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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 353 - (show annotations)
Mon Jul 7 15:44:24 2008 UTC (7 years, 1 month ago) by ph10
File size: 156894 byte(s)
Error occurred while calculating annotation data.
Fix SKIPZERO bug in pcre_study().
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
22 Version 7.7 07-May-08
23 ---------------------
24
25 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
26 a string to a long long, pretend we don't even have a long long." This is
27 done by checking for the strtoq, strtoll, and _strtoi64 functions.
28
29 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
30 pre-7.6 versions, which defined a global no_arg variable instead of putting
31 it in the RE class. (See also #8 below.)
32
33 3. Remove a line of dead code, identified by coverity and reported by Nuno
34 Lopes.
35
36 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
37
38 (1) The include/exclude patterns were being applied to the whole pathnames
39 of files, instead of just to the final components.
40
41 (2) If there was more than one level of directory, the subdirectories were
42 skipped unless they satisfied the include/exclude conditions. This is
43 inconsistent with GNU grep (and could even be seen as contrary to the
44 pcregrep specification - which I improved to make it absolutely clear).
45 The action now is always to scan all levels of directory, and just
46 apply the include/exclude patterns to regular files.
47
48 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
49 --exclude_dir in the tests to avoid scanning .svn directories.
50
51 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
52 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
53 doesn't support NULs in patterns.
54
55 7. Added some missing "const"s to declarations of static tables in
56 pcre_compile.c and pcre_dfa_exec.c.
57
58 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
59 caused by fix #2 above. (Subsequently also a second patch to fix the
60 first patch. And a third patch - this was a messy problem.)
61
62 9. Applied Craig's patch to remove the use of push_back().
63
64 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
65 matching function regexec().
66
67 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
68 which, however, unlike Perl's \g{...}, are subroutine calls, not back
69 references. PCRE supports relative numbers with this syntax (I don't think
70 Oniguruma does).
71
72 12. Previously, a group with a zero repeat such as (...){0} was completely
73 omitted from the compiled regex. However, this means that if the group
74 was called as a subroutine from elsewhere in the pattern, things went wrong
75 (an internal error was given). Such groups are now left in the compiled
76 pattern, with a new opcode that causes them to be skipped at execution
77 time.
78
79 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
80 to the way PCRE behaves:
81
82 (a) A lone ] character is dis-allowed (Perl treats it as data).
83
84 (b) A back reference to an unmatched subpattern matches an empty string
85 (Perl fails the current match path).
86
87 (c) A data ] in a character class must be notated as \] because if the
88 first data character in a class is ], it defines an empty class. (In
89 Perl it is not possible to have an empty class.) The empty class []
90 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
91 The negative empty class [^] matches any one character, independently
92 of the DOTALL setting.
93
94 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
95 non-existent subpattern following a character class starting with ']' and
96 containing () gave an internal compiling error instead of "reference to
97 non-existent subpattern". Fortunately, when the pattern did exist, the
98 compiled code was correct. (When scanning forwards to check for the
99 existencd of the subpattern, it was treating the data ']' as terminating
100 the class, so got the count wrong. When actually compiling, the reference
101 was subsequently set up correctly.)
102
103 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
104 it was being rejected as not supported by pcre_dfa_exec(), even though
105 other assertions are supported. I have made pcre_dfa_exec() support
106 (*FAIL).
107
108 16. The implementation of 13c above involved the invention of a new opcode,
109 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
110 cannot be changed at match time, I realized I could make a small
111 improvement to matching performance by compiling OP_ALLANY instead of
112 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
113 on the OP_ANY path.
114
115 17. Compiling pcretest on Windows with readline support failed without the
116 following two fixes: (1) Make the unistd.h include conditional on
117 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
118
119 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
120 ncurses library to be included for pcretest when ReadLine support is
121 requested, but also to allow for it to be overridden. This patch came from
122 Daniel Bergström.
123
124 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
125 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
126 any errors with the current Unicode tables. Thanks to Peter Kankowski for
127 spotting this.
128
129
130 Version 7.6 28-Jan-08
131 ---------------------
132
133 1. A character class containing a very large number of characters with
134 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
135 overflow.
136
137 2. Patch to cut out the "long long" test in pcrecpp_unittest when
138 HAVE_LONG_LONG is not defined.
139
140 3. Applied Christian Ehrlicher's patch to update the CMake build files to
141 bring them up to date and include new features. This patch includes:
142
143 - Fixed PH's badly added libz and libbz2 support.
144 - Fixed a problem with static linking.
145 - Added pcredemo. [But later removed - see 7 below.]
146 - Fixed dftables problem and added an option.
147 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
148 HAVE_LONG_LONG.
149 - Added readline support for pcretest.
150 - Added an listing of the option settings after cmake has run.
151
152 4. A user submitted a patch to Makefile that makes it easy to create
153 "pcre.dll" under mingw when using Configure/Make. I added stuff to
154 Makefile.am that cause it to include this special target, without
155 affecting anything else. Note that the same mingw target plus all
156 the other distribution libraries and programs are now supported
157 when configuring with CMake (see 6 below) instead of with
158 Configure/Make.
159
160 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
161 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
162 exported in the Windows port". It has not yet been confirmed that the patch
163 solves the problem, but it does no harm.
164
165 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
166 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
167 with CMake, and also correct the comment about stack recursion.
168
169 7. Remove the automatic building of pcredemo from the ./configure system and
170 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
171 of a program that users should build themselves after PCRE is installed, so
172 building it automatically is not really right. What is more, it gave
173 trouble in some build environments.
174
175 8. Further tidies to CMakeLists.txt from Sheri and Christian.
176
177
178 Version 7.5 10-Jan-08
179 ---------------------
180
181 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
182 values in parens when parsing an RE using the C++ wrapper."
183
184 2. Negative specials like \S did not work in character classes in UTF-8 mode.
185 Characters greater than 255 were excluded from the class instead of being
186 included.
187
188 3. The same bug as (2) above applied to negated POSIX classes such as
189 [:^space:].
190
191 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
192 defined or documented. It seems to have been a typo for PCRE_STATIC, so
193 I have changed it.
194
195 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
196 first named subpattern) and a construct such as (?&a) would reference the
197 first named subpattern whose name started with "a" (in other words, the
198 length check was missing). Both these problems are fixed. "Subpattern name
199 expected" is now given for (?&) (a zero-length name), and this patch also
200 makes it give the same error for \k'' (previously it complained that that
201 was a reference to a non-existent subpattern).
202
203 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
204 this is right because (?- can be followed by option settings as well as by
205 digits. I have, however, made the messages clearer.
206
207 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
208 than the number used in the conditional) now cause a compile-time error.
209 This is actually not compatible with Perl, which accepts such patterns, but
210 treats the conditional as always being FALSE (as PCRE used to), but it
211 seems to me that giving a diagnostic is better.
212
213 8. Change "alphameric" to the more common word "alphanumeric" in comments
214 and messages.
215
216 9. Fix two occurrences of "backslash" in comments that should have been
217 "backspace".
218
219 10. Remove two redundant lines of code that can never be obeyed (their function
220 was moved elsewhere).
221
222 11. The program that makes PCRE's Unicode character property table had a bug
223 which caused it to generate incorrect table entries for sequences of
224 characters that have the same character type, but are in different scripts.
225 It amalgamated them into a single range, with the script of the first of
226 them. In other words, some characters were in the wrong script. There were
227 thirteen such cases, affecting characters in the following ranges:
228
229 U+002b0 - U+002c1
230 U+0060c - U+0060d
231 U+0061e - U+00612
232 U+0064b - U+0065e
233 U+0074d - U+0076d
234 U+01800 - U+01805
235 U+01d00 - U+01d77
236 U+01d9b - U+01dbf
237 U+0200b - U+0200f
238 U+030fc - U+030fe
239 U+03260 - U+0327f
240 U+0fb46 - U+0fbb1
241 U+10450 - U+1049d
242
243 12. The -o option (show only the matching part of a line) for pcregrep was not
244 compatible with GNU grep in that, if there was more than one match in a
245 line, it showed only the first of them. It now behaves in the same way as
246 GNU grep.
247
248 13. If the -o and -v options were combined for pcregrep, it printed a blank
249 line for every non-matching line. GNU grep prints nothing, and pcregrep now
250 does the same. The return code can be used to tell if there were any
251 non-matching lines.
252
253 14. Added --file-offsets and --line-offsets to pcregrep.
254
255 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
256 infinitely looping recursion. The bug was that positive lookaheads were not
257 being skipped when checking for a possible empty match (negative lookaheads
258 and both kinds of lookbehind were skipped).
259
260 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
261 inclusion of <windows.h> to before rather than after the definition of
262 INVALID_FILE_ATTRIBUTES (patch from David Byron).
263
264 17. Specifying a possessive quantifier with a specific limit for a Unicode
265 character property caused pcre_compile() to compile bad code, which led at
266 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
267 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
268 caused the error; without that there was no problem.
269
270 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
271
272 19. Added --enable-pcretest-libreadline.
273
274 20. In pcrecpp.cc, the variable 'count' was incremented twice in
275 RE::GlobalReplace(). As a result, the number of replacements returned was
276 double what it should be. I removed one of the increments, but Craig sent a
277 later patch that removed the other one (the right fix) and added unit tests
278 that check the return values (which was not done before).
279
280 21. Several CMake things:
281
282 (1) Arranged that, when cmake is used on Unix, the libraries end up with
283 the names libpcre and libpcreposix, not just pcre and pcreposix.
284
285 (2) The above change means that pcretest and pcregrep are now correctly
286 linked with the newly-built libraries, not previously installed ones.
287
288 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
289
290 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
291 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
292 UTF-8 newline character). The key issue is that the pattern starts .*;
293 this means that the match must be either at the beginning, or after a
294 newline. The bug was in the code for advancing after a failed match and
295 checking that the new position followed a newline. It was not taking
296 account of UTF-8 characters correctly.
297
298 23. PCRE was behaving differently from Perl in the way it recognized POSIX
299 character classes. PCRE was not treating the sequence [:...:] as a
300 character class unless the ... were all letters. Perl, however, seems to
301 allow any characters between [: and :], though of course it rejects as
302 unknown any "names" that contain non-letters, because all the known class
303 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
304 for example, whereas PCRE did not - it did not recognize a POSIX character
305 class. This seemed a bit dangerous, so the code has been changed to be
306 closer to Perl. The behaviour is not identical to Perl, because PCRE will
307 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
308 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
309 Perl does, and where it didn't before.
310
311 24. Rewrite so as to remove the single use of %n from pcregrep because in some
312 Windows environments %n is disabled by default.
313
314
315 Version 7.4 21-Sep-07
316 ---------------------
317
318 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
319 means that a class such as [\s] counted as "explicit reference to CR or
320 LF". That isn't really right - the whole point of the change was to try to
321 help when there was an actual mention of one of the two characters. So now
322 the change happens only if \r or \n (or a literal CR or LF) character is
323 encountered.
324
325 2. The 32-bit options word was also used for 6 internal flags, but the numbers
326 of both had grown to the point where there were only 3 bits left.
327 Fortunately, there was spare space in the data structure, and so I have
328 moved the internal flags into a new 16-bit field to free up more option
329 bits.
330
331 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
332 but did not set the internal JCHANGED flag - either of these is enough to
333 control the way the "get" function works - but the PCRE_INFO_JCHANGED
334 facility is supposed to tell if (?J) was ever used, so now (?J) at the
335 start sets both bits.
336
337 4. Added options (at build time, compile time, exec time) to change \R from
338 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
339
340 5. doc/pcresyntax.html was missing from the distribution.
341
342 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
343 compatibility, even though it is no longer used.
344
345 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
346 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
347 windows.h file is present (where different names are used). [This was
348 reversed later after testing - see 16 below.]
349
350 8. Changed all #include <config.h> to #include "config.h". There were also
351 some further <pcre.h> cases that I changed to "pcre.h".
352
353 9. When pcregrep was used with the --colour option, it missed the line ending
354 sequence off the lines that it output.
355
356 10. It was pointed out to me that arrays of string pointers cause lots of
357 relocations when a shared library is dynamically loaded. A technique of
358 using a single long string with a table of offsets can drastically reduce
359 these. I have refactored PCRE in four places to do this. The result is
360 dramatic:
361
362 Originally: 290
363 After changing UCP table: 187
364 After changing error message table: 43
365 After changing table of "verbs" 36
366 After changing table of Posix names 22
367
368 Thanks to the folks working on Gregex for glib for this insight.
369
370 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
371 unicode-properties was also set.
372
373 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
374
375 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
376 checked only for CRLF.
377
378 14. Added casts to pcretest.c to avoid compiler warnings.
379
380 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
381
382 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
383 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
384 entirely. This removes changes made in 7 above.
385
386 17. The CMake files have been updated, and there is now more information about
387 building with CMake in the NON-UNIX-USE document.
388
389
390 Version 7.3 28-Aug-07
391 ---------------------
392
393 1. In the rejigging of the build system that eventually resulted in 7.1, the
394 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
395 brackets there is not right, since it causes compilers to look for an
396 installed pcre.h, not the version that is in the source that is being
397 compiled (which of course may be different). I have changed it back to:
398
399 #include "pcre.h"
400
401 I have a vague recollection that the change was concerned with compiling in
402 different directories, but in the new build system, that is taken care of
403 by the VPATH setting the Makefile.
404
405 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
406 when the subject happened to end in the byte 0x85 (e.g. if the last
407 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
408 characters but of course it shouldn't be taken as a newline when it is part
409 of another character. The bug was that, for an unlimited repeat of . in
410 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
411 characters when looking for a newline.
412
413 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
414
415 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
416 in debug output.
417
418 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
419 long printing in the pcrecpp unittest when running under MinGW.
420
421 6. ESC_K was left out of the EBCDIC table.
