the real compile function in a "fake" mode that
741 enables it to compute how much memory it would need, while actually only
742 ever using a few hundred bytes of working memory and without too many
743 tests of the mode. This should make future maintenance and development
744 easier. A side effect of this work is that the limit of 200 on the nesting
745 depth of parentheses has been removed (though this was never a serious
746 limitation, I suspect). However, there is a downside: pcre_compile() now
747 runs more slowly than before (30% or more, depending on the pattern). I
748 hope this isn't a big issue. There is no effect on runtime performance.
749
750 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
751 newline (only possible for the last line of a file) and it was a
752 pattern that set a locale (followed by /Lsomething), pcretest crashed.
753
754 19. Added additional timing features to pcretest. (1) The -tm option now times
755 matching only, not compiling. (2) Both -t and -tm can be followed, as a
756 separate command line item, by a number that specifies the number of
757 repeats to use when timing. The default is 50000; this gives better
758 precision, but takes uncomfortably long for very large patterns.
759
760 20. Extended pcre_study() to be more clever in cases where a branch of a
761 subpattern has no definite first character. For example, (a*|b*)[cd] would
762 previously give no result from pcre_study(). Now it recognizes that the
763 first character must be a, b, c, or d.
764
765 21. There was an incorrect error "recursive call could loop indefinitely" if
766 a subpattern (or the entire pattern) that was being tested for matching an
767 empty string contained only one non-empty item after a nested subpattern.
768 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
769 incorrectly, because the \d was being skipped in the check.
770
771 22. The pcretest program now has a new pattern option /B and a command line
772 option -b, which is equivalent to adding /B to every pattern. This causes
773 it to show the compiled bytecode, without the additional information that
774 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
775 is the same as /B/I).
776
777 23. A new optimization is now able automatically to treat some sequences such
778 as a*b as a*+b. More specifically, if something simple (such as a character
779 or a simple class like \d) has an unlimited quantifier, and is followed by
780 something that cannot possibly match the quantified thing, the quantifier
781 is automatically "possessified".
782
783 24. A recursive reference to a subpattern whose number was greater than 39
784 went wrong under certain circumstances in UTF-8 mode. This bug could also
785 have affected the operation of pcre_study().
786
787 25. Realized that a little bit of performance could be had by replacing
788 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
789
790 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
791
792 27. Possessive quantifiers such as a++ were previously implemented by turning
793 them into atomic groups such as ($>a+). Now they have their own opcodes,
794 which improves performance. This includes the automatically created ones
795 from 23 above.
796
797 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
798 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
799 the first matched character to be a colon. This applied both to named and
800 numbered groups.
801
802 29. The ucpinternal.h header file was missing its idempotency #ifdef.
803
804 30. I was sent a "project" file called libpcre.a.dev which I understand makes
805 building PCRE on Windows easier, so I have included it in the distribution.
806
807 31. There is now a check in pcretest against a ridiculously large number being
808 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
809 loop, the loop is abandoned.
810
811 32. Forward references to subpatterns in conditions such as (?(2)...) where
812 subpattern 2 is defined later cause pcre_compile() to search forwards in
813 the pattern for the relevant set of parentheses. This search went wrong
814 when there were unescaped parentheses in a character class, parentheses
815 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
816
817 33. "Subroutine" calls and backreferences were previously restricted to
818 referencing subpatterns earlier in the regex. This restriction has now
819 been removed.
820
821 34. Added a number of extra features that are going to be in Perl 5.10. On the
822 whole, these are just syntactic alternatives for features that PCRE had
823 previously implemented using the Python syntax or my own invention. The
824 other formats are all retained for compatibility.
825
826 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
827 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
828 also .NET compatible.
829
830 (b) A recursion or subroutine call to a named group can now be defined as
831 (?&name) as well as (?P>name).
832
833 (c) A backreference to a named group can now be defined as \k<name> or
834 \k'name' as well as (?P=name). The new forms, as well as being in Perl
835 5.10, are also .NET compatible.
836
837 (d) A conditional reference to a named group can now use the syntax
838 (?(<name>) or (?('name') as well as (?(name).
839
840 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
841 groups (named and numbered) that are never evaluated inline, but can be
842 called as "subroutines" from elsewhere. In effect, the DEFINE condition
843 is always false. There may be only one alternative in such a group.
844
845 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
846 as the simple (?(R). The condition is true only if the most recent
847 recursion is that of the given number or name. It does not search out
848 through the entire recursion stack.
849
850 (g) The escape \gN or \g{N} has been added, where N is a positive or
851 negative number, specifying an absolute or relative reference.
852
853 35. Tidied to get rid of some further signed/unsigned compiler warnings and
854 some "unreachable code" warnings.
855
856 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
857 things, this adds five new scripts.
858
859 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
860 There were also incompatibilities regarding the handling of \Q..\E inside
861 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
862 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
863
864 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
865 matches an empty string, and forcibly breaks the loop. There were bugs in
866 this code in non-simple cases. For a pattern such as ^(a()*)* matched
867 against aaaa the result was just "a" rather than "aaaa", for example. Two
868 separate and independent bugs (that affected different cases) have been
869 fixed.
870
871 39. Refactored the code to abolish the use of different opcodes for small
872 capturing bracket numbers. This is a tidy that I avoided doing when I
873 removed the limit on the number of capturing brackets for 3.5 back in 2001.
874 The new approach is not only tidier, it makes it possible to reduce the
875 memory needed to fix the previous bug (38).
876
877 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
878 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
879 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
880 mode.
881
882 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
883 report.
884
885 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
886 copy construction and assignment in the C++ wrapper.
887
888 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
889 couple of bugs that could have given wrong results in the "--newline=crlf"
890 case.
891
892 44. Added a number of casts and did some reorganization of signed/unsigned int
893 variables following suggestions from Dair Grant. Also renamed the variable
894 "this" as "item" because it is a C++ keyword.
895
896 45. Arranged for dftables to add
897
898 #include "pcre_internal.h"
899
900 to pcre_chartables.c because without it, gcc 4.x may remove the array
901 definition from the final binary if PCRE is built into a static library and
902 dead code stripping is activated.
903
904 46. For an unanchored pattern, if a match attempt fails at the start of a
905 newline sequence, and the newline setting is CRLF or ANY, and the next two
906 characters are CRLF, advance by two characters instead of one.
907
908
909 Version 6.7 04-Jul-06
910 ---------------------
911
912 1. In order to handle tests when input lines are enormously long, pcretest has
913 been re-factored so that it automatically extends its buffers when
914 necessary. The code is crude, but this _is_ just a test program. The
915 default size has been increased from 32K to 50K.
916
917 2. The code in pcre_study() was using the value of the re argument before
918 testing it for NULL. (Of course, in any sensible call of the function, it
919 won't be NULL.)
920
921 3. The memmove() emulation function in pcre_internal.h, which is used on
922 systems that lack both memmove() and bcopy() - that is, hardly ever -
923 was missing a "static" storage class specifier.
924
925 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
926 containing an extended class (one that cannot be represented by a bitmap
927 because it contains high-valued characters or Unicode property items, e.g.
928 [\pZ]). Almost always one would set UTF-8 mode when processing such a
929 pattern, but PCRE should not loop if you do not (it no longer does).
930 [Detail: two cases were found: (a) a repeated subpattern containing an
931 extended class; (b) a recursive reference to a subpattern that followed a
932 previous extended class. It wasn't skipping over the extended class
933 correctly when UTF-8 mode was not set.]
934
935 5. A negated single-character class was not being recognized as fixed-length
936 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
937 compile error "lookbehind assertion is not fixed length".
938
939 6. The RunPerlTest auxiliary script was showing an unexpected difference
940 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
941 write a Perl script that can interpret lines of an input file either as
942 byte characters or as UTF-8, which is what "perltest" was being required to
943 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
944 can't do is switch easily at run time between having the "use utf8;" pragma
945 or not. In the end, I fudged it by using the RunPerlTest script to insert
946 "use utf8;" explicitly for the UTF-8 tests.
947
948 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
949 the end of the subject string, contrary to the documentation and to what
950 Perl does. This was true of both matching functions. Now it matches only at
951 the start of the subject and immediately after *internal* newlines.
952
953 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
954 a pointer to an int instead of a pointer to an unsigned long int. This
955 caused problems on 64-bit systems.
956
957 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
958 instance of the 'standard' template library not being so standard".
959
960 10. There was no check on the number of named subpatterns nor the maximum
961 length of a subpattern name. The product of these values is used to compute
962 the size of the memory block for a compiled pattern. By supplying a very
963 long subpattern name and a large number of named subpatterns, the size
964 computation could be caused to overflow. This is now prevented by limiting
965 the length of names to 32 characters, and the number of named subpatterns
966 to 10,000.
967
968 11. Subpatterns that are repeated with specific counts have to be replicated in
969 the compiled pattern. The size of memory for this was computed from the
970 length of the subpattern and the repeat count. The latter is limited to
971 65535, but there was no limit on the former, meaning that integer overflow
972 could in principle occur. The compiled length of a repeated subpattern is
973 now limited to 30,000 bytes in order to prevent this.
974
975 12. Added the optional facility to have named substrings with the same name.
976
977 13. Added the ability to use a named substring as a condition, using the
978 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
979 are numbers (not recommended). Forward references are permitted.
980
981 14. Added forward references in named backreferences (if you see what I mean).
982
983 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
984 pattern could run off the end of the subject. For example, the pattern
985 "(?s)(.{1,5})"8 did this with the subject "ab".
986
987 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
988 PCRE_CASELESS was set when matching characters that were quantified with ?
989 or *.
990
991 17. A character class other than a single negated character that had a minimum
992 but no maximum quantifier - for example [ab]{6,} - was not handled
993 correctly by pce_dfa_exec(). It would match only one character.
994
995 18. A valid (though odd) pattern that looked like a POSIX character
996 class but used an invalid character after [ (for example [[,abc,]]) caused
997 pcre_compile() to give the error "Failed: internal error: code overflow" or
998 in some cases to crash with a glibc free() error. This could even happen if
999 the pattern terminated after [[ but there just happened to be a sequence of
1000 letters, a binary zero, and a closing ] in the memory that followed.
1001
1002 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
1003 over the years. Originally (before any Unicode support), just the bottom 8
1004 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
1005 output from "man perlunicode" includes this:
1006
1007 The regular expression compiler produces polymorphic opcodes. That
1008 is, the pattern adapts to the data and automatically switches to
1009 the Unicode character scheme when presented with Unicode data--or
1010 instead uses a traditional byte scheme when presented with byte
1011 data.
1012
1013 Sadly, a wide octal escape does not cause a switch, and in a string with
1014 no other multibyte characters, these octal escapes are treated as before.
1015 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
1016 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
1017 Unicode string.
1018
1019 I have not perpetrated such confusion in PCRE. Up till now, it took just
1020 the bottom 8 bits, as in old Perl. I have now made octal escapes with
1021 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
1022 translate to the appropriate multibyte character.
1023
1024 29. Applied some refactoring to reduce the number of warnings from Microsoft
1025 and Borland compilers. This has included removing the fudge introduced
1026 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
1027 a warning about an unused variable.
