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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 356 - (show annotations) (download)
Tue Jul 8 14:14:34 2008 UTC (6 years, 5 months ago) by ph10
File size: 157437 byte(s)
Minor bugfix in pcrecpp.cc.

1 ChangeLog for PCRE
2 ------------------
3
4 Version 8.0 02 Jul-08
5 ---------------------
6
7 1. Replaced UCP searching code with optimized version as implemented for Ad
8 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
9 stage table and inline lookup instead of a function, giving speed ups of 2
10 to 5 times on some simple patterns that I tested. Permission was given to
11 distribute the MultiStage2.py script that generates the tables (it's not in
12 the tarball, but is in the Subversion repository).
13
14 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
15 scripts.
16
17 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
18 a group with a zero qualifier. The result of the study could be incorrect,
19 or the function might crash, depending on the pattern.
20
21 4. Caseless matching was not working for non-ASCII characters in back
22 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
23 It now works when Unicode Property Support is available.
24
25 5. In pcretest, an escape such as \x{de} in the data was always generating
26 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
27 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
28 truncation.
29
30 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
31
32
33 Version 7.7 07-May-08
34 ---------------------
35
36 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
37 a string to a long long, pretend we don't even have a long long." This is
38 done by checking for the strtoq, strtoll, and _strtoi64 functions.
39
40 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
41 pre-7.6 versions, which defined a global no_arg variable instead of putting
42 it in the RE class. (See also #8 below.)
43
44 3. Remove a line of dead code, identified by coverity and reported by Nuno
45 Lopes.
46
47 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
48
49 (1) The include/exclude patterns were being applied to the whole pathnames
50 of files, instead of just to the final components.
51
52 (2) If there was more than one level of directory, the subdirectories were
53 skipped unless they satisfied the include/exclude conditions. This is
54 inconsistent with GNU grep (and could even be seen as contrary to the
55 pcregrep specification - which I improved to make it absolutely clear).
56 The action now is always to scan all levels of directory, and just
57 apply the include/exclude patterns to regular files.
58
59 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
60 --exclude_dir in the tests to avoid scanning .svn directories.
61
62 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
63 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
64 doesn't support NULs in patterns.
65
66 7. Added some missing "const"s to declarations of static tables in
67 pcre_compile.c and pcre_dfa_exec.c.
68
69 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
70 caused by fix #2 above. (Subsequently also a second patch to fix the
71 first patch. And a third patch - this was a messy problem.)
72
73 9. Applied Craig's patch to remove the use of push_back().
74
75 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
76 matching function regexec().
77
78 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
79 which, however, unlike Perl's \g{...}, are subroutine calls, not back
80 references. PCRE supports relative numbers with this syntax (I don't think
81 Oniguruma does).
82
83 12. Previously, a group with a zero repeat such as (...){0} was completely
84 omitted from the compiled regex. However, this means that if the group
85 was called as a subroutine from elsewhere in the pattern, things went wrong
86 (an internal error was given). Such groups are now left in the compiled
87 pattern, with a new opcode that causes them to be skipped at execution
88 time.
89
90 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
91 to the way PCRE behaves:
92
93 (a) A lone ] character is dis-allowed (Perl treats it as data).
94
95 (b) A back reference to an unmatched subpattern matches an empty string
96 (Perl fails the current match path).
97
98 (c) A data ] in a character class must be notated as \] because if the
99 first data character in a class is ], it defines an empty class. (In
100 Perl it is not possible to have an empty class.) The empty class []
101 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
102 The negative empty class [^] matches any one character, independently
103 of the DOTALL setting.
104
105 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
106 non-existent subpattern following a character class starting with ']' and
107 containing () gave an internal compiling error instead of "reference to
108 non-existent subpattern". Fortunately, when the pattern did exist, the
109 compiled code was correct. (When scanning forwards to check for the
110 existencd of the subpattern, it was treating the data ']' as terminating
111 the class, so got the count wrong. When actually compiling, the reference
112 was subsequently set up correctly.)
113
114 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
115 it was being rejected as not supported by pcre_dfa_exec(), even though
116 other assertions are supported. I have made pcre_dfa_exec() support
117 (*FAIL).
118
119 16. The implementation of 13c above involved the invention of a new opcode,
120 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
121 cannot be changed at match time, I realized I could make a small
122 improvement to matching performance by compiling OP_ALLANY instead of
123 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
124 on the OP_ANY path.
125
126 17. Compiling pcretest on Windows with readline support failed without the
127 following two fixes: (1) Make the unistd.h include conditional on
128 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
129
130 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
131 ncurses library to be included for pcretest when ReadLine support is
132 requested, but also to allow for it to be overridden. This patch came from
133 Daniel Bergström.
134
135 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
136 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
137 any errors with the current Unicode tables. Thanks to Peter Kankowski for
138 spotting this.
139
140
141 Version 7.6 28-Jan-08
142 ---------------------
143
144 1. A character class containing a very large number of characters with
145 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
146 overflow.
147
148 2. Patch to cut out the "long long" test in pcrecpp_unittest when
149 HAVE_LONG_LONG is not defined.
150
151 3. Applied Christian Ehrlicher's patch to update the CMake build files to
152 bring them up to date and include new features. This patch includes:
153
154 - Fixed PH's badly added libz and libbz2 support.
155 - Fixed a problem with static linking.
156 - Added pcredemo. [But later removed - see 7 below.]
157 - Fixed dftables problem and added an option.
158 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
159 HAVE_LONG_LONG.
160 - Added readline support for pcretest.
161 - Added an listing of the option settings after cmake has run.
162
163 4. A user submitted a patch to Makefile that makes it easy to create
164 "pcre.dll" under mingw when using Configure/Make. I added stuff to
165 Makefile.am that cause it to include this special target, without
166 affecting anything else. Note that the same mingw target plus all
167 the other distribution libraries and programs are now supported
168 when configuring with CMake (see 6 below) instead of with
169 Configure/Make.
170
171 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
172 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
173 exported in the Windows port". It has not yet been confirmed that the patch
174 solves the problem, but it does no harm.
175
176 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
177 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
178 with CMake, and also correct the comment about stack recursion.
179
180 7. Remove the automatic building of pcredemo from the ./configure system and
181 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
182 of a program that users should build themselves after PCRE is installed, so
183 building it automatically is not really right. What is more, it gave
184 trouble in some build environments.
185
186 8. Further tidies to CMakeLists.txt from Sheri and Christian.
187
188
189 Version 7.5 10-Jan-08
190 ---------------------
191
192 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
193 values in parens when parsing an RE using the C++ wrapper."
194
195 2. Negative specials like \S did not work in character classes in UTF-8 mode.
196 Characters greater than 255 were excluded from the class instead of being
197 included.
198
199 3. The same bug as (2) above applied to negated POSIX classes such as
200 [:^space:].
201
202 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
203 defined or documented. It seems to have been a typo for PCRE_STATIC, so
204 I have changed it.
205
206 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
207 first named subpattern) and a construct such as (?&a) would reference the
208 first named subpattern whose name started with "a" (in other words, the
209 length check was missing). Both these problems are fixed. "Subpattern name
210 expected" is now given for (?&) (a zero-length name), and this patch also
211 makes it give the same error for \k'' (previously it complained that that
212 was a reference to a non-existent subpattern).
213
214 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
215 this is right because (?- can be followed by option settings as well as by
216 digits. I have, however, made the messages clearer.
217
218 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
219 than the number used in the conditional) now cause a compile-time error.
220 This is actually not compatible with Perl, which accepts such patterns, but
221 treats the conditional as always being FALSE (as PCRE used to), but it
222 seems to me that giving a diagnostic is better.
223
224 8. Change "alphameric" to the more common word "alphanumeric" in comments
225 and messages.
226
227 9. Fix two occurrences of "backslash" in comments that should have been
228 "backspace".
229
230 10. Remove two redundant lines of code that can never be obeyed (their function
231 was moved elsewhere).
232
233 11. The program that makes PCRE's Unicode character property table had a bug
234 which caused it to generate incorrect table entries for sequences of
235 characters that have the same character type, but are in different scripts.
236 It amalgamated them into a single range, with the script of the first of
237 them. In other words, some characters were in the wrong script. There were
238 thirteen such cases, affecting characters in the following ranges:
239
240 U+002b0 - U+002c1
241 U+0060c - U+0060d
242 U+0061e - U+00612
243 U+0064b - U+0065e
244 U+0074d - U+0076d
245 U+01800 - U+01805
246 U+01d00 - U+01d77
247 U+01d9b - U+01dbf
248 U+0200b - U+0200f
249 U+030fc - U+030fe
250 U+03260 - U+0327f
251 U+0fb46 - U+0fbb1
252 U+10450 - U+1049d
253
254 12. The -o option (show only the matching part of a line) for pcregrep was not
255 compatible with GNU grep in that, if there was more than one match in a
256 line, it showed only the first of them. It now behaves in the same way as
257 GNU grep.
258
259 13. If the -o and -v options were combined for pcregrep, it printed a blank
260 line for every non-matching line. GNU grep prints nothing, and pcregrep now
261 does the same. The return code can be used to tell if there were any
262 non-matching lines.
263
264 14. Added --file-offsets and --line-offsets to pcregrep.
265
266 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
267 infinitely looping recursion. The bug was that positive lookaheads were not
268 being skipped when checking for a possible empty match (negative lookaheads
269 and both kinds of lookbehind were skipped).
270
271 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
272 inclusion of <windows.h> to before rather than after the definition of
273 INVALID_FILE_ATTRIBUTES (patch from David Byron).
274
275 17. Specifying a possessive quantifier with a specific limit for a Unicode
276 character property caused pcre_compile() to compile bad code, which led at
277 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
278 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
279 caused the error; without that there was no problem.
280
281 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
282
283 19. Added --enable-pcretest-libreadline.
284
285 20. In pcrecpp.cc, the variable 'count' was incremented twice in
286 RE::GlobalReplace(). As a result, the number of replacements returned was
287 double what it should be. I removed one of the increments, but Craig sent a
288 later patch that removed the other one (the right fix) and added unit tests
289 that check the return values (which was not done before).
290
291 21. Several CMake things:
292
293 (1) Arranged that, when cmake is used on Unix, the libraries end up with
294 the names libpcre and libpcreposix, not just pcre and pcreposix.
295
296 (2) The above change means that pcretest and pcregrep are now correctly
297 linked with the newly-built libraries, not previously installed ones.
298
299 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
300
301 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
302 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
303 UTF-8 newline character). The key issue is that the pattern starts .*;
304 this means that the match must be either at the beginning, or after a
305 newline. The bug was in the code for advancing after a failed match and
306 checking that the new position followed a newline. It was not taking
307 account of UTF-8 characters correctly.
308
309 23. PCRE was behaving differently from Perl in the way it recognized POSIX
310 character classes. PCRE was not treating the sequence [:...:] as a
311 character class unless the ... were all letters. Perl, however, seems to
312 allow any characters between [: and :], though of course it rejects as
313 unknown any "names" that contain non-letters, because all the known class
314 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
315 for example, whereas PCRE did not - it did not recognize a POSIX character
316 class. This seemed a bit dangerous, so the code has been changed to be
317 closer to Perl. The behaviour is not identical to Perl, because PCRE will
318 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
319 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
320 Perl does, and where it didn't before.
321
322 24. Rewrite so as to remove the single use of %n from pcregrep because in some
323 Windows environments %n is disabled by default.
324
325
326 Version 7.4 21-Sep-07
327 ---------------------
328
329 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
330 means that a class such as [\s] counted as "explicit reference to CR or
331 LF". That isn't really right - the whole point of the change was to try to
332 help when there was an actual mention of one of the two characters. So now
333 the change happens only if \r or \n (or a literal CR or LF) character is
334 encountered.
335
336 2. The 32-bit options word was also used for 6 internal flags, but the numbers
337 of both had grown to the point where there were only 3 bits left.
338 Fortunately, there was spare space in the data structure, and so I have
339 moved the internal flags into a new 16-bit field to free up more option
340 bits.
341
342 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
343 but did not set the internal JCHANGED flag - either of these is enough to
344 control the way the "get" function works - but the PCRE_INFO_JCHANGED
345 facility is supposed to tell if (?J) was ever used, so now (?J) at the
346 start sets both bits.
347
348 4. Added options (at build time, compile time, exec time) to change \R from
349 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
350
351 5. doc/pcresyntax.html was missing from the distribution.
352
353 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
354 compatibility, even though it is no longer used.
355
356 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
357 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
358 windows.h file is present (where different names are used). [This was
359 reversed later after testing - see 16 below.]
360
361 8. Changed all #include <config.h> to #include "config.h". There were also
362 some further <pcre.h> cases that I changed to "pcre.h".
363
364 9. When pcregrep was used with the --colour option, it missed the line ending
365 sequence off the lines that it output.
366
367 10. It was pointed out to me that arrays of string pointers cause lots of
368 relocations when a shared library is dynamically loaded. A technique of
369 using a single long string with a table of offsets can drastically reduce
370 these. I have refactored PCRE in four places to do this. The result is
371 dramatic:
372
373 Originally: 290
374 After changing UCP table: 187
375 After changing error message table: 43
376 After changing table of "verbs" 36
377 After changing table of Posix names 22
378
379 Thanks to the folks working on Gregex for glib for this insight.
380
381 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
382 unicode-properties was also set.
383
384 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
385
386 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
387 checked only for CRLF.
388
389 14. Added casts to pcretest.c to avoid compiler warnings.
390
391 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
392
393 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
394 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
395 entirely. This removes changes made in 7 above.
396
397 17. The CMake files have been updated, and there is now more information about
398 building with CMake in the NON-UNIX-USE document.
399
400
401 Version 7.3 28-Aug-07
402 ---------------------
403
404 1. In the rejigging of the build system that eventually resulted in 7.1, the
405 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
406 brackets there is not right, since it causes compilers to look for an
407 installed pcre.h, not the version that is in the source that is being
408 compiled (which of course may be different). I have changed it back to:
409
410 #include "pcre.h"
411
412 I have a vague recollection that the change was concerned with compiling in
413 different directories, but in the new build system, that is taken care of
414 by the VPATH setting the Makefile.
415
416 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
417 when the subject happened to end in the byte 0x85 (e.g. if the last
418 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
419 characters but of course it shouldn't be taken as a newline when it is part
420 of another character. The bug was that, for an unlimited repeat of . in
421 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
422 characters when looking for a newline.
423
424 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
425
426 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
427 in debug output.
428
429 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
430 long printing in the pcrecpp unittest when running under MinGW.
431
432 6. ESC_K was left out of the EBCDIC table.