422
423 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
424 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
425 limit also applies to "virtual nesting" when a pattern is recursive, and in
426 this case 1000 isn't so big. I have been able to remove this limit at the
427 expense of backing off one optimization in certain circumstances. Normally,
428 when pcre_exec() would call its internal match() function recursively and
429 immediately return the result unconditionally, it uses a "tail recursion"
430 feature to save stack. However, when a subpattern that can match an empty
431 string has an unlimited repetition quantifier, it no longer makes this
432 optimization. That gives it a stack frame in which to save the data for
433 checking that an empty string has been matched. Previously this was taken
434 from the 1000-entry workspace that had been reserved. So now there is no
435 explicit limit, but more stack is used.
436
437 8. Applied Daniel's patches to solve problems with the import/export magic
438 syntax that is required for Windows, and which was going wrong for the
439 pcreposix and pcrecpp parts of the library. These were overlooked when this
440 problem was solved for the main library.
441
442 9. There were some crude static tests to avoid integer overflow when computing
443 the size of patterns that contain repeated groups with explicit upper
444 limits. As the maximum quantifier is 65535, the maximum group length was
445 set at 30,000 so that the product of these two numbers did not overflow a
446 32-bit integer. However, it turns out that people want to use groups that
447 are longer than 30,000 bytes (though not repeat them that many times).
448 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
449 made it possible to implement the integer overflow checks in a much more
450 dynamic way, which I have now done. The artificial limitation on group
451 length has been removed - we now have only the limit on the total length of
452 the compiled pattern, which depends on the LINK_SIZE setting.
453
454 10. Fixed a bug in the documentation for get/copy named substring when
455 duplicate names are permitted. If none of the named substrings are set, the
456 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
457 empty string.
458
459 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
460 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
461 because the ] is interpreted as the first data character and the
462 terminating ] is not found. PCRE has been made compatible with Perl in this
463 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
464 cause memory overwriting.
465
466 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
467 string has been matched (to stop an infinite loop). It was not recognizing
468 a conditional subpattern that could match an empty string if that
469 subpattern was within another subpattern. For example, it looped when
470 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
471 condition was not nested. This bug has been fixed.
472
473 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
474 past the start of the subject in the presence of bytes with the top bit
475 set, for example "\x8aBCD".
476
477 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
478 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
479
480 14. Optimized (?!) to (*FAIL).
481
482 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
483 This restricts code points to be within the range 0 to 0x10FFFF, excluding
484 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
485 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
486 does: it's just the validity check that is more restrictive.
487
488 16. Inserted checks for integer overflows during escape sequence (backslash)
489 processing, and also fixed erroneous offset values for syntax errors during
490 backslash processing.
491
492 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
493 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
494
495 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
496 caused an overrun.
497
498 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
499 something other than just ASCII characters) inside a group that had an
500 unlimited repeat caused a loop at compile time (while checking to see
501 whether the group could match an empty string).
502
503 20. Debugging a pattern containing \p or \P could cause a crash. For example,
504 [\P{Any}] did so. (Error in the code for printing property names.)
505
506 21. An orphan \E inside a character class could cause a crash.
507
508 22. A repeated capturing bracket such as (A)? could cause a wild memory
509 reference during compilation.
510
511 23. There are several functions in pcre_compile() that scan along a compiled
512 expression for various reasons (e.g. to see if it's fixed length for look
513 behind). There were bugs in these functions when a repeated \p or \P was
514 present in the pattern. These operators have additional parameters compared
515 with \d, etc, and these were not being taken into account when moving along
516 the compiled data. Specifically:
517
518 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
519 length.
520
521 (b) An item such as \pL+ within a repeated group could cause crashes or
522 loops.
523
524 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
525 "reference to non-existent subpattern" error.
526
527 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
528
529 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
530 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
531
532 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
533
534 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
535 character were causing crashes (broken optimization).
536
537 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
538 \p or \P) caused a compile-time loop.
539
540 28. More problems have arisen in unanchored patterns when CRLF is a valid line
541 break. For example, the unstudied pattern [\r\n]A does not match the string
542 "\r\nA" because change 7.0/46 below moves the current point on by two
543 characters after failing to match at the start. However, the pattern \nA
544 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
545 the same is true. There doesn't seem any very clean way out of this, but
546 what I have chosen to do makes the common cases work: PCRE now takes note
547 of whether there can be an explicit match for \r or \n anywhere in the
548 pattern, and if so, 7.0/46 no longer applies. As part of this change,
549 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
550 pattern has explicit CR or LF references.
551
552 29. Added (*CR) etc for changing newline setting at start of pattern.
553
554
555 Version 7.2 19-Jun-07
556 ---------------------
557
558 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
559 which is apparently normally available under Windows.
560
561 2. Re-jig the pcregrep tests with different newline settings in an attempt
562 to make them independent of the local environment's newline setting.
563
564 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
565
566 4. Some of the "internals" tests were previously cut out when the link size
567 was not 2, because the output contained actual offsets. The recent new
568 "Z" feature of pcretest means that these can be cut out, making the tests
569 usable with all link sizes.
570
571 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
572 stack recursion. This gives a massive performance boost under BSD, but just
573 a small improvement under Linux. However, it saves one field in the frame
574 in all cases.
575
576 6. Added more features from the forthcoming Perl 5.10:
577
578 (a) (?-n) (where n is a string of digits) is a relative subroutine or
579 recursion call. It refers to the nth most recently opened parentheses.
580
581 (b) (?+n) is also a relative subroutine call; it refers to the nth next
582 to be opened parentheses.
583
584 (c) Conditions that refer to capturing parentheses can be specified
585 relatively, for example, (?(-2)... or (?(+3)...
586
587 (d) \K resets the start of the current match so that everything before
588 is not part of it.
589
590 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
591
592 (f) \g{name} is another synonym - part of Perl 5.10's unification of
593 reference syntax.
594
595 (g) (?| introduces a group in which the numbering of parentheses in each
596 alternative starts with the same number.
597
598 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
599
600 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
601 PCRE_INFO_JCHANGED.
602
603 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
604 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
605 for detecting groups that can match an empty string.
606
607 9. A pattern with a very large number of alternatives (more than several
608 hundred) was running out of internal workspace during the pre-compile
609 phase, where pcre_compile() figures out how much memory will be needed. A
610 bit of new cunning has reduced the workspace needed for groups with
611 alternatives. The 1000-alternative test pattern now uses 12 bytes of
612 workspace instead of running out of the 4096 that are available.
613
614 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
615
616 11. Applied patch from Google to remove an optimization that didn't quite work.
617 The report of the bug said:
618
619 pcrecpp::RE("a*").FullMatch("aaa") matches, while
620 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
621 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
622
623 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
624 it matched the wrong number of bytes.
625
626
627 Version 7.1 24-Apr-07
628 ---------------------
629
630 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
631 that is more "standard", making use of automake and other Autotools. There
632 is some re-arrangement of the files and adjustment of comments consequent
633 on this.
634
635 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
636 for recursive directory scanning broke on some systems because the files
637 are not scanned in any specific order and on different systems the order
638 was different. A call to "sort" has been inserted into RunGrepTest for the
639 approprate test as a short-term fix. In the longer term there may be an
640 alternative.
641
642 3. I had an email from Eric Raymond about problems translating some of PCRE's
643 man pages to HTML (despite the fact that I distribute HTML pages, some
644 people do their own conversions for various reasons). The problems
645 concerned the use of low-level troff macros .br and .in. I have therefore
646 removed all such uses from the man pages (some were redundant, some could
647 be replaced by .nf/.fi pairs). The 132html script that I use to generate
648 HTML has been updated to handle .nf/.fi and to complain if it encounters
649 .br or .in.
650
651 4. Updated comments in configure.ac that get placed in config.h.in and also
652 arranged for config.h to be included in the distribution, with the name
653 config.h.generic, for the benefit of those who have to compile without
654 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
655
656 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
657 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
658 makevp.bat for the current PCRE, using the additional files
659 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
660
661 6. A Windows user reported a minor discrepancy with test 2, which turned out
662 to be caused by a trailing space on an input line that had got lost in his
663 copy. The trailing space was an accident, so I've just removed it.
664
665 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
666 that is needed.
667
668 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
669 as "const" (a) because they are and (b) because it helps the PHP
670 maintainers who have recently made a script to detect big data structures
671 in the php code that should be moved to the .rodata section. I remembered
672 to update Builducptable as well, so it won't revert if ucptable.h is ever
673 re-created.
674
675 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
676 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
677 order to be able to cut out the UTF-8 tables in the latter when UTF-8
678 support is not required. This saves 1.5-2K of code, which is important in
679 some applications.
680
681 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
682 so as not to refer to the tables, even though these functions will never be
683 called when UTF-8 support is disabled. Otherwise there are problems with a
684 shared library.
685
686 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
687
688 (a) It was defining its arguments as char * instead of void *.
689
690 (b) It was assuming that all moves were upwards in memory; this was true
691 a long time ago when I wrote it, but is no longer the case.
692
693 The emulated memove() is provided for those environments that have neither
694 memmove() nor bcopy(). I didn't think anyone used it these days, but that
695 is clearly not the case, as these two bugs were recently reported.
696
697 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
698 and Detrail to create the HTML documentation, the .txt form of the man
699 pages, and it removes trailing spaces from listed files. It also creates
700 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
701 case, it wraps all the #defines with #ifndefs. This script should be run
702 before "make dist".
703
704 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
705 with Unicode property support.
706
707 (a) For a maximizing quantifier, if the two different cases of the
708 character were of different lengths in their UTF-8 codings (there are
709 some cases like this - I found 11), and the matching function had to
710 back up over a mixture of the two cases, it incorrectly assumed they
711 were both the same length.
712
713 (b) When PCRE was configured to use the heap rather than the stack for
714 recursion during matching, it was not correctly preserving the data for
715 the other case of a UTF-8 character when checking ahead for a match
716 while processing a minimizing repeat. If the check also involved
717 matching a wide character, but failed, corruption could cause an
718 erroneous result when trying to check for a repeat of the original
719 character.
720
721 13. Some tidying changes to the testing mechanism:
722
723 (a) The RunTest script now detects the internal link size and whether there
724 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
725 values substituted by "configure". (The RunGrepTest script already did
726 this for UTF-8.) The configure.ac script no longer substitutes the
727 relevant variables.
728
729 (b) The debugging options /B and /D in pcretest show the compiled bytecode
730 with length and offset values. This means that the output is different
731 for different internal link sizes. Test 2 is skipped for link sizes
732 other than 2 because of this, bypassing the problem. Unfortunately,
733 there was also a test in test 3 (the locale tests) that used /B and
734 failed for link sizes other than 2. Rather than cut the whole test out,
735 I have added a new /Z option to pcretest that replaces the length and
736 offset values with spaces. This is now used to make test 3 independent
737 of link size. (Test 2 will be tidied up later.)
738
739 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
740 segmentation fault instead of returning the appropriate error message.
741
742 15. In multiline mode when the newline sequence was set to "any", the pattern
743 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
744 This doesn't seem right; it now treats the CRLF combination as the line
745 ending, and so does not match in that case. It's only a pattern such as ^$
746 that would hit this one: something like ^ABC$ would have failed after \r
747 and then tried again after \r\n.
748
749 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
750 in an attempt to make files that differ only in their line terminators
751 compare equal. This works on Linux.
752
753 17. Under certain error circumstances pcregrep might try to free random memory
754 as it exited. This is now fixed, thanks to valgrind.
755
756 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
757 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
758 was because its rules for how to advance for /g after matching an empty
759 string at the end of a line did not allow for this case. They now check for
760 it specially.
761
762 20. pcretest is supposed to handle patterns and data of any length, by
763 extending its buffers when necessary. It was getting this wrong when the
764 buffer for a data line had to be extended.
765
766 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
767 CRLF as a newline sequence.
768
769 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
770 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
771 I have nevertheless tidied it up.
772
773 23. Added some casts to kill warnings from HP-UX ia64 compiler.
774
775 24. Added a man page for pcre-config.
776
777
778 Version 7.0 19-Dec-06
779 ---------------------
780
781 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
782 moving to gcc 4.1.1.
783
784 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
785 sys/time.h, which is documented as needed for this function. It doesn't
786 seem to matter on Linux, but it showed up on some releases of OS X.
787
788 3. It seems that there are systems where bytes whose values are greater than
789 127 match isprint() in the "C" locale. The "C" locale should be the
790 default when a C program starts up. In most systems, only ASCII printing
791 characters match isprint(). This difference caused the output from pcretest
792 to vary, making some of the tests fail. I have changed pcretest so that:
793
794 (a) When it is outputting text in the compiled version of a pattern, bytes
795 other than 32-126 are always shown as hex escapes.
796
797 (b) When it is outputting text that is a matched part of a subject string,
798 it does the same, unless a different locale has been set for the match
799 (using the /L modifier). In this case, it uses isprint() to decide.
800
801 4. Fixed a major bug that caused incorrect computation of the amount of memory
802 required for a compiled pattern when options that changed within the
803 pattern affected the logic of the preliminary scan that determines the
804 length. The relevant options are -x, and -i in UTF-8 mode. The result was
805 that the computed length was too small. The symptoms of this bug were
806 either the PCRE error "internal error: code overflow" from pcre_compile(),
807 or a glibc crash with a message such as "pcretest: free(): invalid next
808 size (fast)". Examples of patterns that provoked this bug (shown in
809 pcretest format) are:
810
811 /(?-x: )/x
812 /(?x)(?-x: \s*#\s*)/
813 /((?i)[\x{c0}])/8
814 /(?i:[\x{c0}])/8
815
816 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
817 is now done differently.
818
819 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
820 wrapper classes; (b) implement a new function in the C++ scanner that is
821 more efficient than the old way of doing things because it avoids levels of
822 recursion in the regex matching; (c) add a paragraph to the documentation
823 for the FullMatch() function.
824
825 6. The escape sequence \n was being treated as whatever was defined as
826 "newline". Not only was this contrary to the documentation, which states
827 that \n is character 10 (hex 0A), but it also went horribly wrong when
828 "newline" was defined as CRLF. This has been fixed.
829
830 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
831 was being set to -1 for the "end of line" case (supposedly a value that no
832 character can have). Though this value is never used (the check for end of
833 line is "zero bytes in current character"), it caused compiler complaints.
