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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 332 - (show annotations) (download)
Sat Apr 5 16:11:05 2008 UTC (6 years ago) by ph10
File size: 152852 byte(s)
Alan Lehotsky's patch for REG_STARTEND.

1 ChangeLog for PCRE
2 ------------------
3
4 Version 7.7 05-Mar-08
5 ---------------------
6
7 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
8 a string to a long long, pretend we don't even have a long long." This is
9 done by checking for the strtoq, strtoll, and _strtoi64 functions.
10
11 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
12 pre-7.6 versions, which defined a global no_arg variable instead of putting
13 it in the RE class. (See also #8 below.)
14
15 3. Remove a line of dead code, identified by coverity and reported by Nuno
16 Lopes.
17
18 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
19
20 (1) The include/exclude patterns were being applied to the whole pathnames
21 of files, instead of just to the final components.
22
23 (2) If there was more than one level of directory, the subdirectories were
24 skipped unless they satisfied the include/exclude conditions. This is
25 inconsistent with GNU grep (and could even be seen as contrary to the
26 pcregrep specification - which I improved to make it absolutely clear).
27 The action now is always to scan all levels of directory, and just
28 apply the include/exclude patterns to regular files.
29
30 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
31 --exclude_dir in the tests to avoid scanning .svn directories.
32
33 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
34 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
35 doesn't support NULs in patterns.
36
37 7. Added some missing "const"s to declarations of static tables in
38 pcre_compile.c and pcre_dfa_exec.c.
39
40 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
41 caused by fix #2 above. (Subsequently also a second patch to fix the
42 first patch. And a third patch - this was a messy problem.)
43
44 9. Applied Craig's patch to remove the use of push_back().
45
46 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
47 matching function regexec().
48
49
50 Version 7.6 28-Jan-08
51 ---------------------
52
53 1. A character class containing a very large number of characters with
54 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
55 overflow.
56
57 2. Patch to cut out the "long long" test in pcrecpp_unittest when
58 HAVE_LONG_LONG is not defined.
59
60 3. Applied Christian Ehrlicher's patch to update the CMake build files to
61 bring them up to date and include new features. This patch includes:
62
63 - Fixed PH's badly added libz and libbz2 support.
64 - Fixed a problem with static linking.
65 - Added pcredemo. [But later removed - see 7 below.]
66 - Fixed dftables problem and added an option.
67 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
68 HAVE_LONG_LONG.
69 - Added readline support for pcretest.
70 - Added an listing of the option settings after cmake has run.
71
72 4. A user submitted a patch to Makefile that makes it easy to create
73 "pcre.dll" under mingw when using Configure/Make. I added stuff to
74 Makefile.am that cause it to include this special target, without
75 affecting anything else. Note that the same mingw target plus all
76 the other distribution libraries and programs are now supported
77 when configuring with CMake (see 6 below) instead of with
78 Configure/Make.
79
80 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
81 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
82 exported in the Windows port". It has not yet been confirmed that the patch
83 solves the problem, but it does no harm.
84
85 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
86 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
87 with CMake, and also correct the comment about stack recursion.
88
89 7. Remove the automatic building of pcredemo from the ./configure system and
90 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
91 of a program that users should build themselves after PCRE is installed, so
92 building it automatically is not really right. What is more, it gave
93 trouble in some build environments.
94
95 8. Further tidies to CMakeLists.txt from Sheri and Christian.
96
97
98 Version 7.5 10-Jan-08
99 ---------------------
100
101 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
102 values in parens when parsing an RE using the C++ wrapper."
103
104 2. Negative specials like \S did not work in character classes in UTF-8 mode.
105 Characters greater than 255 were excluded from the class instead of being
106 included.
107
108 3. The same bug as (2) above applied to negated POSIX classes such as
109 [:^space:].
110
111 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
112 defined or documented. It seems to have been a typo for PCRE_STATIC, so
113 I have changed it.
114
115 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
116 first named subpattern) and a construct such as (?&a) would reference the
117 first named subpattern whose name started with "a" (in other words, the
118 length check was missing). Both these problems are fixed. "Subpattern name
119 expected" is now given for (?&) (a zero-length name), and this patch also
120 makes it give the same error for \k'' (previously it complained that that
121 was a reference to a non-existent subpattern).
122
123 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
124 this is right because (?- can be followed by option settings as well as by
125 digits. I have, however, made the messages clearer.
126
127 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
128 than the number used in the conditional) now cause a compile-time error.
129 This is actually not compatible with Perl, which accepts such patterns, but
130 treats the conditional as always being FALSE (as PCRE used to), but it
131 seems to me that giving a diagnostic is better.
132
133 8. Change "alphameric" to the more common word "alphanumeric" in comments
134 and messages.
135
136 9. Fix two occurrences of "backslash" in comments that should have been
137 "backspace".
138
139 10. Remove two redundant lines of code that can never be obeyed (their function
140 was moved elsewhere).
141
142 11. The program that makes PCRE's Unicode character property table had a bug
143 which caused it to generate incorrect table entries for sequences of
144 characters that have the same character type, but are in different scripts.
145 It amalgamated them into a single range, with the script of the first of
146 them. In other words, some characters were in the wrong script. There were
147 thirteen such cases, affecting characters in the following ranges:
148
149 U+002b0 - U+002c1
150 U+0060c - U+0060d
151 U+0061e - U+00612
152 U+0064b - U+0065e
153 U+0074d - U+0076d
154 U+01800 - U+01805
155 U+01d00 - U+01d77
156 U+01d9b - U+01dbf
157 U+0200b - U+0200f
158 U+030fc - U+030fe
159 U+03260 - U+0327f
160 U+0fb46 - U+0fbb1
161 U+10450 - U+1049d
162
163 12. The -o option (show only the matching part of a line) for pcregrep was not
164 compatible with GNU grep in that, if there was more than one match in a
165 line, it showed only the first of them. It now behaves in the same way as
166 GNU grep.
167
168 13. If the -o and -v options were combined for pcregrep, it printed a blank
169 line for every non-matching line. GNU grep prints nothing, and pcregrep now
170 does the same. The return code can be used to tell if there were any
171 non-matching lines.
172
173 14. Added --file-offsets and --line-offsets to pcregrep.
174
175 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
176 infinitely looping recursion. The bug was that positive lookaheads were not
177 being skipped when checking for a possible empty match (negative lookaheads
178 and both kinds of lookbehind were skipped).
179
180 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
181 inclusion of <windows.h> to before rather than after the definition of
182 INVALID_FILE_ATTRIBUTES (patch from David Byron).
183
184 17. Specifying a possessive quantifier with a specific limit for a Unicode
185 character property caused pcre_compile() to compile bad code, which led at
186 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
187 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
188 caused the error; without that there was no problem.
189
190 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
191
192 19. Added --enable-pcretest-libreadline.
193
194 20. In pcrecpp.cc, the variable 'count' was incremented twice in
195 RE::GlobalReplace(). As a result, the number of replacements returned was
196 double what it should be. I removed one of the increments, but Craig sent a
197 later patch that removed the other one (the right fix) and added unit tests
198 that check the return values (which was not done before).
199
200 21. Several CMake things:
201
202 (1) Arranged that, when cmake is used on Unix, the libraries end up with
203 the names libpcre and libpcreposix, not just pcre and pcreposix.
204
205 (2) The above change means that pcretest and pcregrep are now correctly
206 linked with the newly-built libraries, not previously installed ones.
207
208 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
209
210 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
211 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
212 UTF-8 newline character). The key issue is that the pattern starts .*;
213 this means that the match must be either at the beginning, or after a
214 newline. The bug was in the code for advancing after a failed match and
215 checking that the new position followed a newline. It was not taking
216 account of UTF-8 characters correctly.
217
218 23. PCRE was behaving differently from Perl in the way it recognized POSIX
219 character classes. PCRE was not treating the sequence [:...:] as a
220 character class unless the ... were all letters. Perl, however, seems to
221 allow any characters between [: and :], though of course it rejects as
222 unknown any "names" that contain non-letters, because all the known class
223 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
224 for example, whereas PCRE did not - it did not recognize a POSIX character
225 class. This seemed a bit dangerous, so the code has been changed to be
226 closer to Perl. The behaviour is not identical to Perl, because PCRE will
227 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
228 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
229 Perl does, and where it didn't before.
230
231 24. Rewrite so as to remove the single use of %n from pcregrep because in some
232 Windows environments %n is disabled by default.
233
234
235 Version 7.4 21-Sep-07
236 ---------------------
237
238 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
239 means that a class such as [\s] counted as "explicit reference to CR or
240 LF". That isn't really right - the whole point of the change was to try to
241 help when there was an actual mention of one of the two characters. So now
242 the change happens only if \r or \n (or a literal CR or LF) character is
243 encountered.
244
245 2. The 32-bit options word was also used for 6 internal flags, but the numbers
246 of both had grown to the point where there were only 3 bits left.
247 Fortunately, there was spare space in the data structure, and so I have
248 moved the internal flags into a new 16-bit field to free up more option
249 bits.
250
251 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
252 but did not set the internal JCHANGED flag - either of these is enough to
253 control the way the "get" function works - but the PCRE_INFO_JCHANGED
254 facility is supposed to tell if (?J) was ever used, so now (?J) at the
255 start sets both bits.
256
257 4. Added options (at build time, compile time, exec time) to change \R from
258 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
259
260 5. doc/pcresyntax.html was missing from the distribution.
261
262 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
263 compatibility, even though it is no longer used.
264
265 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
266 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
267 windows.h file is present (where different names are used). [This was
268 reversed later after testing - see 16 below.]
269
270 8. Changed all #include <config.h> to #include "config.h". There were also
271 some further <pcre.h> cases that I changed to "pcre.h".
272
273 9. When pcregrep was used with the --colour option, it missed the line ending
274 sequence off the lines that it output.
275
276 10. It was pointed out to me that arrays of string pointers cause lots of
277 relocations when a shared library is dynamically loaded. A technique of
278 using a single long string with a table of offsets can drastically reduce
279 these. I have refactored PCRE in four places to do this. The result is
280 dramatic:
281
282 Originally: 290
283 After changing UCP table: 187
284 After changing error message table: 43
285 After changing table of "verbs" 36
286 After changing table of Posix names 22
287
288 Thanks to the folks working on Gregex for glib for this insight.
289
290 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
291 unicode-properties was also set.
292
293 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
294
295 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
296 checked only for CRLF.
297
298 14. Added casts to pcretest.c to avoid compiler warnings.
299
300 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
301
302 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
303 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
304 entirely. This removes changes made in 7 above.
305
306 17. The CMake files have been updated, and there is now more information about
307 building with CMake in the NON-UNIX-USE document.
308
309
310 Version 7.3 28-Aug-07
311 ---------------------
312
313 1. In the rejigging of the build system that eventually resulted in 7.1, the
314 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
315 brackets there is not right, since it causes compilers to look for an
316 installed pcre.h, not the version that is in the source that is being
317 compiled (which of course may be different). I have changed it back to:
318
319 #include "pcre.h"
320
321 I have a vague recollection that the change was concerned with compiling in
322 different directories, but in the new build system, that is taken care of
323 by the VPATH setting the Makefile.
324
325 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
326 when the subject happened to end in the byte 0x85 (e.g. if the last
327 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
328 characters but of course it shouldn't be taken as a newline when it is part
329 of another character. The bug was that, for an unlimited repeat of . in
330 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
331 characters when looking for a newline.
332
333 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
334
335 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
336 in debug output.
337
338 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
339 long printing in the pcrecpp unittest when running under MinGW.
340
341 6. ESC_K was left out of the EBCDIC table.
342
343 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
344 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
345 limit also applies to "virtual nesting" when a pattern is recursive, and in
346 this case 1000 isn't so big. I have been able to remove this limit at the
347 expense of backing off one optimization in certain circumstances. Normally,
348 when pcre_exec() would call its internal match() function recursively and
349 immediately return the result unconditionally, it uses a "tail recursion"
350 feature to save stack. However, when a subpattern that can match an empty
351 string has an unlimited repetition quantifier, it no longer makes this
352 optimization. That gives it a stack frame in which to save the data for
353 checking that an empty string has been matched. Previously this was taken
354 from the 1000-entry workspace that had been reserved. So now there is no
355 explicit limit, but more stack is used.
356
357 8. Applied Daniel's patches to solve problems with the import/export magic
358 syntax that is required for Windows, and which was going wrong for the
359 pcreposix and pcrecpp parts of the library. These were overlooked when this
360 problem was solved for the main library.
361
362 9. There were some crude static tests to avoid integer overflow when computing
363 the size of patterns that contain repeated groups with explicit upper
364 limits. As the maximum quantifier is 65535, the maximum group length was
365 set at 30,000 so that the product of these two numbers did not overflow a
366 32-bit integer. However, it turns out that people want to use groups that
367 are longer than 30,000 bytes (though not repeat them that many times).
368 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
369 made it possible to implement the integer overflow checks in a much more
370 dynamic way, which I have now done. The artificial limitation on group
371 length has been removed - we now have only the limit on the total length of
372 the compiled pattern, which depends on the LINK_SIZE setting.
373
374 10. Fixed a bug in the documentation for get/copy named substring when
375 duplicate names are permitted. If none of the named substrings are set, the
376 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
377 empty string.
378
379 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
380 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
381 because the ] is interpreted as the first data character and the
382 terminating ] is not found. PCRE has been made compatible with Perl in this
383 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
384 cause memory overwriting.
385
386 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
387 string has been matched (to stop an infinite loop). It was not recognizing
388 a conditional subpattern that could match an empty string if that
389 subpattern was within another subpattern. For example, it looped when
390 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
391 condition was not nested. This bug has been fixed.
392
393 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
394 past the start of the subject in the presence of bytes with the top bit
395 set, for example "\x8aBCD".
396
397 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
398 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
399
400 14. Optimized (?!) to (*FAIL).
401
402 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
403 This restricts code points to be within the range 0 to 0x10FFFF, excluding
404 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
405 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
406 does: it's just the validity check that is more restrictive.
407
408 16. Inserted checks for integer overflows during escape sequence (backslash)
409 processing, and also fixed erroneous offset values for syntax errors during
410 backslash processing.
411
412 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
413 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
414
415 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
416 caused an overrun.
417
418 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
419 something other than just ASCII characters) inside a group that had an
420 unlimited repeat caused a loop at compile time (while checking to see
421 whether the group could match an empty string).
422
423 20. Debugging a pattern containing \p or \P could cause a crash. For example,
424 [\P{Any}] did so. (Error in the code for printing property names.)