1028
1029 21. PCRE has not included VT (character 0x0b) in the set of whitespace
1030 characters since release 4.0, because Perl (from release 5.004) does not.
1031 [Or at least, is documented not to: some releases seem to be in conflict
1032 with the documentation.] However, when a pattern was studied with
1033 pcre_study() and all its branches started with \s, PCRE still included VT
1034 as a possible starting character. Of course, this did no harm; it just
1035 caused an unnecessary match attempt.
1036
1037 22. Removed a now-redundant internal flag bit that recorded the fact that case
1038 dependency changed within the pattern. This was once needed for "required
1039 byte" processing, but is no longer used. This recovers a now-scarce options
1040 bit. Also moved the least significant internal flag bit to the most-
1041 significant bit of the word, which was not previously used (hangover from
1042 the days when it was an int rather than a uint) to free up another bit for
1043 the future.
1044
1045 23. Added support for CRLF line endings as well as CR and LF. As well as the
1046 default being selectable at build time, it can now be changed at runtime
1047 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
1048 specify that it is scanning data with non-default line endings.
1049
1050 24. Changed the definition of CXXLINK to make it agree with the definition of
1051 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
1052
1053 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
1054 recursions. This makes a big different to stack usage for some patterns.
1055
1056 26. If a subpattern containing a named recursion or subroutine reference such
1057 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
1058 the space required for the compiled pattern went wrong and gave too small a
1059 value. Depending on the environment, this could lead to "Failed: internal
1060 error: code overflow at offset 49" or "glibc detected double free or
1061 corruption" errors.
1062
1063 27. Applied patches from Google (a) to support the new newline modes and (b) to
1064 advance over multibyte UTF-8 characters in GlobalReplace.
1065
1066 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
1067 difference for some implementation of PCRE in some Windows version.
1068
1069 29. Added some extra testing facilities to pcretest:
1070
1071 \q<number> in a data line sets the "match limit" value
1072 \Q<number> in a data line sets the "match recursion limt" value
1073 -S <number> sets the stack size, where <number> is in megabytes
1074
1075 The -S option isn't available for Windows.
1076
1077
1078 Version 6.6 06-Feb-06
1079 ---------------------
1080
1081 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
1082 in pcreposix.h. I have copied the definition from pcre.h.
1083
1084 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
1085 because pcre.h is no longer a built file.
1086
1087 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
1088 not normally included in the compiled code.
1089
1090
1091 Version 6.5 01-Feb-06
1092 ---------------------
1093
1094 1. When using the partial match feature with pcre_dfa_exec(), it was not
1095 anchoring the second and subsequent partial matches at the new starting
1096 point. This could lead to incorrect results. For example, with the pattern
1097 /1234/, partially matching against "123" and then "a4" gave a match.
1098
1099 2. Changes to pcregrep:
1100
1101 (a) All non-match returns from pcre_exec() were being treated as failures
1102 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
1103 error message is output. Some extra information is given for the
1104 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
1105 probably the only errors that are likely to be caused by users (by
1106 specifying a regex that has nested indefinite repeats, for instance).
1107 If there are more than 20 of these errors, pcregrep is abandoned.
1108
1109 (b) A binary zero was treated as data while matching, but terminated the
1110 output line if it was written out. This has been fixed: binary zeroes
1111 are now no different to any other data bytes.
1112
1113 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
1114 used to set a locale for matching. The --locale=xxxx long option has
1115 been added (no short equivalent) to specify a locale explicitly on the
1116 pcregrep command, overriding the environment variables.
1117
1118 (d) When -B was used with -n, some line numbers in the output were one less
1119 than they should have been.
1120
1121 (e) Added the -o (--only-matching) option.
1122
1123 (f) If -A or -C was used with -c (count only), some lines of context were
1124 accidentally printed for the final match.
1125
1126 (g) Added the -H (--with-filename) option.
1127
1128 (h) The combination of options -rh failed to suppress file names for files
1129 that were found from directory arguments.
1130
1131 (i) Added the -D (--devices) and -d (--directories) options.
1132
1133 (j) Added the -F (--fixed-strings) option.
1134
1135 (k) Allow "-" to be used as a file name for -f as well as for a data file.
1136
1137 (l) Added the --colo(u)r option.
1138
1139 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
1140 is not present by default.
1141
1142 3. A nasty bug was discovered in the handling of recursive patterns, that is,
1143 items such as (?R) or (?1), when the recursion could match a number of
1144 alternatives. If it matched one of the alternatives, but subsequently,
1145 outside the recursion, there was a failure, the code tried to back up into
1146 the recursion. However, because of the way PCRE is implemented, this is not
1147 possible, and the result was an incorrect result from the match.
1148
1149 In order to prevent this happening, the specification of recursion has
1150 been changed so that all such subpatterns are automatically treated as
1151 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
1152
1153 4. I had overlooked the fact that, in some locales, there are characters for
1154 which isalpha() is true but neither isupper() nor islower() are true. In
1155 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
1156 and ordfeminine) are like this. This affected the treatment of \w and \W
1157 when they appeared in character classes, but not when they appeared outside
1158 a character class. The bit map for "word" characters is now created
1159 separately from the results of isalnum() instead of just taking it from the
1160 upper, lower, and digit maps. (Plus the underscore character, of course.)
1161
1162 5. The above bug also affected the handling of POSIX character classes such as
1163 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
1164 permanent tables. Instead, the bit maps for such a class were previously
1165 created as the appropriate unions of the upper, lower, and digit bitmaps.
1166 Now they are created by subtraction from the [[:word:]] class, which has
1167 its own bitmap.
1168
1169 6. The [[:blank:]] character class matches horizontal, but not vertical space.
1170 It is created by subtracting the vertical space characters (\x09, \x0a,
1171 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
1172 subtraction was done in the overall bitmap for a character class, meaning
1173 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
1174 be recognized. This bug has been fixed.
1175
1176 7. Patches from the folks at Google:
1177
1178 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
1179 real life, but is still worth protecting against".
1180
1181 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
1182 regular expressions".
1183
1184 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
1185 have it.
1186
1187 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
1188 "configure" and the latter not, in order to fix a problem somebody had
1189 with compiling the Arg class on HP-UX.
1190
1191 (e) Improve the error-handling of the C++ wrapper a little bit.
1192
1193 (f) New tests for checking recursion limiting.
1194
1195 8. The pcre_memmove() function, which is used only if the environment does not
1196 have a standard memmove() function (and is therefore rarely compiled),
1197 contained two bugs: (a) use of int instead of size_t, and (b) it was not
1198 returning a result (though PCRE never actually uses the result).
1199
1200 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
1201 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
1202 returned instead of calling malloc() with an overflowing number that would
1203 most likely cause subsequent chaos.
1204
1205 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
1206
1207 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
1208 with this option is matched, the nmatch and pmatch options of regexec() are
1209 ignored.
1210
1211 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
1212 provided in case anyone wants to the the POSIX interface with UTF-8
1213 strings.
1214
1215 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
1216 C++ linking (needed for some HP-UX environments).
1217
1218 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
1219 (unused parameter) and in the pcre_printint() function (omitted "default"
1220 switch label when the default is to do nothing).
1221
1222 15. Added some code to make it possible, when PCRE is compiled as a C++
1223 library, to replace subject pointers for pcre_exec() with a smart pointer
1224 class, thus making it possible to process discontinuous strings.
1225
1226 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
1227 much the same function. They were added by different people who were trying
1228 to make PCRE easy to compile on non-Unix systems. It has been suggested
1229 that PCRE_EXPORT be abolished now that there is more automatic apparatus
1230 for compiling on Windows systems. I have therefore replaced it with
1231 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
1232 defaults to "extern" for C or "extern C" for C++, which works fine on
1233 Unix-like systems. It is now possible to override the value of PCRE_DATA_
1234 SCOPE with something explicit in config.h. In addition:
1235
1236 (a) pcreposix.h still had just "extern" instead of either of these macros;
1237 I have replaced it with PCRE_DATA_SCOPE.
1238
1239 (b) Functions such as _pcre_xclass(), which are internal to the library,
1240 but external in the C sense, all had PCRE_EXPORT in their definitions.
1241 This is apparently wrong for the Windows case, so I have removed it.
1242 (It makes no difference on Unix-like systems.)
1243
1244 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
1245 of recursive calls to match(). This is different to MATCH_LIMIT because
1246 that limits the total number of calls to match(), not all of which increase
1247 the depth of recursion. Limiting the recursion depth limits the amount of
1248 stack (or heap if NO_RECURSE is set) that is used. The default can be set
1249 when PCRE is compiled, and changed at run time. A patch from Google adds
1250 this functionality to the C++ interface.
1251
1252 18. Changes to the handling of Unicode character properties:
1253
1254 (a) Updated the table to Unicode 4.1.0.
1255
1256 (b) Recognize characters that are not in the table as "Cn" (undefined).
1257
1258 (c) I revised the way the table is implemented to a much improved format
1259 which includes recognition of ranges. It now supports the ranges that
1260 are defined in UnicodeData.txt, and it also amalgamates other
1261 characters into ranges. This has reduced the number of entries in the
1262 table from around 16,000 to around 3,000, thus reducing its size
1263 considerably. I realized I did not need to use a tree structure after
1264 all - a binary chop search is just as efficient. Having reduced the
1265 number of entries, I extended their size from 6 bytes to 8 bytes to
1266 allow for more data.
1267
1268 (d) Added support for Unicode script names via properties such as \p{Han}.
1269
1270 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
1271 matching that character.
1272
1273 20. When matching a repeated Unicode property with a minimum greater than zero,
1274 (for example \pL{2,}), PCRE could look past the end of the subject if it
1275 reached it while seeking the minimum number of characters. This could
1276 happen only if some of the characters were more than one byte long, because
1277 there is a check for at least the minimum number of bytes.
1278
1279 21. Refactored the implementation of \p and \P so as to be more general, to
1280 allow for more different types of property in future. This has changed the
1281 compiled form incompatibly. Anybody with saved compiled patterns that use
1282 \p or \P will have to recompile them.
1283
1284 22. Added "Any" and "L&" to the supported property types.
1285
1286 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
1287 but give a compile time error if the value is greater than 0xff.
1288
1289 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
1290 accidentally not being installed or uninstalled.
1291
1292 25. The pcre.h file was built from pcre.h.in, but the only changes that were
1293 made were to insert the current release number. This seemed silly, because
1294 it made things harder for people building PCRE on systems that don't run
1295 "configure". I have turned pcre.h into a distributed file, no longer built
1296 by "configure", with the version identification directly included. There is
1297 no longer a pcre.h.in file.
1298
1299 However, this change necessitated a change to the pcre-config script as
1300 well. It is built from pcre-config.in, and one of the substitutions was the
1301 release number. I have updated configure.ac so that ./configure now finds
1302 the release number by grepping pcre.h.
1303
1304 26. Added the ability to run the tests under valgrind.