834 I've changed it to 0xffffffff.
835
836 8. In pcre_version.c, the version string was being built by a sequence of
837 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
838 string (as it is for production releases) called a macro with an empty
839 argument. The C standard says the result of this is undefined. The gcc
840 compiler treats it as an empty string (which was what was wanted) but it is
841 reported that Visual C gives an error. The source has been hacked around to
842 avoid this problem.
843
844 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
845 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
846 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
847 of them did).
848
849 10. Originally, pcretest opened its input and output without "b"; then I was
850 told that "b" was needed in some environments, so it was added for release
851 5.0 to both the input and output. (It makes no difference on Unix-like
852 systems.) Later I was told that it is wrong for the input on Windows. I've
853 now abstracted the modes into two macros, to make it easier to fiddle with
854 them, and removed "b" from the input mode under Windows.
855
856 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
857
858 12. Added -help and --help to pcretest as an official way of being reminded
859 of the options.
860
861 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
862 and pcrecpp.cc because they annoy compilers at high warning levels.
863
864 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
865
866 15. Fixed an occurrence of == in configure.ac that should have been = (shell
867 scripts are not C programs :-) and which was not noticed because it works
868 on Linux.
869
870 16. pcretest is supposed to handle any length of pattern and data line (as one
871 line or as a continued sequence of lines) by extending its input buffer if
872 necessary. This feature was broken for very long pattern lines, leading to
873 a string of junk being passed to pcre_compile() if the pattern was longer
874 than about 50K.
875
876 17. I have done a major re-factoring of the way pcre_compile() computes the
877 amount of memory needed for a compiled pattern. Previously, there was code
878 that made a preliminary scan of the pattern in order to do this. That was
879 OK when PCRE was new, but as the facilities have expanded, it has become
880 harder and harder to keep it in step with the real compile phase, and there
881 have been a number of bugs (see for example, 4 above). I have now found a
882 cunning way of running the real compile function in a "fake" mode that
883 enables it to compute how much memory it would need, while actually only
884 ever using a few hundred bytes of working memory and without too many
885 tests of the mode. This should make future maintenance and development
886 easier. A side effect of this work is that the limit of 200 on the nesting
887 depth of parentheses has been removed (though this was never a serious
888 limitation, I suspect). However, there is a downside: pcre_compile() now
889 runs more slowly than before (30% or more, depending on the pattern). I
890 hope this isn't a big issue. There is no effect on runtime performance.
891
892 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
893 newline (only possible for the last line of a file) and it was a
894 pattern that set a locale (followed by /Lsomething), pcretest crashed.
895
896 19. Added additional timing features to pcretest. (1) The -tm option now times
897 matching only, not compiling. (2) Both -t and -tm can be followed, as a
898 separate command line item, by a number that specifies the number of
899 repeats to use when timing. The default is 50000; this gives better
900 precision, but takes uncomfortably long for very large patterns.
901
902 20. Extended pcre_study() to be more clever in cases where a branch of a
903 subpattern has no definite first character. For example, (a*|b*)[cd] would
904 previously give no result from pcre_study(). Now it recognizes that the
905 first character must be a, b, c, or d.
906
907 21. There was an incorrect error "recursive call could loop indefinitely" if
908 a subpattern (or the entire pattern) that was being tested for matching an
909 empty string contained only one non-empty item after a nested subpattern.
910 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
911 incorrectly, because the \d was being skipped in the check.
912
913 22. The pcretest program now has a new pattern option /B and a command line
914 option -b, which is equivalent to adding /B to every pattern. This causes
915 it to show the compiled bytecode, without the additional information that
916 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
917 is the same as /B/I).
918
919 23. A new optimization is now able automatically to treat some sequences such
920 as a*b as a*+b. More specifically, if something simple (such as a character
921 or a simple class like \d) has an unlimited quantifier, and is followed by
922 something that cannot possibly match the quantified thing, the quantifier
923 is automatically "possessified".
924
925 24. A recursive reference to a subpattern whose number was greater than 39
926 went wrong under certain circumstances in UTF-8 mode. This bug could also
927 have affected the operation of pcre_study().
928
929 25. Realized that a little bit of performance could be had by replacing
930 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
931
932 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
933
934 27. Possessive quantifiers such as a++ were previously implemented by turning
935 them into atomic groups such as ($>a+). Now they have their own opcodes,
936 which improves performance. This includes the automatically created ones
937 from 23 above.
938
939 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
940 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
941 the first matched character to be a colon. This applied both to named and
942 numbered groups.
943
944 29. The ucpinternal.h header file was missing its idempotency #ifdef.
945
946 30. I was sent a "project" file called libpcre.a.dev which I understand makes
947 building PCRE on Windows easier, so I have included it in the distribution.
948
949 31. There is now a check in pcretest against a ridiculously large number being
950 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
951 loop, the loop is abandoned.
952
953 32. Forward references to subpatterns in conditions such as (?(2)...) where
954 subpattern 2 is defined later cause pcre_compile() to search forwards in
955 the pattern for the relevant set of parentheses. This search went wrong
956 when there were unescaped parentheses in a character class, parentheses
957 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
958
959 33. "Subroutine" calls and backreferences were previously restricted to
960 referencing subpatterns earlier in the regex. This restriction has now
961 been removed.
962
963 34. Added a number of extra features that are going to be in Perl 5.10. On the
964 whole, these are just syntactic alternatives for features that PCRE had
965 previously implemented using the Python syntax or my own invention. The
966 other formats are all retained for compatibility.
967
968 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
969 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
970 also .NET compatible.
971
972 (b) A recursion or subroutine call to a named group can now be defined as
973 (?&name) as well as (?P>name).
974
975 (c) A backreference to a named group can now be defined as \k<name> or
976 \k'name' as well as (?P=name). The new forms, as well as being in Perl
977 5.10, are also .NET compatible.
978
979 (d) A conditional reference to a named group can now use the syntax
980 (?(<name>) or (?('name') as well as (?(name).
981
982 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
983 groups (named and numbered) that are never evaluated inline, but can be
984 called as "subroutines" from elsewhere. In effect, the DEFINE condition
985 is always false. There may be only one alternative in such a group.
986
987 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
988 as the simple (?(R). The condition is true only if the most recent
989 recursion is that of the given number or name. It does not search out
990 through the entire recursion stack.
991
992 (g) The escape \gN or \g{N} has been added, where N is a positive or
993 negative number, specifying an absolute or relative reference.
994
995 35. Tidied to get rid of some further signed/unsigned compiler warnings and
996 some "unreachable code" warnings.
997
998 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
999 things, this adds five new scripts.
1000
1001 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
1002 There were also incompatibilities regarding the handling of \Q..\E inside
1003 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
1004 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
1005
1006 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
1007 matches an empty string, and forcibly breaks the loop. There were bugs in
1008 this code in non-simple cases. For a pattern such as ^(a()*)* matched
1009 against aaaa the result was just "a" rather than "aaaa", for example. Two
1010 separate and independent bugs (that affected different cases) have been
1011 fixed.
1012
1013 39. Refactored the code to abolish the use of different opcodes for small
1014 capturing bracket numbers. This is a tidy that I avoided doing when I
1015 removed the limit on the number of capturing brackets for 3.5 back in 2001.
1016 The new approach is not only tidier, it makes it possible to reduce the
1017 memory needed to fix the previous bug (38).
1018
1019 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
1020 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
1021 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
1022 mode.
1023
1024 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
1025 report.
1026
1027 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
1028 copy construction and assignment in the C++ wrapper.
1029
1030 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
1031 couple of bugs that could have given wrong results in the "--newline=crlf"
1032 case.
1033
1034 44. Added a number of casts and did some reorganization of signed/unsigned int
1035 variables following suggestions from Dair Grant. Also renamed the variable
1036 "this" as "item" because it is a C++ keyword.
1037
1038 45. Arranged for dftables to add
1039
1040 #include "pcre_internal.h"
1041
1042 to pcre_chartables.c because without it, gcc 4.x may remove the array
1043 definition from the final binary if PCRE is built into a static library and
1044 dead code stripping is activated.
1045
1046 46. For an unanchored pattern, if a match attempt fails at the start of a
1047 newline sequence, and the newline setting is CRLF or ANY, and the next two
1048 characters are CRLF, advance by two characters instead of one.
1049
1050
1051 Version 6.7 04-Jul-06
1052 ---------------------
1053
1054 1. In order to handle tests when input lines are enormously long, pcretest has
1055 been re-factored so that it automatically extends its buffers when
1056 necessary. The code is crude, but this _is_ just a test program. The
1057 default size has been increased from 32K to 50K.
1058
1059 2. The code in pcre_study() was using the value of the re argument before
1060 testing it for NULL. (Of course, in any sensible call of the function, it
1061 won't be NULL.)
1062
1063 3. The memmove() emulation function in pcre_internal.h, which is used on
1064 systems that lack both memmove() and bcopy() - that is, hardly ever -
1065 was missing a "static" storage class specifier.
1066
1067 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
1068 containing an extended class (one that cannot be represented by a bitmap
1069 because it contains high-valued characters or Unicode property items, e.g.
1070 [\pZ]). Almost always one would set UTF-8 mode when processing such a
1071 pattern, but PCRE should not loop if you do not (it no longer does).
1072 [Detail: two cases were found: (a) a repeated subpattern containing an
1073 extended class; (b) a recursive reference to a subpattern that followed a
1074 previous extended class. It wasn't skipping over the extended class
1075 correctly when UTF-8 mode was not set.]
1076
1077 5. A negated single-character class was not being recognized as fixed-length
1078 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
1079 compile error "lookbehind assertion is not fixed length".
1080
1081 6. The RunPerlTest auxiliary script was showing an unexpected difference
1082 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
1083 write a Perl script that can interpret lines of an input file either as
1084 byte characters or as UTF-8, which is what "perltest" was being required to
1085 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
1086 can't do is switch easily at run time between having the "use utf8;" pragma
1087 or not. In the end, I fudged it by using the RunPerlTest script to insert
1088 "use utf8;" explicitly for the UTF-8 tests.
1089
1090 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
1091 the end of the subject string, contrary to the documentation and to what
1092 Perl does. This was true of both matching functions. Now it matches only at
1093 the start of the subject and immediately after *internal* newlines.
1094
1095 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
1096 a pointer to an int instead of a pointer to an unsigned long int. This
1097 caused problems on 64-bit systems.
1098
1099 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
1100 instance of the 'standard' template library not being so standard".
1101
1102 10. There was no check on the number of named subpatterns nor the maximum
1103 length of a subpattern name. The product of these values is used to compute
1104 the size of the memory block for a compiled pattern. By supplying a very
1105 long subpattern name and a large number of named subpatterns, the size
1106 computation could be caused to overflow. This is now prevented by limiting
1107 the length of names to 32 characters, and the number of named subpatterns
1108 to 10,000.
1109
1110 11. Subpatterns that are repeated with specific counts have to be replicated in
1111 the compiled pattern. The size of memory for this was computed from the
1112 length of the subpattern and the repeat count. The latter is limited to
1113 65535, but there was no limit on the former, meaning that integer overflow
1114 could in principle occur. The compiled length of a repeated subpattern is
1115 now limited to 30,000 bytes in order to prevent this.
1116
1117 12. Added the optional facility to have named substrings with the same name.
1118
1119 13. Added the ability to use a named substring as a condition, using the
1120 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
1121 are numbers (not recommended). Forward references are permitted.
1122
1123 14. Added forward references in named backreferences (if you see what I mean).
1124
1125 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
1126 pattern could run off the end of the subject. For example, the pattern
1127 "(?s)(.{1,5})"8 did this with the subject "ab".
1128
1129 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
1130 PCRE_CASELESS was set when matching characters that were quantified with ?
1131 or *.
1132
1133 17. A character class other than a single negated character that had a minimum
1134 but no maximum quantifier - for example [ab]{6,} - was not handled
1135 correctly by pce_dfa_exec(). It would match only one character.
1136
1137 18. A valid (though odd) pattern that looked like a POSIX character
1138 class but used an invalid character after [ (for example [[,abc,]]) caused
1139 pcre_compile() to give the error "Failed: internal error: code overflow" or
1140 in some cases to crash with a glibc free() error. This could even happen if
1141 the pattern terminated after [[ but there just happened to be a sequence of
1142 letters, a binary zero, and a closing ] in the memory that followed.
1143
1144 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
1145 over the years. Originally (before any Unicode support), just the bottom 8
1146 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
1147 output from "man perlunicode" includes this:
1148
1149 The regular expression compiler produces polymorphic opcodes. That
1150 is, the pattern adapts to the data and automatically switches to
1151 the Unicode character scheme when presented with Unicode data--or
1152 instead uses a traditional byte scheme when presented with byte
1153 data.
1154
1155 Sadly, a wide octal escape does not cause a switch, and in a string with
1156 no other multibyte characters, these octal escapes are treated as before.
1157 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
1158 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
1159 Unicode string.
1160
1161 I have not perpetrated such confusion in PCRE. Up till now, it took just
1162 the bottom 8 bits, as in old Perl. I have now made octal escapes with
1163 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
1164 translate to the appropriate multibyte character.
1165
1166 29. Applied some refactoring to reduce the number of warnings from Microsoft
1167 and Borland compilers. This has included removing the fudge introduced
1168 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
1169 a warning about an unused variable.
1170
1171 21. PCRE has not included VT (character 0x0b) in the set of whitespace
1172 characters since release 4.0, because Perl (from release 5.004) does not.
1173 [Or at least, is documented not to: some releases seem to be in conflict
1174 with the documentation.] However, when a pattern was studied with
1175 pcre_study() and all its branches started with \s, PCRE still included VT
1176 as a possible starting character. Of course, this did no harm; it just
1177 caused an unnecessary match attempt.
1178
1179 22. Removed a now-redundant internal flag bit that recorded the fact that case
1180 dependency changed within the pattern. This was once needed for "required
1181 byte" processing, but is no longer used. This recovers a now-scarce options
1182 bit. Also moved the least significant internal flag bit to the most-
1183 significant bit of the word, which was not previously used (hangover from
1184 the days when it was an int rather than a uint) to free up another bit for
1185 the future.