425
426 21. An orphan \E inside a character class could cause a crash.
427
428 22. A repeated capturing bracket such as (A)? could cause a wild memory
429 reference during compilation.
430
431 23. There are several functions in pcre_compile() that scan along a compiled
432 expression for various reasons (e.g. to see if it's fixed length for look
433 behind). There were bugs in these functions when a repeated \p or \P was
434 present in the pattern. These operators have additional parameters compared
435 with \d, etc, and these were not being taken into account when moving along
436 the compiled data. Specifically:
437
438 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
439 length.
440
441 (b) An item such as \pL+ within a repeated group could cause crashes or
442 loops.
443
444 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
445 "reference to non-existent subpattern" error.
446
447 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
448
449 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
450 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
451
452 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
453
454 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
455 character were causing crashes (broken optimization).
456
457 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
458 \p or \P) caused a compile-time loop.
459
460 28. More problems have arisen in unanchored patterns when CRLF is a valid line
461 break. For example, the unstudied pattern [\r\n]A does not match the string
462 "\r\nA" because change 7.0/46 below moves the current point on by two
463 characters after failing to match at the start. However, the pattern \nA
464 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
465 the same is true. There doesn't seem any very clean way out of this, but
466 what I have chosen to do makes the common cases work: PCRE now takes note
467 of whether there can be an explicit match for \r or \n anywhere in the
468 pattern, and if so, 7.0/46 no longer applies. As part of this change,
469 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
470 pattern has explicit CR or LF references.
471
472 29. Added (*CR) etc for changing newline setting at start of pattern.
473
474
475 Version 7.2 19-Jun-07
476 ---------------------
477
478 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
479 which is apparently normally available under Windows.
480
481 2. Re-jig the pcregrep tests with different newline settings in an attempt
482 to make them independent of the local environment's newline setting.
483
484 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
485
486 4. Some of the "internals" tests were previously cut out when the link size
487 was not 2, because the output contained actual offsets. The recent new
488 "Z" feature of pcretest means that these can be cut out, making the tests
489 usable with all link sizes.
490
491 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
492 stack recursion. This gives a massive performance boost under BSD, but just
493 a small improvement under Linux. However, it saves one field in the frame
494 in all cases.
495
496 6. Added more features from the forthcoming Perl 5.10:
497
498 (a) (?-n) (where n is a string of digits) is a relative subroutine or
499 recursion call. It refers to the nth most recently opened parentheses.
500
501 (b) (?+n) is also a relative subroutine call; it refers to the nth next
502 to be opened parentheses.
503
504 (c) Conditions that refer to capturing parentheses can be specified
505 relatively, for example, (?(-2)... or (?(+3)...
506
507 (d) \K resets the start of the current match so that everything before
508 is not part of it.
509
510 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
511
512 (f) \g{name} is another synonym - part of Perl 5.10's unification of
513 reference syntax.
514
515 (g) (?| introduces a group in which the numbering of parentheses in each
516 alternative starts with the same number.
517
518 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
519
520 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
521 PCRE_INFO_JCHANGED.
522
523 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
524 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
525 for detecting groups that can match an empty string.
526
527 9. A pattern with a very large number of alternatives (more than several
528 hundred) was running out of internal workspace during the pre-compile
529 phase, where pcre_compile() figures out how much memory will be needed. A
530 bit of new cunning has reduced the workspace needed for groups with
531 alternatives. The 1000-alternative test pattern now uses 12 bytes of
532 workspace instead of running out of the 4096 that are available.
533
534 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
535
536 11. Applied patch from Google to remove an optimization that didn't quite work.
537 The report of the bug said:
538
539 pcrecpp::RE("a*").FullMatch("aaa") matches, while
540 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
541 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
542
543 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
544 it matched the wrong number of bytes.
545
546
547 Version 7.1 24-Apr-07
548 ---------------------
549
550 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
551 that is more "standard", making use of automake and other Autotools. There
552 is some re-arrangement of the files and adjustment of comments consequent
553 on this.
554
555 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
556 for recursive directory scanning broke on some systems because the files
557 are not scanned in any specific order and on different systems the order
558 was different. A call to "sort" has been inserted into RunGrepTest for the
559 approprate test as a short-term fix. In the longer term there may be an
560 alternative.
561
562 3. I had an email from Eric Raymond about problems translating some of PCRE's
563 man pages to HTML (despite the fact that I distribute HTML pages, some
564 people do their own conversions for various reasons). The problems
565 concerned the use of low-level troff macros .br and .in. I have therefore
566 removed all such uses from the man pages (some were redundant, some could
567 be replaced by .nf/.fi pairs). The 132html script that I use to generate
568 HTML has been updated to handle .nf/.fi and to complain if it encounters
569 .br or .in.
570
571 4. Updated comments in configure.ac that get placed in config.h.in and also
572 arranged for config.h to be included in the distribution, with the name
573 config.h.generic, for the benefit of those who have to compile without
574 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
575
576 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
577 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
578 makevp.bat for the current PCRE, using the additional files
579 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
580
581 6. A Windows user reported a minor discrepancy with test 2, which turned out
582 to be caused by a trailing space on an input line that had got lost in his
583 copy. The trailing space was an accident, so I've just removed it.
584
585 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
586 that is needed.
587
588 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
589 as "const" (a) because they are and (b) because it helps the PHP
590 maintainers who have recently made a script to detect big data structures
591 in the php code that should be moved to the .rodata section. I remembered
592 to update Builducptable as well, so it won't revert if ucptable.h is ever
593 re-created.
594
595 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
596 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
597 order to be able to cut out the UTF-8 tables in the latter when UTF-8
598 support is not required. This saves 1.5-2K of code, which is important in
599 some applications.
600
601 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
602 so as not to refer to the tables, even though these functions will never be
603 called when UTF-8 support is disabled. Otherwise there are problems with a
604 shared library.
605
606 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
607
608 (a) It was defining its arguments as char * instead of void *.
609
610 (b) It was assuming that all moves were upwards in memory; this was true
611 a long time ago when I wrote it, but is no longer the case.
612
613 The emulated memove() is provided for those environments that have neither
614 memmove() nor bcopy(). I didn't think anyone used it these days, but that
615 is clearly not the case, as these two bugs were recently reported.
616
617 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
618 and Detrail to create the HTML documentation, the .txt form of the man
619 pages, and it removes trailing spaces from listed files. It also creates
620 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
621 case, it wraps all the #defines with #ifndefs. This script should be run
622 before "make dist".
623
624 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
625 with Unicode property support.
626
627 (a) For a maximizing quantifier, if the two different cases of the
628 character were of different lengths in their UTF-8 codings (there are
629 some cases like this - I found 11), and the matching function had to
630 back up over a mixture of the two cases, it incorrectly assumed they
631 were both the same length.
632
633 (b) When PCRE was configured to use the heap rather than the stack for
634 recursion during matching, it was not correctly preserving the data for
635 the other case of a UTF-8 character when checking ahead for a match
636 while processing a minimizing repeat. If the check also involved
637 matching a wide character, but failed, corruption could cause an
638 erroneous result when trying to check for a repeat of the original
639 character.
640
641 13. Some tidying changes to the testing mechanism:
642
643 (a) The RunTest script now detects the internal link size and whether there
644 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
645 values substituted by "configure". (The RunGrepTest script already did
646 this for UTF-8.) The configure.ac script no longer substitutes the
647 relevant variables.
648
649 (b) The debugging options /B and /D in pcretest show the compiled bytecode
650 with length and offset values. This means that the output is different
651 for different internal link sizes. Test 2 is skipped for link sizes
652 other than 2 because of this, bypassing the problem. Unfortunately,
653 there was also a test in test 3 (the locale tests) that used /B and
654 failed for link sizes other than 2. Rather than cut the whole test out,
655 I have added a new /Z option to pcretest that replaces the length and
656 offset values with spaces. This is now used to make test 3 independent
657 of link size. (Test 2 will be tidied up later.)
658
659 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
660 segmentation fault instead of returning the appropriate error message.
661
662 15. In multiline mode when the newline sequence was set to "any", the pattern
663 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
664 This doesn't seem right; it now treats the CRLF combination as the line
665 ending, and so does not match in that case. It's only a pattern such as ^$
666 that would hit this one: something like ^ABC$ would have failed after \r
667 and then tried again after \r\n.
668
669 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
670 in an attempt to make files that differ only in their line terminators
671 compare equal. This works on Linux.
672
673 17. Under certain error circumstances pcregrep might try to free random memory
674 as it exited. This is now fixed, thanks to valgrind.
675
676 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
677 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
678 was because its rules for how to advance for /g after matching an empty
679 string at the end of a line did not allow for this case. They now check for
680 it specially.
681
682 20. pcretest is supposed to handle patterns and data of any length, by
683 extending its buffers when necessary. It was getting this wrong when the
684 buffer for a data line had to be extended.
685
686 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
687 CRLF as a newline sequence.
688
689 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
690 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
691 I have nevertheless tidied it up.
692
693 23. Added some casts to kill warnings from HP-UX ia64 compiler.
694
695 24. Added a man page for pcre-config.
696
697
698 Version 7.0 19-Dec-06
699 ---------------------
700
701 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
702 moving to gcc 4.1.1.
703
704 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
705 sys/time.h, which is documented as needed for this function. It doesn't
706 seem to matter on Linux, but it showed up on some releases of OS X.
707
708 3. It seems that there are systems where bytes whose values are greater than
709 127 match isprint() in the "C" locale. The "C" locale should be the
710 default when a C program starts up. In most systems, only ASCII printing
711 characters match isprint(). This difference caused the output from pcretest
712 to vary, making some of the tests fail. I have changed pcretest so that:
713
714 (a) When it is outputting text in the compiled version of a pattern, bytes
715 other than 32-126 are always shown as hex escapes.
716
717 (b) When it is outputting text that is a matched part of a subject string,
718 it does the same, unless a different locale has been set for the match
719 (using the /L modifier). In this case, it uses isprint() to decide.
720
721 4. Fixed a major bug that caused incorrect computation of the amount of memory
722 required for a compiled pattern when options that changed within the
723 pattern affected the logic of the preliminary scan that determines the
724 length. The relevant options are -x, and -i in UTF-8 mode. The result was
725 that the computed length was too small. The symptoms of this bug were
726 either the PCRE error "internal error: code overflow" from pcre_compile(),
727 or a glibc crash with a message such as "pcretest: free(): invalid next
728 size (fast)". Examples of patterns that provoked this bug (shown in
729 pcretest format) are:
730
731 /(?-x: )/x
732 /(?x)(?-x: \s*#\s*)/
733 /((?i)[\x{c0}])/8
734 /(?i:[\x{c0}])/8
735
736 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
737 is now done differently.
738
739 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
740 wrapper classes; (b) implement a new function in the C++ scanner that is
741 more efficient than the old way of doing things because it avoids levels of
742 recursion in the regex matching; (c) add a paragraph to the documentation
743 for the FullMatch() function.
744
745 6. The escape sequence \n was being treated as whatever was defined as
746 "newline". Not only was this contrary to the documentation, which states
747 that \n is character 10 (hex 0A), but it also went horribly wrong when
748 "newline" was defined as CRLF. This has been fixed.
749
750 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
751 was being set to -1 for the "end of line" case (supposedly a value that no
752 character can have). Though this value is never used (the check for end of
753 line is "zero bytes in current character"), it caused compiler complaints.
754 I've changed it to 0xffffffff.
755
756 8. In pcre_version.c, the version string was being built by a sequence of
757 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
758 string (as it is for production releases) called a macro with an empty
759 argument. The C standard says the result of this is undefined. The gcc
760 compiler treats it as an empty string (which was what was wanted) but it is
761 reported that Visual C gives an error. The source has been hacked around to
762 avoid this problem.
763
764 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
765 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
766 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
767 of them did).
768
769 10. Originally, pcretest opened its input and output without "b"; then I was
770 told that "b" was needed in some environments, so it was added for release
771 5.0 to both the input and output. (It makes no difference on Unix-like
772 systems.) Later I was told that it is wrong for the input on Windows. I've
773 now abstracted the modes into two macros, to make it easier to fiddle with
774 them, and removed "b" from the input mode under Windows.
775
776 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
777
778 12. Added -help and --help to pcretest as an official way of being reminded
779 of the options.
780
781 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
782 and pcrecpp.cc because they annoy compilers at high warning levels.
783
784 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
785
786 15. Fixed an occurrence of == in configure.ac that should have been = (shell
787 scripts are not C programs :-) and which was not noticed because it works
788 on Linux.
789
790 16. pcretest is supposed to handle any length of pattern and data line (as one
791 line or as a continued sequence of lines) by extending its input buffer if
792 necessary. This feature was broken for very long pattern lines, leading to
793 a string of junk being passed to pcre_compile() if the pattern was longer
794 than about 50K.
795
796 17. I have done a major re-factoring of the way pcre_compile() computes the
797 amount of memory needed for a compiled pattern. Previously, there was code
798 that made a preliminary scan of the pattern in order to do this. That was
799 OK when PCRE was new, but as the facilities have expanded, it has become
800 harder and harder to keep it in step with the real compile phase, and there
801 have been a number of bugs (see for example, 4 above). I have now found a
802 cunning way of running the real compile function in a "fake" mode that
803 enables it to compute how much memory it would need, while actually only
804 ever using a few hundred bytes of working memory and without too many
805 tests of the mode. This should make future maintenance and development
806 easier. A side effect of this work is that the limit of 200 on the nesting
807 depth of parentheses has been removed (though this was never a serious
808 limitation, I suspect). However, there is a downside: pcre_compile() now
809 runs more slowly than before (30% or more, depending on the pattern). I
810 hope this isn't a big issue. There is no effect on runtime performance.
811
812 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
813 newline (only possible for the last line of a file) and it was a
814 pattern that set a locale (followed by /Lsomething), pcretest crashed.
815
816 19. Added additional timing features to pcretest. (1) The -tm option now times
817 matching only, not compiling. (2) Both -t and -tm can be followed, as a
818 separate command line item, by a number that specifies the number of
819 repeats to use when timing. The default is 50000; this gives better
820 precision, but takes uncomfortably long for very large patterns.
821
822 20. Extended pcre_study() to be more clever in cases where a branch of a
823 subpattern has no definite first character. For example, (a*|b*)[cd] would
824 previously give no result from pcre_study(). Now it recognizes that the
825 first character must be a, b, c, or d.