1305
1306
1307 Version 6.4 05-Sep-05
1308 ---------------------
1309
1310 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
1311 "--" to be printed when multiple files were scanned, even when none of the
1312 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
1313 consider it to be a bug, and have restored the previous behaviour.
1314
1315 2. A couple of code tidies to get rid of compiler warnings.
1316
1317 3. The pcretest program used to cheat by referring to symbols in the library
1318 whose names begin with _pcre_. These are internal symbols that are not
1319 really supposed to be visible externally, and in some environments it is
1320 possible to suppress them. The cheating is now confined to including
1321 certain files from the library's source, which is a bit cleaner.
1322
1323 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
1324 file's purpose clearer.
1325
1326 5. Reorganized pcre_ucp_findchar().
1327
1328
1329 Version 6.3 15-Aug-05
1330 ---------------------
1331
1332 1. The file libpcre.pc.in did not have general read permission in the tarball.
1333
1334 2. There were some problems when building without C++ support:
1335
1336 (a) If C++ support was not built, "make install" and "make test" still
1337 tried to test it.
1338
1339 (b) There were problems when the value of CXX was explicitly set. Some
1340 changes have been made to try to fix these, and ...
1341
1342 (c) --disable-cpp can now be used to explicitly disable C++ support.
1343
1344 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
1345 backslash in a target when C++ was disabled. This confuses some
1346 versions of "make", apparently. Using an intermediate variable solves
1347 this. (Same for CPP_LOBJ.)
1348
1349 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
1350 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
1351 necessary on certain architectures.
1352
1353 4. Added a setting of -export-symbols-regex to the link command to remove
1354 those symbols that are exported in the C sense, but actually are local
1355 within the library, and not documented. Their names all begin with
1356 "_pcre_". This is not a perfect job, because (a) we have to except some
1357 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
1358 available (and never for static libraries). I have made a note to try to
1359 find a way round (a) in the future.
1360
1361
1362 Version 6.2 01-Aug-05
1363 ---------------------
1364
1365 1. There was no test for integer overflow of quantifier values. A construction
1366 such as {1111111111111111} would give undefined results. What is worse, if
1367 a minimum quantifier for a parenthesized subpattern overflowed and became
1368 negative, the calculation of the memory size went wrong. This could have
1369 led to memory overwriting.
1370
1371 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
1372
1373 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
1374 operating environments where this matters.
1375
1376 4. Applied Giuseppe Maxia's patch to add additional features for controlling
1377 PCRE options from within the C++ wrapper.
1378
1379 5. Named capturing subpatterns were not being correctly counted when a pattern
1380 was compiled. This caused two problems: (a) If there were more than 100
1381 such subpatterns, the calculation of the memory needed for the whole
1382 compiled pattern went wrong, leading to an overflow error. (b) Numerical
1383 back references of the form \12, where the number was greater than 9, were
1384 not recognized as back references, even though there were sufficient
1385 previous subpatterns.
1386
1387 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
1388 versions of gcc, e.g. 2.95.4.
1389
1390
1391 Version 6.1 21-Jun-05
1392 ---------------------
1393
1394 1. There was one reference to the variable "posix" in pcretest.c that was not
1395 surrounded by "#if !defined NOPOSIX".
1396
1397 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
1398 the cross-check on the old pcre_info() function, for the benefit of the
1399 cut-down version of PCRE that is currently imported into Exim.
1400
1401 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
1402 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
1403 patterns that start (?i) but I don't think that matters. The use of (?i) is
1404 just an example; this all applies to the other options as well.
1405
1406 4. Since libtool seems to echo the compile commands it is issuing, the output
1407 from "make" can be reduced a bit by putting "@" in front of each libtool
1408 compile command.
1409
1410 5. Patch from the folks at Google for configure.in to be a bit more thorough
1411 in checking for a suitable C++ installation before trying to compile the
1412 C++ stuff. This should fix a reported problem when a compiler was present,
1413 but no suitable headers.
1414
1415 6. The man pages all had just "PCRE" as their title. I have changed them to
1416 be the relevant file name. I have also arranged that these names are
1417 retained in the file doc/pcre.txt, which is a concatenation in text format
1418 of all the man pages except the little individual ones for each function.
1419
1420 7. The NON-UNIX-USE file had not been updated for the different set of source
1421 files that come with release 6. I also added a few comments about the C++
1422 wrapper.
1423
1424
1425 Version 6.0 07-Jun-05
1426 ---------------------
1427
1428 1. Some minor internal re-organization to help with my DFA experiments.
1429
1430 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
1431 didn't matter for the library itself when fully configured, but did matter
1432 when compiling without UCP support, or within Exim, where the ucp files are
1433 not imported.
1434
1435 3. Refactoring of the library code to split up the various functions into
1436 different source modules. The addition of the new DFA matching code (see
1437 below) to a single monolithic source would have made it really too
1438 unwieldy, quite apart from causing all the code to be include in a
1439 statically linked application, when only some functions are used. This is
1440 relevant even without the DFA addition now that patterns can be compiled in
1441 one application and matched in another.
1442
1443 The downside of splitting up is that there have to be some external
1444 functions and data tables that are used internally in different modules of
1445 the library but which are not part of the API. These have all had their
1446 names changed to start with "_pcre_" so that they are unlikely to clash
1447 with other external names.
1448
1449 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
1450 a different (DFA) algorithm. Although it is slower than the original
1451 function, it does have some advantages for certain types of matching
1452 problem.
1453
1454 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
1455 including restarting after a partial match.
1456
1457 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
1458 defined when compiling for Windows was sent to me. I have put it into the
1459 code, though I have no means of testing or verifying it.
1460
1461 7. Added the pcre_refcount() auxiliary function.
1462
1463 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
1464 match before or at the first newline in the subject string. In pcretest,
1465 the /f option on a pattern can be used to set this.
1466
1467 9. A repeated \w when used in UTF-8 mode with characters greater than 256
1468 would behave wrongly. This has been present in PCRE since release 4.0.
1469
1470 10. A number of changes to the pcregrep command:
1471
1472 (a) Refactored how -x works; insert ^(...)$ instead of setting
1473 PCRE_ANCHORED and checking the length, in preparation for adding
1474 something similar for -w.
1475
1476 (b) Added the -w (match as a word) option.
1477
1478 (c) Refactored the way lines are read and buffered so as to have more
1479 than one at a time available.
1480
1481 (d) Implemented a pcregrep test script.
1482
1483 (e) Added the -M (multiline match) option. This allows patterns to match
1484 over several lines of the subject. The buffering ensures that at least
1485 8K, or the rest of the document (whichever is the shorter) is available
1486 for matching (and similarly the previous 8K for lookbehind assertions).
1487
1488 (f) Changed the --help output so that it now says
1489
1490 -w, --word-regex(p)
1491
1492 instead of two lines, one with "regex" and the other with "regexp"
1493 because that confused at least one person since the short forms are the
1494 same. (This required a bit of code, as the output is generated
1495 automatically from a table. It wasn't just a text change.)
1496
1497 (g) -- can be used to terminate pcregrep options if the next thing isn't an
1498 option but starts with a hyphen. Could be a pattern or a path name
1499 starting with a hyphen, for instance.
1500
1501 (h) "-" can be given as a file name to represent stdin.
1502
1503 (i) When file names are being printed, "(standard input)" is used for
1504 the standard input, for compatibility with GNU grep. Previously
1505 "<stdin>" was used.
1506
1507 (j) The option --label=xxx can be used to supply a name to be used for
1508 stdin when file names are being printed. There is no short form.
1509
1510 (k) Re-factored the options decoding logic because we are going to add
1511 two more options that take data. Such options can now be given in four
1512 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
1513
1514 (l) Added the -A, -B, and -C options for requesting that lines of context
1515 around matches be printed.
1516
1517 (m) Added the -L option to print the names of files that do not contain
1518 any matching lines, that is, the complement of -l.
1519
1520 (n) The return code is 2 if any file cannot be opened, but pcregrep does
1521 continue to scan other files.
1522
1523 (o) The -s option was incorrectly implemented. For compatibility with other
1524 greps, it now suppresses the error message for a non-existent or non-
1525 accessible file (but not the return code). There is a new option called
1526 -q that suppresses the output of matching lines, which was what -s was
1527 previously doing.
1528
1529 (p) Added --include and --exclude options to specify files for inclusion
1530 and exclusion when recursing.
1531
1532 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
1533 Hopefully, it now does.
1534
1535 12. Missing cast in pcre_study().
1536
1537 13. Added an "uninstall" target to the makefile.
1538
1539 14. Replaced "extern" in the function prototypes in Makefile.in with
1540 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
1541 world, but is set differently for Windows.
1542
1543 15. Added a second compiling function called pcre_compile2(). The only
1544 difference is that it has an extra argument, which is a pointer to an
1545 integer error code. When there is a compile-time failure, this is set
1546 non-zero, in addition to the error test pointer being set to point to an
1547 error message. The new argument may be NULL if no error number is required
1548 (but then you may as well call pcre_compile(), which is now just a
1549 wrapper). This facility is provided because some applications need a
1550 numeric error indication, but it has also enabled me to tidy up the way
1551 compile-time errors are handled in the POSIX wrapper.
1552
1553 16. Added VPATH=.libs to the makefile; this should help when building with one
1554 prefix path and installing with another. (Or so I'm told by someone who
1555 knows more about this stuff than I do.)
1556
1557 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
1558 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
1559 match everything, including newlines. This is not POSIX-compatible, but
1560 somebody wanted the feature. From pcretest it can be activated by using
1561 both the P and the s flags.
1562
1563 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
1564
1565 19. libpcre.pc was being incorrectly installed as executable.
1566
1567 20. A couple of places in pcretest check for end-of-line by looking for '\n';
1568 it now also looks for '\r' so that it will work unmodified on Windows.
1569
1570 21. Added Google's contributed C++ wrapper to the distribution.
1571
1572 22. Added some untidy missing memory free() calls in pcretest, to keep
1573 Electric Fence happy when testing.
1574
1575
1576
1577 Version 5.0 13-Sep-04
1578 ---------------------
1579
1580 1. Internal change: literal characters are no longer packed up into items
1581 containing multiple characters in a single byte-string. Each character
1582 is now matched using a separate opcode. However, there may be more than one
1583 byte in the character in UTF-8 mode.
1584
1585 2. The pcre_callout_block structure has two new fields: pattern_position and
1586 next_item_length. These contain the offset in the pattern to the next match
1587 item, and its length, respectively.
1588
1589 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
1590 insertion of callouts before each pattern item. Added the /C option to
1591 pcretest to make use of this.
1592
1593 4. On the advice of a Windows user, the lines
1594
1595 #if defined(_WIN32) || defined(WIN32)
1596 _setmode( _fileno( stdout ), 0x8000 );
1597 #endif /* defined(_WIN32) || defined(WIN32) */
1598
1599 have been added to the source of pcretest. This apparently does useful
1600 magic in relation to line terminators.
1601
1602 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
1603 for the benefit of those environments where the "b" makes a difference.