433
434 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
435 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
436 limit also applies to "virtual nesting" when a pattern is recursive, and in
437 this case 1000 isn't so big. I have been able to remove this limit at the
438 expense of backing off one optimization in certain circumstances. Normally,
439 when pcre_exec() would call its internal match() function recursively and
440 immediately return the result unconditionally, it uses a "tail recursion"
441 feature to save stack. However, when a subpattern that can match an empty
442 string has an unlimited repetition quantifier, it no longer makes this
443 optimization. That gives it a stack frame in which to save the data for
444 checking that an empty string has been matched. Previously this was taken
445 from the 1000-entry workspace that had been reserved. So now there is no
446 explicit limit, but more stack is used.
447
448 8. Applied Daniel's patches to solve problems with the import/export magic
449 syntax that is required for Windows, and which was going wrong for the
450 pcreposix and pcrecpp parts of the library. These were overlooked when this
451 problem was solved for the main library.
452
453 9. There were some crude static tests to avoid integer overflow when computing
454 the size of patterns that contain repeated groups with explicit upper
455 limits. As the maximum quantifier is 65535, the maximum group length was
456 set at 30,000 so that the product of these two numbers did not overflow a
457 32-bit integer. However, it turns out that people want to use groups that
458 are longer than 30,000 bytes (though not repeat them that many times).
459 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
460 made it possible to implement the integer overflow checks in a much more
461 dynamic way, which I have now done. The artificial limitation on group
462 length has been removed - we now have only the limit on the total length of
463 the compiled pattern, which depends on the LINK_SIZE setting.
464
465 10. Fixed a bug in the documentation for get/copy named substring when
466 duplicate names are permitted. If none of the named substrings are set, the
467 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
468 empty string.
469
470 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
471 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
472 because the ] is interpreted as the first data character and the
473 terminating ] is not found. PCRE has been made compatible with Perl in this
474 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
475 cause memory overwriting.
476
477 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
478 string has been matched (to stop an infinite loop). It was not recognizing
479 a conditional subpattern that could match an empty string if that
480 subpattern was within another subpattern. For example, it looped when
481 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
482 condition was not nested. This bug has been fixed.
483
484 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
485 past the start of the subject in the presence of bytes with the top bit
486 set, for example "\x8aBCD".
487
488 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
489 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
490
491 14. Optimized (?!) to (*FAIL).
492
493 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
494 This restricts code points to be within the range 0 to 0x10FFFF, excluding
495 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
496 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
497 does: it's just the validity check that is more restrictive.
498
499 16. Inserted checks for integer overflows during escape sequence (backslash)
500 processing, and also fixed erroneous offset values for syntax errors during
501 backslash processing.
502
503 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
504 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
505
506 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
507 caused an overrun.
508
509 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
510 something other than just ASCII characters) inside a group that had an
511 unlimited repeat caused a loop at compile time (while checking to see
512 whether the group could match an empty string).
513
514 20. Debugging a pattern containing \p or \P could cause a crash. For example,
515 [\P{Any}] did so. (Error in the code for printing property names.)
516
517 21. An orphan \E inside a character class could cause a crash.
518
519 22. A repeated capturing bracket such as (A)? could cause a wild memory
520 reference during compilation.
521
522 23. There are several functions in pcre_compile() that scan along a compiled
523 expression for various reasons (e.g. to see if it's fixed length for look
524 behind). There were bugs in these functions when a repeated \p or \P was
525 present in the pattern. These operators have additional parameters compared
526 with \d, etc, and these were not being taken into account when moving along
527 the compiled data. Specifically:
528
529 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
530 length.
531
532 (b) An item such as \pL+ within a repeated group could cause crashes or
533 loops.
534
535 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
536 "reference to non-existent subpattern" error.
537
538 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
539
540 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
541 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
542
543 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
544
545 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
546 character were causing crashes (broken optimization).
547
548 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
549 \p or \P) caused a compile-time loop.
550
551 28. More problems have arisen in unanchored patterns when CRLF is a valid line
552 break. For example, the unstudied pattern [\r\n]A does not match the string
553 "\r\nA" because change 7.0/46 below moves the current point on by two
554 characters after failing to match at the start. However, the pattern \nA
555 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
556 the same is true. There doesn't seem any very clean way out of this, but
557 what I have chosen to do makes the common cases work: PCRE now takes note
558 of whether there can be an explicit match for \r or \n anywhere in the
559 pattern, and if so, 7.0/46 no longer applies. As part of this change,
560 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
561 pattern has explicit CR or LF references.
562
563 29. Added (*CR) etc for changing newline setting at start of pattern.
564
565
566 Version 7.2 19-Jun-07
567 ---------------------
568
569 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
570 which is apparently normally available under Windows.
571
572 2. Re-jig the pcregrep tests with different newline settings in an attempt
573 to make them independent of the local environment's newline setting.
574
575 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
576
577 4. Some of the "internals" tests were previously cut out when the link size
578 was not 2, because the output contained actual offsets. The recent new
579 "Z" feature of pcretest means that these can be cut out, making the tests
580 usable with all link sizes.
581
582 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
583 stack recursion. This gives a massive performance boost under BSD, but just
584 a small improvement under Linux. However, it saves one field in the frame
585 in all cases.
586
587 6. Added more features from the forthcoming Perl 5.10:
588
589 (a) (?-n) (where n is a string of digits) is a relative subroutine or
590 recursion call. It refers to the nth most recently opened parentheses.
591
592 (b) (?+n) is also a relative subroutine call; it refers to the nth next
593 to be opened parentheses.
594
595 (c) Conditions that refer to capturing parentheses can be specified
596 relatively, for example, (?(-2)... or (?(+3)...
597
598 (d) \K resets the start of the current match so that everything before
599 is not part of it.
600
601 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
602
603 (f) \g{name} is another synonym - part of Perl 5.10's unification of
604 reference syntax.
605
606 (g) (?| introduces a group in which the numbering of parentheses in each
607 alternative starts with the same number.
608
609 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
610
611 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
612 PCRE_INFO_JCHANGED.
613
614 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
615 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
616 for detecting groups that can match an empty string.
617
618 9. A pattern with a very large number of alternatives (more than several
619 hundred) was running out of internal workspace during the pre-compile
620 phase, where pcre_compile() figures out how much memory will be needed. A
621 bit of new cunning has reduced the workspace needed for groups with
622 alternatives. The 1000-alternative test pattern now uses 12 bytes of
623 workspace instead of running out of the 4096 that are available.
624
625 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
626
627 11. Applied patch from Google to remove an optimization that didn't quite work.
628 The report of the bug said:
629
630 pcrecpp::RE("a*").FullMatch("aaa") matches, while
631 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
632 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
633
634 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
635 it matched the wrong number of bytes.
636
637
638 Version 7.1 24-Apr-07
639 ---------------------
640
641 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
642 that is more "standard", making use of automake and other Autotools. There
643 is some re-arrangement of the files and adjustment of comments consequent
644 on this.
645
646 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
647 for recursive directory scanning broke on some systems because the files
648 are not scanned in any specific order and on different systems the order
649 was different. A call to "sort" has been inserted into RunGrepTest for the
650 approprate test as a short-term fix. In the longer term there may be an
651 alternative.
652
653 3. I had an email from Eric Raymond about problems translating some of PCRE's
654 man pages to HTML (despite the fact that I distribute HTML pages, some
655 people do their own conversions for various reasons). The problems
656 concerned the use of low-level troff macros .br and .in. I have therefore
657 removed all such uses from the man pages (some were redundant, some could
658 be replaced by .nf/.fi pairs). The 132html script that I use to generate
659 HTML has been updated to handle .nf/.fi and to complain if it encounters
660 .br or .in.
661
662 4. Updated comments in configure.ac that get placed in config.h.in and also
663 arranged for config.h to be included in the distribution, with the name
664 config.h.generic, for the benefit of those who have to compile without
665 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
666
667 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
668 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
669 makevp.bat for the current PCRE, using the additional files
670 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
671
672 6. A Windows user reported a minor discrepancy with test 2, which turned out
673 to be caused by a trailing space on an input line that had got lost in his
674 copy. The trailing space was an accident, so I've just removed it.
675
676 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
677 that is needed.
678
679 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
680 as "const" (a) because they are and (b) because it helps the PHP
681 maintainers who have recently made a script to detect big data structures
682 in the php code that should be moved to the .rodata section. I remembered
683 to update Builducptable as well, so it won't revert if ucptable.h is ever
684 re-created.
685
686 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
687 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
688 order to be able to cut out the UTF-8 tables in the latter when UTF-8
689 support is not required. This saves 1.5-2K of code, which is important in
690 some applications.
691
692 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
693 so as not to refer to the tables, even though these functions will never be
694 called when UTF-8 support is disabled. Otherwise there are problems with a
695 shared library.
696
697 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
698
699 (a) It was defining its arguments as char * instead of void *.
700
701 (b) It was assuming that all moves were upwards in memory; this was true
702 a long time ago when I wrote it, but is no longer the case.
703
704 The emulated memove() is provided for those environments that have neither
705 memmove() nor bcopy(). I didn't think anyone used it these days, but that
706 is clearly not the case, as these two bugs were recently reported.
707
708 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
709 and Detrail to create the HTML documentation, the .txt form of the man
710 pages, and it removes trailing spaces from listed files. It also creates
711 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
712 case, it wraps all the #defines with #ifndefs. This script should be run
713 before "make dist".
714
715 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
716 with Unicode property support.
717
718 (a) For a maximizing quantifier, if the two different cases of the
719 character were of different lengths in their UTF-8 codings (there are
720 some cases like this - I found 11), and the matching function had to
721 back up over a mixture of the two cases, it incorrectly assumed they
722 were both the same length.
723
724 (b) When PCRE was configured to use the heap rather than the stack for
725 recursion during matching, it was not correctly preserving the data for
726 the other case of a UTF-8 character when checking ahead for a match
727 while processing a minimizing repeat. If the check also involved
728 matching a wide character, but failed, corruption could cause an
729 erroneous result when trying to check for a repeat of the original
730 character.
731
732 13. Some tidying changes to the testing mechanism:
733
734 (a) The RunTest script now detects the internal link size and whether there
735 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
736 values substituted by "configure". (The RunGrepTest script already did
737 this for UTF-8.) The configure.ac script no longer substitutes the
738 relevant variables.
739
740 (b) The debugging options /B and /D in pcretest show the compiled bytecode
741 with length and offset values. This means that the output is different
742 for different internal link sizes. Test 2 is skipped for link sizes
743 other than 2 because of this, bypassing the problem. Unfortunately,
744 there was also a test in test 3 (the locale tests) that used /B and
745 failed for link sizes other than 2. Rather than cut the whole test out,
746 I have added a new /Z option to pcretest that replaces the length and
747 offset values with spaces. This is now used to make test 3 independent
748 of link size. (Test 2 will be tidied up later.)
749
750 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
751 segmentation fault instead of returning the appropriate error message.
752
753 15. In multiline mode when the newline sequence was set to "any", the pattern
754 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
755 This doesn't seem right; it now treats the CRLF combination as the line
756 ending, and so does not match in that case. It's only a pattern such as ^$
757 that would hit this one: something like ^ABC$ would have failed after \r
758 and then tried again after \r\n.
759
760 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
761 in an attempt to make files that differ only in their line terminators
762 compare equal. This works on Linux.
763
764 17. Under certain error circumstances pcregrep might try to free random memory
765 as it exited. This is now fixed, thanks to valgrind.
766
767 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
768 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
769 was because its rules for how to advance for /g after matching an empty
770 string at the end of a line did not allow for this case. They now check for
771 it specially.
772
773 20. pcretest is supposed to handle patterns and data of any length, by
774 extending its buffers when necessary. It was getting this wrong when the
775 buffer for a data line had to be extended.
776
777 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
778 CRLF as a newline sequence.
779
780 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
781 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
782 I have nevertheless tidied it up.
783
784 23. Added some casts to kill warnings from HP-UX ia64 compiler.
785
786 24. Added a man page for pcre-config.
787
788
789 Version 7.0 19-Dec-06
790 ---------------------
791
792 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
793 moving to gcc 4.1.1.
794
795 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
796 sys/time.h, which is documented as needed for this function. It doesn't
797 seem to matter on Linux, but it showed up on some releases of OS X.
798
799 3. It seems that there are systems where bytes whose values are greater than
800 127 match isprint() in the "C" locale. The "C" locale should be the
801 default when a C program starts up. In most systems, only ASCII printing
802 characters match isprint(). This difference caused the output from pcretest
803 to vary, making some of the tests fail. I have changed pcretest so that:
804
805 (a) When it is outputting text in the compiled version of a pattern, bytes
806 other than 32-126 are always shown as hex escapes.
807
808 (b) When it is outputting text that is a matched part of a subject string,
809 it does the same, unless a different locale has been set for the match
810 (using the /L modifier). In this case, it uses isprint() to decide.
811
812 4. Fixed a major bug that caused incorrect computation of the amount of memory
813 required for a compiled pattern when options that changed within the
814 pattern affected the logic of the preliminary scan that determines the
815 length. The relevant options are -x, and -i in UTF-8 mode. The result was
816 that the computed length was too small. The symptoms of this bug were
817 either the PCRE error "internal error: code overflow" from pcre_compile(),
818 or a glibc crash with a message such as "pcretest: free(): invalid next
819 size (fast)". Examples of patterns that provoked this bug (shown in
820 pcretest format) are:
821
822 /(?-x: )/x
823 /(?x)(?-x: \s*#\s*)/
824 /((?i)[\x{c0}])/8
825 /(?i:[\x{c0}])/8
826
827 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
828 is now done differently.
829
830 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
831 wrapper classes; (b) implement a new function in the C++ scanner that is
832 more efficient than the old way of doing things because it avoids levels of
833 recursion in the regex matching; (c) add a paragraph to the documentation
834 for the FullMatch() function.
835
836 6. The escape sequence \n was being treated as whatever was defined as
837 "newline". Not only was this contrary to the documentation, which states
838 that \n is character 10 (hex 0A), but it also went horribly wrong when
839 "newline" was defined as CRLF. This has been fixed.
840
841 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
842 was being set to -1 for the "end of line" case (supposedly a value that no
843 character can have). Though this value is never used (the check for end of
844 line is "zero bytes in current character"), it caused compiler complaints.
845 I've changed it to 0xffffffff.
846
847 8. In pcre_version.c, the version string was being built by a sequence of
848 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
849 string (as it is for production releases) called a macro with an empty
850 argument. The C standard says the result of this is undefined. The gcc
851 compiler treats it as an empty string (which was what was wanted) but it is
852 reported that Visual C gives an error. The source has been hacked around to
853 avoid this problem.
854
855 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
856 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
857 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
858 of them did).
859
860 10. Originally, pcretest opened its input and output without "b"; then I was
861 told that "b" was needed in some environments, so it was added for release
862 5.0 to both the input and output. (It makes no difference on Unix-like
863 systems.) Later I was told that it is wrong for the input on Windows. I've
864 now abstracted the modes into two macros, to make it easier to fiddle with
865 them, and removed "b" from the input mode under Windows.
866
867 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
868
869 12. Added -help and --help to pcretest as an official way of being reminded
870 of the options.