826
827 21. There was an incorrect error "recursive call could loop indefinitely" if
828 a subpattern (or the entire pattern) that was being tested for matching an
829 empty string contained only one non-empty item after a nested subpattern.
830 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
831 incorrectly, because the \d was being skipped in the check.
832
833 22. The pcretest program now has a new pattern option /B and a command line
834 option -b, which is equivalent to adding /B to every pattern. This causes
835 it to show the compiled bytecode, without the additional information that
836 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
837 is the same as /B/I).
838
839 23. A new optimization is now able automatically to treat some sequences such
840 as a*b as a*+b. More specifically, if something simple (such as a character
841 or a simple class like \d) has an unlimited quantifier, and is followed by
842 something that cannot possibly match the quantified thing, the quantifier
843 is automatically "possessified".
844
845 24. A recursive reference to a subpattern whose number was greater than 39
846 went wrong under certain circumstances in UTF-8 mode. This bug could also
847 have affected the operation of pcre_study().
848
849 25. Realized that a little bit of performance could be had by replacing
850 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
851
852 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
853
854 27. Possessive quantifiers such as a++ were previously implemented by turning
855 them into atomic groups such as ($>a+). Now they have their own opcodes,
856 which improves performance. This includes the automatically created ones
857 from 23 above.
858
859 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
860 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
861 the first matched character to be a colon. This applied both to named and
862 numbered groups.
863
864 29. The ucpinternal.h header file was missing its idempotency #ifdef.
865
866 30. I was sent a "project" file called libpcre.a.dev which I understand makes
867 building PCRE on Windows easier, so I have included it in the distribution.
868
869 31. There is now a check in pcretest against a ridiculously large number being
870 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
871 loop, the loop is abandoned.
872
873 32. Forward references to subpatterns in conditions such as (?(2)...) where
874 subpattern 2 is defined later cause pcre_compile() to search forwards in
875 the pattern for the relevant set of parentheses. This search went wrong
876 when there were unescaped parentheses in a character class, parentheses
877 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
878
879 33. "Subroutine" calls and backreferences were previously restricted to
880 referencing subpatterns earlier in the regex. This restriction has now
881 been removed.
882
883 34. Added a number of extra features that are going to be in Perl 5.10. On the
884 whole, these are just syntactic alternatives for features that PCRE had
885 previously implemented using the Python syntax or my own invention. The
886 other formats are all retained for compatibility.
887
888 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
889 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
890 also .NET compatible.
891
892 (b) A recursion or subroutine call to a named group can now be defined as
893 (?&name) as well as (?P>name).
894
895 (c) A backreference to a named group can now be defined as \k<name> or
896 \k'name' as well as (?P=name). The new forms, as well as being in Perl
897 5.10, are also .NET compatible.
898
899 (d) A conditional reference to a named group can now use the syntax
900 (?(<name>) or (?('name') as well as (?(name).
901
902 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
903 groups (named and numbered) that are never evaluated inline, but can be
904 called as "subroutines" from elsewhere. In effect, the DEFINE condition
905 is always false. There may be only one alternative in such a group.
906
907 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
908 as the simple (?(R). The condition is true only if the most recent
909 recursion is that of the given number or name. It does not search out
910 through the entire recursion stack.
911
912 (g) The escape \gN or \g{N} has been added, where N is a positive or
913 negative number, specifying an absolute or relative reference.
914
915 35. Tidied to get rid of some further signed/unsigned compiler warnings and
916 some "unreachable code" warnings.
917
918 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
919 things, this adds five new scripts.
920
921 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
922 There were also incompatibilities regarding the handling of \Q..\E inside
923 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
924 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
925
926 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
927 matches an empty string, and forcibly breaks the loop. There were bugs in
928 this code in non-simple cases. For a pattern such as ^(a()*)* matched
929 against aaaa the result was just "a" rather than "aaaa", for example. Two
930 separate and independent bugs (that affected different cases) have been
931 fixed.
932
933 39. Refactored the code to abolish the use of different opcodes for small
934 capturing bracket numbers. This is a tidy that I avoided doing when I
935 removed the limit on the number of capturing brackets for 3.5 back in 2001.
936 The new approach is not only tidier, it makes it possible to reduce the
937 memory needed to fix the previous bug (38).
938
939 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
940 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
941 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
942 mode.
943
944 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
945 report.
946
947 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
948 copy construction and assignment in the C++ wrapper.
949
950 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
951 couple of bugs that could have given wrong results in the "--newline=crlf"
952 case.
953
954 44. Added a number of casts and did some reorganization of signed/unsigned int
955 variables following suggestions from Dair Grant. Also renamed the variable
956 "this" as "item" because it is a C++ keyword.
957
958 45. Arranged for dftables to add
959
960 #include "pcre_internal.h"
961
962 to pcre_chartables.c because without it, gcc 4.x may remove the array
963 definition from the final binary if PCRE is built into a static library and
964 dead code stripping is activated.
965
966 46. For an unanchored pattern, if a match attempt fails at the start of a
967 newline sequence, and the newline setting is CRLF or ANY, and the next two
968 characters are CRLF, advance by two characters instead of one.
969
970
971 Version 6.7 04-Jul-06
972 ---------------------
973
974 1. In order to handle tests when input lines are enormously long, pcretest has
975 been re-factored so that it automatically extends its buffers when
976 necessary. The code is crude, but this _is_ just a test program. The
977 default size has been increased from 32K to 50K.
978
979 2. The code in pcre_study() was using the value of the re argument before
980 testing it for NULL. (Of course, in any sensible call of the function, it
981 won't be NULL.)
982
983 3. The memmove() emulation function in pcre_internal.h, which is used on
984 systems that lack both memmove() and bcopy() - that is, hardly ever -
985 was missing a "static" storage class specifier.
986
987 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
988 containing an extended class (one that cannot be represented by a bitmap
989 because it contains high-valued characters or Unicode property items, e.g.
990 [\pZ]). Almost always one would set UTF-8 mode when processing such a
991 pattern, but PCRE should not loop if you do not (it no longer does).
992 [Detail: two cases were found: (a) a repeated subpattern containing an
993 extended class; (b) a recursive reference to a subpattern that followed a
994 previous extended class. It wasn't skipping over the extended class
995 correctly when UTF-8 mode was not set.]
996
997 5. A negated single-character class was not being recognized as fixed-length
998 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
999 compile error "lookbehind assertion is not fixed length".
1000
1001 6. The RunPerlTest auxiliary script was showing an unexpected difference
1002 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
1003 write a Perl script that can interpret lines of an input file either as
1004 byte characters or as UTF-8, which is what "perltest" was being required to
1005 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
1006 can't do is switch easily at run time between having the "use utf8;" pragma
1007 or not. In the end, I fudged it by using the RunPerlTest script to insert
1008 "use utf8;" explicitly for the UTF-8 tests.
1009
1010 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
1011 the end of the subject string, contrary to the documentation and to what
1012 Perl does. This was true of both matching functions. Now it matches only at
1013 the start of the subject and immediately after *internal* newlines.
1014
1015 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
1016 a pointer to an int instead of a pointer to an unsigned long int. This
1017 caused problems on 64-bit systems.
1018
1019 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
1020 instance of the 'standard' template library not being so standard".
1021
1022 10. There was no check on the number of named subpatterns nor the maximum
1023 length of a subpattern name. The product of these values is used to compute
1024 the size of the memory block for a compiled pattern. By supplying a very
1025 long subpattern name and a large number of named subpatterns, the size
1026 computation could be caused to overflow. This is now prevented by limiting
1027 the length of names to 32 characters, and the number of named subpatterns
1028 to 10,000.
1029
1030 11. Subpatterns that are repeated with specific counts have to be replicated in
1031 the compiled pattern. The size of memory for this was computed from the
1032 length of the subpattern and the repeat count. The latter is limited to
1033 65535, but there was no limit on the former, meaning that integer overflow
1034 could in principle occur. The compiled length of a repeated subpattern is
1035 now limited to 30,000 bytes in order to prevent this.
1036
1037 12. Added the optional facility to have named substrings with the same name.
1038
1039 13. Added the ability to use a named substring as a condition, using the
1040 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
1041 are numbers (not recommended). Forward references are permitted.
1042
1043 14. Added forward references in named backreferences (if you see what I mean).
1044
1045 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
1046 pattern could run off the end of the subject. For example, the pattern
1047 "(?s)(.{1,5})"8 did this with the subject "ab".
1048
1049 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
1050 PCRE_CASELESS was set when matching characters that were quantified with ?
1051 or *.
1052
1053 17. A character class other than a single negated character that had a minimum
1054 but no maximum quantifier - for example [ab]{6,} - was not handled
1055 correctly by pce_dfa_exec(). It would match only one character.
1056
1057 18. A valid (though odd) pattern that looked like a POSIX character
1058 class but used an invalid character after [ (for example [[,abc,]]) caused
1059 pcre_compile() to give the error "Failed: internal error: code overflow" or
1060 in some cases to crash with a glibc free() error. This could even happen if
1061 the pattern terminated after [[ but there just happened to be a sequence of
1062 letters, a binary zero, and a closing ] in the memory that followed.
1063
1064 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
1065 over the years. Originally (before any Unicode support), just the bottom 8
1066 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
1067 output from "man perlunicode" includes this:
1068
1069 The regular expression compiler produces polymorphic opcodes. That
1070 is, the pattern adapts to the data and automatically switches to
1071 the Unicode character scheme when presented with Unicode data--or
1072 instead uses a traditional byte scheme when presented with byte
1073 data.
1074
1075 Sadly, a wide octal escape does not cause a switch, and in a string with
1076 no other multibyte characters, these octal escapes are treated as before.
1077 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
1078 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
1079 Unicode string.
1080
1081 I have not perpetrated such confusion in PCRE. Up till now, it took just
1082 the bottom 8 bits, as in old Perl. I have now made octal escapes with
1083 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
1084 translate to the appropriate multibyte character.
1085
1086 29. Applied some refactoring to reduce the number of warnings from Microsoft
1087 and Borland compilers. This has included removing the fudge introduced
1088 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
1089 a warning about an unused variable.
1090
1091 21. PCRE has not included VT (character 0x0b) in the set of whitespace
1092 characters since release 4.0, because Perl (from release 5.004) does not.
1093 [Or at least, is documented not to: some releases seem to be in conflict
1094 with the documentation.] However, when a pattern was studied with
1095 pcre_study() and all its branches started with \s, PCRE still included VT
1096 as a possible starting character. Of course, this did no harm; it just
1097 caused an unnecessary match attempt.
1098
1099 22. Removed a now-redundant internal flag bit that recorded the fact that case
1100 dependency changed within the pattern. This was once needed for "required
1101 byte" processing, but is no longer used. This recovers a now-scarce options
1102 bit. Also moved the least significant internal flag bit to the most-
1103 significant bit of the word, which was not previously used (hangover from
1104 the days when it was an int rather than a uint) to free up another bit for
1105 the future.
1106
1107 23. Added support for CRLF line endings as well as CR and LF. As well as the
1108 default being selectable at build time, it can now be changed at runtime
1109 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
1110 specify that it is scanning data with non-default line endings.
1111
1112 24. Changed the definition of CXXLINK to make it agree with the definition of
1113 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
1114
1115 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
1116 recursions. This makes a big different to stack usage for some patterns.
1117
1118 26. If a subpattern containing a named recursion or subroutine reference such
1119 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
1120 the space required for the compiled pattern went wrong and gave too small a
1121 value. Depending on the environment, this could lead to "Failed: internal
1122 error: code overflow at offset 49" or "glibc detected double free or
1123 corruption" errors.
1124
1125 27. Applied patches from Google (a) to support the new newline modes and (b) to
1126 advance over multibyte UTF-8 characters in GlobalReplace.
1127
1128 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
1129 difference for some implementation of PCRE in some Windows version.
1130
1131 29. Added some extra testing facilities to pcretest:
1132
1133 \q<number> in a data line sets the "match limit" value
1134 \Q<number> in a data line sets the "match recursion limt" value
1135 -S <number> sets the stack size, where <number> is in megabytes
1136
1137 The -S option isn't available for Windows.
1138
1139
1140 Version 6.6 06-Feb-06
1141 ---------------------
1142
1143 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
1144 in pcreposix.h. I have copied the definition from pcre.h.
1145
1146 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
1147 because pcre.h is no longer a built file.
1148
1149 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
1150 not normally included in the compiled code.
1151
1152
1153 Version 6.5 01-Feb-06
1154 ---------------------
1155
1156 1. When using the partial match feature with pcre_dfa_exec(), it was not
1157 anchoring the second and subsequent partial matches at the new starting
1158 point. This could lead to incorrect results. For example, with the pattern
1159 /1234/, partially matching against "123" and then "a4" gave a match.
1160
1161 2. Changes to pcregrep:
1162
1163 (a) All non-match returns from pcre_exec() were being treated as failures
1164 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
1165 error message is output. Some extra information is given for the
1166 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
1167 probably the only errors that are likely to be caused by users (by
1168 specifying a regex that has nested indefinite repeats, for instance).
1169 If there are more than 20 of these errors, pcregrep is abandoned.
1170
1171 (b) A binary zero was treated as data while matching, but terminated the
1172 output line if it was written out. This has been fixed: binary zeroes
1173 are now no different to any other data bytes.
1174
1175 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
1176 used to set a locale for matching. The --locale=xxxx long option has
1177 been added (no short equivalent) to specify a locale explicitly on the
1178 pcregrep command, overriding the environment variables.
1179
1180 (d) When -B was used with -n, some line numbers in the output were one less
1181 than they should have been.
1182
1183 (e) Added the -o (--only-matching) option.
1184
1185 (f) If -A or -C was used with -c (count only), some lines of context were
1186 accidentally printed for the final match.
1187
1188 (g) Added the -H (--with-filename) option.
1189
1190 (h) The combination of options -rh failed to suppress file names for files
1191 that were found from directory arguments.
1192
1193 (i) Added the -D (--devices) and -d (--directories) options.
1194
1195 (j) Added the -F (--fixed-strings) option.
1196
1197 (k) Allow "-" to be used as a file name for -f as well as for a data file.
1198
1199 (l) Added the --colo(u)r option.
1200
1201 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
1202 is not present by default.
1203
1204 3. A nasty bug was discovered in the handling of recursive patterns, that is,
1205 items such as (?R) or (?1), when the recursion could match a number of
1206 alternatives. If it matched one of the alternatives, but subsequently,
1207 outside the recursion, there was a failure, the code tried to back up into
1208 the recursion. However, because of the way PCRE is implemented, this is not
1209 possible, and the result was an incorrect result from the match.