1604
1605 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
1606 to know about it. I have put a hack into configure.in that adds in code
1607 to set GCC=yes if CC=icc. This seems to end up at a point in the
1608 generated configure script that is early enough to affect the setting of
1609 compiler options, which is what is needed, but I have no means of testing
1610 whether it really works. (The user who reported this had patched the
1611 generated configure script, which of course I cannot do.)
1612
1613 LATER: After change 22 below (new libtool files), the configure script
1614 seems to know about icc (and also ecc). Therefore, I have commented out
1615 this hack in configure.in.
1616
1617 7. Added support for pkg-config (2 patches were sent in).
1618
1619 8. Negated POSIX character classes that used a combination of internal tables
1620 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
1621 [[:^ascii]]. Typically, they would match almost any characters. The other
1622 POSIX classes were not broken in this way.
1623
1624 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
1625 to find the match, as PCRE was deluded into thinking that the match had to
1626 start at the start point or following a newline. The same bug applied to
1627 patterns with negative forward assertions or any backward assertions
1628 preceding ".*" at the start, unless the pattern required a fixed first
1629 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
1630
1631 10. In UTF-8 mode, when moving forwards in the subject after a failed match
1632 starting at the last subject character, bytes beyond the end of the subject
1633 string were read.
1634
1635 11. Renamed the variable "class" as "classbits" to make life easier for C++
1636 users. (Previously there was a macro definition, but it apparently wasn't
1637 enough.)
1638
1639 12. Added the new field "tables" to the extra data so that tables can be passed
1640 in at exec time, or the internal tables can be re-selected. This allows
1641 a compiled regex to be saved and re-used at a later time by a different
1642 program that might have everything at different addresses.
1643
1644 13. Modified the pcre-config script so that, when run on Solaris, it shows a
1645 -R library as well as a -L library.
1646
1647 14. The debugging options of pcretest (-d on the command line or D on a
1648 pattern) showed incorrect output for anything following an extended class
1649 that contained multibyte characters and which was followed by a quantifier.
1650
1651 15. Added optional support for general category Unicode character properties
1652 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
1653 support. It adds about 90K to the size of the library. The meanings of the
1654 inbuilt class escapes such as \d and \s have NOT been changed.
1655
1656 16. Updated pcredemo.c to include calls to free() to release the memory for the
1657 compiled pattern.
1658
1659 17. The generated file chartables.c was being created in the source directory
1660 instead of in the building directory. This caused the build to fail if the
1661 source directory was different from the building directory, and was
1662 read-only.
1663
1664 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
1665 file. No doubt somebody will tell me if they don't make sense... Also added
1666 Dan Mooney's comments about building on OpenVMS.
1667
1668 19. Added support for partial matching via the PCRE_PARTIAL option for
1669 pcre_exec() and the \P data escape in pcretest.
1670
1671 20. Extended pcretest with 3 new pattern features:
1672
1673 (i) A pattern option of the form ">rest-of-line" causes pcretest to
1674 write the compiled pattern to the file whose name is "rest-of-line".
1675 This is a straight binary dump of the data, with the saved pointer to
1676 the character tables forced to be NULL. The study data, if any, is
1677 written too. After writing, pcretest reads a new pattern.
1678
1679 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
1680 compiled pattern from the given file. There must not be any
1681 occurrences of "<" in the file name (pretty unlikely); if there are,
1682 pcretest will instead treat the initial "<" as a pattern delimiter.
1683 After reading in the pattern, pcretest goes on to read data lines as
1684 usual.
1685
1686 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
1687 and 16-bit fields in a compiled pattern, to simulate a pattern that
1688 was compiled on a host of opposite endianness.
1689
1690 21. The pcre-exec() function can now cope with patterns that were compiled on
1691 hosts of opposite endianness, with this restriction:
1692
1693 As for any compiled expression that is saved and used later, the tables
1694 pointer field cannot be preserved; the extra_data field in the arguments
1695 to pcre_exec() should be used to pass in a tables address if a value
1696 other than the default internal tables were used at compile time.
1697
1698 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
1699 now diagnosed as an error. Previously, most of the time, a negative number
1700 would have been treated as zero, but if in addition "ovector" was passed as
1701 NULL, a crash could occur.
1702
1703 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
1704 new versions from the libtool 1.5 distribution (the last one is a copy of
1705 a file called libtool.m4). This seems to have fixed the need to patch
1706 "configure" to support Darwin 1.3 (which I used to do). However, I still
1707 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
1708 workstation).
1709
1710 24. Changed the PCRE licence to be the more standard "BSD" licence.
1711
1712
1713 Version 4.5 01-Dec-03
1714 ---------------------
1715
1716 1. There has been some re-arrangement of the code for the match() function so
1717 that it can be compiled in a version that does not call itself recursively.
1718 Instead, it keeps those local variables that need separate instances for
1719 each "recursion" in a frame on the heap, and gets/frees frames whenever it
1720 needs to "recurse". Keeping track of where control must go is done by means
1721 of setjmp/longjmp. The whole thing is implemented by a set of macros that
1722 hide most of the details from the main code, and operates only if
1723 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
1724 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
1725 operating.
1726
1727 To make it easier for callers to provide specially tailored get/free
1728 functions for this usage, two new functions, pcre_stack_malloc, and
1729 pcre_stack_free, are used. They are always called in strict stacking order,
1730 and the size of block requested is always the same.
1731
1732 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
1733 PCRE has been compiled to use the stack or the heap for recursion. The
1734 -C option of pcretest uses this to show which version is compiled.
1735
1736 A new data escape \S, is added to pcretest; it causes the amounts of store
1737 obtained and freed by both kinds of malloc/free at match time to be added
1738 to the output.
1739
1740 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
1741 what's available on my current Linux desktop machine.
1742
1743 3. When matching a UTF-8 string, the test for a valid string at the start has
1744 been extended. If start_offset is not zero, PCRE now checks that it points
1745 to a byte that is the start of a UTF-8 character. If not, it returns
1746 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
1747 this is necessary because there may be backward assertions in the pattern.
1748 When matching the same subject several times, it may save resources to use
1749 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
1750
1751 4. The code for checking the validity of UTF-8 strings has been tightened so
1752 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
1753 containing "overlong sequences".
1754
1755 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
1756 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
1757 should have been "&", but it just so happened that all the cases this let
1758 through by mistake were picked up later in the function.
1759
1760 6. I had used a variable called "isblank" - this is a C99 function, causing
1761 some compilers to warn. To avoid this, I renamed it (as "blankclass").
1762
1763 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
1764 prompting; (b) run "./pcretest /dev/null" at the start of the test script
1765 so the version is shown; (c) stop "make test" echoing "./RunTest".
1766
1767 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
1768
1769 9. The prototype for memmove() for systems that don't have it was using
1770 size_t, but the inclusion of the header that defines size_t was later. I've
1771 moved the #includes for the C headers earlier to avoid this.
1772
1773 10. Added some adjustments to the code to make it easier to compiler on certain
1774 special systems:
1775
1776 (a) Some "const" qualifiers were missing.
1777 (b) Added the macro EXPORT before all exported functions; by default this
1778 is defined to be empty.
1779 (c) Changed the dftables auxiliary program (that builds chartables.c) so
1780 that it reads its output file name as an argument instead of writing
1781 to the standard output and assuming this can be redirected.
1782
1783 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
1784 class containing characters with values greater than 255, PCRE compilation
1785 went into a loop.
1786
1787 12. A recursive reference to a subpattern that was within another subpattern
1788 that had a minimum quantifier of zero caused PCRE to crash. For example,
1789 (x(y(?2))z)? provoked this bug with a subject that got as far as the
1790 recursion. If the recursively-called subpattern itself had a zero repeat,
1791 that was OK.
1792
1793 13. In pcretest, the buffer for reading a data line was set at 30K, but the
1794 buffer into which it was copied (for escape processing) was still set at
1795 1024, so long lines caused crashes.
1796
1797 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
1798 "internal error: code overflow...". This applied to any character class
1799 that was followed by a possessive quantifier.
1800
1801 15. Modified the Makefile to add libpcre.la as a prerequisite for
1802 libpcreposix.la because I was told this is needed for a parallel build to
1803 work.
1804
1805 16. If a pattern that contained .* following optional items at the start was
1806 studied, the wrong optimizing data was generated, leading to matching
1807 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
1808 matching string must start with a or b or c. The correct conclusion for
1809 this pattern is that a match can start with any character.
1810
1811
1812 Version 4.4 13-Aug-03
1813 ---------------------
1814
1815 1. In UTF-8 mode, a character class containing characters with values between
1816 127 and 255 was not handled correctly if the compiled pattern was studied.
1817 In fixing this, I have also improved the studying algorithm for such
1818 classes (slightly).
1819
1820 2. Three internal functions had redundant arguments passed to them. Removal
1821 might give a very teeny performance improvement.
1822
1823 3. Documentation bug: the value of the capture_top field in a callout is *one
1824 more than* the number of the hightest numbered captured substring.
1825
1826 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
1827 in incorrectly linking with a previously installed version. They now link
1828 explicitly with libpcre.la.
1829
1830 5. configure.in no longer needs to recognize Cygwin specially.
1831
1832 6. A problem in pcre.in for Windows platforms is fixed.
1833
1834 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
1835 pcretest, it used to include the size of the study block as part of its
1836 output. Unfortunately, the structure contains a field that has a different
1837 size on different hardware architectures. This meant that the tests that
1838 showed this size failed. As the block is currently always of a fixed size,
1839 this information isn't actually particularly useful in pcretest output, so
1840 I have just removed it.
1841
1842 8. Three pre-processor statements accidentally did not start in column 1.
1843 Sadly, there are *still* compilers around that complain, even though
1844 standard C has not required this for well over a decade. Sigh.
1845
1846 9. In pcretest, the code for checking callouts passed small integers in the
1847 callout_data field, which is a void * field. However, some picky compilers
1848 complained about the casts involved for this on 64-bit systems. Now
1849 pcretest passes the address of the small integer instead, which should get
1850 rid of the warnings.
1851
1852 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
1853 both compile and run time, and gives an error if an invalid UTF-8 sequence
1854 is found. There is a option for disabling this check in cases where the
1855 string is known to be correct and/or the maximum performance is wanted.
1856
1857 11. In response to a bug report, I changed one line in Makefile.in from
1858
1859 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
1860 to
1861 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
1862
1863 to look similar to other lines, but I have no way of telling whether this
1864 is the right thing to do, as I do not use Windows. No doubt I'll get told
1865 if it's wrong...
1866
1867
1868 Version 4.3 21-May-03
1869 ---------------------
1870
1871 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
1872 Makefile.
1873
1874 2. Some refactoring to improve the quality of the code:
1875
1876 (i) The utf8_table... variables are now declared "const".
1877
1878 (ii) The code for \cx, which used the "case flipping" table to upper case
1879 lower case letters, now just substracts 32. This is ASCII-specific,
1880 but the whole concept of \cx is ASCII-specific, so it seems
1881 reasonable.