871
872 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
873 and pcrecpp.cc because they annoy compilers at high warning levels.
874
875 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
876
877 15. Fixed an occurrence of == in configure.ac that should have been = (shell
878 scripts are not C programs :-) and which was not noticed because it works
879 on Linux.
880
881 16. pcretest is supposed to handle any length of pattern and data line (as one
882 line or as a continued sequence of lines) by extending its input buffer if
883 necessary. This feature was broken for very long pattern lines, leading to
884 a string of junk being passed to pcre_compile() if the pattern was longer
885 than about 50K.
886
887 17. I have done a major re-factoring of the way pcre_compile() computes the
888 amount of memory needed for a compiled pattern. Previously, there was code
889 that made a preliminary scan of the pattern in order to do this. That was
890 OK when PCRE was new, but as the facilities have expanded, it has become
891 harder and harder to keep it in step with the real compile phase, and there
892 have been a number of bugs (see for example, 4 above). I have now found a
893 cunning way of running the real compile function in a "fake" mode that
894 enables it to compute how much memory it would need, while actually only
895 ever using a few hundred bytes of working memory and without too many
896 tests of the mode. This should make future maintenance and development
897 easier. A side effect of this work is that the limit of 200 on the nesting
898 depth of parentheses has been removed (though this was never a serious
899 limitation, I suspect). However, there is a downside: pcre_compile() now
900 runs more slowly than before (30% or more, depending on the pattern). I
901 hope this isn't a big issue. There is no effect on runtime performance.
902
903 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
904 newline (only possible for the last line of a file) and it was a
905 pattern that set a locale (followed by /Lsomething), pcretest crashed.
906
907 19. Added additional timing features to pcretest. (1) The -tm option now times
908 matching only, not compiling. (2) Both -t and -tm can be followed, as a
909 separate command line item, by a number that specifies the number of
910 repeats to use when timing. The default is 50000; this gives better
911 precision, but takes uncomfortably long for very large patterns.
912
913 20. Extended pcre_study() to be more clever in cases where a branch of a
914 subpattern has no definite first character. For example, (a*|b*)[cd] would
915 previously give no result from pcre_study(). Now it recognizes that the
916 first character must be a, b, c, or d.
917
918 21. There was an incorrect error "recursive call could loop indefinitely" if
919 a subpattern (or the entire pattern) that was being tested for matching an
920 empty string contained only one non-empty item after a nested subpattern.
921 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
922 incorrectly, because the \d was being skipped in the check.
923
924 22. The pcretest program now has a new pattern option /B and a command line
925 option -b, which is equivalent to adding /B to every pattern. This causes
926 it to show the compiled bytecode, without the additional information that
927 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
928 is the same as /B/I).
929
930 23. A new optimization is now able automatically to treat some sequences such
931 as a*b as a*+b. More specifically, if something simple (such as a character
932 or a simple class like \d) has an unlimited quantifier, and is followed by
933 something that cannot possibly match the quantified thing, the quantifier
934 is automatically "possessified".
935
936 24. A recursive reference to a subpattern whose number was greater than 39
937 went wrong under certain circumstances in UTF-8 mode. This bug could also
938 have affected the operation of pcre_study().
939
940 25. Realized that a little bit of performance could be had by replacing
941 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
942
943 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
944
945 27. Possessive quantifiers such as a++ were previously implemented by turning
946 them into atomic groups such as ($>a+). Now they have their own opcodes,
947 which improves performance. This includes the automatically created ones
948 from 23 above.
949
950 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
951 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
952 the first matched character to be a colon. This applied both to named and
953 numbered groups.
954
955 29. The ucpinternal.h header file was missing its idempotency #ifdef.
956
957 30. I was sent a "project" file called libpcre.a.dev which I understand makes
958 building PCRE on Windows easier, so I have included it in the distribution.
959
960 31. There is now a check in pcretest against a ridiculously large number being
961 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
962 loop, the loop is abandoned.
963
964 32. Forward references to subpatterns in conditions such as (?(2)...) where
965 subpattern 2 is defined later cause pcre_compile() to search forwards in
966 the pattern for the relevant set of parentheses. This search went wrong
967 when there were unescaped parentheses in a character class, parentheses
968 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
969
970 33. "Subroutine" calls and backreferences were previously restricted to
971 referencing subpatterns earlier in the regex. This restriction has now
972 been removed.
973
974 34. Added a number of extra features that are going to be in Perl 5.10. On the
975 whole, these are just syntactic alternatives for features that PCRE had
976 previously implemented using the Python syntax or my own invention. The
977 other formats are all retained for compatibility.
978
979 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
980 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
981 also .NET compatible.
982
983 (b) A recursion or subroutine call to a named group can now be defined as
984 (?&name) as well as (?P>name).
985
986 (c) A backreference to a named group can now be defined as \k<name> or
987 \k'name' as well as (?P=name). The new forms, as well as being in Perl
988 5.10, are also .NET compatible.
989
990 (d) A conditional reference to a named group can now use the syntax
991 (?(<name>) or (?('name') as well as (?(name).
992
993 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
994 groups (named and numbered) that are never evaluated inline, but can be
995 called as "subroutines" from elsewhere. In effect, the DEFINE condition
996 is always false. There may be only one alternative in such a group.
997
998 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
999 as the simple (?(R). The condition is true only if the most recent
1000 recursion is that of the given number or name. It does not search out
1001 through the entire recursion stack.
1002
1003 (g) The escape \gN or \g{N} has been added, where N is a positive or
1004 negative number, specifying an absolute or relative reference.
1005
1006 35. Tidied to get rid of some further signed/unsigned compiler warnings and
1007 some "unreachable code" warnings.
1008
1009 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
1010 things, this adds five new scripts.
1011
1012 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
1013 There were also incompatibilities regarding the handling of \Q..\E inside
1014 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
1015 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
1016
1017 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
1018 matches an empty string, and forcibly breaks the loop. There were bugs in
1019 this code in non-simple cases. For a pattern such as ^(a()*)* matched
1020 against aaaa the result was just "a" rather than "aaaa", for example. Two
1021 separate and independent bugs (that affected different cases) have been
1022 fixed.
1023
1024 39. Refactored the code to abolish the use of different opcodes for small
1025 capturing bracket numbers. This is a tidy that I avoided doing when I
1026 removed the limit on the number of capturing brackets for 3.5 back in 2001.
1027 The new approach is not only tidier, it makes it possible to reduce the
1028 memory needed to fix the previous bug (38).
1029
1030 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
1031 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
1032 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
1033 mode.
1034
1035 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
1036 report.
1037
1038 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
1039 copy construction and assignment in the C++ wrapper.
1040
1041 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
1042 couple of bugs that could have given wrong results in the "--newline=crlf"
1043 case.
1044
1045 44. Added a number of casts and did some reorganization of signed/unsigned int
1046 variables following suggestions from Dair Grant. Also renamed the variable
1047 "this" as "item" because it is a C++ keyword.
1048
1049 45. Arranged for dftables to add
1050
1051 #include "pcre_internal.h"
1052
1053 to pcre_chartables.c because without it, gcc 4.x may remove the array
1054 definition from the final binary if PCRE is built into a static library and
1055 dead code stripping is activated.
1056
1057 46. For an unanchored pattern, if a match attempt fails at the start of a
1058 newline sequence, and the newline setting is CRLF or ANY, and the next two
1059 characters are CRLF, advance by two characters instead of one.
1060
1061
1062 Version 6.7 04-Jul-06
1063 ---------------------
1064
1065 1. In order to handle tests when input lines are enormously long, pcretest has
1066 been re-factored so that it automatically extends its buffers when
1067 necessary. The code is crude, but this _is_ just a test program. The
1068 default size has been increased from 32K to 50K.
1069
1070 2. The code in pcre_study() was using the value of the re argument before
1071 testing it for NULL. (Of course, in any sensible call of the function, it
1072 won't be NULL.)
1073
1074 3. The memmove() emulation function in pcre_internal.h, which is used on
1075 systems that lack both memmove() and bcopy() - that is, hardly ever -
1076 was missing a "static" storage class specifier.
1077
1078 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
1079 containing an extended class (one that cannot be represented by a bitmap
1080 because it contains high-valued characters or Unicode property items, e.g.
1081 [\pZ]). Almost always one would set UTF-8 mode when processing such a
1082 pattern, but PCRE should not loop if you do not (it no longer does).
1083 [Detail: two cases were found: (a) a repeated subpattern containing an
1084 extended class; (b) a recursive reference to a subpattern that followed a
1085 previous extended class. It wasn't skipping over the extended class
1086 correctly when UTF-8 mode was not set.]
1087
1088 5. A negated single-character class was not being recognized as fixed-length
1089 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
1090 compile error "lookbehind assertion is not fixed length".
1091
1092 6. The RunPerlTest auxiliary script was showing an unexpected difference
1093 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
1094 write a Perl script that can interpret lines of an input file either as
1095 byte characters or as UTF-8, which is what "perltest" was being required to
1096 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
1097 can't do is switch easily at run time between having the "use utf8;" pragma
1098 or not. In the end, I fudged it by using the RunPerlTest script to insert
1099 "use utf8;" explicitly for the UTF-8 tests.
1100
1101 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
1102 the end of the subject string, contrary to the documentation and to what
1103 Perl does. This was true of both matching functions. Now it matches only at
1104 the start of the subject and immediately after *internal* newlines.
1105
1106 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
1107 a pointer to an int instead of a pointer to an unsigned long int. This
1108 caused problems on 64-bit systems.
1109
1110 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
1111 instance of the 'standard' template library not being so standard".
1112
1113 10. There was no check on the number of named subpatterns nor the maximum
1114 length of a subpattern name. The product of these values is used to compute
1115 the size of the memory block for a compiled pattern. By supplying a very
1116 long subpattern name and a large number of named subpatterns, the size
1117 computation could be caused to overflow. This is now prevented by limiting
1118 the length of names to 32 characters, and the number of named subpatterns
1119 to 10,000.
1120
1121 11. Subpatterns that are repeated with specific counts have to be replicated in
1122 the compiled pattern. The size of memory for this was computed from the
1123 length of the subpattern and the repeat count. The latter is limited to
1124 65535, but there was no limit on the former, meaning that integer overflow
1125 could in principle occur. The compiled length of a repeated subpattern is
1126 now limited to 30,000 bytes in order to prevent this.
1127
1128 12. Added the optional facility to have named substrings with the same name.
1129
1130 13. Added the ability to use a named substring as a condition, using the
1131 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
1132 are numbers (not recommended). Forward references are permitted.
1133
1134 14. Added forward references in named backreferences (if you see what I mean).
1135
1136 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
1137 pattern could run off the end of the subject. For example, the pattern
1138 "(?s)(.{1,5})"8 did this with the subject "ab".
1139
1140 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
1141 PCRE_CASELESS was set when matching characters that were quantified with ?
1142 or *.
1143
1144 17. A character class other than a single negated character that had a minimum
1145 but no maximum quantifier - for example [ab]{6,} - was not handled
1146 correctly by pce_dfa_exec(). It would match only one character.
1147
1148 18. A valid (though odd) pattern that looked like a POSIX character
1149 class but used an invalid character after [ (for example [[,abc,]]) caused
1150 pcre_compile() to give the error "Failed: internal error: code overflow" or
1151 in some cases to crash with a glibc free() error. This could even happen if
1152 the pattern terminated after [[ but there just happened to be a sequence of
1153 letters, a binary zero, and a closing ] in the memory that followed.
1154
1155 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
1156 over the years. Originally (before any Unicode support), just the bottom 8
1157 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
1158 output from "man perlunicode" includes this:
1159
1160 The regular expression compiler produces polymorphic opcodes. That
1161 is, the pattern adapts to the data and automatically switches to
1162 the Unicode character scheme when presented with Unicode data--or
1163 instead uses a traditional byte scheme when presented with byte
1164 data.
1165
1166 Sadly, a wide octal escape does not cause a switch, and in a string with
1167 no other multibyte characters, these octal escapes are treated as before.
1168 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
1169 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
1170 Unicode string.
1171
1172 I have not perpetrated such confusion in PCRE. Up till now, it took just
1173 the bottom 8 bits, as in old Perl. I have now made octal escapes with
1174 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
1175 translate to the appropriate multibyte character.
1176
1177 29. Applied some refactoring to reduce the number of warnings from Microsoft
1178 and Borland compilers. This has included removing the fudge introduced
1179 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
1180 a warning about an unused variable.
1181
1182 21. PCRE has not included VT (character 0x0b) in the set of whitespace
1183 characters since release 4.0, because Perl (from release 5.004) does not.
1184 [Or at least, is documented not to: some releases seem to be in conflict
1185 with the documentation.] However, when a pattern was studied with
1186 pcre_study() and all its branches started with \s, PCRE still included VT
1187 as a possible starting character. Of course, this did no harm; it just
1188 caused an unnecessary match attempt.
1189
1190 22. Removed a now-redundant internal flag bit that recorded the fact that case
1191 dependency changed within the pattern. This was once needed for "required
1192 byte" processing, but is no longer used. This recovers a now-scarce options
1193 bit. Also moved the least significant internal flag bit to the most-
1194 significant bit of the word, which was not previously used (hangover from
1195 the days when it was an int rather than a uint) to free up another bit for
1196 the future.
1197
1198 23. Added support for CRLF line endings as well as CR and LF. As well as the
1199 default being selectable at build time, it can now be changed at runtime
1200 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
1201 specify that it is scanning data with non-default line endings.
1202
1203 24. Changed the definition of CXXLINK to make it agree with the definition of
1204 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
1205
1206 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
1207 recursions. This makes a big different to stack usage for some patterns.
1208
1209 26. If a subpattern containing a named recursion or subroutine reference such
1210 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
1211 the space required for the compiled pattern went wrong and gave too small a
1212 value. Depending on the environment, this could lead to "Failed: internal
1213 error: code overflow at offset 49" or "glibc detected double free or
1214 corruption" errors.
1215
1216 27. Applied patches from Google (a) to support the new newline modes and (b) to
1217 advance over multibyte UTF-8 characters in GlobalReplace.
1218
1219 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
1220 difference for some implementation of PCRE in some Windows version.
1221
1222 29. Added some extra testing facilities to pcretest:
1223
1224 \q<number> in a data line sets the "match limit" value
1225 \Q<number> in a data line sets the "match recursion limt" value
1226 -S <number> sets the stack size, where <number> is in megabytes
1227
1228 The -S option isn't available for Windows.
1229
1230
1231 Version 6.6 06-Feb-06
1232 ---------------------
1233
1234 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
1235 in pcreposix.h. I have copied the definition from pcre.h.
1236
1237 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
1238 because pcre.h is no longer a built file.
1239
1240 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
1241 not normally included in the compiled code.