1210
1211 In order to prevent this happening, the specification of recursion has
1212 been changed so that all such subpatterns are automatically treated as
1213 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
1214
1215 4. I had overlooked the fact that, in some locales, there are characters for
1216 which isalpha() is true but neither isupper() nor islower() are true. In
1217 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
1218 and ordfeminine) are like this. This affected the treatment of \w and \W
1219 when they appeared in character classes, but not when they appeared outside
1220 a character class. The bit map for "word" characters is now created
1221 separately from the results of isalnum() instead of just taking it from the
1222 upper, lower, and digit maps. (Plus the underscore character, of course.)
1223
1224 5. The above bug also affected the handling of POSIX character classes such as
1225 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
1226 permanent tables. Instead, the bit maps for such a class were previously
1227 created as the appropriate unions of the upper, lower, and digit bitmaps.
1228 Now they are created by subtraction from the [[:word:]] class, which has
1229 its own bitmap.
1230
1231 6. The [[:blank:]] character class matches horizontal, but not vertical space.
1232 It is created by subtracting the vertical space characters (\x09, \x0a,
1233 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
1234 subtraction was done in the overall bitmap for a character class, meaning
1235 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
1236 be recognized. This bug has been fixed.
1237
1238 7. Patches from the folks at Google:
1239
1240 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
1241 real life, but is still worth protecting against".
1242
1243 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
1244 regular expressions".
1245
1246 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
1247 have it.
1248
1249 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
1250 "configure" and the latter not, in order to fix a problem somebody had
1251 with compiling the Arg class on HP-UX.
1252
1253 (e) Improve the error-handling of the C++ wrapper a little bit.
1254
1255 (f) New tests for checking recursion limiting.
1256
1257 8. The pcre_memmove() function, which is used only if the environment does not
1258 have a standard memmove() function (and is therefore rarely compiled),
1259 contained two bugs: (a) use of int instead of size_t, and (b) it was not
1260 returning a result (though PCRE never actually uses the result).
1261
1262 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
1263 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
1264 returned instead of calling malloc() with an overflowing number that would
1265 most likely cause subsequent chaos.
1266
1267 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
1268
1269 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
1270 with this option is matched, the nmatch and pmatch options of regexec() are
1271 ignored.
1272
1273 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
1274 provided in case anyone wants to the the POSIX interface with UTF-8
1275 strings.
1276
1277 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
1278 C++ linking (needed for some HP-UX environments).
1279
1280 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
1281 (unused parameter) and in the pcre_printint() function (omitted "default"
1282 switch label when the default is to do nothing).
1283
1284 15. Added some code to make it possible, when PCRE is compiled as a C++
1285 library, to replace subject pointers for pcre_exec() with a smart pointer
1286 class, thus making it possible to process discontinuous strings.
1287
1288 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
1289 much the same function. They were added by different people who were trying
1290 to make PCRE easy to compile on non-Unix systems. It has been suggested
1291 that PCRE_EXPORT be abolished now that there is more automatic apparatus
1292 for compiling on Windows systems. I have therefore replaced it with
1293 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
1294 defaults to "extern" for C or "extern C" for C++, which works fine on
1295 Unix-like systems. It is now possible to override the value of PCRE_DATA_
1296 SCOPE with something explicit in config.h. In addition:
1297
1298 (a) pcreposix.h still had just "extern" instead of either of these macros;
1299 I have replaced it with PCRE_DATA_SCOPE.
1300
1301 (b) Functions such as _pcre_xclass(), which are internal to the library,
1302 but external in the C sense, all had PCRE_EXPORT in their definitions.
1303 This is apparently wrong for the Windows case, so I have removed it.
1304 (It makes no difference on Unix-like systems.)
1305
1306 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
1307 of recursive calls to match(). This is different to MATCH_LIMIT because
1308 that limits the total number of calls to match(), not all of which increase
1309 the depth of recursion. Limiting the recursion depth limits the amount of
1310 stack (or heap if NO_RECURSE is set) that is used. The default can be set
1311 when PCRE is compiled, and changed at run time. A patch from Google adds
1312 this functionality to the C++ interface.
1313
1314 18. Changes to the handling of Unicode character properties:
1315
1316 (a) Updated the table to Unicode 4.1.0.
1317
1318 (b) Recognize characters that are not in the table as "Cn" (undefined).
1319
1320 (c) I revised the way the table is implemented to a much improved format
1321 which includes recognition of ranges. It now supports the ranges that
1322 are defined in UnicodeData.txt, and it also amalgamates other
1323 characters into ranges. This has reduced the number of entries in the
1324 table from around 16,000 to around 3,000, thus reducing its size
1325 considerably. I realized I did not need to use a tree structure after
1326 all - a binary chop search is just as efficient. Having reduced the
1327 number of entries, I extended their size from 6 bytes to 8 bytes to
1328 allow for more data.
1329
1330 (d) Added support for Unicode script names via properties such as \p{Han}.
1331
1332 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
1333 matching that character.
1334
1335 20. When matching a repeated Unicode property with a minimum greater than zero,
1336 (for example \pL{2,}), PCRE could look past the end of the subject if it
1337 reached it while seeking the minimum number of characters. This could
1338 happen only if some of the characters were more than one byte long, because
1339 there is a check for at least the minimum number of bytes.
1340
1341 21. Refactored the implementation of \p and \P so as to be more general, to
1342 allow for more different types of property in future. This has changed the
1343 compiled form incompatibly. Anybody with saved compiled patterns that use
1344 \p or \P will have to recompile them.
1345
1346 22. Added "Any" and "L&" to the supported property types.
1347
1348 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
1349 but give a compile time error if the value is greater than 0xff.
1350
1351 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
1352 accidentally not being installed or uninstalled.
1353
1354 25. The pcre.h file was built from pcre.h.in, but the only changes that were
1355 made were to insert the current release number. This seemed silly, because
1356 it made things harder for people building PCRE on systems that don't run
1357 "configure". I have turned pcre.h into a distributed file, no longer built
1358 by "configure", with the version identification directly included. There is
1359 no longer a pcre.h.in file.
1360
1361 However, this change necessitated a change to the pcre-config script as
1362 well. It is built from pcre-config.in, and one of the substitutions was the
1363 release number. I have updated configure.ac so that ./configure now finds
1364 the release number by grepping pcre.h.
1365
1366 26. Added the ability to run the tests under valgrind.
1367
1368
1369 Version 6.4 05-Sep-05
1370 ---------------------
1371
1372 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
1373 "--" to be printed when multiple files were scanned, even when none of the
1374 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
1375 consider it to be a bug, and have restored the previous behaviour.
1376
1377 2. A couple of code tidies to get rid of compiler warnings.
1378
1379 3. The pcretest program used to cheat by referring to symbols in the library
1380 whose names begin with _pcre_. These are internal symbols that are not
1381 really supposed to be visible externally, and in some environments it is
1382 possible to suppress them. The cheating is now confined to including
1383 certain files from the library's source, which is a bit cleaner.
1384
1385 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
1386 file's purpose clearer.
1387
1388 5. Reorganized pcre_ucp_findchar().
1389
1390
1391 Version 6.3 15-Aug-05
1392 ---------------------
1393
1394 1. The file libpcre.pc.in did not have general read permission in the tarball.
1395
1396 2. There were some problems when building without C++ support:
1397
1398 (a) If C++ support was not built, "make install" and "make test" still
1399 tried to test it.
1400
1401 (b) There were problems when the value of CXX was explicitly set. Some
1402 changes have been made to try to fix these, and ...
1403
1404 (c) --disable-cpp can now be used to explicitly disable C++ support.
1405
1406 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
1407 backslash in a target when C++ was disabled. This confuses some
1408 versions of "make", apparently. Using an intermediate variable solves
1409 this. (Same for CPP_LOBJ.)
1410
1411 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
1412 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
1413 necessary on certain architectures.
1414
1415 4. Added a setting of -export-symbols-regex to the link command to remove
1416 those symbols that are exported in the C sense, but actually are local
1417 within the library, and not documented. Their names all begin with
1418 "_pcre_". This is not a perfect job, because (a) we have to except some
1419 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
1420 available (and never for static libraries). I have made a note to try to
1421 find a way round (a) in the future.
1422
1423
1424 Version 6.2 01-Aug-05
1425 ---------------------
1426
1427 1. There was no test for integer overflow of quantifier values. A construction
1428 such as {1111111111111111} would give undefined results. What is worse, if
1429 a minimum quantifier for a parenthesized subpattern overflowed and became
1430 negative, the calculation of the memory size went wrong. This could have
1431 led to memory overwriting.
1432
1433 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
1434
1435 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
1436 operating environments where this matters.
1437
1438 4. Applied Giuseppe Maxia's patch to add additional features for controlling
1439 PCRE options from within the C++ wrapper.
1440
1441 5. Named capturing subpatterns were not being correctly counted when a pattern
1442 was compiled. This caused two problems: (a) If there were more than 100
1443 such subpatterns, the calculation of the memory needed for the whole
1444 compiled pattern went wrong, leading to an overflow error. (b) Numerical
1445 back references of the form \12, where the number was greater than 9, were
1446 not recognized as back references, even though there were sufficient
1447 previous subpatterns.
1448
1449 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
1450 versions of gcc, e.g. 2.95.4.
1451
1452
1453 Version 6.1 21-Jun-05
1454 ---------------------
1455
1456 1. There was one reference to the variable "posix" in pcretest.c that was not
1457 surrounded by "#if !defined NOPOSIX".
1458
1459 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
1460 the cross-check on the old pcre_info() function, for the benefit of the
1461 cut-down version of PCRE that is currently imported into Exim.
1462
1463 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
1464 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
1465 patterns that start (?i) but I don't think that matters. The use of (?i) is
1466 just an example; this all applies to the other options as well.
1467
1468 4. Since libtool seems to echo the compile commands it is issuing, the output
1469 from "make" can be reduced a bit by putting "@" in front of each libtool
1470 compile command.
1471
1472 5. Patch from the folks at Google for configure.in to be a bit more thorough
1473 in checking for a suitable C++ installation before trying to compile the
1474 C++ stuff. This should fix a reported problem when a compiler was present,
1475 but no suitable headers.
1476
1477 6. The man pages all had just "PCRE" as their title. I have changed them to
1478 be the relevant file name. I have also arranged that these names are
1479 retained in the file doc/pcre.txt, which is a concatenation in text format
1480 of all the man pages except the little individual ones for each function.
1481
1482 7. The NON-UNIX-USE file had not been updated for the different set of source
1483 files that come with release 6. I also added a few comments about the C++
1484 wrapper.
1485
1486
1487 Version 6.0 07-Jun-05
1488 ---------------------
1489
1490 1. Some minor internal re-organization to help with my DFA experiments.
1491
1492 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
1493 didn't matter for the library itself when fully configured, but did matter
1494 when compiling without UCP support, or within Exim, where the ucp files are
1495 not imported.
1496
1497 3. Refactoring of the library code to split up the various functions into
1498 different source modules. The addition of the new DFA matching code (see
1499 below) to a single monolithic source would have made it really too
1500 unwieldy, quite apart from causing all the code to be include in a
1501 statically linked application, when only some functions are used. This is
1502 relevant even without the DFA addition now that patterns can be compiled in
1503 one application and matched in another.
1504
1505 The downside of splitting up is that there have to be some external
1506 functions and data tables that are used internally in different modules of
1507 the library but which are not part of the API. These have all had their
1508 names changed to start with "_pcre_" so that they are unlikely to clash
1509 with other external names.
1510
1511 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
1512 a different (DFA) algorithm. Although it is slower than the original
1513 function, it does have some advantages for certain types of matching
1514 problem.
1515
1516 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
1517 including restarting after a partial match.
1518
1519 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
1520 defined when compiling for Windows was sent to me. I have put it into the
1521 code, though I have no means of testing or verifying it.
1522
1523 7. Added the pcre_refcount() auxiliary function.
1524
1525 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
1526 match before or at the first newline in the subject string. In pcretest,
1527 the /f option on a pattern can be used to set this.
1528
1529 9. A repeated \w when used in UTF-8 mode with characters greater than 256
1530 would behave wrongly. This has been present in PCRE since release 4.0.
1531
1532 10. A number of changes to the pcregrep command:
1533
1534 (a) Refactored how -x works; insert ^(...)$ instead of setting
1535 PCRE_ANCHORED and checking the length, in preparation for adding
1536 something similar for -w.
1537
1538 (b) Added the -w (match as a word) option.
1539
1540 (c) Refactored the way lines are read and buffered so as to have more
1541 than one at a time available.
1542
1543 (d) Implemented a pcregrep test script.
1544
1545 (e) Added the -M (multiline match) option. This allows patterns to match
1546 over several lines of the subject. The buffering ensures that at least
1547 8K, or the rest of the document (whichever is the shorter) is available
1548 for matching (and similarly the previous 8K for lookbehind assertions).
1549
1550 (f) Changed the --help output so that it now says
1551
1552 -w, --word-regex(p)
1553
1554 instead of two lines, one with "regex" and the other with "regexp"
1555 because that confused at least one person since the short forms are the
1556 same. (This required a bit of code, as the output is generated
1557 automatically from a table. It wasn't just a text change.)
1558
1559 (g) -- can be used to terminate pcregrep options if the next thing isn't an
1560 option but starts with a hyphen. Could be a pattern or a path name
1561 starting with a hyphen, for instance.
1562
1563 (h) "-" can be given as a file name to represent stdin.
1564
1565 (i) When file names are being printed, "(standard input)" is used for
1566 the standard input, for compatibility with GNU grep. Previously
1567 "<stdin>" was used.
1568
1569 (j) The option --label=xxx can be used to supply a name to be used for
1570 stdin when file names are being printed. There is no short form.
1571
1572 (k) Re-factored the options decoding logic because we are going to add
1573 two more options that take data. Such options can now be given in four
1574 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
1575
1576 (l) Added the -A, -B, and -C options for requesting that lines of context
1577 around matches be printed.
1578
1579 (m) Added the -L option to print the names of files that do not contain
1580 any matching lines, that is, the complement of -l.
1581
1582 (n) The return code is 2 if any file cannot be opened, but pcregrep does
1583 continue to scan other files.
1584
1585 (o) The -s option was incorrectly implemented. For compatibility with other
1586 greps, it now suppresses the error message for a non-existent or non-
1587 accessible file (but not the return code). There is a new option called
1588 -q that suppresses the output of matching lines, which was what -s was
1589 previously doing.