1882
1883 (iii) PCRE was using its character types table to recognize decimal and
1884 hexadecimal digits in the pattern. This is silly, because it handles
1885 only 0-9, a-f, and A-F, but the character types table is locale-
1886 specific, which means strange things might happen. A private
1887 table is now used for this - though it costs 256 bytes, a table is
1888 much faster than multiple explicit tests. Of course, the standard
1889 character types table is still used for matching digits in subject
1890 strings against \d.
1891
1892 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
1893 ending in _t are). So I've renamed it as ESC_tee.
1894
1895 3. The first argument for regexec() in the POSIX wrapper should have been
1896 defined as "const".
1897
1898 4. Changed pcretest to use malloc() for its buffers so that they can be
1899 Electric Fenced for debugging.
1900
1901 5. There were several places in the code where, in UTF-8 mode, PCRE would try
1902 to read one or more bytes before the start of the subject string. Often this
1903 had no effect on PCRE's behaviour, but in some circumstances it could
1904 provoke a segmentation fault.
1905
1906 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
1907 to try to read one or more bytes before the start of the subject string.
1908
1909 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
1910 UTF-8 support could misbehave in various ways if the subject string
1911 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
1912 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
1913 back over UTF-8 characters.)
1914
1915
1916 Version 4.2 14-Apr-03
1917 ---------------------
1918
1919 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
1920
1921 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
1922 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
1923 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
1924 [WIN_PREFIX]: new variable, "cyg" for Cygwin
1925 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
1926 and BUILD_EXEEXT
1927 Note: automatic setting of the BUILD variables is not yet working
1928 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
1929 compile-time but not at link-time
1930 [LINK]: use for linking executables only
1931 make different versions for Windows and non-Windows
1932 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
1933 libraries
1934 [LINK_FOR_BUILD]: new variable
1935 [OBJEXT]: use throughout
1936 [EXEEXT]: use throughout
1937 <winshared>: new target
1938 <wininstall>: new target
1939 <dftables.o>: use native compiler
1940 <dftables>: use native linker
1941 <install>: handle Windows platform correctly
1942 <clean>: ditto
1943 <check>: ditto
1944 copy DLL to top builddir before testing
1945
1946 As part of these changes, -no-undefined was removed again. This was reported
1947 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
1948 in any case.
1949
1950 3. Some tidies to get rid of compiler warnings:
1951
1952 . In the match_data structure, match_limit was an unsigned long int, whereas
1953 match_call_count was an int. I've made them both unsigned long ints.
1954
1955 . In pcretest the fact that a const uschar * doesn't automatically cast to
1956 a void * provoked a warning.
1957
1958 . Turning on some more compiler warnings threw up some "shadow" variables
1959 and a few more missing casts.
1960
1961 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
1962 option, a class that contained a single character with a value between 128
1963 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
1964
1965 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
1966 option, a class that contained several characters, but with at least one
1967 whose value was between 128 and 255 caused PCRE to crash.
1968
1969
1970 Version 4.1 12-Mar-03
1971 ---------------------
1972
1973 1. Compiling with gcc -pedantic found a couple of places where casts were
1974 needed, and a string in dftables.c that was longer than standard compilers are
1975 required to support.
1976
1977 2. Compiling with Sun's compiler found a few more places where the code could
1978 be tidied up in order to avoid warnings.
1979
1980 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
1981 first of these names is deprecated in the latest Autoconf in favour of the name
1982 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
1983 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
1984 analogy I have changed it to CFLAGS_FOR_BUILD.
1985
1986 4. Added -no-undefined to the linking command in the Makefile, because this is
1987 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
1988 linking step for the pcreposix library.
1989
1990 5. PCRE was failing to diagnose the case of two named groups with the same
1991 name.
1992
1993 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
1994 literal character that is needed in the subject for a match, and scans along to
1995 ensure that it is present before embarking on the full matching process. This
1996 saves time in cases of nested unlimited repeats that are never going to match.
1997 Problem: the scan can take a lot of time if the subject is very long (e.g.
1998 megabytes), thus penalizing straightforward matches. It is now done only if the
1999 amount of subject to be scanned is less than 1000 bytes.
2000
2001 7. A lesser problem with the same optimization is that it was recording the
2002 first character of an anchored pattern as "needed", thus provoking a search
2003 right along the subject, even when the first match of the pattern was going to
2004 fail. The "needed" character is now not set for anchored patterns, unless it
2005 follows something in the pattern that is of non-fixed length. Thus, it still
2006 fulfils its original purpose of finding quick non-matches in cases of nested
2007 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
2008
2009
2010 Version 4.0 17-Feb-03
2011 ---------------------
2012
2013 1. If a comment in an extended regex that started immediately after a meta-item
2014 extended to the end of string, PCRE compiled incorrect data. This could lead to
2015 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
2016
2017 2. Moved to autoconf 2.53 and libtool 1.4.2.
2018
2019 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
2020 the special perltest8 script is no longer needed - all the tests can be run
2021 from a single perltest script.
2022
2023 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
2024 by \s. It has now been removed in PCRE. This means it isn't recognized as
2025 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
2026 class [:space:] *does* include VT, thereby creating a mess.
2027
2028 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
2029 space and tab.
2030
2031 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
2032 its new features into the main test script, reducing the number of scripts.
2033
2034 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
2035 were backward compatible, and made the (?i) apply to the whole pattern, as if
2036 /i were given. Now it behaves more logically, and applies the option setting
2037 only to what follows. PCRE has been changed to follow suit. However, if it
2038 finds options settings right at the start of the pattern, it extracts them into
2039 the global options, as before. Thus, they show up in the info data.
2040
2041 8. Added support for the \Q...\E escape sequence. Characters in between are
2042 treated as literals. This is slightly different from Perl in that $ and @ are
2043 also handled as literals inside the quotes. In Perl, they will cause variable
2044 interpolation. Note the following examples:
2045
2046 Pattern PCRE matches Perl matches
2047
2048 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
2049 \Qabc\$xyz\E abc\$xyz abc\$xyz
2050 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
2051
2052 For compatibility with Perl, \Q...\E sequences are recognized inside character
2053 classes as well as outside them.
2054
2055 9. Re-organized 3 code statements in pcretest to avoid "overflow in
2056 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
2057 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
2058 signed/unsigned warnings.
2059
2060 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
2061 option for pcretest, so I've replaced it by a simple function that does just
2062 that job.
2063
2064 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
2065 "pcregrep -".
2066
2067 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
2068 Java package. This provides some syntactic sugar for simple cases of what my
2069 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
2070 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
2071 item, you can use this simplified notation. Note that only makes sense with
2072 greedy quantifiers. Consequently, the use of the possessive quantifier forces
2073 greediness, whatever the setting of the PCRE_UNGREEDY option.
2074
2075 13. A change of greediness default within a pattern was not taking effect at
2076 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
2077 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
2078 was abstracted outside.
2079
2080 14. PCRE now supports the \G assertion. It is true when the current matching
2081 position is at the start point of the match. This differs from \A when the
2082 starting offset is non-zero. Used with the /g option of pcretest (or similar
2083 code), it works in the same way as it does for Perl's /g option. If all
2084 alternatives of a regex begin with \G, the expression is anchored to the start
2085 match position, and the "anchored" flag is set in the compiled expression.
2086
2087 15. Some bugs concerning the handling of certain option changes within patterns
2088 have been fixed. These applied to options other than (?ims). For example,
2089 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
2090 been the other way round. Some of this was related to change 7 above.
2091
2092 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
2093 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
2094 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
2095 POSIX classes only within a class (e.g. /[[:alpha:]]/).
2096
2097 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
2098 mode. Unlike ".", it always matches newline, whatever the setting of
2099 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
2100 assertions. Perl allows it, but it doesn't (in general) work because it can't
2101 calculate the length of the lookbehind. At least, that's the case for Perl
2102 5.8.0 - I've been told they are going to document that it doesn't work in
2103 future.
2104
2105 18. Added an error diagnosis for escapes that PCRE does not support: these are
2106 \L, \l, \N, \P, \p, \U, \u, and \X.
2107
2108 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
2109 reading past the end of the pattern in cases such as /[abcd/.
2110
2111 20. PCRE was getting more memory than necessary for patterns with classes that
2112 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
2113
2114 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
2115 compiling PCRE for use with Virtual Pascal.
2116
2117 22. Small fix to the Makefile to make it work properly if the build is done
2118 outside the source tree.
2119
2120 23. Added a new extension: a condition to go with recursion. If a conditional
2121 subpattern starts with (?(R) the "true" branch is used if recursion has
2122 happened, whereas the "false" branch is used only at the top level.
2123
2124 24. When there was a very long string of literal characters (over 255 bytes
2125 without UTF support, over 250 bytes with UTF support), the computation of how
2126 much memory was required could be incorrect, leading to segfaults or other
2127 strange effects.
2128
2129 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
2130 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
2131 there was a subsequent back reference to those brackets. This meant that, for
2132 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
2133 possible to check for precisely this case. All we can do is abandon the
2134 optimization if .* occurs inside capturing brackets when there are any back
2135 references whatsoever. (See below for a better fix that came later.)
2136
2137 26. The handling of the optimization for finding the first character of a
2138 non-anchored pattern, and for finding a character that is required later in the
2139 match were failing in some cases. This didn't break the matching; it just
2140 failed to optimize when it could. The way this is done has been re-implemented.
2141
2142 27. Fixed typo in error message for invalid (?R item (it said "(?p").
2143
2144 28. Added a new feature that provides some of the functionality that Perl
2145 provides with (?{...}). The facility is termed a "callout". The way it is done
2146 in PCRE is for the caller to provide an optional function, by setting
2147 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
2148 global variable. By default it is unset, which disables all calling out. To get
2149 the function called, the regex must include (?C) at appropriate points. This
2150 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
2151 This provides a means of identifying different callout points. When PCRE
2152 reaches such a point in the regex, if pcre_callout has been set, the external
2153 function is called. It is provided with data in a structure called
2154 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
2155 matching continues; if it returns a non-zero value, the match at the current
2156 point fails. However, backtracking will occur if possible. [This was changed
2157 later and other features added - see item 49 below.]
2158
2159 29. pcretest is upgraded to test the callout functionality. It provides a
2160 callout function that displays information. By default, it shows the start of
2161 the match and the current position in the text. There are some new data escapes
2162 to vary what happens:
2163
2164 \C+ in addition, show current contents of captured substrings
2165 \C- do not supply a callout function
2166 \C!n return 1 when callout number n is reached
2167 \C!n!m return 1 when callout number n is reached for the mth time
2168
2169 30. If pcregrep was called with the -l option and just a single file name, it
2170 output "<stdin>" if a match was found, instead of the file name.
2171
2172 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
2173 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
2174 pcre_exec(). This saves a malloc/free per call. The default value of
2175 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
2176 when configuring.
2177
2178 32. The default maximum size of a compiled pattern is 64K. There have been a
2179 few cases of people hitting this limit. The code now uses macros to handle the
2180 storing of links as offsets within the compiled pattern. It defaults to 2-byte
2181 links, but this can be changed to 3 or 4 bytes by --with-link-size when
2182 configuring. Tests 2 and 5 work only with 2-byte links because they output
2183 debugging information about compiled patterns.