1242
1243
1244 Version 6.5 01-Feb-06
1245 ---------------------
1246
1247 1. When using the partial match feature with pcre_dfa_exec(), it was not
1248 anchoring the second and subsequent partial matches at the new starting
1249 point. This could lead to incorrect results. For example, with the pattern
1250 /1234/, partially matching against "123" and then "a4" gave a match.
1251
1252 2. Changes to pcregrep:
1253
1254 (a) All non-match returns from pcre_exec() were being treated as failures
1255 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
1256 error message is output. Some extra information is given for the
1257 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
1258 probably the only errors that are likely to be caused by users (by
1259 specifying a regex that has nested indefinite repeats, for instance).
1260 If there are more than 20 of these errors, pcregrep is abandoned.
1261
1262 (b) A binary zero was treated as data while matching, but terminated the
1263 output line if it was written out. This has been fixed: binary zeroes
1264 are now no different to any other data bytes.
1265
1266 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
1267 used to set a locale for matching. The --locale=xxxx long option has
1268 been added (no short equivalent) to specify a locale explicitly on the
1269 pcregrep command, overriding the environment variables.
1270
1271 (d) When -B was used with -n, some line numbers in the output were one less
1272 than they should have been.
1273
1274 (e) Added the -o (--only-matching) option.
1275
1276 (f) If -A or -C was used with -c (count only), some lines of context were
1277 accidentally printed for the final match.
1278
1279 (g) Added the -H (--with-filename) option.
1280
1281 (h) The combination of options -rh failed to suppress file names for files
1282 that were found from directory arguments.
1283
1284 (i) Added the -D (--devices) and -d (--directories) options.
1285
1286 (j) Added the -F (--fixed-strings) option.
1287
1288 (k) Allow "-" to be used as a file name for -f as well as for a data file.
1289
1290 (l) Added the --colo(u)r option.
1291
1292 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
1293 is not present by default.
1294
1295 3. A nasty bug was discovered in the handling of recursive patterns, that is,
1296 items such as (?R) or (?1), when the recursion could match a number of
1297 alternatives. If it matched one of the alternatives, but subsequently,
1298 outside the recursion, there was a failure, the code tried to back up into
1299 the recursion. However, because of the way PCRE is implemented, this is not
1300 possible, and the result was an incorrect result from the match.
1301
1302 In order to prevent this happening, the specification of recursion has
1303 been changed so that all such subpatterns are automatically treated as
1304 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
1305
1306 4. I had overlooked the fact that, in some locales, there are characters for
1307 which isalpha() is true but neither isupper() nor islower() are true. In
1308 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
1309 and ordfeminine) are like this. This affected the treatment of \w and \W
1310 when they appeared in character classes, but not when they appeared outside
1311 a character class. The bit map for "word" characters is now created
1312 separately from the results of isalnum() instead of just taking it from the
1313 upper, lower, and digit maps. (Plus the underscore character, of course.)
1314
1315 5. The above bug also affected the handling of POSIX character classes such as
1316 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
1317 permanent tables. Instead, the bit maps for such a class were previously
1318 created as the appropriate unions of the upper, lower, and digit bitmaps.
1319 Now they are created by subtraction from the [[:word:]] class, which has
1320 its own bitmap.
1321
1322 6. The [[:blank:]] character class matches horizontal, but not vertical space.
1323 It is created by subtracting the vertical space characters (\x09, \x0a,
1324 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
1325 subtraction was done in the overall bitmap for a character class, meaning
1326 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
1327 be recognized. This bug has been fixed.
1328
1329 7. Patches from the folks at Google:
1330
1331 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
1332 real life, but is still worth protecting against".
1333
1334 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
1335 regular expressions".
1336
1337 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
1338 have it.
1339
1340 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
1341 "configure" and the latter not, in order to fix a problem somebody had
1342 with compiling the Arg class on HP-UX.
1343
1344 (e) Improve the error-handling of the C++ wrapper a little bit.
1345
1346 (f) New tests for checking recursion limiting.
1347
1348 8. The pcre_memmove() function, which is used only if the environment does not
1349 have a standard memmove() function (and is therefore rarely compiled),
1350 contained two bugs: (a) use of int instead of size_t, and (b) it was not
1351 returning a result (though PCRE never actually uses the result).
1352
1353 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
1354 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
1355 returned instead of calling malloc() with an overflowing number that would
1356 most likely cause subsequent chaos.
1357
1358 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
1359
1360 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
1361 with this option is matched, the nmatch and pmatch options of regexec() are
1362 ignored.
1363
1364 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
1365 provided in case anyone wants to the the POSIX interface with UTF-8
1366 strings.
1367
1368 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
1369 C++ linking (needed for some HP-UX environments).
1370
1371 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
1372 (unused parameter) and in the pcre_printint() function (omitted "default"
1373 switch label when the default is to do nothing).
1374
1375 15. Added some code to make it possible, when PCRE is compiled as a C++
1376 library, to replace subject pointers for pcre_exec() with a smart pointer
1377 class, thus making it possible to process discontinuous strings.
1378
1379 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
1380 much the same function. They were added by different people who were trying
1381 to make PCRE easy to compile on non-Unix systems. It has been suggested
1382 that PCRE_EXPORT be abolished now that there is more automatic apparatus
1383 for compiling on Windows systems. I have therefore replaced it with
1384 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
1385 defaults to "extern" for C or "extern C" for C++, which works fine on
1386 Unix-like systems. It is now possible to override the value of PCRE_DATA_
1387 SCOPE with something explicit in config.h. In addition:
1388
1389 (a) pcreposix.h still had just "extern" instead of either of these macros;
1390 I have replaced it with PCRE_DATA_SCOPE.
1391
1392 (b) Functions such as _pcre_xclass(), which are internal to the library,
1393 but external in the C sense, all had PCRE_EXPORT in their definitions.
1394 This is apparently wrong for the Windows case, so I have removed it.
1395 (It makes no difference on Unix-like systems.)
1396
1397 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
1398 of recursive calls to match(). This is different to MATCH_LIMIT because
1399 that limits the total number of calls to match(), not all of which increase
1400 the depth of recursion. Limiting the recursion depth limits the amount of
1401 stack (or heap if NO_RECURSE is set) that is used. The default can be set
1402 when PCRE is compiled, and changed at run time. A patch from Google adds
1403 this functionality to the C++ interface.
1404
1405 18. Changes to the handling of Unicode character properties:
1406
1407 (a) Updated the table to Unicode 4.1.0.
1408
1409 (b) Recognize characters that are not in the table as "Cn" (undefined).
1410
1411 (c) I revised the way the table is implemented to a much improved format
1412 which includes recognition of ranges. It now supports the ranges that
1413 are defined in UnicodeData.txt, and it also amalgamates other
1414 characters into ranges. This has reduced the number of entries in the
1415 table from around 16,000 to around 3,000, thus reducing its size
1416 considerably. I realized I did not need to use a tree structure after
1417 all - a binary chop search is just as efficient. Having reduced the
1418 number of entries, I extended their size from 6 bytes to 8 bytes to
1419 allow for more data.
1420
1421 (d) Added support for Unicode script names via properties such as \p{Han}.
1422
1423 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
1424 matching that character.
1425
1426 20. When matching a repeated Unicode property with a minimum greater than zero,
1427 (for example \pL{2,}), PCRE could look past the end of the subject if it
1428 reached it while seeking the minimum number of characters. This could
1429 happen only if some of the characters were more than one byte long, because
1430 there is a check for at least the minimum number of bytes.
1431
1432 21. Refactored the implementation of \p and \P so as to be more general, to
1433 allow for more different types of property in future. This has changed the
1434 compiled form incompatibly. Anybody with saved compiled patterns that use
1435 \p or \P will have to recompile them.
1436
1437 22. Added "Any" and "L&" to the supported property types.
1438
1439 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
1440 but give a compile time error if the value is greater than 0xff.
1441
1442 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
1443 accidentally not being installed or uninstalled.
1444
1445 25. The pcre.h file was built from pcre.h.in, but the only changes that were
1446 made were to insert the current release number. This seemed silly, because
1447 it made things harder for people building PCRE on systems that don't run
1448 "configure". I have turned pcre.h into a distributed file, no longer built
1449 by "configure", with the version identification directly included. There is
1450 no longer a pcre.h.in file.
1451
1452 However, this change necessitated a change to the pcre-config script as
1453 well. It is built from pcre-config.in, and one of the substitutions was the
1454 release number. I have updated configure.ac so that ./configure now finds
1455 the release number by grepping pcre.h.
1456
1457 26. Added the ability to run the tests under valgrind.
1458
1459
1460 Version 6.4 05-Sep-05
1461 ---------------------
1462
1463 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
1464 "--" to be printed when multiple files were scanned, even when none of the
1465 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
1466 consider it to be a bug, and have restored the previous behaviour.
1467
1468 2. A couple of code tidies to get rid of compiler warnings.
1469
1470 3. The pcretest program used to cheat by referring to symbols in the library
1471 whose names begin with _pcre_. These are internal symbols that are not
1472 really supposed to be visible externally, and in some environments it is
1473 possible to suppress them. The cheating is now confined to including
1474 certain files from the library's source, which is a bit cleaner.
1475
1476 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
1477 file's purpose clearer.
1478
1479 5. Reorganized pcre_ucp_findchar().
1480
1481
1482 Version 6.3 15-Aug-05
1483 ---------------------
1484
1485 1. The file libpcre.pc.in did not have general read permission in the tarball.
1486
1487 2. There were some problems when building without C++ support:
1488
1489 (a) If C++ support was not built, "make install" and "make test" still
1490 tried to test it.
1491
1492 (b) There were problems when the value of CXX was explicitly set. Some
1493 changes have been made to try to fix these, and ...
1494
1495 (c) --disable-cpp can now be used to explicitly disable C++ support.
1496
1497 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
1498 backslash in a target when C++ was disabled. This confuses some
1499 versions of "make", apparently. Using an intermediate variable solves
1500 this. (Same for CPP_LOBJ.)
1501
1502 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
1503 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
1504 necessary on certain architectures.
1505
1506 4. Added a setting of -export-symbols-regex to the link command to remove
1507 those symbols that are exported in the C sense, but actually are local
1508 within the library, and not documented. Their names all begin with
1509 "_pcre_". This is not a perfect job, because (a) we have to except some
1510 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
1511 available (and never for static libraries). I have made a note to try to
1512 find a way round (a) in the future.
1513
1514
1515 Version 6.2 01-Aug-05
1516 ---------------------
1517
1518 1. There was no test for integer overflow of quantifier values. A construction
1519 such as {1111111111111111} would give undefined results. What is worse, if
1520 a minimum quantifier for a parenthesized subpattern overflowed and became
1521 negative, the calculation of the memory size went wrong. This could have
1522 led to memory overwriting.
1523
1524 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
1525
1526 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
1527 operating environments where this matters.
1528
1529 4. Applied Giuseppe Maxia's patch to add additional features for controlling
1530 PCRE options from within the C++ wrapper.
1531
1532 5. Named capturing subpatterns were not being correctly counted when a pattern
1533 was compiled. This caused two problems: (a) If there were more than 100
1534 such subpatterns, the calculation of the memory needed for the whole
1535 compiled pattern went wrong, leading to an overflow error. (b) Numerical
1536 back references of the form \12, where the number was greater than 9, were
1537 not recognized as back references, even though there were sufficient
1538 previous subpatterns.
1539
1540 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
1541 versions of gcc, e.g. 2.95.4.
1542
1543
1544 Version 6.1 21-Jun-05
1545 ---------------------
1546
1547 1. There was one reference to the variable "posix" in pcretest.c that was not
1548 surrounded by "#if !defined NOPOSIX".
1549
1550 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
1551 the cross-check on the old pcre_info() function, for the benefit of the
1552 cut-down version of PCRE that is currently imported into Exim.
1553
1554 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
1555 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
1556 patterns that start (?i) but I don't think that matters. The use of (?i) is
1557 just an example; this all applies to the other options as well.
1558
1559 4. Since libtool seems to echo the compile commands it is issuing, the output
1560 from "make" can be reduced a bit by putting "@" in front of each libtool
1561 compile command.
1562
1563 5. Patch from the folks at Google for configure.in to be a bit more thorough
1564 in checking for a suitable C++ installation before trying to compile the
1565 C++ stuff. This should fix a reported problem when a compiler was present,
1566 but no suitable headers.
1567
1568 6. The man pages all had just "PCRE" as their title. I have changed them to
1569 be the relevant file name. I have also arranged that these names are
1570 retained in the file doc/pcre.txt, which is a concatenation in text format
1571 of all the man pages except the little individual ones for each function.
1572
1573 7. The NON-UNIX-USE file had not been updated for the different set of source
1574 files that come with release 6. I also added a few comments about the C++
1575 wrapper.
1576
1577
1578 Version 6.0 07-Jun-05
1579 ---------------------
1580
1581 1. Some minor internal re-organization to help with my DFA experiments.
1582
1583 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
1584 didn't matter for the library itself when fully configured, but did matter
1585 when compiling without UCP support, or within Exim, where the ucp files are
1586 not imported.
1587
1588 3. Refactoring of the library code to split up the various functions into
1589 different source modules. The addition of the new DFA matching code (see
1590 below) to a single monolithic source would have made it really too
1591 unwieldy, quite apart from causing all the code to be include in a
1592 statically linked application, when only some functions are used. This is
1593 relevant even without the DFA addition now that patterns can be compiled in
1594 one application and matched in another.
1595
1596 The downside of splitting up is that there have to be some external
1597 functions and data tables that are used internally in different modules of
1598 the library but which are not part of the API. These have all had their
1599 names changed to start with "_pcre_" so that they are unlikely to clash
1600 with other external names.
1601
1602 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
1603 a different (DFA) algorithm. Although it is slower than the original
1604 function, it does have some advantages for certain types of matching
1605 problem.
1606
1607 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
1608 including restarting after a partial match.
1609
1610 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
1611 defined when compiling for Windows was sent to me. I have put it into the
1612 code, though I have no means of testing or verifying it.
1613
1614 7. Added the pcre_refcount() auxiliary function.
1615
1616 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
1617 match before or at the first newline in the subject string. In pcretest,
1618 the /f option on a pattern can be used to set this.
1619
1620 9. A repeated \w when used in UTF-8 mode with characters greater than 256
1621 would behave wrongly. This has been present in PCRE since release 4.0.
1622
1623 10. A number of changes to the pcregrep command:
1624
1625 (a) Refactored how -x works; insert ^(...)$ instead of setting
1626 PCRE_ANCHORED and checking the length, in preparation for adding
1627 something similar for -w.
1628
1629 (b) Added the -w (match as a word) option.
1630
1631 (c) Refactored the way lines are read and buffered so as to have more
1632 than one at a time available.
1633
1634 (d) Implemented a pcregrep test script.
1635
1636 (e) Added the -M (multiline match) option. This allows patterns to match
1637 over several lines of the subject. The buffering ensures that at least
1638 8K, or the rest of the document (whichever is the shorter) is available
1639 for matching (and similarly the previous 8K for lookbehind assertions).