1186
1187 23. Added support for CRLF line endings as well as CR and LF. As well as the
1188 default being selectable at build time, it can now be changed at runtime
1189 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
1190 specify that it is scanning data with non-default line endings.
1191
1192 24. Changed the definition of CXXLINK to make it agree with the definition of
1193 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
1194
1195 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
1196 recursions. This makes a big different to stack usage for some patterns.
1197
1198 26. If a subpattern containing a named recursion or subroutine reference such
1199 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
1200 the space required for the compiled pattern went wrong and gave too small a
1201 value. Depending on the environment, this could lead to "Failed: internal
1202 error: code overflow at offset 49" or "glibc detected double free or
1203 corruption" errors.
1204
1205 27. Applied patches from Google (a) to support the new newline modes and (b) to
1206 advance over multibyte UTF-8 characters in GlobalReplace.
1207
1208 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
1209 difference for some implementation of PCRE in some Windows version.
1210
1211 29. Added some extra testing facilities to pcretest:
1212
1213 \q<number> in a data line sets the "match limit" value
1214 \Q<number> in a data line sets the "match recursion limt" value
1215 -S <number> sets the stack size, where <number> is in megabytes
1216
1217 The -S option isn't available for Windows.
1218
1219
1220 Version 6.6 06-Feb-06
1221 ---------------------
1222
1223 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
1224 in pcreposix.h. I have copied the definition from pcre.h.
1225
1226 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
1227 because pcre.h is no longer a built file.
1228
1229 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
1230 not normally included in the compiled code.
1231
1232
1233 Version 6.5 01-Feb-06
1234 ---------------------
1235
1236 1. When using the partial match feature with pcre_dfa_exec(), it was not
1237 anchoring the second and subsequent partial matches at the new starting
1238 point. This could lead to incorrect results. For example, with the pattern
1239 /1234/, partially matching against "123" and then "a4" gave a match.
1240
1241 2. Changes to pcregrep:
1242
1243 (a) All non-match returns from pcre_exec() were being treated as failures
1244 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
1245 error message is output. Some extra information is given for the
1246 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
1247 probably the only errors that are likely to be caused by users (by
1248 specifying a regex that has nested indefinite repeats, for instance).
1249 If there are more than 20 of these errors, pcregrep is abandoned.
1250
1251 (b) A binary zero was treated as data while matching, but terminated the
1252 output line if it was written out. This has been fixed: binary zeroes
1253 are now no different to any other data bytes.
1254
1255 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
1256 used to set a locale for matching. The --locale=xxxx long option has
1257 been added (no short equivalent) to specify a locale explicitly on the
1258 pcregrep command, overriding the environment variables.
1259
1260 (d) When -B was used with -n, some line numbers in the output were one less
1261 than they should have been.
1262
1263 (e) Added the -o (--only-matching) option.
1264
1265 (f) If -A or -C was used with -c (count only), some lines of context were
1266 accidentally printed for the final match.
1267
1268 (g) Added the -H (--with-filename) option.
1269
1270 (h) The combination of options -rh failed to suppress file names for files
1271 that were found from directory arguments.
1272
1273 (i) Added the -D (--devices) and -d (--directories) options.
1274
1275 (j) Added the -F (--fixed-strings) option.
1276
1277 (k) Allow "-" to be used as a file name for -f as well as for a data file.
1278
1279 (l) Added the --colo(u)r option.
1280
1281 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
1282 is not present by default.
1283
1284 3. A nasty bug was discovered in the handling of recursive patterns, that is,
1285 items such as (?R) or (?1), when the recursion could match a number of
1286 alternatives. If it matched one of the alternatives, but subsequently,
1287 outside the recursion, there was a failure, the code tried to back up into
1288 the recursion. However, because of the way PCRE is implemented, this is not
1289 possible, and the result was an incorrect result from the match.
1290
1291 In order to prevent this happening, the specification of recursion has
1292 been changed so that all such subpatterns are automatically treated as
1293 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
1294
1295 4. I had overlooked the fact that, in some locales, there are characters for
1296 which isalpha() is true but neither isupper() nor islower() are true. In
1297 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
1298 and ordfeminine) are like this. This affected the treatment of \w and \W
1299 when they appeared in character classes, but not when they appeared outside
1300 a character class. The bit map for "word" characters is now created
1301 separately from the results of isalnum() instead of just taking it from the
1302 upper, lower, and digit maps. (Plus the underscore character, of course.)
1303
1304 5. The above bug also affected the handling of POSIX character classes such as
1305 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
1306 permanent tables. Instead, the bit maps for such a class were previously
1307 created as the appropriate unions of the upper, lower, and digit bitmaps.
1308 Now they are created by subtraction from the [[:word:]] class, which has
1309 its own bitmap.
1310
1311 6. The [[:blank:]] character class matches horizontal, but not vertical space.
1312 It is created by subtracting the vertical space characters (\x09, \x0a,
1313 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
1314 subtraction was done in the overall bitmap for a character class, meaning
1315 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
1316 be recognized. This bug has been fixed.
1317
1318 7. Patches from the folks at Google:
1319
1320 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
1321 real life, but is still worth protecting against".
1322
1323 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
1324 regular expressions".
1325
1326 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
1327 have it.
1328
1329 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
1330 "configure" and the latter not, in order to fix a problem somebody had
1331 with compiling the Arg class on HP-UX.
1332
1333 (e) Improve the error-handling of the C++ wrapper a little bit.
1334
1335 (f) New tests for checking recursion limiting.
1336
1337 8. The pcre_memmove() function, which is used only if the environment does not
1338 have a standard memmove() function (and is therefore rarely compiled),
1339 contained two bugs: (a) use of int instead of size_t, and (b) it was not
1340 returning a result (though PCRE never actually uses the result).
1341
1342 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
1343 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
1344 returned instead of calling malloc() with an overflowing number that would
1345 most likely cause subsequent chaos.
1346
1347 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
1348
1349 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
1350 with this option is matched, the nmatch and pmatch options of regexec() are
1351 ignored.
1352
1353 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
1354 provided in case anyone wants to the the POSIX interface with UTF-8
1355 strings.
1356
1357 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
1358 C++ linking (needed for some HP-UX environments).
1359
1360 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
1361 (unused parameter) and in the pcre_printint() function (omitted "default"
1362 switch label when the default is to do nothing).
1363
1364 15. Added some code to make it possible, when PCRE is compiled as a C++
1365 library, to replace subject pointers for pcre_exec() with a smart pointer
1366 class, thus making it possible to process discontinuous strings.
1367
1368 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
1369 much the same function. They were added by different people who were trying
1370 to make PCRE easy to compile on non-Unix systems. It has been suggested
1371 that PCRE_EXPORT be abolished now that there is more automatic apparatus
1372 for compiling on Windows systems. I have therefore replaced it with
1373 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
1374 defaults to "extern" for C or "extern C" for C++, which works fine on
1375 Unix-like systems. It is now possible to override the value of PCRE_DATA_
1376 SCOPE with something explicit in config.h. In addition:
1377
1378 (a) pcreposix.h still had just "extern" instead of either of these macros;
1379 I have replaced it with PCRE_DATA_SCOPE.
1380
1381 (b) Functions such as _pcre_xclass(), which are internal to the library,
1382 but external in the C sense, all had PCRE_EXPORT in their definitions.
1383 This is apparently wrong for the Windows case, so I have removed it.
1384 (It makes no difference on Unix-like systems.)
1385
1386 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
1387 of recursive calls to match(). This is different to MATCH_LIMIT because
1388 that limits the total number of calls to match(), not all of which increase
1389 the depth of recursion. Limiting the recursion depth limits the amount of
1390 stack (or heap if NO_RECURSE is set) that is used. The default can be set
1391 when PCRE is compiled, and changed at run time. A patch from Google adds
1392 this functionality to the C++ interface.
1393
1394 18. Changes to the handling of Unicode character properties:
1395
1396 (a) Updated the table to Unicode 4.1.0.
1397
1398 (b) Recognize characters that are not in the table as "Cn" (undefined).
1399
1400 (c) I revised the way the table is implemented to a much improved format
1401 which includes recognition of ranges. It now supports the ranges that
1402 are defined in UnicodeData.txt, and it also amalgamates other
1403 characters into ranges. This has reduced the number of entries in the
1404 table from around 16,000 to around 3,000, thus reducing its size
1405 considerably. I realized I did not need to use a tree structure after
1406 all - a binary chop search is just as efficient. Having reduced the
1407 number of entries, I extended their size from 6 bytes to 8 bytes to
1408 allow for more data.
1409
1410 (d) Added support for Unicode script names via properties such as \p{Han}.
1411
1412 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
1413 matching that character.
1414
1415 20. When matching a repeated Unicode property with a minimum greater than zero,
1416 (for example \pL{2,}), PCRE could look past the end of the subject if it
1417 reached it while seeking the minimum number of characters. This could
1418 happen only if some of the characters were more than one byte long, because
1419 there is a check for at least the minimum number of bytes.
1420
1421 21. Refactored the implementation of \p and \P so as to be more general, to
1422 allow for more different types of property in future. This has changed the
1423 compiled form incompatibly. Anybody with saved compiled patterns that use
1424 \p or \P will have to recompile them.
1425
1426 22. Added "Any" and "L&" to the supported property types.
1427
1428 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
1429 but give a compile time error if the value is greater than 0xff.
1430
1431 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
1432 accidentally not being installed or uninstalled.
1433
1434 25. The pcre.h file was built from pcre.h.in, but the only changes that were
1435 made were to insert the current release number. This seemed silly, because
1436 it made things harder for people building PCRE on systems that don't run
1437 "configure". I have turned pcre.h into a distributed file, no longer built
1438 by "configure", with the version identification directly included. There is
1439 no longer a pcre.h.in file.
1440
1441 However, this change necessitated a change to the pcre-config script as
1442 well. It is built from pcre-config.in, and one of the substitutions was the
1443 release number. I have updated configure.ac so that ./configure now finds
1444 the release number by grepping pcre.h.
1445
1446 26. Added the ability to run the tests under valgrind.
1447
1448
1449 Version 6.4 05-Sep-05
1450 ---------------------
1451
1452 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
1453 "--" to be printed when multiple files were scanned, even when none of the
1454 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
1455 consider it to be a bug, and have restored the previous behaviour.
1456
1457 2. A couple of code tidies to get rid of compiler warnings.
1458
1459 3. The pcretest program used to cheat by referring to symbols in the library
1460 whose names begin with _pcre_. These are internal symbols that are not
1461 really supposed to be visible externally, and in some environments it is
1462 possible to suppress them. The cheating is now confined to including
1463 certain files from the library's source, which is a bit cleaner.
1464
1465 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
1466 file's purpose clearer.
1467
1468 5. Reorganized pcre_ucp_findchar().
1469
1470
1471 Version 6.3 15-Aug-05
1472 ---------------------
1473
1474 1. The file libpcre.pc.in did not have general read permission in the tarball.
1475
1476 2. There were some problems when building without C++ support:
1477
1478 (a) If C++ support was not built, "make install" and "make test" still
1479 tried to test it.
1480
1481 (b) There were problems when the value of CXX was explicitly set. Some
1482 changes have been made to try to fix these, and ...
1483
1484 (c) --disable-cpp can now be used to explicitly disable C++ support.
1485
1486 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
1487 backslash in a target when C++ was disabled. This confuses some
1488 versions of "make", apparently. Using an intermediate variable solves
1489 this. (Same for CPP_LOBJ.)
1490
1491 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
1492 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
1493 necessary on certain architectures.
1494
1495 4. Added a setting of -export-symbols-regex to the link command to remove
1496 those symbols that are exported in the C sense, but actually are local
1497 within the library, and not documented. Their names all begin with
1498 "_pcre_". This is not a perfect job, because (a) we have to except some
1499 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
1500 available (and never for static libraries). I have made a note to try to
1501 find a way round (a) in the future.
1502
1503
1504 Version 6.2 01-Aug-05
1505 ---------------------
1506
1507 1. There was no test for integer overflow of quantifier values. A construction
1508 such as {1111111111111111} would give undefined results. What is worse, if
1509 a minimum quantifier for a parenthesized subpattern overflowed and became
1510 negative, the calculation of the memory size went wrong. This could have
1511 led to memory overwriting.
1512
1513 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
1514
1515 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
1516 operating environments where this matters.
1517
1518 4. Applied Giuseppe Maxia's patch to add additional features for controlling
1519 PCRE options from within the C++ wrapper.
1520
1521 5. Named capturing subpatterns were not being correctly counted when a pattern
1522 was compiled. This caused two problems: (a) If there were more than 100
1523 such subpatterns, the calculation of the memory needed for the whole
1524 compiled pattern went wrong, leading to an overflow error. (b) Numerical
1525 back references of the form \12, where the number was greater than 9, were
1526 not recognized as back references, even though there were sufficient
1527 previous subpatterns.
1528
1529 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
1530 versions of gcc, e.g. 2.95.4.
1531
1532
1533 Version 6.1 21-Jun-05
1534 ---------------------
1535
1536 1. There was one reference to the variable "posix" in pcretest.c that was not
1537 surrounded by "#if !defined NOPOSIX".
1538
1539 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
1540 the cross-check on the old pcre_info() function, for the benefit of the
1541 cut-down version of PCRE that is currently imported into Exim.
1542
1543 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
1544 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
1545 patterns that start (?i) but I don't think that matters. The use of (?i) is
1546 just an example; this all applies to the other options as well.
1547
1548 4. Since libtool seems to echo the compile commands it is issuing, the output
1549 from "make" can be reduced a bit by putting "@" in front of each libtool
1550 compile command.
1551
1552 5. Patch from the folks at Google for configure.in to be a bit more thorough
1553 in checking for a suitable C++ installation before trying to compile the
1554 C++ stuff. This should fix a reported problem when a compiler was present,
1555 but no suitable headers.
1556
1557 6. The man pages all had just "PCRE" as their title. I have changed them to
1558 be the relevant file name. I have also arranged that these names are
1559 retained in the file doc/pcre.txt, which is a concatenation in text format
1560 of all the man pages except the little individual ones for each function.