1590
1591 (p) Added --include and --exclude options to specify files for inclusion
1592 and exclusion when recursing.
1593
1594 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
1595 Hopefully, it now does.
1596
1597 12. Missing cast in pcre_study().
1598
1599 13. Added an "uninstall" target to the makefile.
1600
1601 14. Replaced "extern" in the function prototypes in Makefile.in with
1602 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
1603 world, but is set differently for Windows.
1604
1605 15. Added a second compiling function called pcre_compile2(). The only
1606 difference is that it has an extra argument, which is a pointer to an
1607 integer error code. When there is a compile-time failure, this is set
1608 non-zero, in addition to the error test pointer being set to point to an
1609 error message. The new argument may be NULL if no error number is required
1610 (but then you may as well call pcre_compile(), which is now just a
1611 wrapper). This facility is provided because some applications need a
1612 numeric error indication, but it has also enabled me to tidy up the way
1613 compile-time errors are handled in the POSIX wrapper.
1614
1615 16. Added VPATH=.libs to the makefile; this should help when building with one
1616 prefix path and installing with another. (Or so I'm told by someone who
1617 knows more about this stuff than I do.)
1618
1619 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
1620 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
1621 match everything, including newlines. This is not POSIX-compatible, but
1622 somebody wanted the feature. From pcretest it can be activated by using
1623 both the P and the s flags.
1624
1625 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
1626
1627 19. libpcre.pc was being incorrectly installed as executable.
1628
1629 20. A couple of places in pcretest check for end-of-line by looking for '\n';
1630 it now also looks for '\r' so that it will work unmodified on Windows.
1631
1632 21. Added Google's contributed C++ wrapper to the distribution.
1633
1634 22. Added some untidy missing memory free() calls in pcretest, to keep
1635 Electric Fence happy when testing.
1636
1637
1638
1639 Version 5.0 13-Sep-04
1640 ---------------------
1641
1642 1. Internal change: literal characters are no longer packed up into items
1643 containing multiple characters in a single byte-string. Each character
1644 is now matched using a separate opcode. However, there may be more than one
1645 byte in the character in UTF-8 mode.
1646
1647 2. The pcre_callout_block structure has two new fields: pattern_position and
1648 next_item_length. These contain the offset in the pattern to the next match
1649 item, and its length, respectively.
1650
1651 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
1652 insertion of callouts before each pattern item. Added the /C option to
1653 pcretest to make use of this.
1654
1655 4. On the advice of a Windows user, the lines
1656
1657 #if defined(_WIN32) || defined(WIN32)
1658 _setmode( _fileno( stdout ), 0x8000 );
1659 #endif /* defined(_WIN32) || defined(WIN32) */
1660
1661 have been added to the source of pcretest. This apparently does useful
1662 magic in relation to line terminators.
1663
1664 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
1665 for the benefit of those environments where the "b" makes a difference.
1666
1667 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
1668 to know about it. I have put a hack into configure.in that adds in code
1669 to set GCC=yes if CC=icc. This seems to end up at a point in the
1670 generated configure script that is early enough to affect the setting of
1671 compiler options, which is what is needed, but I have no means of testing
1672 whether it really works. (The user who reported this had patched the
1673 generated configure script, which of course I cannot do.)
1674
1675 LATER: After change 22 below (new libtool files), the configure script
1676 seems to know about icc (and also ecc). Therefore, I have commented out
1677 this hack in configure.in.
1678
1679 7. Added support for pkg-config (2 patches were sent in).
1680
1681 8. Negated POSIX character classes that used a combination of internal tables
1682 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
1683 [[:^ascii]]. Typically, they would match almost any characters. The other
1684 POSIX classes were not broken in this way.
1685
1686 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
1687 to find the match, as PCRE was deluded into thinking that the match had to
1688 start at the start point or following a newline. The same bug applied to
1689 patterns with negative forward assertions or any backward assertions
1690 preceding ".*" at the start, unless the pattern required a fixed first
1691 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
1692
1693 10. In UTF-8 mode, when moving forwards in the subject after a failed match
1694 starting at the last subject character, bytes beyond the end of the subject
1695 string were read.
1696
1697 11. Renamed the variable "class" as "classbits" to make life easier for C++
1698 users. (Previously there was a macro definition, but it apparently wasn't
1699 enough.)
1700
1701 12. Added the new field "tables" to the extra data so that tables can be passed
1702 in at exec time, or the internal tables can be re-selected. This allows
1703 a compiled regex to be saved and re-used at a later time by a different
1704 program that might have everything at different addresses.
1705
1706 13. Modified the pcre-config script so that, when run on Solaris, it shows a
1707 -R library as well as a -L library.
1708
1709 14. The debugging options of pcretest (-d on the command line or D on a
1710 pattern) showed incorrect output for anything following an extended class
1711 that contained multibyte characters and which was followed by a quantifier.
1712
1713 15. Added optional support for general category Unicode character properties
1714 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
1715 support. It adds about 90K to the size of the library. The meanings of the
1716 inbuilt class escapes such as \d and \s have NOT been changed.
1717
1718 16. Updated pcredemo.c to include calls to free() to release the memory for the
1719 compiled pattern.
1720
1721 17. The generated file chartables.c was being created in the source directory
1722 instead of in the building directory. This caused the build to fail if the
1723 source directory was different from the building directory, and was
1724 read-only.
1725
1726 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
1727 file. No doubt somebody will tell me if they don't make sense... Also added
1728 Dan Mooney's comments about building on OpenVMS.
1729
1730 19. Added support for partial matching via the PCRE_PARTIAL option for
1731 pcre_exec() and the \P data escape in pcretest.
1732
1733 20. Extended pcretest with 3 new pattern features:
1734
1735 (i) A pattern option of the form ">rest-of-line" causes pcretest to
1736 write the compiled pattern to the file whose name is "rest-of-line".
1737 This is a straight binary dump of the data, with the saved pointer to
1738 the character tables forced to be NULL. The study data, if any, is
1739 written too. After writing, pcretest reads a new pattern.
1740
1741 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
1742 compiled pattern from the given file. There must not be any
1743 occurrences of "<" in the file name (pretty unlikely); if there are,
1744 pcretest will instead treat the initial "<" as a pattern delimiter.
1745 After reading in the pattern, pcretest goes on to read data lines as
1746 usual.
1747
1748 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
1749 and 16-bit fields in a compiled pattern, to simulate a pattern that
1750 was compiled on a host of opposite endianness.
1751
1752 21. The pcre-exec() function can now cope with patterns that were compiled on
1753 hosts of opposite endianness, with this restriction:
1754
1755 As for any compiled expression that is saved and used later, the tables
1756 pointer field cannot be preserved; the extra_data field in the arguments
1757 to pcre_exec() should be used to pass in a tables address if a value
1758 other than the default internal tables were used at compile time.
1759
1760 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
1761 now diagnosed as an error. Previously, most of the time, a negative number
1762 would have been treated as zero, but if in addition "ovector" was passed as
1763 NULL, a crash could occur.
1764
1765 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
1766 new versions from the libtool 1.5 distribution (the last one is a copy of
1767 a file called libtool.m4). This seems to have fixed the need to patch
1768 "configure" to support Darwin 1.3 (which I used to do). However, I still
1769 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
1770 workstation).
1771
1772 24. Changed the PCRE licence to be the more standard "BSD" licence.
1773
1774
1775 Version 4.5 01-Dec-03
1776 ---------------------
1777
1778 1. There has been some re-arrangement of the code for the match() function so
1779 that it can be compiled in a version that does not call itself recursively.
1780 Instead, it keeps those local variables that need separate instances for
1781 each "recursion" in a frame on the heap, and gets/frees frames whenever it
1782 needs to "recurse". Keeping track of where control must go is done by means
1783 of setjmp/longjmp. The whole thing is implemented by a set of macros that
1784 hide most of the details from the main code, and operates only if
1785 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
1786 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
1787 operating.
1788
1789 To make it easier for callers to provide specially tailored get/free
1790 functions for this usage, two new functions, pcre_stack_malloc, and
1791 pcre_stack_free, are used. They are always called in strict stacking order,
1792 and the size of block requested is always the same.
1793
1794 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
1795 PCRE has been compiled to use the stack or the heap for recursion. The
1796 -C option of pcretest uses this to show which version is compiled.
1797
1798 A new data escape \S, is added to pcretest; it causes the amounts of store
1799 obtained and freed by both kinds of malloc/free at match time to be added
1800 to the output.
1801
1802 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
1803 what's available on my current Linux desktop machine.
1804
1805 3. When matching a UTF-8 string, the test for a valid string at the start has
1806 been extended. If start_offset is not zero, PCRE now checks that it points
1807 to a byte that is the start of a UTF-8 character. If not, it returns
1808 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
1809 this is necessary because there may be backward assertions in the pattern.
1810 When matching the same subject several times, it may save resources to use
1811 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
1812
1813 4. The code for checking the validity of UTF-8 strings has been tightened so
1814 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
1815 containing "overlong sequences".
1816
1817 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
1818 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
1819 should have been "&", but it just so happened that all the cases this let
1820 through by mistake were picked up later in the function.
1821
1822 6. I had used a variable called "isblank" - this is a C99 function, causing
1823 some compilers to warn. To avoid this, I renamed it (as "blankclass").
1824
1825 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
1826 prompting; (b) run "./pcretest /dev/null" at the start of the test script
1827 so the version is shown; (c) stop "make test" echoing "./RunTest".
1828
1829 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
1830
1831 9. The prototype for memmove() for systems that don't have it was using
1832 size_t, but the inclusion of the header that defines size_t was later. I've
1833 moved the #includes for the C headers earlier to avoid this.
1834
1835 10. Added some adjustments to the code to make it easier to compiler on certain
1836 special systems:
1837
1838 (a) Some "const" qualifiers were missing.
1839 (b) Added the macro EXPORT before all exported functions; by default this
1840 is defined to be empty.
1841 (c) Changed the dftables auxiliary program (that builds chartables.c) so
1842 that it reads its output file name as an argument instead of writing
1843 to the standard output and assuming this can be redirected.
1844
1845 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
1846 class containing characters with values greater than 255, PCRE compilation
1847 went into a loop.
1848
1849 12. A recursive reference to a subpattern that was within another subpattern
1850 that had a minimum quantifier of zero caused PCRE to crash. For example,
1851 (x(y(?2))z)? provoked this bug with a subject that got as far as the
1852 recursion. If the recursively-called subpattern itself had a zero repeat,
1853 that was OK.
1854
1855 13. In pcretest, the buffer for reading a data line was set at 30K, but the
1856 buffer into which it was copied (for escape processing) was still set at
1857 1024, so long lines caused crashes.
1858
1859 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
1860 "internal error: code overflow...". This applied to any character class
1861 that was followed by a possessive quantifier.
1862
1863 15. Modified the Makefile to add libpcre.la as a prerequisite for
1864 libpcreposix.la because I was told this is needed for a parallel build to
1865 work.
1866
1867 16. If a pattern that contained .* following optional items at the start was
1868 studied, the wrong optimizing data was generated, leading to matching
1869 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
1870 matching string must start with a or b or c. The correct conclusion for
1871 this pattern is that a match can start with any character.
1872
1873
1874 Version 4.4 13-Aug-03
1875 ---------------------
1876
1877 1. In UTF-8 mode, a character class containing characters with values between
1878 127 and 255 was not handled correctly if the compiled pattern was studied.
1879 In fixing this, I have also improved the studying algorithm for such
1880 classes (slightly).
1881
1882 2. Three internal functions had redundant arguments passed to them. Removal
1883 might give a very teeny performance improvement.
1884
1885 3. Documentation bug: the value of the capture_top field in a callout is *one
1886 more than* the number of the hightest numbered captured substring.
1887
1888 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
1889 in incorrectly linking with a previously installed version. They now link
1890 explicitly with libpcre.la.
1891
1892 5. configure.in no longer needs to recognize Cygwin specially.
1893
1894 6. A problem in pcre.in for Windows platforms is fixed.
1895
1896 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
1897 pcretest, it used to include the size of the study block as part of its
1898 output. Unfortunately, the structure contains a field that has a different
1899 size on different hardware architectures. This meant that the tests that
1900 showed this size failed. As the block is currently always of a fixed size,
1901 this information isn't actually particularly useful in pcretest output, so
1902 I have just removed it.
1903
1904 8. Three pre-processor statements accidentally did not start in column 1.
1905 Sadly, there are *still* compilers around that complain, even though
1906 standard C has not required this for well over a decade. Sigh.
1907
1908 9. In pcretest, the code for checking callouts passed small integers in the
1909 callout_data field, which is a void * field. However, some picky compilers
1910 complained about the casts involved for this on 64-bit systems. Now
1911 pcretest passes the address of the small integer instead, which should get
1912 rid of the warnings.
1913
1914 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
1915 both compile and run time, and gives an error if an invalid UTF-8 sequence
1916 is found. There is a option for disabling this check in cases where the
1917 string is known to be correct and/or the maximum performance is wanted.
1918
1919 11. In response to a bug report, I changed one line in Makefile.in from
1920
1921 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
1922 to
1923 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
1924
1925 to look similar to other lines, but I have no way of telling whether this
1926 is the right thing to do, as I do not use Windows. No doubt I'll get told
1927 if it's wrong...
1928
1929
1930 Version 4.3 21-May-03
1931 ---------------------
1932
1933 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
1934 Makefile.
1935
1936 2. Some refactoring to improve the quality of the code:
1937
1938 (i) The utf8_table... variables are now declared "const".
1939
1940 (ii) The code for \cx, which used the "case flipping" table to upper case
1941 lower case letters, now just substracts 32. This is ASCII-specific,
1942 but the whole concept of \cx is ASCII-specific, so it seems
1943 reasonable.
1944
1945 (iii) PCRE was using its character types table to recognize decimal and
1946 hexadecimal digits in the pattern. This is silly, because it handles
1947 only 0-9, a-f, and A-F, but the character types table is locale-
1948 specific, which means strange things might happen. A private
1949 table is now used for this - though it costs 256 bytes, a table is
1950 much faster than multiple explicit tests. Of course, the standard
1951 character types table is still used for matching digits in subject
1952 strings against \d.
1953
1954 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
1955 ending in _t are). So I've renamed it as ESC_tee.
1956
1957 3. The first argument for regexec() in the POSIX wrapper should have been
1958 defined as "const".