1640
1641 (f) Changed the --help output so that it now says
1642
1643 -w, --word-regex(p)
1644
1645 instead of two lines, one with "regex" and the other with "regexp"
1646 because that confused at least one person since the short forms are the
1647 same. (This required a bit of code, as the output is generated
1648 automatically from a table. It wasn't just a text change.)
1649
1650 (g) -- can be used to terminate pcregrep options if the next thing isn't an
1651 option but starts with a hyphen. Could be a pattern or a path name
1652 starting with a hyphen, for instance.
1653
1654 (h) "-" can be given as a file name to represent stdin.
1655
1656 (i) When file names are being printed, "(standard input)" is used for
1657 the standard input, for compatibility with GNU grep. Previously
1658 "<stdin>" was used.
1659
1660 (j) The option --label=xxx can be used to supply a name to be used for
1661 stdin when file names are being printed. There is no short form.
1662
1663 (k) Re-factored the options decoding logic because we are going to add
1664 two more options that take data. Such options can now be given in four
1665 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
1666
1667 (l) Added the -A, -B, and -C options for requesting that lines of context
1668 around matches be printed.
1669
1670 (m) Added the -L option to print the names of files that do not contain
1671 any matching lines, that is, the complement of -l.
1672
1673 (n) The return code is 2 if any file cannot be opened, but pcregrep does
1674 continue to scan other files.
1675
1676 (o) The -s option was incorrectly implemented. For compatibility with other
1677 greps, it now suppresses the error message for a non-existent or non-
1678 accessible file (but not the return code). There is a new option called
1679 -q that suppresses the output of matching lines, which was what -s was
1680 previously doing.
1681
1682 (p) Added --include and --exclude options to specify files for inclusion
1683 and exclusion when recursing.
1684
1685 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
1686 Hopefully, it now does.
1687
1688 12. Missing cast in pcre_study().
1689
1690 13. Added an "uninstall" target to the makefile.
1691
1692 14. Replaced "extern" in the function prototypes in Makefile.in with
1693 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
1694 world, but is set differently for Windows.
1695
1696 15. Added a second compiling function called pcre_compile2(). The only
1697 difference is that it has an extra argument, which is a pointer to an
1698 integer error code. When there is a compile-time failure, this is set
1699 non-zero, in addition to the error test pointer being set to point to an
1700 error message. The new argument may be NULL if no error number is required
1701 (but then you may as well call pcre_compile(), which is now just a
1702 wrapper). This facility is provided because some applications need a
1703 numeric error indication, but it has also enabled me to tidy up the way
1704 compile-time errors are handled in the POSIX wrapper.
1705
1706 16. Added VPATH=.libs to the makefile; this should help when building with one
1707 prefix path and installing with another. (Or so I'm told by someone who
1708 knows more about this stuff than I do.)
1709
1710 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
1711 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
1712 match everything, including newlines. This is not POSIX-compatible, but
1713 somebody wanted the feature. From pcretest it can be activated by using
1714 both the P and the s flags.
1715
1716 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
1717
1718 19. libpcre.pc was being incorrectly installed as executable.
1719
1720 20. A couple of places in pcretest check for end-of-line by looking for '\n';
1721 it now also looks for '\r' so that it will work unmodified on Windows.
1722
1723 21. Added Google's contributed C++ wrapper to the distribution.
1724
1725 22. Added some untidy missing memory free() calls in pcretest, to keep
1726 Electric Fence happy when testing.
1727
1728
1729
1730 Version 5.0 13-Sep-04
1731 ---------------------
1732
1733 1. Internal change: literal characters are no longer packed up into items
1734 containing multiple characters in a single byte-string. Each character
1735 is now matched using a separate opcode. However, there may be more than one
1736 byte in the character in UTF-8 mode.
1737
1738 2. The pcre_callout_block structure has two new fields: pattern_position and
1739 next_item_length. These contain the offset in the pattern to the next match
1740 item, and its length, respectively.
1741
1742 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
1743 insertion of callouts before each pattern item. Added the /C option to
1744 pcretest to make use of this.
1745
1746 4. On the advice of a Windows user, the lines
1747
1748 #if defined(_WIN32) || defined(WIN32)
1749 _setmode( _fileno( stdout ), 0x8000 );
1750 #endif /* defined(_WIN32) || defined(WIN32) */
1751
1752 have been added to the source of pcretest. This apparently does useful
1753 magic in relation to line terminators.
1754
1755 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
1756 for the benefit of those environments where the "b" makes a difference.
1757
1758 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
1759 to know about it. I have put a hack into configure.in that adds in code
1760 to set GCC=yes if CC=icc. This seems to end up at a point in the
1761 generated configure script that is early enough to affect the setting of
1762 compiler options, which is what is needed, but I have no means of testing
1763 whether it really works. (The user who reported this had patched the
1764 generated configure script, which of course I cannot do.)
1765
1766 LATER: After change 22 below (new libtool files), the configure script
1767 seems to know about icc (and also ecc). Therefore, I have commented out
1768 this hack in configure.in.
1769
1770 7. Added support for pkg-config (2 patches were sent in).
1771
1772 8. Negated POSIX character classes that used a combination of internal tables
1773 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
1774 [[:^ascii]]. Typically, they would match almost any characters. The other
1775 POSIX classes were not broken in this way.
1776
1777 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
1778 to find the match, as PCRE was deluded into thinking that the match had to
1779 start at the start point or following a newline. The same bug applied to
1780 patterns with negative forward assertions or any backward assertions
1781 preceding ".*" at the start, unless the pattern required a fixed first
1782 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
1783
1784 10. In UTF-8 mode, when moving forwards in the subject after a failed match
1785 starting at the last subject character, bytes beyond the end of the subject
1786 string were read.
1787
1788 11. Renamed the variable "class" as "classbits" to make life easier for C++
1789 users. (Previously there was a macro definition, but it apparently wasn't
1790 enough.)
1791
1792 12. Added the new field "tables" to the extra data so that tables can be passed
1793 in at exec time, or the internal tables can be re-selected. This allows
1794 a compiled regex to be saved and re-used at a later time by a different
1795 program that might have everything at different addresses.
1796
1797 13. Modified the pcre-config script so that, when run on Solaris, it shows a
1798 -R library as well as a -L library.
1799
1800 14. The debugging options of pcretest (-d on the command line or D on a
1801 pattern) showed incorrect output for anything following an extended class
1802 that contained multibyte characters and which was followed by a quantifier.
1803
1804 15. Added optional support for general category Unicode character properties
1805 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
1806 support. It adds about 90K to the size of the library. The meanings of the
1807 inbuilt class escapes such as \d and \s have NOT been changed.
1808
1809 16. Updated pcredemo.c to include calls to free() to release the memory for the
1810 compiled pattern.
1811
1812 17. The generated file chartables.c was being created in the source directory
1813 instead of in the building directory. This caused the build to fail if the
1814 source directory was different from the building directory, and was
1815 read-only.
1816
1817 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
1818 file. No doubt somebody will tell me if they don't make sense... Also added
1819 Dan Mooney's comments about building on OpenVMS.
1820
1821 19. Added support for partial matching via the PCRE_PARTIAL option for
1822 pcre_exec() and the \P data escape in pcretest.
1823
1824 20. Extended pcretest with 3 new pattern features:
1825
1826 (i) A pattern option of the form ">rest-of-line" causes pcretest to
1827 write the compiled pattern to the file whose name is "rest-of-line".
1828 This is a straight binary dump of the data, with the saved pointer to
1829 the character tables forced to be NULL. The study data, if any, is
1830 written too. After writing, pcretest reads a new pattern.
1831
1832 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
1833 compiled pattern from the given file. There must not be any
1834 occurrences of "<" in the file name (pretty unlikely); if there are,
1835 pcretest will instead treat the initial "<" as a pattern delimiter.
1836 After reading in the pattern, pcretest goes on to read data lines as
1837 usual.
1838
1839 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
1840 and 16-bit fields in a compiled pattern, to simulate a pattern that
1841 was compiled on a host of opposite endianness.
1842
1843 21. The pcre-exec() function can now cope with patterns that were compiled on
1844 hosts of opposite endianness, with this restriction:
1845
1846 As for any compiled expression that is saved and used later, the tables
1847 pointer field cannot be preserved; the extra_data field in the arguments
1848 to pcre_exec() should be used to pass in a tables address if a value
1849 other than the default internal tables were used at compile time.
1850
1851 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
1852 now diagnosed as an error. Previously, most of the time, a negative number
1853 would have been treated as zero, but if in addition "ovector" was passed as
1854 NULL, a crash could occur.
1855
1856 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
1857 new versions from the libtool 1.5 distribution (the last one is a copy of
1858 a file called libtool.m4). This seems to have fixed the need to patch
1859 "configure" to support Darwin 1.3 (which I used to do). However, I still
1860 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
1861 workstation).
1862
1863 24. Changed the PCRE licence to be the more standard "BSD" licence.
1864
1865
1866 Version 4.5 01-Dec-03
1867 ---------------------
1868
1869 1. There has been some re-arrangement of the code for the match() function so
1870 that it can be compiled in a version that does not call itself recursively.
1871 Instead, it keeps those local variables that need separate instances for
1872 each "recursion" in a frame on the heap, and gets/frees frames whenever it
1873 needs to "recurse". Keeping track of where control must go is done by means
1874 of setjmp/longjmp. The whole thing is implemented by a set of macros that
1875 hide most of the details from the main code, and operates only if
1876 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
1877 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
1878 operating.
1879
1880 To make it easier for callers to provide specially tailored get/free
1881 functions for this usage, two new functions, pcre_stack_malloc, and
1882 pcre_stack_free, are used. They are always called in strict stacking order,
1883 and the size of block requested is always the same.
1884
1885 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
1886 PCRE has been compiled to use the stack or the heap for recursion. The
1887 -C option of pcretest uses this to show which version is compiled.
1888
1889 A new data escape \S, is added to pcretest; it causes the amounts of store
1890 obtained and freed by both kinds of malloc/free at match time to be added
1891 to the output.
1892
1893 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
1894 what's available on my current Linux desktop machine.
1895
1896 3. When matching a UTF-8 string, the test for a valid string at the start has
1897 been extended. If start_offset is not zero, PCRE now checks that it points
1898 to a byte that is the start of a UTF-8 character. If not, it returns
1899 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
1900 this is necessary because there may be backward assertions in the pattern.
1901 When matching the same subject several times, it may save resources to use
1902 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
1903
1904 4. The code for checking the validity of UTF-8 strings has been tightened so
1905 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
1906 containing "overlong sequences".
1907
1908 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
1909 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
1910 should have been "&", but it just so happened that all the cases this let
1911 through by mistake were picked up later in the function.
1912
1913 6. I had used a variable called "isblank" - this is a C99 function, causing
1914 some compilers to warn. To avoid this, I renamed it (as "blankclass").
1915
1916 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
1917 prompting; (b) run "./pcretest /dev/null" at the start of the test script
1918 so the version is shown; (c) stop "make test" echoing "./RunTest".
1919
1920 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
1921
1922 9. The prototype for memmove() for systems that don't have it was using
1923 size_t, but the inclusion of the header that defines size_t was later. I've
1924 moved the #includes for the C headers earlier to avoid this.
1925
1926 10. Added some adjustments to the code to make it easier to compiler on certain
1927 special systems:
1928
1929 (a) Some "const" qualifiers were missing.
1930 (b) Added the macro EXPORT before all exported functions; by default this
1931 is defined to be empty.
1932 (c) Changed the dftables auxiliary program (that builds chartables.c) so
1933 that it reads its output file name as an argument instead of writing
1934 to the standard output and assuming this can be redirected.
1935
1936 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
1937 class containing characters with values greater than 255, PCRE compilation
1938 went into a loop.
1939
1940 12. A recursive reference to a subpattern that was within another subpattern
1941 that had a minimum quantifier of zero caused PCRE to crash. For example,
1942 (x(y(?2))z)? provoked this bug with a subject that got as far as the
1943 recursion. If the recursively-called subpattern itself had a zero repeat,
1944 that was OK.
1945
1946 13. In pcretest, the buffer for reading a data line was set at 30K, but the
1947 buffer into which it was copied (for escape processing) was still set at
1948 1024, so long lines caused crashes.
1949
1950 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
1951 "internal error: code overflow...". This applied to any character class
1952 that was followed by a possessive quantifier.
1953
1954 15. Modified the Makefile to add libpcre.la as a prerequisite for
1955 libpcreposix.la because I was told this is needed for a parallel build to
1956 work.
1957
1958 16. If a pattern that contained .* following optional items at the start was
1959 studied, the wrong optimizing data was generated, leading to matching
1960 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
1961 matching string must start with a or b or c. The correct conclusion for
1962 this pattern is that a match can start with any character.
1963
1964
1965 Version 4.4 13-Aug-03
1966 ---------------------
1967
1968 1. In UTF-8 mode, a character class containing characters with values between
1969 127 and 255 was not handled correctly if the compiled pattern was studied.
1970 In fixing this, I have also improved the studying algorithm for such
1971 classes (slightly).
1972
1973 2. Three internal functions had redundant arguments passed to them. Removal
1974 might give a very teeny performance improvement.
1975
1976 3. Documentation bug: the value of the capture_top field in a callout is *one
1977 more than* the number of the hightest numbered captured substring.
1978
1979 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
1980 in incorrectly linking with a previously installed version. They now link
1981 explicitly with libpcre.la.
1982
1983 5. configure.in no longer needs to recognize Cygwin specially.
1984
1985 6. A problem in pcre.in for Windows platforms is fixed.
1986
1987 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
1988 pcretest, it used to include the size of the study block as part of its
1989 output. Unfortunately, the structure contains a field that has a different
1990 size on different hardware architectures. This meant that the tests that
1991 showed this size failed. As the block is currently always of a fixed size,
1992 this information isn't actually particularly useful in pcretest output, so
1993 I have just removed it.
1994
1995 8. Three pre-processor statements accidentally did not start in column 1.
1996 Sadly, there are *still* compilers around that complain, even though
1997 standard C has not required this for well over a decade. Sigh.
1998
1999 9. In pcretest, the code for checking callouts passed small integers in the
2000 callout_data field, which is a void * field. However, some picky compilers
2001 complained about the casts involved for this on 64-bit systems. Now
2002 pcretest passes the address of the small integer instead, which should get
2003 rid of the warnings.
2004
2005 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
2006 both compile and run time, and gives an error if an invalid UTF-8 sequence
2007 is found. There is a option for disabling this check in cases where the
2008 string is known to be correct and/or the maximum performance is wanted.
2009
2010 11. In response to a bug report, I changed one line in Makefile.in from
2011
2012 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
2013 to
2014 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
2015
2016 to look similar to other lines, but I have no way of telling whether this
2017 is the right thing to do, as I do not use Windows. No doubt I'll get told
2018 if it's wrong...