1561
1562 7. The NON-UNIX-USE file had not been updated for the different set of source
1563 files that come with release 6. I also added a few comments about the C++
1564 wrapper.
1565
1566
1567 Version 6.0 07-Jun-05
1568 ---------------------
1569
1570 1. Some minor internal re-organization to help with my DFA experiments.
1571
1572 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
1573 didn't matter for the library itself when fully configured, but did matter
1574 when compiling without UCP support, or within Exim, where the ucp files are
1575 not imported.
1576
1577 3. Refactoring of the library code to split up the various functions into
1578 different source modules. The addition of the new DFA matching code (see
1579 below) to a single monolithic source would have made it really too
1580 unwieldy, quite apart from causing all the code to be include in a
1581 statically linked application, when only some functions are used. This is
1582 relevant even without the DFA addition now that patterns can be compiled in
1583 one application and matched in another.
1584
1585 The downside of splitting up is that there have to be some external
1586 functions and data tables that are used internally in different modules of
1587 the library but which are not part of the API. These have all had their
1588 names changed to start with "_pcre_" so that they are unlikely to clash
1589 with other external names.
1590
1591 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
1592 a different (DFA) algorithm. Although it is slower than the original
1593 function, it does have some advantages for certain types of matching
1594 problem.
1595
1596 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
1597 including restarting after a partial match.
1598
1599 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
1600 defined when compiling for Windows was sent to me. I have put it into the
1601 code, though I have no means of testing or verifying it.
1602
1603 7. Added the pcre_refcount() auxiliary function.
1604
1605 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
1606 match before or at the first newline in the subject string. In pcretest,
1607 the /f option on a pattern can be used to set this.
1608
1609 9. A repeated \w when used in UTF-8 mode with characters greater than 256
1610 would behave wrongly. This has been present in PCRE since release 4.0.
1611
1612 10. A number of changes to the pcregrep command:
1613
1614 (a) Refactored how -x works; insert ^(...)$ instead of setting
1615 PCRE_ANCHORED and checking the length, in preparation for adding
1616 something similar for -w.
1617
1618 (b) Added the -w (match as a word) option.
1619
1620 (c) Refactored the way lines are read and buffered so as to have more
1621 than one at a time available.
1622
1623 (d) Implemented a pcregrep test script.
1624
1625 (e) Added the -M (multiline match) option. This allows patterns to match
1626 over several lines of the subject. The buffering ensures that at least
1627 8K, or the rest of the document (whichever is the shorter) is available
1628 for matching (and similarly the previous 8K for lookbehind assertions).
1629
1630 (f) Changed the --help output so that it now says
1631
1632 -w, --word-regex(p)
1633
1634 instead of two lines, one with "regex" and the other with "regexp"
1635 because that confused at least one person since the short forms are the
1636 same. (This required a bit of code, as the output is generated
1637 automatically from a table. It wasn't just a text change.)
1638
1639 (g) -- can be used to terminate pcregrep options if the next thing isn't an
1640 option but starts with a hyphen. Could be a pattern or a path name
1641 starting with a hyphen, for instance.
1642
1643 (h) "-" can be given as a file name to represent stdin.
1644
1645 (i) When file names are being printed, "(standard input)" is used for
1646 the standard input, for compatibility with GNU grep. Previously
1647 "<stdin>" was used.
1648
1649 (j) The option --label=xxx can be used to supply a name to be used for
1650 stdin when file names are being printed. There is no short form.
1651
1652 (k) Re-factored the options decoding logic because we are going to add
1653 two more options that take data. Such options can now be given in four
1654 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
1655
1656 (l) Added the -A, -B, and -C options for requesting that lines of context
1657 around matches be printed.
1658
1659 (m) Added the -L option to print the names of files that do not contain
1660 any matching lines, that is, the complement of -l.
1661
1662 (n) The return code is 2 if any file cannot be opened, but pcregrep does
1663 continue to scan other files.
1664
1665 (o) The -s option was incorrectly implemented. For compatibility with other
1666 greps, it now suppresses the error message for a non-existent or non-
1667 accessible file (but not the return code). There is a new option called
1668 -q that suppresses the output of matching lines, which was what -s was
1669 previously doing.
1670
1671 (p) Added --include and --exclude options to specify files for inclusion
1672 and exclusion when recursing.
1673
1674 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
1675 Hopefully, it now does.
1676
1677 12. Missing cast in pcre_study().
1678
1679 13. Added an "uninstall" target to the makefile.
1680
1681 14. Replaced "extern" in the function prototypes in Makefile.in with
1682 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
1683 world, but is set differently for Windows.
1684
1685 15. Added a second compiling function called pcre_compile2(). The only
1686 difference is that it has an extra argument, which is a pointer to an
1687 integer error code. When there is a compile-time failure, this is set
1688 non-zero, in addition to the error test pointer being set to point to an
1689 error message. The new argument may be NULL if no error number is required
1690 (but then you may as well call pcre_compile(), which is now just a
1691 wrapper). This facility is provided because some applications need a
1692 numeric error indication, but it has also enabled me to tidy up the way
1693 compile-time errors are handled in the POSIX wrapper.
1694
1695 16. Added VPATH=.libs to the makefile; this should help when building with one
1696 prefix path and installing with another. (Or so I'm told by someone who
1697 knows more about this stuff than I do.)
1698
1699 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
1700 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
1701 match everything, including newlines. This is not POSIX-compatible, but
1702 somebody wanted the feature. From pcretest it can be activated by using
1703 both the P and the s flags.
1704
1705 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
1706
1707 19. libpcre.pc was being incorrectly installed as executable.
1708
1709 20. A couple of places in pcretest check for end-of-line by looking for '\n';
1710 it now also looks for '\r' so that it will work unmodified on Windows.
1711
1712 21. Added Google's contributed C++ wrapper to the distribution.
1713
1714 22. Added some untidy missing memory free() calls in pcretest, to keep
1715 Electric Fence happy when testing.
1716
1717
1718
1719 Version 5.0 13-Sep-04
1720 ---------------------
1721
1722 1. Internal change: literal characters are no longer packed up into items
1723 containing multiple characters in a single byte-string. Each character
1724 is now matched using a separate opcode. However, there may be more than one
1725 byte in the character in UTF-8 mode.
1726
1727 2. The pcre_callout_block structure has two new fields: pattern_position and
1728 next_item_length. These contain the offset in the pattern to the next match
1729 item, and its length, respectively.
1730
1731 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
1732 insertion of callouts before each pattern item. Added the /C option to
1733 pcretest to make use of this.
1734
1735 4. On the advice of a Windows user, the lines
1736
1737 #if defined(_WIN32) || defined(WIN32)
1738 _setmode( _fileno( stdout ), 0x8000 );
1739 #endif /* defined(_WIN32) || defined(WIN32) */
1740
1741 have been added to the source of pcretest. This apparently does useful
1742 magic in relation to line terminators.
1743
1744 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
1745 for the benefit of those environments where the "b" makes a difference.
1746
1747 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
1748 to know about it. I have put a hack into configure.in that adds in code
1749 to set GCC=yes if CC=icc. This seems to end up at a point in the
1750 generated configure script that is early enough to affect the setting of
1751 compiler options, which is what is needed, but I have no means of testing
1752 whether it really works. (The user who reported this had patched the
1753 generated configure script, which of course I cannot do.)
1754
1755 LATER: After change 22 below (new libtool files), the configure script
1756 seems to know about icc (and also ecc). Therefore, I have commented out
1757 this hack in configure.in.
1758
1759 7. Added support for pkg-config (2 patches were sent in).
1760
1761 8. Negated POSIX character classes that used a combination of internal tables
1762 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
1763 [[:^ascii]]. Typically, they would match almost any characters. The other
1764 POSIX classes were not broken in this way.
1765
1766 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
1767 to find the match, as PCRE was deluded into thinking that the match had to
1768 start at the start point or following a newline. The same bug applied to
1769 patterns with negative forward assertions or any backward assertions
1770 preceding ".*" at the start, unless the pattern required a fixed first
1771 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
1772
1773 10. In UTF-8 mode, when moving forwards in the subject after a failed match
1774 starting at the last subject character, bytes beyond the end of the subject
1775 string were read.
1776
1777 11. Renamed the variable "class" as "classbits" to make life easier for C++
1778 users. (Previously there was a macro definition, but it apparently wasn't
1779 enough.)
1780
1781 12. Added the new field "tables" to the extra data so that tables can be passed
1782 in at exec time, or the internal tables can be re-selected. This allows
1783 a compiled regex to be saved and re-used at a later time by a different
1784 program that might have everything at different addresses.
1785
1786 13. Modified the pcre-config script so that, when run on Solaris, it shows a
1787 -R library as well as a -L library.
1788
1789 14. The debugging options of pcretest (-d on the command line or D on a
1790 pattern) showed incorrect output for anything following an extended class
1791 that contained multibyte characters and which was followed by a quantifier.
1792
1793 15. Added optional support for general category Unicode character properties
1794 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
1795 support. It adds about 90K to the size of the library. The meanings of the
1796 inbuilt class escapes such as \d and \s have NOT been changed.
1797
1798 16. Updated pcredemo.c to include calls to free() to release the memory for the
1799 compiled pattern.
1800
1801 17. The generated file chartables.c was being created in the source directory
1802 instead of in the building directory. This caused the build to fail if the
1803 source directory was different from the building directory, and was
1804 read-only.
1805
1806 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
1807 file. No doubt somebody will tell me if they don't make sense... Also added
1808 Dan Mooney's comments about building on OpenVMS.
1809
1810 19. Added support for partial matching via the PCRE_PARTIAL option for
1811 pcre_exec() and the \P data escape in pcretest.
1812
1813 20. Extended pcretest with 3 new pattern features:
1814
1815 (i) A pattern option of the form ">rest-of-line" causes pcretest to
1816 write the compiled pattern to the file whose name is "rest-of-line".
1817 This is a straight binary dump of the data, with the saved pointer to
1818 the character tables forced to be NULL. The study data, if any, is
1819 written too. After writing, pcretest reads a new pattern.
1820
1821 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
1822 compiled pattern from the given file. There must not be any
1823 occurrences of "<" in the file name (pretty unlikely); if there are,
1824 pcretest will instead treat the initial "<" as a pattern delimiter.
1825 After reading in the pattern, pcretest goes on to read data lines as
1826 usual.
1827
1828 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
1829 and 16-bit fields in a compiled pattern, to simulate a pattern that
1830 was compiled on a host of opposite endianness.
1831
1832 21. The pcre-exec() function can now cope with patterns that were compiled on
1833 hosts of opposite endianness, with this restriction:
1834
1835 As for any compiled expression that is saved and used later, the tables
1836 pointer field cannot be preserved; the extra_data field in the arguments
1837 to pcre_exec() should be used to pass in a tables address if a value
1838 other than the default internal tables were used at compile time.
1839
1840 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
1841 now diagnosed as an error. Previously, most of the time, a negative number
1842 would have been treated as zero, but if in addition "ovector" was passed as
1843 NULL, a crash could occur.
1844
1845 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
1846 new versions from the libtool 1.5 distribution (the last one is a copy of
1847 a file called libtool.m4). This seems to have fixed the need to patch
1848 "configure" to support Darwin 1.3 (which I used to do). However, I still
1849 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
1850 workstation).
1851
1852 24. Changed the PCRE licence to be the more standard "BSD" licence.
1853
1854
1855 Version 4.5 01-Dec-03
1856 ---------------------
1857
1858 1. There has been some re-arrangement of the code for the match() function so
1859 that it can be compiled in a version that does not call itself recursively.
1860 Instead, it keeps those local variables that need separate instances for
1861 each "recursion" in a frame on the heap, and gets/frees frames whenever it
1862 needs to "recurse". Keeping track of where control must go is done by means
1863 of setjmp/longjmp. The whole thing is implemented by a set of macros that
1864 hide most of the details from the main code, and operates only if
1865 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
1866 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
1867 operating.
1868
1869 To make it easier for callers to provide specially tailored get/free
1870 functions for this usage, two new functions, pcre_stack_malloc, and
1871 pcre_stack_free, are used. They are always called in strict stacking order,
1872 and the size of block requested is always the same.
1873
1874 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
1875 PCRE has been compiled to use the stack or the heap for recursion. The
1876 -C option of pcretest uses this to show which version is compiled.
1877
1878 A new data escape \S, is added to pcretest; it causes the amounts of store
1879 obtained and freed by both kinds of malloc/free at match time to be added
1880 to the output.
1881
1882 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
1883 what's available on my current Linux desktop machine.
1884
1885 3. When matching a UTF-8 string, the test for a valid string at the start has
1886 been extended. If start_offset is not zero, PCRE now checks that it points
1887 to a byte that is the start of a UTF-8 character. If not, it returns
1888 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
1889 this is necessary because there may be backward assertions in the pattern.
1890 When matching the same subject several times, it may save resources to use
1891 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
1892
1893 4. The code for checking the validity of UTF-8 strings has been tightened so
1894 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
1895 containing "overlong sequences".
1896
1897 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
1898 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
1899 should have been "&", but it just so happened that all the cases this let
1900 through by mistake were picked up later in the function.
1901
1902 6. I had used a variable called "isblank" - this is a C99 function, causing
1903 some compilers to warn. To avoid this, I renamed it (as "blankclass").
1904
1905 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
1906 prompting; (b) run "./pcretest /dev/null" at the start of the test script
1907 so the version is shown; (c) stop "make test" echoing "./RunTest".
1908
1909 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
1910
1911 9. The prototype for memmove() for systems that don't have it was using
1912 size_t, but the inclusion of the header that defines size_t was later. I've
1913 moved the #includes for the C headers earlier to avoid this.
1914
1915 10. Added some adjustments to the code to make it easier to compiler on certain
1916 special systems:
1917
1918 (a) Some "const" qualifiers were missing.
1919 (b) Added the macro EXPORT before all exported functions; by default this
1920 is defined to be empty.