2184
2185 33. Internal code re-arrangements:
2186
2187 (a) Moved the debugging function for printing out a compiled regex into
2188 its own source file (printint.c) and used #include to pull it into
2189 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
2190 separate copies.
2191
2192 (b) Defined the list of op-code names for debugging as a macro in
2193 internal.h so that it is next to the definition of the opcodes.
2194
2195 (c) Defined a table of op-code lengths for simpler skipping along compiled
2196 code. This is again a macro in internal.h so that it is next to the
2197 definition of the opcodes.
2198
2199 34. Added support for recursive calls to individual subpatterns, along the
2200 lines of Robin Houston's patch (but implemented somewhat differently).
2201
2202 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
2203 allow it to read and process whole directories in Win32. This code was
2204 contributed by Lionel Fourquaux; it has not been tested by me.
2205
2206 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
2207 used to name a group. Names consist of alphanumerics and underscores, and must
2208 be unique. Back references use the syntax (?P=name) and recursive calls use
2209 (?P>name) which is a PCRE extension to the Python extension. Groups still have
2210 numbers. The function pcre_fullinfo() can be used after compilation to extract
2211 a name/number map. There are three relevant calls:
2212
2213 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
2214 PCRE_INFO_NAMECOUNT yields the number of entries
2215 PCRE_INFO_NAMETABLE yields a pointer to the map.
2216
2217 The map is a vector of fixed-size entries. The size of each entry depends on
2218 the length of the longest name used. The first two bytes of each entry are the
2219 group number, most significant byte first. There follows the corresponding
2220 name, zero terminated. The names are in alphabetical order.
2221
2222 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
2223 case instead of 255. Making it the same both with and without UTF-8 support
2224 means that the same test output works with both.
2225
2226 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
2227 calling malloc() with a zero argument.
2228
2229 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
2230 optimization. I've improved things by keeping a bitmap of backreferences with
2231 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
2232 fact referenced, the optimization can be applied. It is unlikely that a
2233 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
2234 the match to follow \n) will appear inside brackets with a number greater than
2235 31, but if it does, any back reference > 31 suppresses the optimization.
2236
2237 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
2238 of disabling numbered capturing parentheses. Any opening parenthesis that is
2239 not followed by ? behaves as if it were followed by ?: but named parentheses
2240 can still be used for capturing (and they will acquire numbers in the usual
2241 way).
2242
2243 41. Redesigned the return codes from the match() function into yes/no/error so
2244 that errors can be passed back from deep inside the nested calls. A malloc
2245 failure while inside a recursive subpattern call now causes the
2246 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
2247
2248 42. It is now possible to set a limit on the number of times the match()
2249 function is called in a call to pcre_exec(). This facility makes it possible to
2250 limit the amount of recursion and backtracking, though not in a directly
2251 obvious way, because the match() function is used in a number of different
2252 circumstances. The count starts from zero for each position in the subject
2253 string (for non-anchored patterns). The default limit is, for compatibility, a
2254 large number, namely 10 000 000. You can change this in two ways:
2255
2256 (a) When configuring PCRE before making, you can use --with-match-limit=n
2257 to set a default value for the compiled library.
2258
2259 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
2260 a different value is set. See 45 below.
2261
2262 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
2263
2264 43. Added a new function pcre_config(int, void *) to enable run-time extraction
2265 of things that can be changed at compile time. The first argument specifies
2266 what is wanted and the second points to where the information is to be placed.
2267 The current list of available information is:
2268
2269 PCRE_CONFIG_UTF8
2270
2271 The output is an integer that is set to one if UTF-8 support is available;
2272 otherwise it is set to zero.
2273
2274 PCRE_CONFIG_NEWLINE
2275
2276 The output is an integer that it set to the value of the code that is used for
2277 newline. It is either LF (10) or CR (13).
2278
2279 PCRE_CONFIG_LINK_SIZE
2280
2281 The output is an integer that contains the number of bytes used for internal
2282 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
2283
2284 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
2285
2286 The output is an integer that contains the threshold above which the POSIX
2287 interface uses malloc() for output vectors. See item 31 above.
2288
2289 PCRE_CONFIG_MATCH_LIMIT
2290
2291 The output is an unsigned integer that contains the default limit of the number
2292 of match() calls in a pcre_exec() execution. See 42 above.
2293
2294 44. pcretest has been upgraded by the addition of the -C option. This causes it
2295 to extract all the available output from the new pcre_config() function, and to
2296 output it. The program then exits immediately.
2297
2298 45. A need has arisen to pass over additional data with calls to pcre_exec() in
2299 order to support additional features. One way would have been to define
2300 pcre_exec2() (for example) with extra arguments, but this would not have been
2301 extensible, and would also have required all calls to the original function to
2302 be mapped to the new one. Instead, I have chosen to extend the mechanism that
2303 is used for passing in "extra" data from pcre_study().
2304
2305 The pcre_extra structure is now exposed and defined in pcre.h. It currently
2306 contains the following fields:
2307
2308 flags a bitmap indicating which of the following fields are set
2309 study_data opaque data from pcre_study()
2310 match_limit a way of specifying a limit on match() calls for a specific
2311 call to pcre_exec()
2312 callout_data data for callouts (see 49 below)
2313
2314 The flag bits are also defined in pcre.h, and are
2315
2316 PCRE_EXTRA_STUDY_DATA
2317 PCRE_EXTRA_MATCH_LIMIT
2318 PCRE_EXTRA_CALLOUT_DATA
2319
2320 The pcre_study() function now returns one of these new pcre_extra blocks, with
2321 the actual study data pointed to by the study_data field, and the
2322 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
2323 before. That is, this change is entirely upwards-compatible and requires no
2324 change to existing code.
2325
2326 If you want to pass in additional data to pcre_exec(), you can either place it
2327 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
2328 block.
2329
2330 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
2331 data string contains the escape sequence \M, pcretest calls pcre_exec() several
2332 times with different match limits, until it finds the minimum value needed for
2333 pcre_exec() to complete. The value is then output. This can be instructive; for
2334 most simple matches the number is quite small, but for pathological cases it
2335 gets very large very quickly.
2336
2337 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
2338 returns the size of the data block pointed to by the study_data field in a
2339 pcre_extra block, that is, the value that was passed as the argument to
2340 pcre_malloc() when PCRE was getting memory in which to place the information
2341 created by pcre_study(). The fourth argument should point to a size_t variable.
2342 pcretest has been extended so that this information is shown after a successful
2343 pcre_study() call when information about the compiled regex is being displayed.
2344
2345 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
2346 because what follows is always an absolute path. (Later: it turns out that this
2347 is more than cosmetic for MinGW, because it doesn't like empty path
2348 components.)
2349
2350 49. Some changes have been made to the callout feature (see 28 above):
2351
2352 (i) A callout function now has three choices for what it returns:
2353
2354 0 => success, carry on matching
2355 > 0 => failure at this point, but backtrack if possible
2356 < 0 => serious error, return this value from pcre_exec()
2357
2358 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
2359 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
2360 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
2361 use by callout functions. It will never be used by PCRE itself.
2362
2363 (ii) The pcre_extra structure (see 45 above) has a void * field called
2364 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
2365 pcre_callout_block structure has a field of the same name. The contents of
2366 the field passed in the pcre_extra structure are passed to the callout
2367 function in the corresponding field in the callout block. This makes it
2368 easier to use the same callout-containing regex from multiple threads. For
2369 testing, the pcretest program has a new data escape
2370
2371 \C*n pass the number n (may be negative) as callout_data
2372
2373 If the callout function in pcretest receives a non-zero value as
2374 callout_data, it returns that value.
2375
2376 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
2377 there were some redundant $(CFLAGS) in commands that are now specified as
2378 $(LINK), which already includes $(CFLAGS).
2379
2380 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
2381 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
2382 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
2383 one-byte characters throughout. Note that case-insensitive matching applies
2384 only to characters whose values are less than 256. PCRE doesn't support the
2385 notion of cases for higher-valued characters.
2386
2387 (i) A character class whose characters are all within 0-255 is handled as
2388 a bit map, and the map is inverted for negative classes. Previously, a
2389 character > 255 always failed to match such a class; however it should
2390 match if the class was a negative one (e.g. [^ab]). This has been fixed.
2391
2392 (ii) A negated character class with a single character < 255 is coded as
2393 "not this character" (OP_NOT). This wasn't working properly when the test
2394 character was multibyte, either singly or repeated.
2395
2396 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
2397 mode, for example: \x{100}{2,3}.
2398
2399 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
2400 singly or repeated) now correctly test multibyte characters. However,
2401 PCRE doesn't recognize any characters with values greater than 255 as
2402 digits, spaces, or word characters. Such characters always match \D, \S,
2403 and \W, and never match \d, \s, or \w.
2404
2405 (v) Classes may now contain characters and character ranges with values
2406 greater than 255. For example: [ab\x{100}-\x{400}].
2407
2408 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
2409 PCRE in UTF-8 mode.
2410
2411 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
2412 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
2413 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
2414 value.)
2415
2416 53. The single man page has become too large. I have therefore split it up into
2417 a number of separate man pages. These also give rise to individual HTML pages;
2418 these are now put in a separate directory, and there is an index.html page that
2419 lists them all. Some hyperlinking between the pages has been installed.
2420
2421 54. Added convenience functions for handling named capturing parentheses.
2422
2423 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
2424 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
2425 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
2426 are faulted.
2427
2428 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
2429 calling configure. These values are used when compiling the dftables.c program
2430 which is run to generate the source of the default character tables. They
2431 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
2432 you will need to set these values.
2433
2434 57. Updated the building process for Windows DLL, as provided by Fred Cox.
2435
2436
2437 Version 3.9 02-Jan-02
2438 ---------------------
2439
2440 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
2441
2442 2. If --disable-static was given, the building process failed when trying to
2443 build pcretest and pcregrep. (For some reason it was using libtool to compile
2444 them, which is not right, as they aren't part of the library.)
2445
2446
2447 Version 3.8 18-Dec-01
2448 ---------------------
2449
2450 1. The experimental UTF-8 code was completely screwed up. It was packing the
2451 bytes in the wrong order. How dumb can you get?
2452
2453
2454 Version 3.7 29-Oct-01
2455 ---------------------
2456
2457 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
2458 This caused pcretest, when used on the test data, to segfault. Unfortunately,
2459 this didn't happen under Solaris 8, where I normally test things.
2460
2461 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
2462 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
2463 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
2464 this fix an hour or so after the initial 3.7 release.)
2465
2466
2467 Version 3.6 23-Oct-01
2468 ---------------------
2469
2470 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
2471 offsets passed as NULL with zero offset count.
2472
2473 2. The config.guess and config.sub files had not been updated when I moved to
2474 the latest autoconf.
2475
2476
2477 Version 3.5 15-Aug-01
2478 ---------------------
2479
2480 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
2481 had been forgotten.
2482
2483 2. By using declared but undefined structures, we can avoid using "void"
2484 definitions in pcre.h while keeping the internal definitions of the structures
2485 private.