2019
2020
2021 Version 4.3 21-May-03
2022 ---------------------
2023
2024 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
2025 Makefile.
2026
2027 2. Some refactoring to improve the quality of the code:
2028
2029 (i) The utf8_table... variables are now declared "const".
2030
2031 (ii) The code for \cx, which used the "case flipping" table to upper case
2032 lower case letters, now just substracts 32. This is ASCII-specific,
2033 but the whole concept of \cx is ASCII-specific, so it seems
2034 reasonable.
2035
2036 (iii) PCRE was using its character types table to recognize decimal and
2037 hexadecimal digits in the pattern. This is silly, because it handles
2038 only 0-9, a-f, and A-F, but the character types table is locale-
2039 specific, which means strange things might happen. A private
2040 table is now used for this - though it costs 256 bytes, a table is
2041 much faster than multiple explicit tests. Of course, the standard
2042 character types table is still used for matching digits in subject
2043 strings against \d.
2044
2045 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
2046 ending in _t are). So I've renamed it as ESC_tee.
2047
2048 3. The first argument for regexec() in the POSIX wrapper should have been
2049 defined as "const".
2050
2051 4. Changed pcretest to use malloc() for its buffers so that they can be
2052 Electric Fenced for debugging.
2053
2054 5. There were several places in the code where, in UTF-8 mode, PCRE would try
2055 to read one or more bytes before the start of the subject string. Often this
2056 had no effect on PCRE's behaviour, but in some circumstances it could
2057 provoke a segmentation fault.
2058
2059 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
2060 to try to read one or more bytes before the start of the subject string.
2061
2062 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
2063 UTF-8 support could misbehave in various ways if the subject string
2064 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
2065 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
2066 back over UTF-8 characters.)
2067
2068
2069 Version 4.2 14-Apr-03
2070 ---------------------
2071
2072 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
2073
2074 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
2075 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
2076 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
2077 [WIN_PREFIX]: new variable, "cyg" for Cygwin
2078 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
2079 and BUILD_EXEEXT
2080 Note: automatic setting of the BUILD variables is not yet working
2081 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
2082 compile-time but not at link-time
2083 [LINK]: use for linking executables only
2084 make different versions for Windows and non-Windows
2085 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
2086 libraries
2087 [LINK_FOR_BUILD]: new variable
2088 [OBJEXT]: use throughout
2089 [EXEEXT]: use throughout
2090 <winshared>: new target
2091 <wininstall>: new target
2092 <dftables.o>: use native compiler
2093 <dftables>: use native linker
2094 <install>: handle Windows platform correctly
2095 <clean>: ditto
2096 <check>: ditto
2097 copy DLL to top builddir before testing
2098
2099 As part of these changes, -no-undefined was removed again. This was reported
2100 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
2101 in any case.
2102
2103 3. Some tidies to get rid of compiler warnings:
2104
2105 . In the match_data structure, match_limit was an unsigned long int, whereas
2106 match_call_count was an int. I've made them both unsigned long ints.
2107
2108 . In pcretest the fact that a const uschar * doesn't automatically cast to
2109 a void * provoked a warning.
2110
2111 . Turning on some more compiler warnings threw up some "shadow" variables
2112 and a few more missing casts.
2113
2114 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
2115 option, a class that contained a single character with a value between 128
2116 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
2117
2118 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
2119 option, a class that contained several characters, but with at least one
2120 whose value was between 128 and 255 caused PCRE to crash.
2121
2122
2123 Version 4.1 12-Mar-03
2124 ---------------------
2125
2126 1. Compiling with gcc -pedantic found a couple of places where casts were
2127 needed, and a string in dftables.c that was longer than standard compilers are
2128 required to support.
2129
2130 2. Compiling with Sun's compiler found a few more places where the code could
2131 be tidied up in order to avoid warnings.
2132
2133 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
2134 first of these names is deprecated in the latest Autoconf in favour of the name
2135 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
2136 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
2137 analogy I have changed it to CFLAGS_FOR_BUILD.
2138
2139 4. Added -no-undefined to the linking command in the Makefile, because this is
2140 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
2141 linking step for the pcreposix library.
2142
2143 5. PCRE was failing to diagnose the case of two named groups with the same
2144 name.
2145
2146 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
2147 literal character that is needed in the subject for a match, and scans along to
2148 ensure that it is present before embarking on the full matching process. This
2149 saves time in cases of nested unlimited repeats that are never going to match.
2150 Problem: the scan can take a lot of time if the subject is very long (e.g.
2151 megabytes), thus penalizing straightforward matches. It is now done only if the
2152 amount of subject to be scanned is less than 1000 bytes.
2153
2154 7. A lesser problem with the same optimization is that it was recording the
2155 first character of an anchored pattern as "needed", thus provoking a search
2156 right along the subject, even when the first match of the pattern was going to
2157 fail. The "needed" character is now not set for anchored patterns, unless it
2158 follows something in the pattern that is of non-fixed length. Thus, it still
2159 fulfils its original purpose of finding quick non-matches in cases of nested
2160 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
2161
2162
2163 Version 4.0 17-Feb-03
2164 ---------------------
2165
2166 1. If a comment in an extended regex that started immediately after a meta-item
2167 extended to the end of string, PCRE compiled incorrect data. This could lead to
2168 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
2169
2170 2. Moved to autoconf 2.53 and libtool 1.4.2.
2171
2172 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
2173 the special perltest8 script is no longer needed - all the tests can be run
2174 from a single perltest script.
2175
2176 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
2177 by \s. It has now been removed in PCRE. This means it isn't recognized as
2178 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
2179 class [:space:] *does* include VT, thereby creating a mess.
2180
2181 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
2182 space and tab.
2183
2184 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
2185 its new features into the main test script, reducing the number of scripts.
2186
2187 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
2188 were backward compatible, and made the (?i) apply to the whole pattern, as if
2189 /i were given. Now it behaves more logically, and applies the option setting
2190 only to what follows. PCRE has been changed to follow suit. However, if it
2191 finds options settings right at the start of the pattern, it extracts them into
2192 the global options, as before. Thus, they show up in the info data.
2193
2194 8. Added support for the \Q...\E escape sequence. Characters in between are
2195 treated as literals. This is slightly different from Perl in that $ and @ are
2196 also handled as literals inside the quotes. In Perl, they will cause variable
2197 interpolation. Note the following examples:
2198
2199 Pattern PCRE matches Perl matches
2200
2201 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
2202 \Qabc\$xyz\E abc\$xyz abc\$xyz
2203 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
2204
2205 For compatibility with Perl, \Q...\E sequences are recognized inside character
2206 classes as well as outside them.
2207
2208 9. Re-organized 3 code statements in pcretest to avoid "overflow in
2209 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
2210 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
2211 signed/unsigned warnings.
2212
2213 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
2214 option for pcretest, so I've replaced it by a simple function that does just
2215 that job.
2216
2217 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
2218 "pcregrep -".
2219
2220 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
2221 Java package. This provides some syntactic sugar for simple cases of what my
2222 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
2223 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
2224 item, you can use this simplified notation. Note that only makes sense with
2225 greedy quantifiers. Consequently, the use of the possessive quantifier forces
2226 greediness, whatever the setting of the PCRE_UNGREEDY option.
2227
2228 13. A change of greediness default within a pattern was not taking effect at
2229 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
2230 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
2231 was abstracted outside.
2232
2233 14. PCRE now supports the \G assertion. It is true when the current matching
2234 position is at the start point of the match. This differs from \A when the
2235 starting offset is non-zero. Used with the /g option of pcretest (or similar
2236 code), it works in the same way as it does for Perl's /g option. If all
2237 alternatives of a regex begin with \G, the expression is anchored to the start
2238 match position, and the "anchored" flag is set in the compiled expression.
2239
2240 15. Some bugs concerning the handling of certain option changes within patterns
2241 have been fixed. These applied to options other than (?ims). For example,
2242 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
2243 been the other way round. Some of this was related to change 7 above.
2244
2245 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
2246 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
2247 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
2248 POSIX classes only within a class (e.g. /[[:alpha:]]/).
2249
2250 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
2251 mode. Unlike ".", it always matches newline, whatever the setting of
2252 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
2253 assertions. Perl allows it, but it doesn't (in general) work because it can't
2254 calculate the length of the lookbehind. At least, that's the case for Perl
2255 5.8.0 - I've been told they are going to document that it doesn't work in
2256 future.
2257
2258 18. Added an error diagnosis for escapes that PCRE does not support: these are
2259 \L, \l, \N, \P, \p, \U, \u, and \X.
2260
2261 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
2262 reading past the end of the pattern in cases such as /[abcd/.
2263
2264 20. PCRE was getting more memory than necessary for patterns with classes that
2265 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
2266
2267 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
2268 compiling PCRE for use with Virtual Pascal.
2269
2270 22. Small fix to the Makefile to make it work properly if the build is done
2271 outside the source tree.
2272
2273 23. Added a new extension: a condition to go with recursion. If a conditional
2274 subpattern starts with (?(R) the "true" branch is used if recursion has
2275 happened, whereas the "false" branch is used only at the top level.
2276
2277 24. When there was a very long string of literal characters (over 255 bytes
2278 without UTF support, over 250 bytes with UTF support), the computation of how
2279 much memory was required could be incorrect, leading to segfaults or other
2280 strange effects.
2281
2282 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
2283 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
2284 there was a subsequent back reference to those brackets. This meant that, for
2285 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
2286 possible to check for precisely this case. All we can do is abandon the
2287 optimization if .* occurs inside capturing brackets when there are any back
2288 references whatsoever. (See below for a better fix that came later.)
2289
2290 26. The handling of the optimization for finding the first character of a
2291 non-anchored pattern, and for finding a character that is required later in the
2292 match were failing in some cases. This didn't break the matching; it just
2293 failed to optimize when it could. The way this is done has been re-implemented.
2294
2295 27. Fixed typo in error message for invalid (?R item (it said "(?p").
2296
2297 28. Added a new feature that provides some of the functionality that Perl
2298 provides with (?{...}). The facility is termed a "callout". The way it is done
2299 in PCRE is for the caller to provide an optional function, by setting
2300 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
2301 global variable. By default it is unset, which disables all calling out. To get
2302 the function called, the regex must include (?C) at appropriate points. This
2303 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
2304 This provides a means of identifying different callout points. When PCRE
2305 reaches such a point in the regex, if pcre_callout has been set, the external
2306 function is called. It is provided with data in a structure called
2307 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
2308 matching continues; if it returns a non-zero value, the match at the current
2309 point fails. However, backtracking will occur if possible. [This was changed
2310 later and other features added - see item 49 below.]
2311
2312 29. pcretest is upgraded to test the callout functionality. It provides a
2313 callout function that displays information. By default, it shows the start of
2314 the match and the current position in the text. There are some new data escapes
2315 to vary what happens:
2316
2317 \C+ in addition, show current contents of captured substrings
2318 \C- do not supply a callout function
2319 \C!n return 1 when callout number n is reached
2320 \C!n!m return 1 when callout number n is reached for the mth time
2321
2322 30. If pcregrep was called with the -l option and just a single file name, it
2323 output "<stdin>" if a match was found, instead of the file name.
2324
2325 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
2326 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
2327 pcre_exec(). This saves a malloc/free per call. The default value of
2328 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
2329 when configuring.
2330
2331 32. The default maximum size of a compiled pattern is 64K. There have been a
2332 few cases of people hitting this limit. The code now uses macros to handle the
2333 storing of links as offsets within the compiled pattern. It defaults to 2-byte
2334 links, but this can be changed to 3 or 4 bytes by --with-link-size when
2335 configuring. Tests 2 and 5 work only with 2-byte links because they output
2336 debugging information about compiled patterns.
2337
2338 33. Internal code re-arrangements:
2339
2340 (a) Moved the debugging function for printing out a compiled regex into
2341 its own source file (printint.c) and used #include to pull it into
2342 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
2343 separate copies.
2344
2345 (b) Defined the list of op-code names for debugging as a macro in
2346 internal.h so that it is next to the definition of the opcodes.
2347
2348 (c) Defined a table of op-code lengths for simpler skipping along compiled
2349 code. This is again a macro in internal.h so that it is next to the
2350 definition of the opcodes.
2351
2352 34. Added support for recursive calls to individual subpatterns, along the
2353 lines of Robin Houston's patch (but implemented somewhat differently).
2354
2355 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
2356 allow it to read and process whole directories in Win32. This code was
2357 contributed by Lionel Fourquaux; it has not been tested by me.
2358
2359 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
2360 used to name a group. Names consist of alphanumerics and underscores, and must
2361 be unique. Back references use the syntax (?P=name) and recursive calls use
2362 (?P>name) which is a PCRE extension to the Python extension. Groups still have
2363 numbers. The function pcre_fullinfo() can be used after compilation to extract
2364 a name/number map. There are three relevant calls:
2365
2366 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
2367 PCRE_INFO_NAMECOUNT yields the number of entries
2368 PCRE_INFO_NAMETABLE yields a pointer to the map.
2369
2370 The map is a vector of fixed-size entries. The size of each entry depends on
2371 the length of the longest name used. The first two bytes of each entry are the
2372 group number, most significant byte first. There follows the corresponding
2373 name, zero terminated. The names are in alphabetical order.
2374
2375 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
2376 case instead of 255. Making it the same both with and without UTF-8 support
2377 means that the same test output works with both.
2378
2379 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
2380 calling malloc() with a zero argument.
2381
2382 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
2383 optimization. I've improved things by keeping a bitmap of backreferences with
2384 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
2385 fact referenced, the optimization can be applied. It is unlikely that a
2386 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
2387 the match to follow \n) will appear inside brackets with a number greater than
2388 31, but if it does, any back reference > 31 suppresses the optimization.
2389
2390 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
2391 of disabling numbered capturing parentheses. Any opening parenthesis that is
2392 not followed by ? behaves as if it were followed by ?: but named parentheses
2393 can still be used for capturing (and they will acquire numbers in the usual
2394 way).
2395
2396 41. Redesigned the return codes from the match() function into yes/no/error so
2397 that errors can be passed back from deep inside the nested calls. A malloc
2398 failure while inside a recursive subpattern call now causes the
2399 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
2400
2401 42. It is now possible to set a limit on the number of times the match()
2402 function is called in a call to pcre_exec(). This facility makes it possible to
2403 limit the amount of recursion and backtracking, though not in a directly
2404 obvious way, because the match() function is used in a number of different
2405 circumstances. The count starts from zero for each position in the subject
2406 string (for non-anchored patterns). The default limit is, for compatibility, a
2407 large number, namely 10 000 000. You can change this in two ways:
2408
2409 (a) When configuring PCRE before making, you can use --with-match-limit=n
2410 to set a default value for the compiled library.
2411
2412 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
2413 a different value is set. See 45 below.