1921 (c) Changed the dftables auxiliary program (that builds chartables.c) so
1922 that it reads its output file name as an argument instead of writing
1923 to the standard output and assuming this can be redirected.
1924
1925 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
1926 class containing characters with values greater than 255, PCRE compilation
1927 went into a loop.
1928
1929 12. A recursive reference to a subpattern that was within another subpattern
1930 that had a minimum quantifier of zero caused PCRE to crash. For example,
1931 (x(y(?2))z)? provoked this bug with a subject that got as far as the
1932 recursion. If the recursively-called subpattern itself had a zero repeat,
1933 that was OK.
1934
1935 13. In pcretest, the buffer for reading a data line was set at 30K, but the
1936 buffer into which it was copied (for escape processing) was still set at
1937 1024, so long lines caused crashes.
1938
1939 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
1940 "internal error: code overflow...". This applied to any character class
1941 that was followed by a possessive quantifier.
1942
1943 15. Modified the Makefile to add libpcre.la as a prerequisite for
1944 libpcreposix.la because I was told this is needed for a parallel build to
1945 work.
1946
1947 16. If a pattern that contained .* following optional items at the start was
1948 studied, the wrong optimizing data was generated, leading to matching
1949 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
1950 matching string must start with a or b or c. The correct conclusion for
1951 this pattern is that a match can start with any character.
1952
1953
1954 Version 4.4 13-Aug-03
1955 ---------------------
1956
1957 1. In UTF-8 mode, a character class containing characters with values between
1958 127 and 255 was not handled correctly if the compiled pattern was studied.
1959 In fixing this, I have also improved the studying algorithm for such
1960 classes (slightly).
1961
1962 2. Three internal functions had redundant arguments passed to them. Removal
1963 might give a very teeny performance improvement.
1964
1965 3. Documentation bug: the value of the capture_top field in a callout is *one
1966 more than* the number of the hightest numbered captured substring.
1967
1968 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
1969 in incorrectly linking with a previously installed version. They now link
1970 explicitly with libpcre.la.
1971
1972 5. configure.in no longer needs to recognize Cygwin specially.
1973
1974 6. A problem in pcre.in for Windows platforms is fixed.
1975
1976 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
1977 pcretest, it used to include the size of the study block as part of its
1978 output. Unfortunately, the structure contains a field that has a different
1979 size on different hardware architectures. This meant that the tests that
1980 showed this size failed. As the block is currently always of a fixed size,
1981 this information isn't actually particularly useful in pcretest output, so
1982 I have just removed it.
1983
1984 8. Three pre-processor statements accidentally did not start in column 1.
1985 Sadly, there are *still* compilers around that complain, even though
1986 standard C has not required this for well over a decade. Sigh.
1987
1988 9. In pcretest, the code for checking callouts passed small integers in the
1989 callout_data field, which is a void * field. However, some picky compilers
1990 complained about the casts involved for this on 64-bit systems. Now
1991 pcretest passes the address of the small integer instead, which should get
1992 rid of the warnings.
1993
1994 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
1995 both compile and run time, and gives an error if an invalid UTF-8 sequence
1996 is found. There is a option for disabling this check in cases where the
1997 string is known to be correct and/or the maximum performance is wanted.
1998
1999 11. In response to a bug report, I changed one line in Makefile.in from
2000
2001 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
2002 to
2003 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
2004
2005 to look similar to other lines, but I have no way of telling whether this
2006 is the right thing to do, as I do not use Windows. No doubt I'll get told
2007 if it's wrong...
2008
2009
2010 Version 4.3 21-May-03
2011 ---------------------
2012
2013 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
2014 Makefile.
2015
2016 2. Some refactoring to improve the quality of the code:
2017
2018 (i) The utf8_table... variables are now declared "const".
2019
2020 (ii) The code for \cx, which used the "case flipping" table to upper case
2021 lower case letters, now just substracts 32. This is ASCII-specific,
2022 but the whole concept of \cx is ASCII-specific, so it seems
2023 reasonable.
2024
2025 (iii) PCRE was using its character types table to recognize decimal and
2026 hexadecimal digits in the pattern. This is silly, because it handles
2027 only 0-9, a-f, and A-F, but the character types table is locale-
2028 specific, which means strange things might happen. A private
2029 table is now used for this - though it costs 256 bytes, a table is
2030 much faster than multiple explicit tests. Of course, the standard
2031 character types table is still used for matching digits in subject
2032 strings against \d.
2033
2034 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
2035 ending in _t are). So I've renamed it as ESC_tee.
2036
2037 3. The first argument for regexec() in the POSIX wrapper should have been
2038 defined as "const".
2039
2040 4. Changed pcretest to use malloc() for its buffers so that they can be
2041 Electric Fenced for debugging.
2042
2043 5. There were several places in the code where, in UTF-8 mode, PCRE would try
2044 to read one or more bytes before the start of the subject string. Often this
2045 had no effect on PCRE's behaviour, but in some circumstances it could
2046 provoke a segmentation fault.
2047
2048 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
2049 to try to read one or more bytes before the start of the subject string.
2050
2051 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
2052 UTF-8 support could misbehave in various ways if the subject string
2053 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
2054 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
2055 back over UTF-8 characters.)
2056
2057
2058 Version 4.2 14-Apr-03
2059 ---------------------
2060
2061 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
2062
2063 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
2064 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
2065 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
2066 [WIN_PREFIX]: new variable, "cyg" for Cygwin
2067 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
2068 and BUILD_EXEEXT
2069 Note: automatic setting of the BUILD variables is not yet working
2070 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
2071 compile-time but not at link-time
2072 [LINK]: use for linking executables only
2073 make different versions for Windows and non-Windows
2074 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
2075 libraries
2076 [LINK_FOR_BUILD]: new variable
2077 [OBJEXT]: use throughout
2078 [EXEEXT]: use throughout
2079 <winshared>: new target
2080 <wininstall>: new target
2081 <dftables.o>: use native compiler
2082 <dftables>: use native linker
2083 <install>: handle Windows platform correctly
2084 <clean>: ditto
2085 <check>: ditto
2086 copy DLL to top builddir before testing
2087
2088 As part of these changes, -no-undefined was removed again. This was reported
2089 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
2090 in any case.
2091
2092 3. Some tidies to get rid of compiler warnings:
2093
2094 . In the match_data structure, match_limit was an unsigned long int, whereas
2095 match_call_count was an int. I've made them both unsigned long ints.
2096
2097 . In pcretest the fact that a const uschar * doesn't automatically cast to
2098 a void * provoked a warning.
2099
2100 . Turning on some more compiler warnings threw up some "shadow" variables
2101 and a few more missing casts.
2102
2103 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
2104 option, a class that contained a single character with a value between 128
2105 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
2106
2107 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
2108 option, a class that contained several characters, but with at least one
2109 whose value was between 128 and 255 caused PCRE to crash.
2110
2111
2112 Version 4.1 12-Mar-03
2113 ---------------------
2114
2115 1. Compiling with gcc -pedantic found a couple of places where casts were
2116 needed, and a string in dftables.c that was longer than standard compilers are
2117 required to support.
2118
2119 2. Compiling with Sun's compiler found a few more places where the code could
2120 be tidied up in order to avoid warnings.
2121
2122 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
2123 first of these names is deprecated in the latest Autoconf in favour of the name
2124 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
2125 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
2126 analogy I have changed it to CFLAGS_FOR_BUILD.
2127
2128 4. Added -no-undefined to the linking command in the Makefile, because this is
2129 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
2130 linking step for the pcreposix library.
2131
2132 5. PCRE was failing to diagnose the case of two named groups with the same
2133 name.
2134
2135 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
2136 literal character that is needed in the subject for a match, and scans along to
2137 ensure that it is present before embarking on the full matching process. This
2138 saves time in cases of nested unlimited repeats that are never going to match.
2139 Problem: the scan can take a lot of time if the subject is very long (e.g.
2140 megabytes), thus penalizing straightforward matches. It is now done only if the
2141 amount of subject to be scanned is less than 1000 bytes.
2142
2143 7. A lesser problem with the same optimization is that it was recording the
2144 first character of an anchored pattern as "needed", thus provoking a search
2145 right along the subject, even when the first match of the pattern was going to
2146 fail. The "needed" character is now not set for anchored patterns, unless it
2147 follows something in the pattern that is of non-fixed length. Thus, it still
2148 fulfils its original purpose of finding quick non-matches in cases of nested
2149 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
2150
2151
2152 Version 4.0 17-Feb-03
2153 ---------------------
2154
2155 1. If a comment in an extended regex that started immediately after a meta-item
2156 extended to the end of string, PCRE compiled incorrect data. This could lead to
2157 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
2158
2159 2. Moved to autoconf 2.53 and libtool 1.4.2.
2160
2161 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
2162 the special perltest8 script is no longer needed - all the tests can be run
2163 from a single perltest script.
2164
2165 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
2166 by \s. It has now been removed in PCRE. This means it isn't recognized as
2167 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
2168 class [:space:] *does* include VT, thereby creating a mess.
2169
2170 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
2171 space and tab.
2172
2173 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
2174 its new features into the main test script, reducing the number of scripts.
2175
2176 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
2177 were backward compatible, and made the (?i) apply to the whole pattern, as if
2178 /i were given. Now it behaves more logically, and applies the option setting
2179 only to what follows. PCRE has been changed to follow suit. However, if it
2180 finds options settings right at the start of the pattern, it extracts them into
2181 the global options, as before. Thus, they show up in the info data.
2182
2183 8. Added support for the \Q...\E escape sequence. Characters in between are
2184 treated as literals. This is slightly different from Perl in that $ and @ are
2185 also handled as literals inside the quotes. In Perl, they will cause variable
2186 interpolation. Note the following examples:
2187
2188 Pattern PCRE matches Perl matches
2189
2190 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
2191 \Qabc\$xyz\E abc\$xyz abc\$xyz
2192 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
2193
2194 For compatibility with Perl, \Q...\E sequences are recognized inside character
2195 classes as well as outside them.
2196
2197 9. Re-organized 3 code statements in pcretest to avoid "overflow in
2198 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
2199 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
2200 signed/unsigned warnings.
2201
2202 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
2203 option for pcretest, so I've replaced it by a simple function that does just
2204 that job.
2205
2206 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
2207 "pcregrep -".
2208
2209 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
2210 Java package. This provides some syntactic sugar for simple cases of what my
2211 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
2212 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
2213 item, you can use this simplified notation. Note that only makes sense with
2214 greedy quantifiers. Consequently, the use of the possessive quantifier forces
2215 greediness, whatever the setting of the PCRE_UNGREEDY option.
2216
2217 13. A change of greediness default within a pattern was not taking effect at
2218 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
2219 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
2220 was abstracted outside.
2221
2222 14. PCRE now supports the \G assertion. It is true when the current matching
2223 position is at the start point of the match. This differs from \A when the
2224 starting offset is non-zero. Used with the /g option of pcretest (or similar
2225 code), it works in the same way as it does for Perl's /g option. If all
2226 alternatives of a regex begin with \G, the expression is anchored to the start
2227 match position, and the "anchored" flag is set in the compiled expression.
2228
2229 15. Some bugs concerning the handling of certain option changes within patterns
2230 have been fixed. These applied to options other than (?ims). For example,
2231 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
2232 been the other way round. Some of this was related to change 7 above.
2233
2234 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
2235 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
2236 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
2237 POSIX classes only within a class (e.g. /[[:alpha:]]/).
2238
2239 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
2240 mode. Unlike ".", it always matches newline, whatever the setting of
2241 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
2242 assertions. Perl allows it, but it doesn't (in general) work because it can't
2243 calculate the length of the lookbehind. At least, that's the case for Perl
2244 5.8.0 - I've been told they are going to document that it doesn't work in
2245 future.
2246
2247 18. Added an error diagnosis for escapes that PCRE does not support: these are
2248 \L, \l, \N, \P, \p, \U, \u, and \X.
2249
2250 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
2251 reading past the end of the pattern in cases such as /[abcd/.
2252
2253 20. PCRE was getting more memory than necessary for patterns with classes that
2254 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
2255
2256 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
2257 compiling PCRE for use with Virtual Pascal.
2258
2259 22. Small fix to the Makefile to make it work properly if the build is done
2260 outside the source tree.
2261
2262 23. Added a new extension: a condition to go with recursion. If a conditional
2263 subpattern starts with (?(R) the "true" branch is used if recursion has
2264 happened, whereas the "false" branch is used only at the top level.
2265
2266 24. When there was a very long string of literal characters (over 255 bytes
2267 without UTF support, over 250 bytes with UTF support), the computation of how
2268 much memory was required could be incorrect, leading to segfaults or other
2269 strange effects.
2270
2271 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
2272 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
2273 there was a subsequent back reference to those brackets. This meant that, for
2274 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
2275 possible to check for precisely this case. All we can do is abandon the
2276 optimization if .* occurs inside capturing brackets when there are any back
2277 references whatsoever. (See below for a better fix that came later.)
2278
2279 26. The handling of the optimization for finding the first character of a
2280 non-anchored pattern, and for finding a character that is required later in the
2281 match were failing in some cases. This didn't break the matching; it just
2282 failed to optimize when it could. The way this is done has been re-implemented.
2283
2284 27. Fixed typo in error message for invalid (?R item (it said "(?p").
2285
2286 28. Added a new feature that provides some of the functionality that Perl
2287 provides with (?{...}). The facility is termed a "callout". The way it is done
2288 in PCRE is for the caller to provide an optional function, by setting
2289 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
2290 global variable. By default it is unset, which disables all calling out. To get
2291 the function called, the regex must include (?C) at appropriate points. This
2292 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
2293 This provides a means of identifying different callout points. When PCRE
2294 reaches such a point in the regex, if pcre_callout has been set, the external
2295 function is called. It is provided with data in a structure called
2296 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
2297 matching continues; if it returns a non-zero value, the match at the current
2298 point fails. However, backtracking will occur if possible. [This was changed
2299 later and other features added - see item 49 below.]