1959
1960 4. Changed pcretest to use malloc() for its buffers so that they can be
1961 Electric Fenced for debugging.
1962
1963 5. There were several places in the code where, in UTF-8 mode, PCRE would try
1964 to read one or more bytes before the start of the subject string. Often this
1965 had no effect on PCRE's behaviour, but in some circumstances it could
1966 provoke a segmentation fault.
1967
1968 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
1969 to try to read one or more bytes before the start of the subject string.
1970
1971 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
1972 UTF-8 support could misbehave in various ways if the subject string
1973 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
1974 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
1975 back over UTF-8 characters.)
1976
1977
1978 Version 4.2 14-Apr-03
1979 ---------------------
1980
1981 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
1982
1983 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
1984 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
1985 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
1986 [WIN_PREFIX]: new variable, "cyg" for Cygwin
1987 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
1988 and BUILD_EXEEXT
1989 Note: automatic setting of the BUILD variables is not yet working
1990 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
1991 compile-time but not at link-time
1992 [LINK]: use for linking executables only
1993 make different versions for Windows and non-Windows
1994 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
1995 libraries
1996 [LINK_FOR_BUILD]: new variable
1997 [OBJEXT]: use throughout
1998 [EXEEXT]: use throughout
1999 <winshared>: new target
2000 <wininstall>: new target
2001 <dftables.o>: use native compiler
2002 <dftables>: use native linker
2003 <install>: handle Windows platform correctly
2004 <clean>: ditto
2005 <check>: ditto
2006 copy DLL to top builddir before testing
2007
2008 As part of these changes, -no-undefined was removed again. This was reported
2009 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
2010 in any case.
2011
2012 3. Some tidies to get rid of compiler warnings:
2013
2014 . In the match_data structure, match_limit was an unsigned long int, whereas
2015 match_call_count was an int. I've made them both unsigned long ints.
2016
2017 . In pcretest the fact that a const uschar * doesn't automatically cast to
2018 a void * provoked a warning.
2019
2020 . Turning on some more compiler warnings threw up some "shadow" variables
2021 and a few more missing casts.
2022
2023 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
2024 option, a class that contained a single character with a value between 128
2025 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
2026
2027 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
2028 option, a class that contained several characters, but with at least one
2029 whose value was between 128 and 255 caused PCRE to crash.
2030
2031
2032 Version 4.1 12-Mar-03
2033 ---------------------
2034
2035 1. Compiling with gcc -pedantic found a couple of places where casts were
2036 needed, and a string in dftables.c that was longer than standard compilers are
2037 required to support.
2038
2039 2. Compiling with Sun's compiler found a few more places where the code could
2040 be tidied up in order to avoid warnings.
2041
2042 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
2043 first of these names is deprecated in the latest Autoconf in favour of the name
2044 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
2045 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
2046 analogy I have changed it to CFLAGS_FOR_BUILD.
2047
2048 4. Added -no-undefined to the linking command in the Makefile, because this is
2049 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
2050 linking step for the pcreposix library.
2051
2052 5. PCRE was failing to diagnose the case of two named groups with the same
2053 name.
2054
2055 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
2056 literal character that is needed in the subject for a match, and scans along to
2057 ensure that it is present before embarking on the full matching process. This
2058 saves time in cases of nested unlimited repeats that are never going to match.
2059 Problem: the scan can take a lot of time if the subject is very long (e.g.
2060 megabytes), thus penalizing straightforward matches. It is now done only if the
2061 amount of subject to be scanned is less than 1000 bytes.
2062
2063 7. A lesser problem with the same optimization is that it was recording the
2064 first character of an anchored pattern as "needed", thus provoking a search
2065 right along the subject, even when the first match of the pattern was going to
2066 fail. The "needed" character is now not set for anchored patterns, unless it
2067 follows something in the pattern that is of non-fixed length. Thus, it still
2068 fulfils its original purpose of finding quick non-matches in cases of nested
2069 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
2070
2071
2072 Version 4.0 17-Feb-03
2073 ---------------------
2074
2075 1. If a comment in an extended regex that started immediately after a meta-item
2076 extended to the end of string, PCRE compiled incorrect data. This could lead to
2077 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
2078
2079 2. Moved to autoconf 2.53 and libtool 1.4.2.
2080
2081 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
2082 the special perltest8 script is no longer needed - all the tests can be run
2083 from a single perltest script.
2084
2085 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
2086 by \s. It has now been removed in PCRE. This means it isn't recognized as
2087 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
2088 class [:space:] *does* include VT, thereby creating a mess.
2089
2090 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
2091 space and tab.
2092
2093 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
2094 its new features into the main test script, reducing the number of scripts.
2095
2096 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
2097 were backward compatible, and made the (?i) apply to the whole pattern, as if
2098 /i were given. Now it behaves more logically, and applies the option setting
2099 only to what follows. PCRE has been changed to follow suit. However, if it
2100 finds options settings right at the start of the pattern, it extracts them into
2101 the global options, as before. Thus, they show up in the info data.
2102
2103 8. Added support for the \Q...\E escape sequence. Characters in between are
2104 treated as literals. This is slightly different from Perl in that $ and @ are
2105 also handled as literals inside the quotes. In Perl, they will cause variable
2106 interpolation. Note the following examples:
2107
2108 Pattern PCRE matches Perl matches
2109
2110 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
2111 \Qabc\$xyz\E abc\$xyz abc\$xyz
2112 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
2113
2114 For compatibility with Perl, \Q...\E sequences are recognized inside character
2115 classes as well as outside them.
2116
2117 9. Re-organized 3 code statements in pcretest to avoid "overflow in
2118 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
2119 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
2120 signed/unsigned warnings.
2121
2122 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
2123 option for pcretest, so I've replaced it by a simple function that does just
2124 that job.
2125
2126 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
2127 "pcregrep -".
2128
2129 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
2130 Java package. This provides some syntactic sugar for simple cases of what my
2131 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
2132 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
2133 item, you can use this simplified notation. Note that only makes sense with
2134 greedy quantifiers. Consequently, the use of the possessive quantifier forces
2135 greediness, whatever the setting of the PCRE_UNGREEDY option.
2136
2137 13. A change of greediness default within a pattern was not taking effect at
2138 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
2139 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
2140 was abstracted outside.
2141
2142 14. PCRE now supports the \G assertion. It is true when the current matching
2143 position is at the start point of the match. This differs from \A when the
2144 starting offset is non-zero. Used with the /g option of pcretest (or similar
2145 code), it works in the same way as it does for Perl's /g option. If all
2146 alternatives of a regex begin with \G, the expression is anchored to the start
2147 match position, and the "anchored" flag is set in the compiled expression.
2148
2149 15. Some bugs concerning the handling of certain option changes within patterns
2150 have been fixed. These applied to options other than (?ims). For example,
2151 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
2152 been the other way round. Some of this was related to change 7 above.
2153
2154 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
2155 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
2156 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
2157 POSIX classes only within a class (e.g. /[[:alpha:]]/).
2158
2159 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
2160 mode. Unlike ".", it always matches newline, whatever the setting of
2161 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
2162 assertions. Perl allows it, but it doesn't (in general) work because it can't
2163 calculate the length of the lookbehind. At least, that's the case for Perl
2164 5.8.0 - I've been told they are going to document that it doesn't work in
2165 future.
2166
2167 18. Added an error diagnosis for escapes that PCRE does not support: these are
2168 \L, \l, \N, \P, \p, \U, \u, and \X.
2169
2170 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
2171 reading past the end of the pattern in cases such as /[abcd/.
2172
2173 20. PCRE was getting more memory than necessary for patterns with classes that
2174 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
2175
2176 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
2177 compiling PCRE for use with Virtual Pascal.
2178
2179 22. Small fix to the Makefile to make it work properly if the build is done
2180 outside the source tree.
2181
2182 23. Added a new extension: a condition to go with recursion. If a conditional
2183 subpattern starts with (?(R) the "true" branch is used if recursion has
2184 happened, whereas the "false" branch is used only at the top level.
2185
2186 24. When there was a very long string of literal characters (over 255 bytes
2187 without UTF support, over 250 bytes with UTF support), the computation of how
2188 much memory was required could be incorrect, leading to segfaults or other
2189 strange effects.
2190
2191 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
2192 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
2193 there was a subsequent back reference to those brackets. This meant that, for
2194 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
2195 possible to check for precisely this case. All we can do is abandon the
2196 optimization if .* occurs inside capturing brackets when there are any back
2197 references whatsoever. (See below for a better fix that came later.)
2198
2199 26. The handling of the optimization for finding the first character of a
2200 non-anchored pattern, and for finding a character that is required later in the
2201 match were failing in some cases. This didn't break the matching; it just
2202 failed to optimize when it could. The way this is done has been re-implemented.
2203
2204 27. Fixed typo in error message for invalid (?R item (it said "(?p").
2205
2206 28. Added a new feature that provides some of the functionality that Perl
2207 provides with (?{...}). The facility is termed a "callout". The way it is done
2208 in PCRE is for the caller to provide an optional function, by setting
2209 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
2210 global variable. By default it is unset, which disables all calling out. To get
2211 the function called, the regex must include (?C) at appropriate points. This
2212 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
2213 This provides a means of identifying different callout points. When PCRE
2214 reaches such a point in the regex, if pcre_callout has been set, the external
2215 function is called. It is provided with data in a structure called
2216 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
2217 matching continues; if it returns a non-zero value, the match at the current
2218 point fails. However, backtracking will occur if possible. [This was changed
2219 later and other features added - see item 49 below.]
2220
2221 29. pcretest is upgraded to test the callout functionality. It provides a
2222 callout function that displays information. By default, it shows the start of
2223 the match and the current position in the text. There are some new data escapes
2224 to vary what happens:
2225
2226 \C+ in addition, show current contents of captured substrings
2227 \C- do not supply a callout function
2228 \C!n return 1 when callout number n is reached
2229 \C!n!m return 1 when callout number n is reached for the mth time
2230
2231 30. If pcregrep was called with the -l option and just a single file name, it
2232 output "<stdin>" if a match was found, instead of the file name.
2233
2234 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
2235 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
2236 pcre_exec(). This saves a malloc/free per call. The default value of
2237 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
2238 when configuring.
2239
2240 32. The default maximum size of a compiled pattern is 64K. There have been a
2241 few cases of people hitting this limit. The code now uses macros to handle the
2242 storing of links as offsets within the compiled pattern. It defaults to 2-byte
2243 links, but this can be changed to 3 or 4 bytes by --with-link-size when
2244 configuring. Tests 2 and 5 work only with 2-byte links because they output
2245 debugging information about compiled patterns.
2246
2247 33. Internal code re-arrangements:
2248
2249 (a) Moved the debugging function for printing out a compiled regex into
2250 its own source file (printint.c) and used #include to pull it into
2251 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
2252 separate copies.
2253
2254 (b) Defined the list of op-code names for debugging as a macro in
2255 internal.h so that it is next to the definition of the opcodes.
2256
2257 (c) Defined a table of op-code lengths for simpler skipping along compiled
2258 code. This is again a macro in internal.h so that it is next to the
2259 definition of the opcodes.
2260
2261 34. Added support for recursive calls to individual subpatterns, along the
2262 lines of Robin Houston's patch (but implemented somewhat differently).
2263
2264 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
2265 allow it to read and process whole directories in Win32. This code was
2266 contributed by Lionel Fourquaux; it has not been tested by me.
2267
2268 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
2269 used to name a group. Names consist of alphanumerics and underscores, and must
2270 be unique. Back references use the syntax (?P=name) and recursive calls use
2271 (?P>name) which is a PCRE extension to the Python extension. Groups still have
2272 numbers. The function pcre_fullinfo() can be used after compilation to extract
2273 a name/number map. There are three relevant calls:
2274
2275 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
2276 PCRE_INFO_NAMECOUNT yields the number of entries
2277 PCRE_INFO_NAMETABLE yields a pointer to the map.
2278
2279 The map is a vector of fixed-size entries. The size of each entry depends on
2280 the length of the longest name used. The first two bytes of each entry are the
2281 group number, most significant byte first. There follows the corresponding
2282 name, zero terminated. The names are in alphabetical order.
2283
2284 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
2285 case instead of 255. Making it the same both with and without UTF-8 support
2286 means that the same test output works with both.
2287
2288 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
2289 calling malloc() with a zero argument.
2290
2291 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
2292 optimization. I've improved things by keeping a bitmap of backreferences with
2293 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
2294 fact referenced, the optimization can be applied. It is unlikely that a
2295 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
2296 the match to follow \n) will appear inside brackets with a number greater than
2297 31, but if it does, any back reference > 31 suppresses the optimization.
2298
2299 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
2300 of disabling numbered capturing parentheses. Any opening parenthesis that is
2301 not followed by ? behaves as if it were followed by ?: but named parentheses
2302 can still be used for capturing (and they will acquire numbers in the usual
2303 way).
2304
2305 41. Redesigned the return codes from the match() function into yes/no/error so
2306 that errors can be passed back from deep inside the nested calls. A malloc
2307 failure while inside a recursive subpattern call now causes the
2308 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
2309
2310 42. It is now possible to set a limit on the number of times the match()
2311 function is called in a call to pcre_exec(). This facility makes it possible to
2312 limit the amount of recursion and backtracking, though not in a directly
2313 obvious way, because the match() function is used in a number of different
2314 circumstances. The count starts from zero for each position in the subject
2315 string (for non-anchored patterns). The default limit is, for compatibility, a
2316 large number, namely 10 000 000. You can change this in two ways:
2317
2318 (a) When configuring PCRE before making, you can use --with-match-limit=n
2319 to set a default value for the compiled library.
2320
2321 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
2322 a different value is set. See 45 below.
2323
2324 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
2325
2326 43. Added a new function pcre_config(int, void *) to enable run-time extraction
2327 of things that can be changed at compile time. The first argument specifies
2328 what is wanted and the second points to where the information is to be placed.
2329 The current list of available information is:
2330
2331 PCRE_CONFIG_UTF8
2332
2333 The output is an integer that is set to one if UTF-8 support is available;
2334 otherwise it is set to zero.
2335
2336 PCRE_CONFIG_NEWLINE
2337
2338 The output is an integer that it set to the value of the code that is used for
2339 newline. It is either LF (10) or CR (13).
2340
2341 PCRE_CONFIG_LINK_SIZE
2342
2343 The output is an integer that contains the number of bytes used for internal
2344 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
2345
2346 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
2347
2348 The output is an integer that contains the threshold above which the POSIX
2349 interface uses malloc() for output vectors. See item 31 above.