2414
2415 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
2416
2417 43. Added a new function pcre_config(int, void *) to enable run-time extraction
2418 of things that can be changed at compile time. The first argument specifies
2419 what is wanted and the second points to where the information is to be placed.
2420 The current list of available information is:
2421
2422 PCRE_CONFIG_UTF8
2423
2424 The output is an integer that is set to one if UTF-8 support is available;
2425 otherwise it is set to zero.
2426
2427 PCRE_CONFIG_NEWLINE
2428
2429 The output is an integer that it set to the value of the code that is used for
2430 newline. It is either LF (10) or CR (13).
2431
2432 PCRE_CONFIG_LINK_SIZE
2433
2434 The output is an integer that contains the number of bytes used for internal
2435 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
2436
2437 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
2438
2439 The output is an integer that contains the threshold above which the POSIX
2440 interface uses malloc() for output vectors. See item 31 above.
2441
2442 PCRE_CONFIG_MATCH_LIMIT
2443
2444 The output is an unsigned integer that contains the default limit of the number
2445 of match() calls in a pcre_exec() execution. See 42 above.
2446
2447 44. pcretest has been upgraded by the addition of the -C option. This causes it
2448 to extract all the available output from the new pcre_config() function, and to
2449 output it. The program then exits immediately.
2450
2451 45. A need has arisen to pass over additional data with calls to pcre_exec() in
2452 order to support additional features. One way would have been to define
2453 pcre_exec2() (for example) with extra arguments, but this would not have been
2454 extensible, and would also have required all calls to the original function to
2455 be mapped to the new one. Instead, I have chosen to extend the mechanism that
2456 is used for passing in "extra" data from pcre_study().
2457
2458 The pcre_extra structure is now exposed and defined in pcre.h. It currently
2459 contains the following fields:
2460
2461 flags a bitmap indicating which of the following fields are set
2462 study_data opaque data from pcre_study()
2463 match_limit a way of specifying a limit on match() calls for a specific
2464 call to pcre_exec()
2465 callout_data data for callouts (see 49 below)
2466
2467 The flag bits are also defined in pcre.h, and are
2468
2469 PCRE_EXTRA_STUDY_DATA
2470 PCRE_EXTRA_MATCH_LIMIT
2471 PCRE_EXTRA_CALLOUT_DATA
2472
2473 The pcre_study() function now returns one of these new pcre_extra blocks, with
2474 the actual study data pointed to by the study_data field, and the
2475 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
2476 before. That is, this change is entirely upwards-compatible and requires no
2477 change to existing code.
2478
2479 If you want to pass in additional data to pcre_exec(), you can either place it
2480 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
2481 block.
2482
2483 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
2484 data string contains the escape sequence \M, pcretest calls pcre_exec() several
2485 times with different match limits, until it finds the minimum value needed for
2486 pcre_exec() to complete. The value is then output. This can be instructive; for
2487 most simple matches the number is quite small, but for pathological cases it
2488 gets very large very quickly.
2489
2490 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
2491 returns the size of the data block pointed to by the study_data field in a
2492 pcre_extra block, that is, the value that was passed as the argument to
2493 pcre_malloc() when PCRE was getting memory in which to place the information
2494 created by pcre_study(). The fourth argument should point to a size_t variable.
2495 pcretest has been extended so that this information is shown after a successful
2496 pcre_study() call when information about the compiled regex is being displayed.
2497
2498 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
2499 because what follows is always an absolute path. (Later: it turns out that this
2500 is more than cosmetic for MinGW, because it doesn't like empty path
2501 components.)
2502
2503 49. Some changes have been made to the callout feature (see 28 above):
2504
2505 (i) A callout function now has three choices for what it returns:
2506
2507 0 => success, carry on matching
2508 > 0 => failure at this point, but backtrack if possible
2509 < 0 => serious error, return this value from pcre_exec()
2510
2511 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
2512 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
2513 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
2514 use by callout functions. It will never be used by PCRE itself.
2515
2516 (ii) The pcre_extra structure (see 45 above) has a void * field called
2517 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
2518 pcre_callout_block structure has a field of the same name. The contents of
2519 the field passed in the pcre_extra structure are passed to the callout
2520 function in the corresponding field in the callout block. This makes it
2521 easier to use the same callout-containing regex from multiple threads. For
2522 testing, the pcretest program has a new data escape
2523
2524 \C*n pass the number n (may be negative) as callout_data
2525
2526 If the callout function in pcretest receives a non-zero value as
2527 callout_data, it returns that value.
2528
2529 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
2530 there were some redundant $(CFLAGS) in commands that are now specified as
2531 $(LINK), which already includes $(CFLAGS).
2532
2533 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
2534 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
2535 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
2536 one-byte characters throughout. Note that case-insensitive matching applies
2537 only to characters whose values are less than 256. PCRE doesn't support the
2538 notion of cases for higher-valued characters.
2539
2540 (i) A character class whose characters are all within 0-255 is handled as
2541 a bit map, and the map is inverted for negative classes. Previously, a
2542 character > 255 always failed to match such a class; however it should
2543 match if the class was a negative one (e.g. [^ab]). This has been fixed.
2544
2545 (ii) A negated character class with a single character < 255 is coded as
2546 "not this character" (OP_NOT). This wasn't working properly when the test
2547 character was multibyte, either singly or repeated.
2548
2549 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
2550 mode, for example: \x{100}{2,3}.
2551
2552 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
2553 singly or repeated) now correctly test multibyte characters. However,
2554 PCRE doesn't recognize any characters with values greater than 255 as
2555 digits, spaces, or word characters. Such characters always match \D, \S,
2556 and \W, and never match \d, \s, or \w.
2557
2558 (v) Classes may now contain characters and character ranges with values
2559 greater than 255. For example: [ab\x{100}-\x{400}].
2560
2561 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
2562 PCRE in UTF-8 mode.
2563
2564 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
2565 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
2566 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
2567 value.)
2568
2569 53. The single man page has become too large. I have therefore split it up into
2570 a number of separate man pages. These also give rise to individual HTML pages;
2571 these are now put in a separate directory, and there is an index.html page that
2572 lists them all. Some hyperlinking between the pages has been installed.
2573
2574 54. Added convenience functions for handling named capturing parentheses.
2575
2576 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
2577 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
2578 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
2579 are faulted.
2580
2581 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
2582 calling configure. These values are used when compiling the dftables.c program
2583 which is run to generate the source of the default character tables. They
2584 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
2585 you will need to set these values.
2586
2587 57. Updated the building process for Windows DLL, as provided by Fred Cox.
2588
2589
2590 Version 3.9 02-Jan-02
2591 ---------------------
2592
2593 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
2594
2595 2. If --disable-static was given, the building process failed when trying to
2596 build pcretest and pcregrep. (For some reason it was using libtool to compile
2597 them, which is not right, as they aren't part of the library.)
2598
2599
2600 Version 3.8 18-Dec-01
2601 ---------------------
2602
2603 1. The experimental UTF-8 code was completely screwed up. It was packing the
2604 bytes in the wrong order. How dumb can you get?
2605
2606
2607 Version 3.7 29-Oct-01
2608 ---------------------
2609
2610 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
2611 This caused pcretest, when used on the test data, to segfault. Unfortunately,
2612 this didn't happen under Solaris 8, where I normally test things.
2613
2614 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
2615 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
2616 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
2617 this fix an hour or so after the initial 3.7 release.)
2618
2619
2620 Version 3.6 23-Oct-01
2621 ---------------------
2622
2623 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
2624 offsets passed as NULL with zero offset count.
2625
2626 2. The config.guess and config.sub files had not been updated when I moved to
2627 the latest autoconf.
2628
2629
2630 Version 3.5 15-Aug-01
2631 ---------------------
2632
2633 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
2634 had been forgotten.
2635
2636 2. By using declared but undefined structures, we can avoid using "void"
2637 definitions in pcre.h while keeping the internal definitions of the structures
2638 private.
2639
2640 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
2641 user point of view, this means that both static and shared libraries are built
2642 by default, but this can be individually controlled. More of the work of
2643 handling this static/shared cases is now inside libtool instead of PCRE's make
2644 file.
2645
2646 4. The pcretest utility is now installed along with pcregrep because it is
2647 useful for users (to test regexs) and by doing this, it automatically gets
2648 relinked by libtool. The documentation has been turned into a man page, so
2649 there are now .1, .txt, and .html versions in /doc.
2650
2651 5. Upgrades to pcregrep:
2652 (i) Added long-form option names like gnu grep.
2653 (ii) Added --help to list all options with an explanatory phrase.
2654 (iii) Added -r, --recursive to recurse into sub-directories.
2655 (iv) Added -f, --file to read patterns from a file.
2656
2657 6. pcre_exec() was referring to its "code" argument before testing that
2658 argument for NULL (and giving an error if it was NULL).
2659
2660 7. Upgraded Makefile.in to allow for compiling in a different directory from
2661 the source directory.
2662
2663 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
2664 options bits, the pointer it was passed was to an int instead of to an unsigned
2665 long int. This mattered only on 64-bit systems.
2666
2667 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
2668 generated) instead of pcre.in, which it its source. Also made the same change
2669 in several of the .c files.
2670
2671 10. A new release of gcc defines printf() as a macro, which broke pcretest
2672 because it had an ifdef in the middle of a string argument for printf(). Fixed
2673 by using separate calls to printf().
2674
2675 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
2676 script, to force use of CR or LF instead of \n in the source. On non-Unix
2677 systems, the value can be set in config.h.
2678
2679 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
2680 absolute limit. Changed the text of the error message to make this clear, and
2681 likewise updated the man page.
2682
2683 13. The limit of 99 on the number of capturing subpatterns has been removed.
2684 The new limit is 65535, which I hope will not be a "real" limit.
2685
2686
2687 Version 3.4 22-Aug-00
2688 ---------------------
2689
2690 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
2691
2692 2. Diagnose condition (?(0) as an error instead of crashing on matching.
2693
2694
2695 Version 3.3 01-Aug-00
2696 ---------------------
2697
2698 1. If an octal character was given, but the value was greater than \377, it
2699 was not getting masked to the least significant bits, as documented. This could
2700 lead to crashes in some systems.
2701
2702 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
2703 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
2704
2705 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
2706 These just pass their arguments on to (pcre_free)(), but they are provided
2707 because some uses of PCRE bind it to non-C systems that can call its functions,
2708 but cannot call free() or pcre_free() directly.
2709
2710 4. Add "make test" as a synonym for "make check". Corrected some comments in
2711 the Makefile.
2712
2713 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
2714 Makefile.
2715
2716 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
2717 command called pgrep for grepping around the active processes.
2718
2719 7. Added the beginnings of support for UTF-8 character strings.
2720
2721 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
2722 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
2723 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
2724 out for the ar command.)
2725
2726
2727 Version 3.2 12-May-00
2728 ---------------------
2729
2730 This is purely a bug fixing release.
2731
2732 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
2733 of ZA. This was just one example of several cases that could provoke this bug,
2734 which was introduced by change 9 of version 2.00. The code for breaking
2735 infinite loops after an iteration that matches an empty string was't working
2736 correctly.
2737
2738 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
2739 when matched against abbab (for example). After matching an empty string, it
2740 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
2741 caused it to match further down the string than it should.
2742
2743 3. The code contained an inclusion of sys/types.h. It isn't clear why this
2744 was there because it doesn't seem to be needed, and it causes trouble on some
2745 systems, as it is not a Standard C header. It has been removed.
2746
2747 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
2748 were reported on the Macintosh. The changes were from
2749
2750 while ((c = *(++ptr)) != 0 && c != '\n');
2751 to
2752 while ((c = *(++ptr)) != 0 && c != '\n') ;
2753
2754 Totally extraordinary, but if that's what it takes...
2755
2756 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
2757 available. Added HAVE_BCOPY and an autoconf test for it; if neither
2758 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
2759 assumes the way PCRE uses memmove() (always moving upwards).
2760
2761 6. PCRE is being used in one environment where strchr() is not available. There
2762 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
2763 faster code anyway.
2764
2765
2766 Version 3.1 09-Feb-00
2767 ---------------------
2768
2769 The only change in this release is the fixing of some bugs in Makefile.in for
2770 the "install" target:
2771
2772 (1) It was failing to install pcreposix.h.
2773
2774 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
2775
2776
2777 Version 3.0 01-Feb-00
2778 ---------------------
2779
2780 1. Add support for the /+ modifier to perltest (to output $` like it does in
2781 pcretest).
2782
2783 2. Add support for the /g modifier to perltest.
2784
2785 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
2786 matches null strings.
2787
2788 4. Fix perltest so that it doesn't do unwanted things when fed an empty
2789 pattern. Perl treats empty patterns specially - it reuses the most recent
2790 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
2791 effect.
2792
2793 5. The POSIX interface was broken in that it was just handing over the POSIX
2794 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
2795 required a bigger vector, with some working space on the end. This means that
2796 the POSIX wrapper now has to get and free some memory, and copy the results.
2797
2798 6. Added some simple autoconf support, placing the test data and the
2799 documentation in separate directories, re-organizing some of the
2800 information files, and making it build pcre-config (a GNU standard). Also added
2801 libtool support for building PCRE as a shared library, which is now the
2802 default.
2803
2804 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
2805 09 are not valid octal constants. Single digits will be used for minor values
2806 less than 10.
2807
2808 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
2809 existing programs that set these in the POSIX interface can use PCRE without
2810 modification.
2811
2812 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
2813 return all that pcre_info() returns, plus additional data. The pcre_info()
2814 function is retained for compatibility, but is considered to be obsolete.
2815
2816 10. Added experimental recursion feature (?R) to handle one common case that
2817 Perl 5.6 will be able to do with (?p{...}).
2818
2819 11. Added support for POSIX character classes like [:alpha:], which Perl is
2820 adopting.
2821
2822
2823 Version 2.08 31-Aug-99
2824 ----------------------
2825
2826 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
2827 trying to match at the startoffset position, but instead was moving forward to
2828 the next newline as if a previous match had failed.
2829
2830 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
2831 and could get into a loop if a null string was matched other than at the start
2832 of the subject.
2833
2834 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
2835 be distinguished at compile time, and for completeness also added PCRE_DATE.
2836
2837 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
2838 in GnuWin32 environments.
2839
2840
2841 Version 2.07 29-Jul-99
2842 ----------------------
2843
2844 1. The documentation is now supplied in plain text form and HTML as well as in
2845 the form of man page sources.
2846
2847 2. C++ compilers don't like assigning (void *) values to other pointer types.
2848 In particular this affects malloc(). Although there is no problem in Standard
2849 C, I've put in casts to keep C++ compilers happy.
2850
2851 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
2852 should be (const char *).
2853
2854 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
2855 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
2856 However, I haven't made this a standard facility. The documentation doesn't
2857 mention it, and the Makefile doesn't support it.
2858
2859 5. The Makefile now contains an "install" target, with editable destinations at
2860 the top of the file. The pcretest program is not installed.