2300
2301 29. pcretest is upgraded to test the callout functionality. It provides a
2302 callout function that displays information. By default, it shows the start of
2303 the match and the current position in the text. There are some new data escapes
2304 to vary what happens:
2305
2306 \C+ in addition, show current contents of captured substrings
2307 \C- do not supply a callout function
2308 \C!n return 1 when callout number n is reached
2309 \C!n!m return 1 when callout number n is reached for the mth time
2310
2311 30. If pcregrep was called with the -l option and just a single file name, it
2312 output "<stdin>" if a match was found, instead of the file name.
2313
2314 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
2315 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
2316 pcre_exec(). This saves a malloc/free per call. The default value of
2317 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
2318 when configuring.
2319
2320 32. The default maximum size of a compiled pattern is 64K. There have been a
2321 few cases of people hitting this limit. The code now uses macros to handle the
2322 storing of links as offsets within the compiled pattern. It defaults to 2-byte
2323 links, but this can be changed to 3 or 4 bytes by --with-link-size when
2324 configuring. Tests 2 and 5 work only with 2-byte links because they output
2325 debugging information about compiled patterns.
2326
2327 33. Internal code re-arrangements:
2328
2329 (a) Moved the debugging function for printing out a compiled regex into
2330 its own source file (printint.c) and used #include to pull it into
2331 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
2332 separate copies.
2333
2334 (b) Defined the list of op-code names for debugging as a macro in
2335 internal.h so that it is next to the definition of the opcodes.
2336
2337 (c) Defined a table of op-code lengths for simpler skipping along compiled
2338 code. This is again a macro in internal.h so that it is next to the
2339 definition of the opcodes.
2340
2341 34. Added support for recursive calls to individual subpatterns, along the
2342 lines of Robin Houston's patch (but implemented somewhat differently).
2343
2344 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
2345 allow it to read and process whole directories in Win32. This code was
2346 contributed by Lionel Fourquaux; it has not been tested by me.
2347
2348 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
2349 used to name a group. Names consist of alphanumerics and underscores, and must
2350 be unique. Back references use the syntax (?P=name) and recursive calls use
2351 (?P>name) which is a PCRE extension to the Python extension. Groups still have
2352 numbers. The function pcre_fullinfo() can be used after compilation to extract
2353 a name/number map. There are three relevant calls:
2354
2355 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
2356 PCRE_INFO_NAMECOUNT yields the number of entries
2357 PCRE_INFO_NAMETABLE yields a pointer to the map.
2358
2359 The map is a vector of fixed-size entries. The size of each entry depends on
2360 the length of the longest name used. The first two bytes of each entry are the
2361 group number, most significant byte first. There follows the corresponding
2362 name, zero terminated. The names are in alphabetical order.
2363
2364 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
2365 case instead of 255. Making it the same both with and without UTF-8 support
2366 means that the same test output works with both.
2367
2368 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
2369 calling malloc() with a zero argument.
2370
2371 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
2372 optimization. I've improved things by keeping a bitmap of backreferences with
2373 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
2374 fact referenced, the optimization can be applied. It is unlikely that a
2375 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
2376 the match to follow \n) will appear inside brackets with a number greater than
2377 31, but if it does, any back reference > 31 suppresses the optimization.
2378
2379 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
2380 of disabling numbered capturing parentheses. Any opening parenthesis that is
2381 not followed by ? behaves as if it were followed by ?: but named parentheses
2382 can still be used for capturing (and they will acquire numbers in the usual
2383 way).
2384
2385 41. Redesigned the return codes from the match() function into yes/no/error so
2386 that errors can be passed back from deep inside the nested calls. A malloc
2387 failure while inside a recursive subpattern call now causes the
2388 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
2389
2390 42. It is now possible to set a limit on the number of times the match()
2391 function is called in a call to pcre_exec(). This facility makes it possible to
2392 limit the amount of recursion and backtracking, though not in a directly
2393 obvious way, because the match() function is used in a number of different
2394 circumstances. The count starts from zero for each position in the subject
2395 string (for non-anchored patterns). The default limit is, for compatibility, a
2396 large number, namely 10 000 000. You can change this in two ways:
2397
2398 (a) When configuring PCRE before making, you can use --with-match-limit=n
2399 to set a default value for the compiled library.
2400
2401 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
2402 a different value is set. See 45 below.
2403
2404 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
2405
2406 43. Added a new function pcre_config(int, void *) to enable run-time extraction
2407 of things that can be changed at compile time. The first argument specifies
2408 what is wanted and the second points to where the information is to be placed.
2409 The current list of available information is:
2410
2411 PCRE_CONFIG_UTF8
2412
2413 The output is an integer that is set to one if UTF-8 support is available;
2414 otherwise it is set to zero.
2415
2416 PCRE_CONFIG_NEWLINE
2417
2418 The output is an integer that it set to the value of the code that is used for
2419 newline. It is either LF (10) or CR (13).
2420
2421 PCRE_CONFIG_LINK_SIZE
2422
2423 The output is an integer that contains the number of bytes used for internal
2424 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
2425
2426 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
2427
2428 The output is an integer that contains the threshold above which the POSIX
2429 interface uses malloc() for output vectors. See item 31 above.
2430
2431 PCRE_CONFIG_MATCH_LIMIT
2432
2433 The output is an unsigned integer that contains the default limit of the number
2434 of match() calls in a pcre_exec() execution. See 42 above.
2435
2436 44. pcretest has been upgraded by the addition of the -C option. This causes it
2437 to extract all the available output from the new pcre_config() function, and to
2438 output it. The program then exits immediately.
2439
2440 45. A need has arisen to pass over additional data with calls to pcre_exec() in
2441 order to support additional features. One way would have been to define
2442 pcre_exec2() (for example) with extra arguments, but this would not have been
2443 extensible, and would also have required all calls to the original function to
2444 be mapped to the new one. Instead, I have chosen to extend the mechanism that
2445 is used for passing in "extra" data from pcre_study().
2446
2447 The pcre_extra structure is now exposed and defined in pcre.h. It currently
2448 contains the following fields:
2449
2450 flags a bitmap indicating which of the following fields are set
2451 study_data opaque data from pcre_study()
2452 match_limit a way of specifying a limit on match() calls for a specific
2453 call to pcre_exec()
2454 callout_data data for callouts (see 49 below)
2455
2456 The flag bits are also defined in pcre.h, and are
2457
2458 PCRE_EXTRA_STUDY_DATA
2459 PCRE_EXTRA_MATCH_LIMIT
2460 PCRE_EXTRA_CALLOUT_DATA
2461
2462 The pcre_study() function now returns one of these new pcre_extra blocks, with
2463 the actual study data pointed to by the study_data field, and the
2464 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
2465 before. That is, this change is entirely upwards-compatible and requires no
2466 change to existing code.
2467
2468 If you want to pass in additional data to pcre_exec(), you can either place it
2469 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
2470 block.
2471
2472 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
2473 data string contains the escape sequence \M, pcretest calls pcre_exec() several
2474 times with different match limits, until it finds the minimum value needed for
2475 pcre_exec() to complete. The value is then output. This can be instructive; for
2476 most simple matches the number is quite small, but for pathological cases it
2477 gets very large very quickly.
2478
2479 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
2480 returns the size of the data block pointed to by the study_data field in a
2481 pcre_extra block, that is, the value that was passed as the argument to
2482 pcre_malloc() when PCRE was getting memory in which to place the information
2483 created by pcre_study(). The fourth argument should point to a size_t variable.
2484 pcretest has been extended so that this information is shown after a successful
2485 pcre_study() call when information about the compiled regex is being displayed.
2486
2487 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
2488 because what follows is always an absolute path. (Later: it turns out that this
2489 is more than cosmetic for MinGW, because it doesn't like empty path
2490 components.)
2491
2492 49. Some changes have been made to the callout feature (see 28 above):
2493
2494 (i) A callout function now has three choices for what it returns:
2495
2496 0 => success, carry on matching
2497 > 0 => failure at this point, but backtrack if possible
2498 < 0 => serious error, return this value from pcre_exec()
2499
2500 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
2501 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
2502 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
2503 use by callout functions. It will never be used by PCRE itself.
2504
2505 (ii) The pcre_extra structure (see 45 above) has a void * field called
2506 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
2507 pcre_callout_block structure has a field of the same name. The contents of
2508 the field passed in the pcre_extra structure are passed to the callout
2509 function in the corresponding field in the callout block. This makes it
2510 easier to use the same callout-containing regex from multiple threads. For
2511 testing, the pcretest program has a new data escape
2512
2513 \C*n pass the number n (may be negative) as callout_data
2514
2515 If the callout function in pcretest receives a non-zero value as
2516 callout_data, it returns that value.
2517
2518 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
2519 there were some redundant $(CFLAGS) in commands that are now specified as
2520 $(LINK), which already includes $(CFLAGS).
2521
2522 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
2523 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
2524 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
2525 one-byte characters throughout. Note that case-insensitive matching applies
2526 only to characters whose values are less than 256. PCRE doesn't support the
2527 notion of cases for higher-valued characters.
2528
2529 (i) A character class whose characters are all within 0-255 is handled as
2530 a bit map, and the map is inverted for negative classes. Previously, a
2531 character > 255 always failed to match such a class; however it should
2532 match if the class was a negative one (e.g. [^ab]). This has been fixed.
2533
2534 (ii) A negated character class with a single character < 255 is coded as
2535 "not this character" (OP_NOT). This wasn't working properly when the test
2536 character was multibyte, either singly or repeated.
2537
2538 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
2539 mode, for example: \x{100}{2,3}.
2540
2541 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
2542 singly or repeated) now correctly test multibyte characters. However,
2543 PCRE doesn't recognize any characters with values greater than 255 as
2544 digits, spaces, or word characters. Such characters always match \D, \S,
2545 and \W, and never match \d, \s, or \w.
2546
2547 (v) Classes may now contain characters and character ranges with values
2548 greater than 255. For example: [ab\x{100}-\x{400}].
2549
2550 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
2551 PCRE in UTF-8 mode.
2552
2553 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
2554 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
2555 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
2556 value.)
2557
2558 53. The single man page has become too large. I have therefore split it up into
2559 a number of separate man pages. These also give rise to individual HTML pages;
2560 these are now put in a separate directory, and there is an index.html page that
2561 lists them all. Some hyperlinking between the pages has been installed.
2562
2563 54. Added convenience functions for handling named capturing parentheses.
2564
2565 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
2566 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
2567 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
2568 are faulted.
2569
2570 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
2571 calling configure. These values are used when compiling the dftables.c program
2572 which is run to generate the source of the default character tables. They
2573 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
2574 you will need to set these values.
2575
2576 57. Updated the building process for Windows DLL, as provided by Fred Cox.
2577
2578
2579 Version 3.9 02-Jan-02
2580 ---------------------
2581
2582 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
2583
2584 2. If --disable-static was given, the building process failed when trying to
2585 build pcretest and pcregrep. (For some reason it was using libtool to compile
2586 them, which is not right, as they aren't part of the library.)
2587
2588
2589 Version 3.8 18-Dec-01
2590 ---------------------
2591
2592 1. The experimental UTF-8 code was completely screwed up. It was packing the
2593 bytes in the wrong order. How dumb can you get?
2594
2595
2596 Version 3.7 29-Oct-01
2597 ---------------------
2598
2599 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
2600 This caused pcretest, when used on the test data, to segfault. Unfortunately,
2601 this didn't happen under Solaris 8, where I normally test things.
2602
2603 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
2604 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
2605 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
2606 this fix an hour or so after the initial 3.7 release.)
2607
2608
2609 Version 3.6 23-Oct-01
2610 ---------------------
2611
2612 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
2613 offsets passed as NULL with zero offset count.
2614
2615 2. The config.guess and config.sub files had not been updated when I moved to
2616 the latest autoconf.
2617
2618
2619 Version 3.5 15-Aug-01
2620 ---------------------
2621
2622 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
2623 had been forgotten.
2624
2625 2. By using declared but undefined structures, we can avoid using "void"
2626 definitions in pcre.h while keeping the internal definitions of the structures
2627 private.
2628
2629 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
2630 user point of view, this means that both static and shared libraries are built
2631 by default, but this can be individually controlled. More of the work of
2632 handling this static/shared cases is now inside libtool instead of PCRE's make
2633 file.
2634
2635 4. The pcretest utility is now installed along with pcregrep because it is
2636 useful for users (to test regexs) and by doing this, it automatically gets
2637 relinked by libtool. The documentation has been turned into a man page, so
2638 there are now .1, .txt, and .html versions in /doc.
2639
2640 5. Upgrades to pcregrep:
2641 (i) Added long-form option names like gnu grep.
2642 (ii) Added --help to list all options with an explanatory phrase.
2643 (iii) Added -r, --recursive to recurse into sub-directories.
2644 (iv) Added -f, --file to read patterns from a file.
2645
2646 6. pcre_exec() was referring to its "code" argument before testing that
2647 argument for NULL (and giving an error if it was NULL).
2648
2649 7. Upgraded Makefile.in to allow for compiling in a different directory from
2650 the source directory.
2651
2652 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
2653 options bits, the pointer it was passed was to an int instead of to an unsigned
2654 long int. This mattered only on 64-bit systems.
2655
2656 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
2657 generated) instead of pcre.in, which it its source. Also made the same change
2658 in several of the .c files.
2659
2660 10. A new release of gcc defines printf() as a macro, which broke pcretest
2661 because it had an ifdef in the middle of a string argument for printf(). Fixed
2662 by using separate calls to printf().
2663
2664 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
2665 script, to force use of CR or LF instead of \n in the source. On non-Unix
2666 systems, the value can be set in config.h.
2667
2668 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
2669 absolute limit. Changed the text of the error message to make this clear, and
2670 likewise updated the man page.
2671
2672 13. The limit of 99 on the number of capturing subpatterns has been removed.
2673 The new limit is 65535, which I hope will not be a "real" limit.
2674
2675
2676 Version 3.4 22-Aug-00
2677 ---------------------
2678
2679 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
2680
2681 2. Diagnose condition (?(0) as an error instead of crashing on matching.