2486
2487 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
2488 user point of view, this means that both static and shared libraries are built
2489 by default, but this can be individually controlled. More of the work of
2490 handling this static/shared cases is now inside libtool instead of PCRE's make
2491 file.
2492
2493 4. The pcretest utility is now installed along with pcregrep because it is
2494 useful for users (to test regexs) and by doing this, it automatically gets
2495 relinked by libtool. The documentation has been turned into a man page, so
2496 there are now .1, .txt, and .html versions in /doc.
2497
2498 5. Upgrades to pcregrep:
2499 (i) Added long-form option names like gnu grep.
2500 (ii) Added --help to list all options with an explanatory phrase.
2501 (iii) Added -r, --recursive to recurse into sub-directories.
2502 (iv) Added -f, --file to read patterns from a file.
2503
2504 6. pcre_exec() was referring to its "code" argument before testing that
2505 argument for NULL (and giving an error if it was NULL).
2506
2507 7. Upgraded Makefile.in to allow for compiling in a different directory from
2508 the source directory.
2509
2510 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
2511 options bits, the pointer it was passed was to an int instead of to an unsigned
2512 long int. This mattered only on 64-bit systems.
2513
2514 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
2515 generated) instead of pcre.in, which it its source. Also made the same change
2516 in several of the .c files.
2517
2518 10. A new release of gcc defines printf() as a macro, which broke pcretest
2519 because it had an ifdef in the middle of a string argument for printf(). Fixed
2520 by using separate calls to printf().
2521
2522 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
2523 script, to force use of CR or LF instead of \n in the source. On non-Unix
2524 systems, the value can be set in config.h.
2525
2526 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
2527 absolute limit. Changed the text of the error message to make this clear, and
2528 likewise updated the man page.
2529
2530 13. The limit of 99 on the number of capturing subpatterns has been removed.
2531 The new limit is 65535, which I hope will not be a "real" limit.
2532
2533
2534 Version 3.4 22-Aug-00
2535 ---------------------
2536
2537 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
2538
2539 2. Diagnose condition (?(0) as an error instead of crashing on matching.
2540
2541
2542 Version 3.3 01-Aug-00
2543 ---------------------
2544
2545 1. If an octal character was given, but the value was greater than \377, it
2546 was not getting masked to the least significant bits, as documented. This could
2547 lead to crashes in some systems.
2548
2549 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
2550 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
2551
2552 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
2553 These just pass their arguments on to (pcre_free)(), but they are provided
2554 because some uses of PCRE bind it to non-C systems that can call its functions,
2555 but cannot call free() or pcre_free() directly.
2556
2557 4. Add "make test" as a synonym for "make check". Corrected some comments in
2558 the Makefile.
2559
2560 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
2561 Makefile.
2562
2563 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
2564 command called pgrep for grepping around the active processes.
2565
2566 7. Added the beginnings of support for UTF-8 character strings.
2567
2568 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
2569 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
2570 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
2571 out for the ar command.)
2572
2573
2574 Version 3.2 12-May-00
2575 ---------------------
2576
2577 This is purely a bug fixing release.
2578
2579 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
2580 of ZA. This was just one example of several cases that could provoke this bug,
2581 which was introduced by change 9 of version 2.00. The code for breaking
2582 infinite loops after an iteration that matches an empty string was't working
2583 correctly.
2584
2585 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
2586 when matched against abbab (for example). After matching an empty string, it
2587 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
2588 caused it to match further down the string than it should.
2589
2590 3. The code contained an inclusion of sys/types.h. It isn't clear why this
2591 was there because it doesn't seem to be needed, and it causes trouble on some
2592 systems, as it is not a Standard C header. It has been removed.
2593
2594 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
2595 were reported on the Macintosh. The changes were from
2596
2597 while ((c = *(++ptr)) != 0 && c != '\n');
2598 to
2599 while ((c = *(++ptr)) != 0 && c != '\n') ;
2600
2601 Totally extraordinary, but if that's what it takes...
2602
2603 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
2604 available. Added HAVE_BCOPY and an autoconf test for it; if neither
2605 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
2606 assumes the way PCRE uses memmove() (always moving upwards).
2607
2608 6. PCRE is being used in one environment where strchr() is not available. There
2609 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
2610 faster code anyway.
2611
2612
2613 Version 3.1 09-Feb-00
2614 ---------------------
2615
2616 The only change in this release is the fixing of some bugs in Makefile.in for
2617 the "install" target:
2618
2619 (1) It was failing to install pcreposix.h.
2620
2621 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
2622
2623
2624 Version 3.0 01-Feb-00
2625 ---------------------
2626
2627 1. Add support for the /+ modifier to perltest (to output $` like it does in
2628 pcretest).
2629
2630 2. Add support for the /g modifier to perltest.
2631
2632 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
2633 matches null strings.
2634
2635 4. Fix perltest so that it doesn't do unwanted things when fed an empty
2636 pattern. Perl treats empty patterns specially - it reuses the most recent
2637 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
2638 effect.
2639
2640 5. The POSIX interface was broken in that it was just handing over the POSIX
2641 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
2642 required a bigger vector, with some working space on the end. This means that
2643 the POSIX wrapper now has to get and free some memory, and copy the results.
2644
2645 6. Added some simple autoconf support, placing the test data and the
2646 documentation in separate directories, re-organizing some of the
2647 information files, and making it build pcre-config (a GNU standard). Also added
2648 libtool support for building PCRE as a shared library, which is now the
2649 default.
2650
2651 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
2652 09 are not valid octal constants. Single digits will be used for minor values
2653 less than 10.
2654
2655 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
2656 existing programs that set these in the POSIX interface can use PCRE without
2657 modification.
2658
2659 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
2660 return all that pcre_info() returns, plus additional data. The pcre_info()
2661 function is retained for compatibility, but is considered to be obsolete.
2662
2663 10. Added experimental recursion feature (?R) to handle one common case that
2664 Perl 5.6 will be able to do with (?p{...}).
2665
2666 11. Added support for POSIX character classes like [:alpha:], which Perl is
2667 adopting.
2668
2669
2670 Version 2.08 31-Aug-99
2671 ----------------------
2672
2673 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
2674 trying to match at the startoffset position, but instead was moving forward to
2675 the next newline as if a previous match had failed.
2676
2677 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
2678 and could get into a loop if a null string was matched other than at the start
2679 of the subject.
2680
2681 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
2682 be distinguished at compile time, and for completeness also added PCRE_DATE.
2683
2684 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
2685 in GnuWin32 environments.
2686
2687
2688 Version 2.07 29-Jul-99
2689 ----------------------
2690
2691 1. The documentation is now supplied in plain text form and HTML as well as in
2692 the form of man page sources.
2693
2694 2. C++ compilers don't like assigning (void *) values to other pointer types.
2695 In particular this affects malloc(). Although there is no problem in Standard
2696 C, I've put in casts to keep C++ compilers happy.
2697
2698 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
2699 should be (const char *).
2700
2701 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
2702 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
2703 However, I haven't made this a standard facility. The documentation doesn't
2704 mention it, and the Makefile doesn't support it.
2705
2706 5. The Makefile now contains an "install" target, with editable destinations at
2707 the top of the file. The pcretest program is not installed.
2708
2709 6. pgrep -V now gives the PCRE version number and date.
2710
2711 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
2712 causing the entire string to be ignored, instead of just the last character.
2713
2714 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
2715 non-matching string, it can take a very, very long time, even for strings of
2716 quite modest length, because of the nested recursion. PCRE now does better in
2717 some of these cases. It does this by remembering the last required literal
2718 character in the pattern, and pre-searching the subject to ensure it is present
2719 before running the real match. In other words, it applies a heuristic to detect
2720 some types of certain failure quickly, and in the above example, if presented
2721 with a string that has no trailing " it gives "no match" very quickly.
2722
2723 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
2724 other alternatives are tried instead.
2725
2726
2727 Version 2.06 09-Jun-99
2728 ----------------------
2729
2730 1. Change pcretest's output for amount of store used to show just the code
2731 space, because the remainder (the data block) varies in size between 32-bit and
2732 64-bit systems.
2733
2734 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
2735 start matching at. This allows lookbehinds to work when searching for multiple
2736 occurrences in a string.
2737
2738 3. Added additional options to pcretest for testing multiple occurrences:
2739
2740 /+ outputs the rest of the string that follows a match
2741 /g loops for multiple occurrences, using the new startoffset argument
2742 /G loops for multiple occurrences by passing an incremented pointer
2743
2744 4. PCRE wasn't doing the "first character" optimization for patterns starting
2745 with \b or \B, though it was doing it for other lookbehind assertions. That is,
2746 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
2747 the letter 'x'. On long subject strings, this gives a significant speed-up.
2748
2749
2750 Version 2.05 21-Apr-99
2751 ----------------------
2752
2753 1. Changed the type of magic_number from int to long int so that it works
2754 properly on 16-bit systems.
2755
2756 2. Fixed a bug which caused patterns starting with .* not to work correctly
2757 when the subject string contained newline characters. PCRE was assuming
2758 anchoring for such patterns in all cases, which is not correct because .* will
2759 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
2760 DOTALL is set at top level; otherwise it knows that patterns starting with .*
2761 must be retried after every newline in the subject.
2762
2763
2764 Version 2.04 18-Feb-99
2765 ----------------------
2766
2767 1. For parenthesized subpatterns with repeats whose minimum was zero, the
2768 computation of the store needed to hold the pattern was incorrect (too large).
2769 If such patterns were nested a few deep, this could multiply and become a real
2770 problem.
2771
2772 2. Added /M option to pcretest to show the memory requirement of a specific
2773 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
2774
2775 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
2776 compiled in such a way that the backtracking after subsequent failure was
2777 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
2778 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
2779
2780
2781 Version 2.03 02-Feb-99
2782 ----------------------
2783
2784 1. Fixed typo and small mistake in man page.
2785
2786 2. Added 4th condition (GPL supersedes if conflict) and created separate
2787 LICENCE file containing the conditions.
2788
2789 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
2790 Perl, that is the internal \ allows the delimiter to be included in the
2791 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
2792 the final delimiter, add \ to the end of the pattern (to test the error).
2793
2794 4. Added the convenience functions for extracting substrings after a successful
2795 match. Updated pcretest to make it able to test these functions.
2796
2797
2798 Version 2.02 14-Jan-99
2799 ----------------------
2800
2801 1. Initialized the working variables associated with each extraction so that
2802 their saving and restoring doesn't refer to uninitialized store.
2803
2804 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
2805 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
2806 fix the problem.
2807
2808 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
2809 calls, and wasn't printing the correct value for compiling calls. Increased the
2810 default value of LOOPREPEAT, and the number of significant figures in the
2811 times.
2812
2813 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
2814
2815 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
2816 a building problem on Windows NT with a FAT file system.
2817
2818
2819 Version 2.01 21-Oct-98
2820 ----------------------
2821
2822 1. Changed the API for pcre_compile() to allow for the provision of a pointer
2823 to character tables built by pcre_maketables() in the current locale. If NULL
2824 is passed, the default tables are used.