2861
2862 6. pgrep -V now gives the PCRE version number and date.
2863
2864 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
2865 causing the entire string to be ignored, instead of just the last character.
2866
2867 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
2868 non-matching string, it can take a very, very long time, even for strings of
2869 quite modest length, because of the nested recursion. PCRE now does better in
2870 some of these cases. It does this by remembering the last required literal
2871 character in the pattern, and pre-searching the subject to ensure it is present
2872 before running the real match. In other words, it applies a heuristic to detect
2873 some types of certain failure quickly, and in the above example, if presented
2874 with a string that has no trailing " it gives "no match" very quickly.
2875
2876 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
2877 other alternatives are tried instead.
2878
2879
2880 Version 2.06 09-Jun-99
2881 ----------------------
2882
2883 1. Change pcretest's output for amount of store used to show just the code
2884 space, because the remainder (the data block) varies in size between 32-bit and
2885 64-bit systems.
2886
2887 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
2888 start matching at. This allows lookbehinds to work when searching for multiple
2889 occurrences in a string.
2890
2891 3. Added additional options to pcretest for testing multiple occurrences:
2892
2893 /+ outputs the rest of the string that follows a match
2894 /g loops for multiple occurrences, using the new startoffset argument
2895 /G loops for multiple occurrences by passing an incremented pointer
2896
2897 4. PCRE wasn't doing the "first character" optimization for patterns starting
2898 with \b or \B, though it was doing it for other lookbehind assertions. That is,
2899 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
2900 the letter 'x'. On long subject strings, this gives a significant speed-up.
2901
2902
2903 Version 2.05 21-Apr-99
2904 ----------------------
2905
2906 1. Changed the type of magic_number from int to long int so that it works
2907 properly on 16-bit systems.
2908
2909 2. Fixed a bug which caused patterns starting with .* not to work correctly
2910 when the subject string contained newline characters. PCRE was assuming
2911 anchoring for such patterns in all cases, which is not correct because .* will
2912 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
2913 DOTALL is set at top level; otherwise it knows that patterns starting with .*
2914 must be retried after every newline in the subject.
2915
2916
2917 Version 2.04 18-Feb-99
2918 ----------------------
2919
2920 1. For parenthesized subpatterns with repeats whose minimum was zero, the
2921 computation of the store needed to hold the pattern was incorrect (too large).
2922 If such patterns were nested a few deep, this could multiply and become a real
2923 problem.
2924
2925 2. Added /M option to pcretest to show the memory requirement of a specific
2926 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
2927
2928 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
2929 compiled in such a way that the backtracking after subsequent failure was
2930 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
2931 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
2932
2933
2934 Version 2.03 02-Feb-99
2935 ----------------------
2936
2937 1. Fixed typo and small mistake in man page.
2938
2939 2. Added 4th condition (GPL supersedes if conflict) and created separate
2940 LICENCE file containing the conditions.
2941
2942 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
2943 Perl, that is the internal \ allows the delimiter to be included in the
2944 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
2945 the final delimiter, add \ to the end of the pattern (to test the error).
2946
2947 4. Added the convenience functions for extracting substrings after a successful
2948 match. Updated pcretest to make it able to test these functions.
2949
2950
2951 Version 2.02 14-Jan-99
2952 ----------------------
2953
2954 1. Initialized the working variables associated with each extraction so that
2955 their saving and restoring doesn't refer to uninitialized store.
2956
2957 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
2958 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
2959 fix the problem.
2960
2961 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
2962 calls, and wasn't printing the correct value for compiling calls. Increased the
2963 default value of LOOPREPEAT, and the number of significant figures in the
2964 times.
2965
2966 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
2967
2968 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
2969 a building problem on Windows NT with a FAT file system.
2970
2971
2972 Version 2.01 21-Oct-98
2973 ----------------------
2974
2975 1. Changed the API for pcre_compile() to allow for the provision of a pointer
2976 to character tables built by pcre_maketables() in the current locale. If NULL
2977 is passed, the default tables are used.
2978
2979
2980 Version 2.00 24-Sep-98
2981 ----------------------
2982
2983 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
2984 it any more.
2985
2986 2. Allow quantification of (?>) groups, and make it work correctly.
2987
2988 3. The first character computation wasn't working for (?>) groups.
2989
2990 4. Correct the implementation of \Z (it is permitted to match on the \n at the
2991 end of the subject) and add 5.005's \z, which really does match only at the
2992 very end of the subject.
2993
2994 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
2995
2996 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
2997 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
2998 localized options. All options to pcre_study() were also removed.
2999
3000 7. Add other new features from 5.005:
3001
3002 $(?<= positive lookbehind
3003 $(?<! negative lookbehind
3004 (?imsx-imsx) added the unsetting capability
3005 such a setting is global if at outer level; local otherwise
3006 (?imsx-imsx:) non-capturing groups with option setting
3007 (?(cond)re|re) conditional pattern matching
3008
3009 A backreference to itself in a repeated group matches the previous
3010 captured string.
3011
3012 8. General tidying up of studying (both automatic and via "study")
3013 consequential on the addition of new assertions.
3014
3015 9. As in 5.005, unlimited repeated groups that could match an empty substring
3016 are no longer faulted at compile time. Instead, the loop is forcibly broken at
3017 runtime if any iteration does actually match an empty substring.
3018
3019 10. Include the RunTest script in the distribution.
3020
3021 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
3022 discrepancies, some of which were old and were also with respect to 5.004. They
3023 have now been fixed.
3024
3025
3026 Version 1.09 28-Apr-98
3027 ----------------------
3028
3029 1. A negated single character class followed by a quantifier with a minimum
3030 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
3031 program crashes, or just wrong answers. This did not apply to negated classes
3032 containing more than one character, or to minima other than one.
3033
3034
3035 Version 1.08 27-Mar-98
3036 ----------------------
3037
3038 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
3039
3040 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
3041 latter must appear before anything that relies on it in the pattern.
3042
3043
3044 Version 1.07 16-Feb-98
3045 ----------------------
3046
3047 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
3048 repeat of a potentially empty string).
3049
3050
3051 Version 1.06 23-Jan-98
3052 ----------------------
3053
3054 1. Added Markus Oberhumer's little patches for C++.
3055
3056 2. Literal strings longer than 255 characters were broken.
3057
3058
3059 Version 1.05 23-Dec-97
3060 ----------------------
3061
3062 1. Negated character classes containing more than one character were failing if
3063 PCRE_CASELESS was set at run time.
3064
3065
3066 Version 1.04 19-Dec-97
3067 ----------------------
3068
3069 1. Corrected the man page, where some "const" qualifiers had been omitted.
3070
3071 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
3072 input syntax.
3073
3074 3. Fixed memory leak which occurred when a regex with back references was
3075 matched with an offsets vector that wasn't big enough. The temporary memory
3076 that is used in this case wasn't being freed if the match failed.
3077
3078 4. Tidied pcretest to ensure it frees memory that it gets.
3079
3080 5. Temporary memory was being obtained in the case where the passed offsets
3081 vector was exactly big enough.
3082
3083 6. Corrected definition of offsetof() from change 5 below.
3084
3085 7. I had screwed up change 6 below and broken the rules for the use of
3086 setjmp(). Now fixed.
3087
3088
3089 Version 1.03 18-Dec-97
3090 ----------------------
3091
3092 1. A erroneous regex with a missing opening parenthesis was correctly
3093 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
3094 on some systems.
3095
3096 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
3097 it was reported that one broken compiler failed on the former because "code" is
3098 also an independent variable.
3099
3100 3. The erroneous regex a[]b caused an array overrun reference.
3101
3102 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
3103 fail on data ending with that character. (It was going on too far, and checking
3104 the next character, typically a binary zero.) This was specific to the
3105 optimized code for single-character negative classes.
3106
3107 5. Added a contributed patch from the TIN world which does the following:
3108
3109 + Add an undef for memmove, in case the the system defines a macro for it.
3110
3111 + Add a definition of offsetof(), in case there isn't one. (I don't know
3112 the reason behind this - offsetof() is part of the ANSI standard - but
3113 it does no harm).
3114
3115 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
3116 most of the places where whitespace preceded '#'. I have given up and
3117 allowed the remaining 2 cases to be at the margin.
3118
3119 + Rename some variables in pcre to eliminate shadowing. This seems very
3120 pedantic, but does no harm, of course.
3121
3122 6. Moved the call to setjmp() into its own function, to get rid of warnings
3123 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
3124
3125 7. Constructs such as \d{8,} were compiling into the equivalent of
3126 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
3127 outcome, but in this particular case used more store than had been allocated,
3128 which caused the bug to be discovered because it threw up an internal error.
3129
3130 8. The debugging code in both pcre and pcretest for outputting the compiled
3131 form of a regex was going wrong in the case of back references followed by
3132 curly-bracketed repeats.
3133
3134
3135 Version 1.02 12-Dec-97
3136 ----------------------
3137
3138 1. Typos in pcre.3 and comments in the source fixed.
3139
3140 2. Applied a contributed patch to get rid of places where it used to remove
3141 'const' from variables, and fixed some signed/unsigned and uninitialized
3142 variable warnings.
3143
3144 3. Added the "runtest" target to Makefile.
3145
3146 4. Set default compiler flag to -O2 rather than just -O.
3147
3148
3149 Version 1.01 19-Nov-97
3150 ----------------------
3151
3152 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
3153 like /([ab]*)*/, that is, for classes with more than one character in them.
3154
3155 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
3156 as /((?>a*))*/ (a PCRE_EXTRA facility).
3157
3158
3159 Version 1.00 18-Nov-97
3160 ----------------------
3161
3162 1. Added compile-time macros to support systems such as SunOS4 which don't have
3163 memmove() or strerror() but have other things that can be used instead.
3164
3165 2. Arranged that "make clean" removes the executables.
3166
3167
3168 Version 0.99 27-Oct-97
3169 ----------------------
3170
3171 1. Fixed bug in code for optimizing classes with only one character. It was
3172 initializing a 32-byte map regardless, which could cause it to run off the end
3173 of the memory it had got.
3174
3175 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
3176
3177
3178 Version 0.98 22-Oct-97
3179 ----------------------
3180
3181 1. Fixed bug in code for handling temporary memory usage when there are more
3182 back references than supplied space in the ovector. This could cause segfaults.
3183
3184
3185 Version 0.97 21-Oct-97
3186 ----------------------
3187
3188 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
3189
3190 2. Optimized negated single characters not to use a bit map.
3191
3192 3. Brought error texts together as macro definitions; clarified some of them;
3193 fixed one that was wrong - it said "range out of order" when it meant "invalid
3194 escape sequence".
3195
3196 4. Changed some char * arguments to const char *.
3197
3198 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
3199
3200 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
3201 pcretest.
3202
3203
3204 Version 0.96 16-Oct-97
3205 ----------------------
3206
3207 1. Added a simple "pgrep" utility to the distribution.
3208
3209 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
3210 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
3211 where "ddd" means "one or more decimal digits".
3212
3213 3. Fixed serious bug. If a pattern had a back reference, but the call to
3214 pcre_exec() didn't supply a large enough ovector to record the related
3215 identifying subpattern, the match always failed. PCRE now remembers the number
3216 of the largest back reference, and gets some temporary memory in which to save
3217 the offsets during matching if necessary, in order to ensure that
3218 backreferences always work.
3219
3220 4. Increased the compatibility with Perl in a number of ways:
3221
3222 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
3223 to request this handling. The option can be set at compile or exec time.
3224
3225 (b) $ matches before a terminating newline by default; an option
3226 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
3227 mode). The option can be set at compile or exec time.
3228
3229 (c) The handling of \ followed by a digit other than 0 is now supposed to be
3230 the same as Perl's. If the decimal number it represents is less than 10
3231 or there aren't that many previous left capturing parentheses, an octal
3232 escape is read. Inside a character class, it's always an octal escape,
3233 even if it is a single digit.
3234
3235 (d) An escaped but undefined alphabetic character is taken as a literal,
3236 unless PCRE_EXTRA is set. Currently this just reserves the remaining
3237 escapes.
3238
3239 (e) {0} is now permitted. (The previous item is removed from the compiled
3240 pattern).
3241
3242 5. Changed all the names of code files so that the basic parts are no longer
3243 than 10 characters, and abolished the teeny "globals.c" file.
3244
3245 6. Changed the handling of character classes; they are now done with a 32-byte
3246 bit map always.
3247
3248 7. Added the -d and /D options to pcretest to make it possible to look at the
3249 internals of compilation without having to recompile pcre.
3250
3251
3252 Version 0.95 23-Sep-97
3253 ----------------------
3254
3255 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
3256 \x20 at the start of a run of normal characters. These were being treated as
3257 real characters, instead of the source characters being re-checked.
3258
3259
3260 Version 0.94 18-Sep-97
3261 ----------------------
3262
3263 1. The functions are now thread-safe, with the caveat that the global variables
3264 containing pointers to malloc() and free() or alternative functions are the
3265 same for all threads.
3266
3267 2. Get pcre_study() to generate a bitmap of initial characters for non-
3268 anchored patterns when this is possible, and use it if passed to pcre_exec().
3269
3270
3271 Version 0.93 15-Sep-97
3272 ----------------------
3273
3274 1. /(b)|(:+)/ was computing an incorrect first character.
3275
3276 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
3277 but not actually doing anything yet.
3278
3279 3. Treat "-" characters in classes that cannot be part of ranges as literals,
3280 as Perl does (e.g. [-az] or [az-]).
3281
3282 4. Set the anchored flag if a branch starts with .* or .*? because that tests
3283 all possible positions.
3284
3285 5. Split up into different modules to avoid including unneeded functions in a
3286 compiled binary. However, compile and exec are still in one module. The "study"
3287 function is split off.
3288
3289 6. The character tables are now in a separate module whose source is generated
3290 by an auxiliary program - but can then be edited by hand if required. There are
3291 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
3292 toupper() in the code.
3293
3294 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
3295 make them global. Abolish the function for setting them, as the caller can now
3296 set them directly.
3297
3298
3299 Version 0.92 11-Sep-97
3300 ----------------------
3301
3302 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
3303 (e.g. /a{1,3}/) was broken (I mis-optimized it).
3304
3305 2. Caseless matching was not working in character classes if the characters in
3306 the pattern were in upper case.
3307
3308 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
3309
3310 4. Make PCRE_ANCHORED public and accept as a compile option.
3311
3312 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
3313 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
3314 pass them.
3315
3316 6. Give an error if bad option bits passed at compile or run time.
3317
3318 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
3319 pcretest to cause it to pass that flag.
3320
3321 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
3322 options, and the first character, if set.
3323
3324 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
3325
3326
3327 Version 0.91 10-Sep-97
3328 ----------------------
3329
3330 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
3331 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
3332
3333 2. PCRE was looping on encountering an indefinitely repeated back reference to
3334 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
3335 Perl does - treats the match as successful.
3336
3337 ****

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