2350
2351 PCRE_CONFIG_MATCH_LIMIT
2352
2353 The output is an unsigned integer that contains the default limit of the number
2354 of match() calls in a pcre_exec() execution. See 42 above.
2355
2356 44. pcretest has been upgraded by the addition of the -C option. This causes it
2357 to extract all the available output from the new pcre_config() function, and to
2358 output it. The program then exits immediately.
2359
2360 45. A need has arisen to pass over additional data with calls to pcre_exec() in
2361 order to support additional features. One way would have been to define
2362 pcre_exec2() (for example) with extra arguments, but this would not have been
2363 extensible, and would also have required all calls to the original function to
2364 be mapped to the new one. Instead, I have chosen to extend the mechanism that
2365 is used for passing in "extra" data from pcre_study().
2366
2367 The pcre_extra structure is now exposed and defined in pcre.h. It currently
2368 contains the following fields:
2369
2370 flags a bitmap indicating which of the following fields are set
2371 study_data opaque data from pcre_study()
2372 match_limit a way of specifying a limit on match() calls for a specific
2373 call to pcre_exec()
2374 callout_data data for callouts (see 49 below)
2375
2376 The flag bits are also defined in pcre.h, and are
2377
2378 PCRE_EXTRA_STUDY_DATA
2379 PCRE_EXTRA_MATCH_LIMIT
2380 PCRE_EXTRA_CALLOUT_DATA
2381
2382 The pcre_study() function now returns one of these new pcre_extra blocks, with
2383 the actual study data pointed to by the study_data field, and the
2384 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
2385 before. That is, this change is entirely upwards-compatible and requires no
2386 change to existing code.
2387
2388 If you want to pass in additional data to pcre_exec(), you can either place it
2389 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
2390 block.
2391
2392 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
2393 data string contains the escape sequence \M, pcretest calls pcre_exec() several
2394 times with different match limits, until it finds the minimum value needed for
2395 pcre_exec() to complete. The value is then output. This can be instructive; for
2396 most simple matches the number is quite small, but for pathological cases it
2397 gets very large very quickly.
2398
2399 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
2400 returns the size of the data block pointed to by the study_data field in a
2401 pcre_extra block, that is, the value that was passed as the argument to
2402 pcre_malloc() when PCRE was getting memory in which to place the information
2403 created by pcre_study(). The fourth argument should point to a size_t variable.
2404 pcretest has been extended so that this information is shown after a successful
2405 pcre_study() call when information about the compiled regex is being displayed.
2406
2407 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
2408 because what follows is always an absolute path. (Later: it turns out that this
2409 is more than cosmetic for MinGW, because it doesn't like empty path
2410 components.)
2411
2412 49. Some changes have been made to the callout feature (see 28 above):
2413
2414 (i) A callout function now has three choices for what it returns:
2415
2416 0 => success, carry on matching
2417 > 0 => failure at this point, but backtrack if possible
2418 < 0 => serious error, return this value from pcre_exec()
2419
2420 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
2421 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
2422 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
2423 use by callout functions. It will never be used by PCRE itself.
2424
2425 (ii) The pcre_extra structure (see 45 above) has a void * field called
2426 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
2427 pcre_callout_block structure has a field of the same name. The contents of
2428 the field passed in the pcre_extra structure are passed to the callout
2429 function in the corresponding field in the callout block. This makes it
2430 easier to use the same callout-containing regex from multiple threads. For
2431 testing, the pcretest program has a new data escape
2432
2433 \C*n pass the number n (may be negative) as callout_data
2434
2435 If the callout function in pcretest receives a non-zero value as
2436 callout_data, it returns that value.
2437
2438 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
2439 there were some redundant $(CFLAGS) in commands that are now specified as
2440 $(LINK), which already includes $(CFLAGS).
2441
2442 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
2443 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
2444 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
2445 one-byte characters throughout. Note that case-insensitive matching applies
2446 only to characters whose values are less than 256. PCRE doesn't support the
2447 notion of cases for higher-valued characters.
2448
2449 (i) A character class whose characters are all within 0-255 is handled as
2450 a bit map, and the map is inverted for negative classes. Previously, a
2451 character > 255 always failed to match such a class; however it should
2452 match if the class was a negative one (e.g. [^ab]). This has been fixed.
2453
2454 (ii) A negated character class with a single character < 255 is coded as
2455 "not this character" (OP_NOT). This wasn't working properly when the test
2456 character was multibyte, either singly or repeated.
2457
2458 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
2459 mode, for example: \x{100}{2,3}.
2460
2461 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
2462 singly or repeated) now correctly test multibyte characters. However,
2463 PCRE doesn't recognize any characters with values greater than 255 as
2464 digits, spaces, or word characters. Such characters always match \D, \S,
2465 and \W, and never match \d, \s, or \w.
2466
2467 (v) Classes may now contain characters and character ranges with values
2468 greater than 255. For example: [ab\x{100}-\x{400}].
2469
2470 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
2471 PCRE in UTF-8 mode.
2472
2473 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
2474 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
2475 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
2476 value.)
2477
2478 53. The single man page has become too large. I have therefore split it up into
2479 a number of separate man pages. These also give rise to individual HTML pages;
2480 these are now put in a separate directory, and there is an index.html page that
2481 lists them all. Some hyperlinking between the pages has been installed.
2482
2483 54. Added convenience functions for handling named capturing parentheses.
2484
2485 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
2486 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
2487 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
2488 are faulted.
2489
2490 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
2491 calling configure. These values are used when compiling the dftables.c program
2492 which is run to generate the source of the default character tables. They
2493 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
2494 you will need to set these values.
2495
2496 57. Updated the building process for Windows DLL, as provided by Fred Cox.
2497
2498
2499 Version 3.9 02-Jan-02
2500 ---------------------
2501
2502 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
2503
2504 2. If --disable-static was given, the building process failed when trying to
2505 build pcretest and pcregrep. (For some reason it was using libtool to compile
2506 them, which is not right, as they aren't part of the library.)
2507
2508
2509 Version 3.8 18-Dec-01
2510 ---------------------
2511
2512 1. The experimental UTF-8 code was completely screwed up. It was packing the
2513 bytes in the wrong order. How dumb can you get?
2514
2515
2516 Version 3.7 29-Oct-01
2517 ---------------------
2518
2519 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
2520 This caused pcretest, when used on the test data, to segfault. Unfortunately,
2521 this didn't happen under Solaris 8, where I normally test things.
2522
2523 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
2524 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
2525 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
2526 this fix an hour or so after the initial 3.7 release.)
2527
2528
2529 Version 3.6 23-Oct-01
2530 ---------------------
2531
2532 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
2533 offsets passed as NULL with zero offset count.
2534
2535 2. The config.guess and config.sub files had not been updated when I moved to
2536 the latest autoconf.
2537
2538
2539 Version 3.5 15-Aug-01
2540 ---------------------
2541
2542 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
2543 had been forgotten.
2544
2545 2. By using declared but undefined structures, we can avoid using "void"
2546 definitions in pcre.h while keeping the internal definitions of the structures
2547 private.
2548
2549 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
2550 user point of view, this means that both static and shared libraries are built
2551 by default, but this can be individually controlled. More of the work of
2552 handling this static/shared cases is now inside libtool instead of PCRE's make
2553 file.
2554
2555 4. The pcretest utility is now installed along with pcregrep because it is
2556 useful for users (to test regexs) and by doing this, it automatically gets
2557 relinked by libtool. The documentation has been turned into a man page, so
2558 there are now .1, .txt, and .html versions in /doc.
2559
2560 5. Upgrades to pcregrep:
2561 (i) Added long-form option names like gnu grep.
2562 (ii) Added --help to list all options with an explanatory phrase.
2563 (iii) Added -r, --recursive to recurse into sub-directories.
2564 (iv) Added -f, --file to read patterns from a file.
2565
2566 6. pcre_exec() was referring to its "code" argument before testing that
2567 argument for NULL (and giving an error if it was NULL).
2568
2569 7. Upgraded Makefile.in to allow for compiling in a different directory from
2570 the source directory.
2571
2572 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
2573 options bits, the pointer it was passed was to an int instead of to an unsigned
2574 long int. This mattered only on 64-bit systems.
2575
2576 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
2577 generated) instead of pcre.in, which it its source. Also made the same change
2578 in several of the .c files.
2579
2580 10. A new release of gcc defines printf() as a macro, which broke pcretest
2581 because it had an ifdef in the middle of a string argument for printf(). Fixed
2582 by using separate calls to printf().
2583
2584 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
2585 script, to force use of CR or LF instead of \n in the source. On non-Unix
2586 systems, the value can be set in config.h.
2587
2588 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
2589 absolute limit. Changed the text of the error message to make this clear, and
2590 likewise updated the man page.
2591
2592 13. The limit of 99 on the number of capturing subpatterns has been removed.
2593 The new limit is 65535, which I hope will not be a "real" limit.
2594
2595
2596 Version 3.4 22-Aug-00
2597 ---------------------
2598
2599 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
2600
2601 2. Diagnose condition (?(0) as an error instead of crashing on matching.
2602
2603
2604 Version 3.3 01-Aug-00
2605 ---------------------
2606
2607 1. If an octal character was given, but the value was greater than \377, it
2608 was not getting masked to the least significant bits, as documented. This could
2609 lead to crashes in some systems.
2610
2611 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
2612 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
2613
2614 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
2615 These just pass their arguments on to (pcre_free)(), but they are provided
2616 because some uses of PCRE bind it to non-C systems that can call its functions,
2617 but cannot call free() or pcre_free() directly.
2618
2619 4. Add "make test" as a synonym for "make check". Corrected some comments in
2620 the Makefile.
2621
2622 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
2623 Makefile.
2624
2625 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
2626 command called pgrep for grepping around the active processes.
2627
2628 7. Added the beginnings of support for UTF-8 character strings.
2629
2630 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
2631 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
2632 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
2633 out for the ar command.)
2634
2635
2636 Version 3.2 12-May-00
2637 ---------------------
2638
2639 This is purely a bug fixing release.
2640
2641 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
2642 of ZA. This was just one example of several cases that could provoke this bug,
2643 which was introduced by change 9 of version 2.00. The code for breaking
2644 infinite loops after an iteration that matches an empty string was't working
2645 correctly.
2646
2647 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
2648 when matched against abbab (for example). After matching an empty string, it
2649 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
2650 caused it to match further down the string than it should.
2651
2652 3. The code contained an inclusion of sys/types.h. It isn't clear why this
2653 was there because it doesn't seem to be needed, and it causes trouble on some
2654 systems, as it is not a Standard C header. It has been removed.
2655
2656 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
2657 were reported on the Macintosh. The changes were from
2658
2659 while ((c = *(++ptr)) != 0 && c != '\n');
2660 to
2661 while ((c = *(++ptr)) != 0 && c != '\n') ;
2662
2663 Totally extraordinary, but if that's what it takes...
2664
2665 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
2666 available. Added HAVE_BCOPY and an autoconf test for it; if neither
2667 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
2668 assumes the way PCRE uses memmove() (always moving upwards).
2669
2670 6. PCRE is being used in one environment where strchr() is not available. There
2671 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
2672 faster code anyway.
2673
2674
2675 Version 3.1 09-Feb-00
2676 ---------------------
2677
2678 The only change in this release is the fixing of some bugs in Makefile.in for
2679 the "install" target:
2680
2681 (1) It was failing to install pcreposix.h.
2682
2683 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
2684
2685
2686 Version 3.0 01-Feb-00
2687 ---------------------
2688
2689 1. Add support for the /+ modifier to perltest (to output $` like it does in
2690 pcretest).
2691
2692 2. Add support for the /g modifier to perltest.
2693
2694 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
2695 matches null strings.
2696
2697 4. Fix perltest so that it doesn't do unwanted things when fed an empty
2698 pattern. Perl treats empty patterns specially - it reuses the most recent
2699 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
2700 effect.
2701
2702 5. The POSIX interface was broken in that it was just handing over the POSIX
2703 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
2704 required a bigger vector, with some working space on the end. This means that
2705 the POSIX wrapper now has to get and free some memory, and copy the results.
2706
2707 6. Added some simple autoconf support, placing the test data and the
2708 documentation in separate directories, re-organizing some of the
2709 information files, and making it build pcre-config (a GNU standard). Also added
2710 libtool support for building PCRE as a shared library, which is now the
2711 default.
2712
2713 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
2714 09 are not valid octal constants. Single digits will be used for minor values
2715 less than 10.
2716
2717 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
2718 existing programs that set these in the POSIX interface can use PCRE without
2719 modification.
2720
2721 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
2722 return all that pcre_info() returns, plus additional data. The pcre_info()
2723 function is retained for compatibility, but is considered to be obsolete.
2724
2725 10. Added experimental recursion feature (?R) to handle one common case that
2726 Perl 5.6 will be able to do with (?p{...}).
2727
2728 11. Added support for POSIX character classes like [:alpha:], which Perl is
2729 adopting.
2730
2731
2732 Version 2.08 31-Aug-99
2733 ----------------------
2734
2735 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
2736 trying to match at the startoffset position, but instead was moving forward to
2737 the next newline as if a previous match had failed.
2738
2739 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
2740 and could get into a loop if a null string was matched other than at the start
2741 of the subject.
2742
2743 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
2744 be distinguished at compile time, and for completeness also added PCRE_DATE.
2745
2746 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
2747 in GnuWin32 environments.
2748
2749
2750 Version 2.07 29-Jul-99
2751 ----------------------
2752
2753 1. The documentation is now supplied in plain text form and HTML as well as in
2754 the form of man page sources.
2755
2756 2. C++ compilers don't like assigning (void *) values to other pointer types.
2757 In particular this affects malloc(). Although there is no problem in Standard
2758 C, I've put in casts to keep C++ compilers happy.
2759
2760 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
2761 should be (const char *).
2762
2763 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
2764 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
2765 However, I haven't made this a standard facility. The documentation doesn't
2766 mention it, and the Makefile doesn't support it.
2767
2768 5. The Makefile now contains an "install" target, with editable destinations at
2769 the top of the file. The pcretest program is not installed.
2770
2771 6. pgrep -V now gives the PCRE version number and date.
2772
2773 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
2774 causing the entire string to be ignored, instead of just the last character.
2775
2776 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
2777 non-matching string, it can take a very, very long time, even for strings of
2778 quite modest length, because of the nested recursion. PCRE now does better in
2779 some of these cases. It does this by remembering the last required literal
2780 character in the pattern, and pre-searching the subject to ensure it is present
2781 before running the real match. In other words, it applies a heuristic to detect
2782 some types of certain failure quickly, and in the above example, if presented
2783 with a string that has no trailing " it gives "no match" very quickly.