2825
2826
2827 Version 2.00 24-Sep-98
2828 ----------------------
2829
2830 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
2831 it any more.
2832
2833 2. Allow quantification of (?>) groups, and make it work correctly.
2834
2835 3. The first character computation wasn't working for (?>) groups.
2836
2837 4. Correct the implementation of \Z (it is permitted to match on the \n at the
2838 end of the subject) and add 5.005's \z, which really does match only at the
2839 very end of the subject.
2840
2841 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
2842
2843 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
2844 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
2845 localized options. All options to pcre_study() were also removed.
2846
2847 7. Add other new features from 5.005:
2848
2849 $(?<= positive lookbehind
2850 $(?<! negative lookbehind
2851 (?imsx-imsx) added the unsetting capability
2852 such a setting is global if at outer level; local otherwise
2853 (?imsx-imsx:) non-capturing groups with option setting
2854 (?(cond)re|re) conditional pattern matching
2855
2856 A backreference to itself in a repeated group matches the previous
2857 captured string.
2858
2859 8. General tidying up of studying (both automatic and via "study")
2860 consequential on the addition of new assertions.
2861
2862 9. As in 5.005, unlimited repeated groups that could match an empty substring
2863 are no longer faulted at compile time. Instead, the loop is forcibly broken at
2864 runtime if any iteration does actually match an empty substring.
2865
2866 10. Include the RunTest script in the distribution.
2867
2868 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
2869 discrepancies, some of which were old and were also with respect to 5.004. They
2870 have now been fixed.
2871
2872
2873 Version 1.09 28-Apr-98
2874 ----------------------
2875
2876 1. A negated single character class followed by a quantifier with a minimum
2877 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
2878 program crashes, or just wrong answers. This did not apply to negated classes
2879 containing more than one character, or to minima other than one.
2880
2881
2882 Version 1.08 27-Mar-98
2883 ----------------------
2884
2885 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
2886
2887 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
2888 latter must appear before anything that relies on it in the pattern.
2889
2890
2891 Version 1.07 16-Feb-98
2892 ----------------------
2893
2894 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
2895 repeat of a potentially empty string).
2896
2897
2898 Version 1.06 23-Jan-98
2899 ----------------------
2900
2901 1. Added Markus Oberhumer's little patches for C++.
2902
2903 2. Literal strings longer than 255 characters were broken.
2904
2905
2906 Version 1.05 23-Dec-97
2907 ----------------------
2908
2909 1. Negated character classes containing more than one character were failing if
2910 PCRE_CASELESS was set at run time.
2911
2912
2913 Version 1.04 19-Dec-97
2914 ----------------------
2915
2916 1. Corrected the man page, where some "const" qualifiers had been omitted.
2917
2918 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
2919 input syntax.
2920
2921 3. Fixed memory leak which occurred when a regex with back references was
2922 matched with an offsets vector that wasn't big enough. The temporary memory
2923 that is used in this case wasn't being freed if the match failed.
2924
2925 4. Tidied pcretest to ensure it frees memory that it gets.
2926
2927 5. Temporary memory was being obtained in the case where the passed offsets
2928 vector was exactly big enough.
2929
2930 6. Corrected definition of offsetof() from change 5 below.
2931
2932 7. I had screwed up change 6 below and broken the rules for the use of
2933 setjmp(). Now fixed.
2934
2935
2936 Version 1.03 18-Dec-97
2937 ----------------------
2938
2939 1. A erroneous regex with a missing opening parenthesis was correctly
2940 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
2941 on some systems.
2942
2943 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
2944 it was reported that one broken compiler failed on the former because "code" is
2945 also an independent variable.
2946
2947 3. The erroneous regex a[]b caused an array overrun reference.
2948
2949 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
2950 fail on data ending with that character. (It was going on too far, and checking
2951 the next character, typically a binary zero.) This was specific to the
2952 optimized code for single-character negative classes.
2953
2954 5. Added a contributed patch from the TIN world which does the following:
2955
2956 + Add an undef for memmove, in case the the system defines a macro for it.
2957
2958 + Add a definition of offsetof(), in case there isn't one. (I don't know
2959 the reason behind this - offsetof() is part of the ANSI standard - but
2960 it does no harm).
2961
2962 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
2963 most of the places where whitespace preceded '#'. I have given up and
2964 allowed the remaining 2 cases to be at the margin.
2965
2966 + Rename some variables in pcre to eliminate shadowing. This seems very
2967 pedantic, but does no harm, of course.
2968
2969 6. Moved the call to setjmp() into its own function, to get rid of warnings
2970 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
2971
2972 7. Constructs such as \d{8,} were compiling into the equivalent of
2973 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
2974 outcome, but in this particular case used more store than had been allocated,
2975 which caused the bug to be discovered because it threw up an internal error.
2976
2977 8. The debugging code in both pcre and pcretest for outputting the compiled
2978 form of a regex was going wrong in the case of back references followed by
2979 curly-bracketed repeats.
2980
2981
2982 Version 1.02 12-Dec-97
2983 ----------------------
2984
2985 1. Typos in pcre.3 and comments in the source fixed.
2986
2987 2. Applied a contributed patch to get rid of places where it used to remove
2988 'const' from variables, and fixed some signed/unsigned and uninitialized
2989 variable warnings.
2990
2991 3. Added the "runtest" target to Makefile.
2992
2993 4. Set default compiler flag to -O2 rather than just -O.
2994
2995
2996 Version 1.01 19-Nov-97
2997 ----------------------
2998
2999 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
3000 like /([ab]*)*/, that is, for classes with more than one character in them.
3001
3002 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
3003 as /((?>a*))*/ (a PCRE_EXTRA facility).
3004
3005
3006 Version 1.00 18-Nov-97
3007 ----------------------
3008
3009 1. Added compile-time macros to support systems such as SunOS4 which don't have
3010 memmove() or strerror() but have other things that can be used instead.
3011
3012 2. Arranged that "make clean" removes the executables.
3013
3014
3015 Version 0.99 27-Oct-97
3016 ----------------------
3017
3018 1. Fixed bug in code for optimizing classes with only one character. It was
3019 initializing a 32-byte map regardless, which could cause it to run off the end
3020 of the memory it had got.
3021
3022 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
3023
3024
3025 Version 0.98 22-Oct-97
3026 ----------------------
3027
3028 1. Fixed bug in code for handling temporary memory usage when there are more
3029 back references than supplied space in the ovector. This could cause segfaults.
3030
3031
3032 Version 0.97 21-Oct-97
3033 ----------------------
3034
3035 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
3036
3037 2. Optimized negated single characters not to use a bit map.
3038
3039 3. Brought error texts together as macro definitions; clarified some of them;
3040 fixed one that was wrong - it said "range out of order" when it meant "invalid
3041 escape sequence".
3042
3043 4. Changed some char * arguments to const char *.
3044
3045 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
3046
3047 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
3048 pcretest.
3049
3050
3051 Version 0.96 16-Oct-97
3052 ----------------------
3053
3054 1. Added a simple "pgrep" utility to the distribution.
3055
3056 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
3057 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
3058 where "ddd" means "one or more decimal digits".
3059
3060 3. Fixed serious bug. If a pattern had a back reference, but the call to
3061 pcre_exec() didn't supply a large enough ovector to record the related
3062 identifying subpattern, the match always failed. PCRE now remembers the number
3063 of the largest back reference, and gets some temporary memory in which to save
3064 the offsets during matching if necessary, in order to ensure that
3065 backreferences always work.
3066
3067 4. Increased the compatibility with Perl in a number of ways:
3068
3069 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
3070 to request this handling. The option can be set at compile or exec time.
3071
3072 (b) $ matches before a terminating newline by default; an option
3073 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
3074 mode). The option can be set at compile or exec time.
3075
3076 (c) The handling of \ followed by a digit other than 0 is now supposed to be
3077 the same as Perl's. If the decimal number it represents is less than 10
3078 or there aren't that many previous left capturing parentheses, an octal
3079 escape is read. Inside a character class, it's always an octal escape,
3080 even if it is a single digit.
3081
3082 (d) An escaped but undefined alphabetic character is taken as a literal,
3083 unless PCRE_EXTRA is set. Currently this just reserves the remaining
3084 escapes.
3085
3086 (e) {0} is now permitted. (The previous item is removed from the compiled
3087 pattern).
3088
3089 5. Changed all the names of code files so that the basic parts are no longer
3090 than 10 characters, and abolished the teeny "globals.c" file.
3091
3092 6. Changed the handling of character classes; they are now done with a 32-byte
3093 bit map always.
3094
3095 7. Added the -d and /D options to pcretest to make it possible to look at the
3096 internals of compilation without having to recompile pcre.
3097
3098
3099 Version 0.95 23-Sep-97
3100 ----------------------
3101
3102 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
3103 \x20 at the start of a run of normal characters. These were being treated as
3104 real characters, instead of the source characters being re-checked.
3105
3106
3107 Version 0.94 18-Sep-97
3108 ----------------------
3109
3110 1. The functions are now thread-safe, with the caveat that the global variables
3111 containing pointers to malloc() and free() or alternative functions are the
3112 same for all threads.
3113
3114 2. Get pcre_study() to generate a bitmap of initial characters for non-
3115 anchored patterns when this is possible, and use it if passed to pcre_exec().
3116
3117
3118 Version 0.93 15-Sep-97
3119 ----------------------
3120
3121 1. /(b)|(:+)/ was computing an incorrect first character.
3122
3123 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
3124 but not actually doing anything yet.
3125
3126 3. Treat "-" characters in classes that cannot be part of ranges as literals,
3127 as Perl does (e.g. [-az] or [az-]).
3128
3129 4. Set the anchored flag if a branch starts with .* or .*? because that tests
3130 all possible positions.
3131
3132 5. Split up into different modules to avoid including unneeded functions in a
3133 compiled binary. However, compile and exec are still in one module. The "study"
3134 function is split off.
3135
3136 6. The character tables are now in a separate module whose source is generated
3137 by an auxiliary program - but can then be edited by hand if required. There are
3138 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
3139 toupper() in the code.
3140
3141 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
3142 make them global. Abolish the function for setting them, as the caller can now
3143 set them directly.
3144
3145
3146 Version 0.92 11-Sep-97
3147 ----------------------
3148
3149 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
3150 (e.g. /a{1,3}/) was broken (I mis-optimized it).
3151
3152 2. Caseless matching was not working in character classes if the characters in
3153 the pattern were in upper case.
3154
3155 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
3156
3157 4. Make PCRE_ANCHORED public and accept as a compile option.
3158
3159 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
3160 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
3161 pass them.
3162
3163 6. Give an error if bad option bits passed at compile or run time.
3164
3165 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
3166 pcretest to cause it to pass that flag.
3167
3168 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
3169 options, and the first character, if set.
3170
3171 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
3172
3173
3174 Version 0.91 10-Sep-97
3175 ----------------------
3176
3177 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
3178 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
3179
3180 2. PCRE was looping on encountering an indefinitely repeated back reference to
3181 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
3182 Perl does - treats the match as successful.
3183
3184 ****

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