2784
2785 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
2786 other alternatives are tried instead.
2787
2788
2789 Version 2.06 09-Jun-99
2790 ----------------------
2791
2792 1. Change pcretest's output for amount of store used to show just the code
2793 space, because the remainder (the data block) varies in size between 32-bit and
2794 64-bit systems.
2795
2796 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
2797 start matching at. This allows lookbehinds to work when searching for multiple
2798 occurrences in a string.
2799
2800 3. Added additional options to pcretest for testing multiple occurrences:
2801
2802 /+ outputs the rest of the string that follows a match
2803 /g loops for multiple occurrences, using the new startoffset argument
2804 /G loops for multiple occurrences by passing an incremented pointer
2805
2806 4. PCRE wasn't doing the "first character" optimization for patterns starting
2807 with \b or \B, though it was doing it for other lookbehind assertions. That is,
2808 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
2809 the letter 'x'. On long subject strings, this gives a significant speed-up.
2810
2811
2812 Version 2.05 21-Apr-99
2813 ----------------------
2814
2815 1. Changed the type of magic_number from int to long int so that it works
2816 properly on 16-bit systems.
2817
2818 2. Fixed a bug which caused patterns starting with .* not to work correctly
2819 when the subject string contained newline characters. PCRE was assuming
2820 anchoring for such patterns in all cases, which is not correct because .* will
2821 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
2822 DOTALL is set at top level; otherwise it knows that patterns starting with .*
2823 must be retried after every newline in the subject.
2824
2825
2826 Version 2.04 18-Feb-99
2827 ----------------------
2828
2829 1. For parenthesized subpatterns with repeats whose minimum was zero, the
2830 computation of the store needed to hold the pattern was incorrect (too large).
2831 If such patterns were nested a few deep, this could multiply and become a real
2832 problem.
2833
2834 2. Added /M option to pcretest to show the memory requirement of a specific
2835 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
2836
2837 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
2838 compiled in such a way that the backtracking after subsequent failure was
2839 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
2840 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
2841
2842
2843 Version 2.03 02-Feb-99
2844 ----------------------
2845
2846 1. Fixed typo and small mistake in man page.
2847
2848 2. Added 4th condition (GPL supersedes if conflict) and created separate
2849 LICENCE file containing the conditions.
2850
2851 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
2852 Perl, that is the internal \ allows the delimiter to be included in the
2853 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
2854 the final delimiter, add \ to the end of the pattern (to test the error).
2855
2856 4. Added the convenience functions for extracting substrings after a successful
2857 match. Updated pcretest to make it able to test these functions.
2858
2859
2860 Version 2.02 14-Jan-99
2861 ----------------------
2862
2863 1. Initialized the working variables associated with each extraction so that
2864 their saving and restoring doesn't refer to uninitialized store.
2865
2866 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
2867 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
2868 fix the problem.
2869
2870 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
2871 calls, and wasn't printing the correct value for compiling calls. Increased the
2872 default value of LOOPREPEAT, and the number of significant figures in the
2873 times.
2874
2875 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
2876
2877 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
2878 a building problem on Windows NT with a FAT file system.
2879
2880
2881 Version 2.01 21-Oct-98
2882 ----------------------
2883
2884 1. Changed the API for pcre_compile() to allow for the provision of a pointer
2885 to character tables built by pcre_maketables() in the current locale. If NULL
2886 is passed, the default tables are used.
2887
2888
2889 Version 2.00 24-Sep-98
2890 ----------------------
2891
2892 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
2893 it any more.
2894
2895 2. Allow quantification of (?>) groups, and make it work correctly.
2896
2897 3. The first character computation wasn't working for (?>) groups.
2898
2899 4. Correct the implementation of \Z (it is permitted to match on the \n at the
2900 end of the subject) and add 5.005's \z, which really does match only at the
2901 very end of the subject.
2902
2903 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
2904
2905 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
2906 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
2907 localized options. All options to pcre_study() were also removed.
2908
2909 7. Add other new features from 5.005:
2910
2911 $(?<= positive lookbehind
2912 $(?<! negative lookbehind
2913 (?imsx-imsx) added the unsetting capability
2914 such a setting is global if at outer level; local otherwise
2915 (?imsx-imsx:) non-capturing groups with option setting
2916 (?(cond)re|re) conditional pattern matching
2917
2918 A backreference to itself in a repeated group matches the previous
2919 captured string.
2920
2921 8. General tidying up of studying (both automatic and via "study")
2922 consequential on the addition of new assertions.
2923
2924 9. As in 5.005, unlimited repeated groups that could match an empty substring
2925 are no longer faulted at compile time. Instead, the loop is forcibly broken at
2926 runtime if any iteration does actually match an empty substring.
2927
2928 10. Include the RunTest script in the distribution.
2929
2930 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
2931 discrepancies, some of which were old and were also with respect to 5.004. They
2932 have now been fixed.
2933
2934
2935 Version 1.09 28-Apr-98
2936 ----------------------
2937
2938 1. A negated single character class followed by a quantifier with a minimum
2939 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
2940 program crashes, or just wrong answers. This did not apply to negated classes
2941 containing more than one character, or to minima other than one.
2942
2943
2944 Version 1.08 27-Mar-98
2945 ----------------------
2946
2947 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
2948
2949 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
2950 latter must appear before anything that relies on it in the pattern.
2951
2952
2953 Version 1.07 16-Feb-98
2954 ----------------------
2955
2956 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
2957 repeat of a potentially empty string).
2958
2959
2960 Version 1.06 23-Jan-98
2961 ----------------------
2962
2963 1. Added Markus Oberhumer's little patches for C++.
2964
2965 2. Literal strings longer than 255 characters were broken.
2966
2967
2968 Version 1.05 23-Dec-97
2969 ----------------------
2970
2971 1. Negated character classes containing more than one character were failing if
2972 PCRE_CASELESS was set at run time.
2973
2974
2975 Version 1.04 19-Dec-97
2976 ----------------------
2977
2978 1. Corrected the man page, where some "const" qualifiers had been omitted.
2979
2980 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
2981 input syntax.
2982
2983 3. Fixed memory leak which occurred when a regex with back references was
2984 matched with an offsets vector that wasn't big enough. The temporary memory
2985 that is used in this case wasn't being freed if the match failed.
2986
2987 4. Tidied pcretest to ensure it frees memory that it gets.
2988
2989 5. Temporary memory was being obtained in the case where the passed offsets
2990 vector was exactly big enough.
2991
2992 6. Corrected definition of offsetof() from change 5 below.
2993
2994 7. I had screwed up change 6 below and broken the rules for the use of
2995 setjmp(). Now fixed.
2996
2997
2998 Version 1.03 18-Dec-97
2999 ----------------------
3000
3001 1. A erroneous regex with a missing opening parenthesis was correctly
3002 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
3003 on some systems.
3004
3005 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
3006 it was reported that one broken compiler failed on the former because "code" is
3007 also an independent variable.
3008
3009 3. The erroneous regex a[]b caused an array overrun reference.
3010
3011 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
3012 fail on data ending with that character. (It was going on too far, and checking
3013 the next character, typically a binary zero.) This was specific to the
3014 optimized code for single-character negative classes.
3015
3016 5. Added a contributed patch from the TIN world which does the following:
3017
3018 + Add an undef for memmove, in case the the system defines a macro for it.
3019
3020 + Add a definition of offsetof(), in case there isn't one. (I don't know
3021 the reason behind this - offsetof() is part of the ANSI standard - but
3022 it does no harm).
3023
3024 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
3025 most of the places where whitespace preceded '#'. I have given up and
3026 allowed the remaining 2 cases to be at the margin.
3027
3028 + Rename some variables in pcre to eliminate shadowing. This seems very
3029 pedantic, but does no harm, of course.
3030
3031 6. Moved the call to setjmp() into its own function, to get rid of warnings
3032 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
3033
3034 7. Constructs such as \d{8,} were compiling into the equivalent of
3035 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
3036 outcome, but in this particular case used more store than had been allocated,
3037 which caused the bug to be discovered because it threw up an internal error.
3038
3039 8. The debugging code in both pcre and pcretest for outputting the compiled
3040 form of a regex was going wrong in the case of back references followed by
3041 curly-bracketed repeats.
3042
3043
3044 Version 1.02 12-Dec-97
3045 ----------------------
3046
3047 1. Typos in pcre.3 and comments in the source fixed.
3048
3049 2. Applied a contributed patch to get rid of places where it used to remove
3050 'const' from variables, and fixed some signed/unsigned and uninitialized
3051 variable warnings.
3052
3053 3. Added the "runtest" target to Makefile.
3054
3055 4. Set default compiler flag to -O2 rather than just -O.
3056
3057
3058 Version 1.01 19-Nov-97
3059 ----------------------
3060
3061 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
3062 like /([ab]*)*/, that is, for classes with more than one character in them.
3063
3064 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
3065 as /((?>a*))*/ (a PCRE_EXTRA facility).
3066
3067
3068 Version 1.00 18-Nov-97
3069 ----------------------
3070
3071 1. Added compile-time macros to support systems such as SunOS4 which don't have
3072 memmove() or strerror() but have other things that can be used instead.
3073
3074 2. Arranged that "make clean" removes the executables.
3075
3076
3077 Version 0.99 27-Oct-97
3078 ----------------------
3079
3080 1. Fixed bug in code for optimizing classes with only one character. It was
3081 initializing a 32-byte map regardless, which could cause it to run off the end
3082 of the memory it had got.
3083
3084 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
3085
3086
3087 Version 0.98 22-Oct-97
3088 ----------------------
3089
3090 1. Fixed bug in code for handling temporary memory usage when there are more
3091 back references than supplied space in the ovector. This could cause segfaults.
3092
3093
3094 Version 0.97 21-Oct-97
3095 ----------------------
3096
3097 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
3098
3099 2. Optimized negated single characters not to use a bit map.
3100
3101 3. Brought error texts together as macro definitions; clarified some of them;
3102 fixed one that was wrong - it said "range out of order" when it meant "invalid
3103 escape sequence".
3104
3105 4. Changed some char * arguments to const char *.
3106
3107 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
3108
3109 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
3110 pcretest.
3111
3112
3113 Version 0.96 16-Oct-97
3114 ----------------------
3115
3116 1. Added a simple "pgrep" utility to the distribution.
3117
3118 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
3119 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
3120 where "ddd" means "one or more decimal digits".
3121
3122 3. Fixed serious bug. If a pattern had a back reference, but the call to
3123 pcre_exec() didn't supply a large enough ovector to record the related
3124 identifying subpattern, the match always failed. PCRE now remembers the number
3125 of the largest back reference, and gets some temporary memory in which to save
3126 the offsets during matching if necessary, in order to ensure that
3127 backreferences always work.
3128
3129 4. Increased the compatibility with Perl in a number of ways:
3130
3131 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
3132 to request this handling. The option can be set at compile or exec time.
3133
3134 (b) $ matches before a terminating newline by default; an option
3135 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
3136 mode). The option can be set at compile or exec time.
3137
3138 (c) The handling of \ followed by a digit other than 0 is now supposed to be
3139 the same as Perl's. If the decimal number it represents is less than 10
3140 or there aren't that many previous left capturing parentheses, an octal
3141 escape is read. Inside a character class, it's always an octal escape,
3142 even if it is a single digit.
3143
3144 (d) An escaped but undefined alphabetic character is taken as a literal,
3145 unless PCRE_EXTRA is set. Currently this just reserves the remaining
3146 escapes.
3147
3148 (e) {0} is now permitted. (The previous item is removed from the compiled
3149 pattern).
3150
3151 5. Changed all the names of code files so that the basic parts are no longer
3152 than 10 characters, and abolished the teeny "globals.c" file.
3153
3154 6. Changed the handling of character classes; they are now done with a 32-byte
3155 bit map always.
3156
3157 7. Added the -d and /D options to pcretest to make it possible to look at the
3158 internals of compilation without having to recompile pcre.
3159
3160
3161 Version 0.95 23-Sep-97
3162 ----------------------
3163
3164 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
3165 \x20 at the start of a run of normal characters. These were being treated as
3166 real characters, instead of the source characters being re-checked.
3167
3168
3169 Version 0.94 18-Sep-97
3170 ----------------------
3171
3172 1. The functions are now thread-safe, with the caveat that the global variables
3173 containing pointers to malloc() and free() or alternative functions are the
3174 same for all threads.
3175
3176 2. Get pcre_study() to generate a bitmap of initial characters for non-
3177 anchored patterns when this is possible, and use it if passed to pcre_exec().
3178
3179
3180 Version 0.93 15-Sep-97
3181 ----------------------
3182
3183 1. /(b)|(:+)/ was computing an incorrect first character.
3184
3185 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
3186 but not actually doing anything yet.
3187
3188 3. Treat "-" characters in classes that cannot be part of ranges as literals,
3189 as Perl does (e.g. [-az] or [az-]).
3190
3191 4. Set the anchored flag if a branch starts with .* or .*? because that tests
3192 all possible positions.
3193
3194 5. Split up into different modules to avoid including unneeded functions in a
3195 compiled binary. However, compile and exec are still in one module. The "study"
3196 function is split off.
3197
3198 6. The character tables are now in a separate module whose source is generated
3199 by an auxiliary program - but can then be edited by hand if required. There are
3200 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
3201 toupper() in the code.
3202
3203 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
3204 make them global. Abolish the function for setting them, as the caller can now
3205 set them directly.
3206
3207
3208 Version 0.92 11-Sep-97
3209 ----------------------
3210
3211 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
3212 (e.g. /a{1,3}/) was broken (I mis-optimized it).
3213
3214 2. Caseless matching was not working in character classes if the characters in
3215 the pattern were in upper case.
3216
3217 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
3218
3219 4. Make PCRE_ANCHORED public and accept as a compile option.
3220
3221 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
3222 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
3223 pass them.
3224
3225 6. Give an error if bad option bits passed at compile or run time.
3226
3227 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
3228 pcretest to cause it to pass that flag.
3229
3230 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
3231 options, and the first character, if set.
3232
3233 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
3234
3235
3236 Version 0.91 10-Sep-97
3237 ----------------------
3238
3239 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
3240 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
3241
3242 2. PCRE was looping on encountering an indefinitely repeated back reference to
3243 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
3244 Perl does - treats the match as successful.
3245
3246 ****

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