2682
2683
2684 Version 3.3 01-Aug-00
2685 ---------------------
2686
2687 1. If an octal character was given, but the value was greater than \377, it
2688 was not getting masked to the least significant bits, as documented. This could
2689 lead to crashes in some systems.
2690
2691 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
2692 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
2693
2694 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
2695 These just pass their arguments on to (pcre_free)(), but they are provided
2696 because some uses of PCRE bind it to non-C systems that can call its functions,
2697 but cannot call free() or pcre_free() directly.
2698
2699 4. Add "make test" as a synonym for "make check". Corrected some comments in
2700 the Makefile.
2701
2702 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
2703 Makefile.
2704
2705 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
2706 command called pgrep for grepping around the active processes.
2707
2708 7. Added the beginnings of support for UTF-8 character strings.
2709
2710 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
2711 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
2712 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
2713 out for the ar command.)
2714
2715
2716 Version 3.2 12-May-00
2717 ---------------------
2718
2719 This is purely a bug fixing release.
2720
2721 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
2722 of ZA. This was just one example of several cases that could provoke this bug,
2723 which was introduced by change 9 of version 2.00. The code for breaking
2724 infinite loops after an iteration that matches an empty string was't working
2725 correctly.
2726
2727 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
2728 when matched against abbab (for example). After matching an empty string, it
2729 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
2730 caused it to match further down the string than it should.
2731
2732 3. The code contained an inclusion of sys/types.h. It isn't clear why this
2733 was there because it doesn't seem to be needed, and it causes trouble on some
2734 systems, as it is not a Standard C header. It has been removed.
2735
2736 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
2737 were reported on the Macintosh. The changes were from
2738
2739 while ((c = *(++ptr)) != 0 && c != '\n');
2740 to
2741 while ((c = *(++ptr)) != 0 && c != '\n') ;
2742
2743 Totally extraordinary, but if that's what it takes...
2744
2745 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
2746 available. Added HAVE_BCOPY and an autoconf test for it; if neither
2747 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
2748 assumes the way PCRE uses memmove() (always moving upwards).
2749
2750 6. PCRE is being used in one environment where strchr() is not available. There
2751 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
2752 faster code anyway.
2753
2754
2755 Version 3.1 09-Feb-00
2756 ---------------------
2757
2758 The only change in this release is the fixing of some bugs in Makefile.in for
2759 the "install" target:
2760
2761 (1) It was failing to install pcreposix.h.
2762
2763 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
2764
2765
2766 Version 3.0 01-Feb-00
2767 ---------------------
2768
2769 1. Add support for the /+ modifier to perltest (to output $` like it does in
2770 pcretest).
2771
2772 2. Add support for the /g modifier to perltest.
2773
2774 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
2775 matches null strings.
2776
2777 4. Fix perltest so that it doesn't do unwanted things when fed an empty
2778 pattern. Perl treats empty patterns specially - it reuses the most recent
2779 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
2780 effect.
2781
2782 5. The POSIX interface was broken in that it was just handing over the POSIX
2783 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
2784 required a bigger vector, with some working space on the end. This means that
2785 the POSIX wrapper now has to get and free some memory, and copy the results.
2786
2787 6. Added some simple autoconf support, placing the test data and the
2788 documentation in separate directories, re-organizing some of the
2789 information files, and making it build pcre-config (a GNU standard). Also added
2790 libtool support for building PCRE as a shared library, which is now the
2791 default.
2792
2793 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
2794 09 are not valid octal constants. Single digits will be used for minor values
2795 less than 10.
2796
2797 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
2798 existing programs that set these in the POSIX interface can use PCRE without
2799 modification.
2800
2801 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
2802 return all that pcre_info() returns, plus additional data. The pcre_info()
2803 function is retained for compatibility, but is considered to be obsolete.
2804
2805 10. Added experimental recursion feature (?R) to handle one common case that
2806 Perl 5.6 will be able to do with (?p{...}).
2807
2808 11. Added support for POSIX character classes like [:alpha:], which Perl is
2809 adopting.
2810
2811
2812 Version 2.08 31-Aug-99
2813 ----------------------
2814
2815 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
2816 trying to match at the startoffset position, but instead was moving forward to
2817 the next newline as if a previous match had failed.
2818
2819 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
2820 and could get into a loop if a null string was matched other than at the start
2821 of the subject.
2822
2823 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
2824 be distinguished at compile time, and for completeness also added PCRE_DATE.
2825
2826 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
2827 in GnuWin32 environments.
2828
2829
2830 Version 2.07 29-Jul-99
2831 ----------------------
2832
2833 1. The documentation is now supplied in plain text form and HTML as well as in
2834 the form of man page sources.
2835
2836 2. C++ compilers don't like assigning (void *) values to other pointer types.
2837 In particular this affects malloc(). Although there is no problem in Standard
2838 C, I've put in casts to keep C++ compilers happy.
2839
2840 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
2841 should be (const char *).
2842
2843 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
2844 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
2845 However, I haven't made this a standard facility. The documentation doesn't
2846 mention it, and the Makefile doesn't support it.
2847
2848 5. The Makefile now contains an "install" target, with editable destinations at
2849 the top of the file. The pcretest program is not installed.
2850
2851 6. pgrep -V now gives the PCRE version number and date.
2852
2853 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
2854 causing the entire string to be ignored, instead of just the last character.
2855
2856 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
2857 non-matching string, it can take a very, very long time, even for strings of
2858 quite modest length, because of the nested recursion. PCRE now does better in
2859 some of these cases. It does this by remembering the last required literal
2860 character in the pattern, and pre-searching the subject to ensure it is present
2861 before running the real match. In other words, it applies a heuristic to detect
2862 some types of certain failure quickly, and in the above example, if presented
2863 with a string that has no trailing " it gives "no match" very quickly.
2864
2865 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
2866 other alternatives are tried instead.
2867
2868
2869 Version 2.06 09-Jun-99
2870 ----------------------
2871
2872 1. Change pcretest's output for amount of store used to show just the code
2873 space, because the remainder (the data block) varies in size between 32-bit and
2874 64-bit systems.
2875
2876 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
2877 start matching at. This allows lookbehinds to work when searching for multiple
2878 occurrences in a string.
2879
2880 3. Added additional options to pcretest for testing multiple occurrences:
2881
2882 /+ outputs the rest of the string that follows a match
2883 /g loops for multiple occurrences, using the new startoffset argument
2884 /G loops for multiple occurrences by passing an incremented pointer
2885
2886 4. PCRE wasn't doing the "first character" optimization for patterns starting
2887 with \b or \B, though it was doing it for other lookbehind assertions. That is,
2888 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
2889 the letter 'x'. On long subject strings, this gives a significant speed-up.
2890
2891
2892 Version 2.05 21-Apr-99
2893 ----------------------
2894
2895 1. Changed the type of magic_number from int to long int so that it works
2896 properly on 16-bit systems.
2897
2898 2. Fixed a bug which caused patterns starting with .* not to work correctly
2899 when the subject string contained newline characters. PCRE was assuming
2900 anchoring for such patterns in all cases, which is not correct because .* will
2901 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
2902 DOTALL is set at top level; otherwise it knows that patterns starting with .*
2903 must be retried after every newline in the subject.
2904
2905
2906 Version 2.04 18-Feb-99
2907 ----------------------
2908
2909 1. For parenthesized subpatterns with repeats whose minimum was zero, the
2910 computation of the store needed to hold the pattern was incorrect (too large).
2911 If such patterns were nested a few deep, this could multiply and become a real
2912 problem.
2913
2914 2. Added /M option to pcretest to show the memory requirement of a specific
2915 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
2916
2917 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
2918 compiled in such a way that the backtracking after subsequent failure was
2919 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
2920 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
2921
2922
2923 Version 2.03 02-Feb-99
2924 ----------------------
2925
2926 1. Fixed typo and small mistake in man page.
2927
2928 2. Added 4th condition (GPL supersedes if conflict) and created separate
2929 LICENCE file containing the conditions.
2930
2931 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
2932 Perl, that is the internal \ allows the delimiter to be included in the
2933 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
2934 the final delimiter, add \ to the end of the pattern (to test the error).
2935
2936 4. Added the convenience functions for extracting substrings after a successful
2937 match. Updated pcretest to make it able to test these functions.
2938
2939
2940 Version 2.02 14-Jan-99
2941 ----------------------
2942
2943 1. Initialized the working variables associated with each extraction so that
2944 their saving and restoring doesn't refer to uninitialized store.
2945
2946 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
2947 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
2948 fix the problem.
2949
2950 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
2951 calls, and wasn't printing the correct value for compiling calls. Increased the
2952 default value of LOOPREPEAT, and the number of significant figures in the
2953 times.
2954
2955 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
2956
2957 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
2958 a building problem on Windows NT with a FAT file system.
2959
2960
2961 Version 2.01 21-Oct-98
2962 ----------------------
2963
2964 1. Changed the API for pcre_compile() to allow for the provision of a pointer
2965 to character tables built by pcre_maketables() in the current locale. If NULL
2966 is passed, the default tables are used.
2967
2968
2969 Version 2.00 24-Sep-98
2970 ----------------------
2971
2972 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
2973 it any more.
2974
2975 2. Allow quantification of (?>) groups, and make it work correctly.
2976
2977 3. The first character computation wasn't working for (?>) groups.
2978
2979 4. Correct the implementation of \Z (it is permitted to match on the \n at the
2980 end of the subject) and add 5.005's \z, which really does match only at the
2981 very end of the subject.
2982
2983 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
2984
2985 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
2986 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
2987 localized options. All options to pcre_study() were also removed.
2988
2989 7. Add other new features from 5.005:
2990
2991 $(?<= positive lookbehind
2992 $(?<! negative lookbehind
2993 (?imsx-imsx) added the unsetting capability
2994 such a setting is global if at outer level; local otherwise
2995 (?imsx-imsx:) non-capturing groups with option setting
2996 (?(cond)re|re) conditional pattern matching
2997
2998 A backreference to itself in a repeated group matches the previous
2999 captured string.
3000
3001 8. General tidying up of studying (both automatic and via "study")
3002 consequential on the addition of new assertions.
3003
3004 9. As in 5.005, unlimited repeated groups that could match an empty substring
3005 are no longer faulted at compile time. Instead, the loop is forcibly broken at
3006 runtime if any iteration does actually match an empty substring.
3007
3008 10. Include the RunTest script in the distribution.
3009
3010 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
3011 discrepancies, some of which were old and were also with respect to 5.004. They
3012 have now been fixed.
3013
3014
3015 Version 1.09 28-Apr-98
3016 ----------------------
3017
3018 1. A negated single character class followed by a quantifier with a minimum
3019 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
3020 program crashes, or just wrong answers. This did not apply to negated classes
3021 containing more than one character, or to minima other than one.
3022
3023
3024 Version 1.08 27-Mar-98
3025 ----------------------
3026
3027 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
3028
3029 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
3030 latter must appear before anything that relies on it in the pattern.
3031
3032
3033 Version 1.07 16-Feb-98
3034 ----------------------
3035
3036 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
3037 repeat of a potentially empty string).
3038
3039
3040 Version 1.06 23-Jan-98
3041 ----------------------
3042
3043 1. Added Markus Oberhumer's little patches for C++.
3044
3045 2. Literal strings longer than 255 characters were broken.
3046
3047
3048 Version 1.05 23-Dec-97
3049 ----------------------
3050
3051 1. Negated character classes containing more than one character were failing if
3052 PCRE_CASELESS was set at run time.
3053
3054
3055 Version 1.04 19-Dec-97
3056 ----------------------
3057
3058 1. Corrected the man page, where some "const" qualifiers had been omitted.
3059
3060 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
3061 input syntax.
3062
3063 3. Fixed memory leak which occurred when a regex with back references was
3064 matched with an offsets vector that wasn't big enough. The temporary memory
3065 that is used in this case wasn't being freed if the match failed.
3066
3067 4. Tidied pcretest to ensure it frees memory that it gets.
3068
3069 5. Temporary memory was being obtained in the case where the passed offsets
3070 vector was exactly big enough.
3071
3072 6. Corrected definition of offsetof() from change 5 below.
3073
3074 7. I had screwed up change 6 below and broken the rules for the use of
3075 setjmp(). Now fixed.
3076
3077
3078 Version 1.03 18-Dec-97
3079 ----------------------
3080
3081 1. A erroneous regex with a missing opening parenthesis was correctly
3082 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
3083 on some systems.
3084
3085 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
3086 it was reported that one broken compiler failed on the former because "code" is
3087 also an independent variable.
3088
3089 3. The erroneous regex a[]b caused an array overrun reference.
3090
3091 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
3092 fail on data ending with that character. (It was going on too far, and checking
3093 the next character, typically a binary zero.) This was specific to the
3094 optimized code for single-character negative classes.
3095
3096 5. Added a contributed patch from the TIN world which does the following:
3097
3098 + Add an undef for memmove, in case the the system defines a macro for it.
3099
3100 + Add a definition of offsetof(), in case there isn't one. (I don't know
3101 the reason behind this - offsetof() is part of the ANSI standard - but
3102 it does no harm).
3103
3104 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
3105 most of the places where whitespace preceded '#'. I have given up and
3106 allowed the remaining 2 cases to be at the margin.
3107
3108 + Rename some variables in pcre to eliminate shadowing. This seems very
3109 pedantic, but does no harm, of course.
3110
3111 6. Moved the call to setjmp() into its own function, to get rid of warnings
3112 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
3113
3114 7. Constructs such as \d{8,} were compiling into the equivalent of
3115 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
3116 outcome, but in this particular case used more store than had been allocated,
3117 which caused the bug to be discovered because it threw up an internal error.
3118
3119 8. The debugging code in both pcre and pcretest for outputting the compiled
3120 form of a regex was going wrong in the case of back references followed by
3121 curly-bracketed repeats.
3122
3123
3124 Version 1.02 12-Dec-97
3125 ----------------------
3126
3127 1. Typos in pcre.3 and comments in the source fixed.
3128
3129 2. Applied a contributed patch to get rid of places where it used to remove
3130 'const' from variables, and fixed some signed/unsigned and uninitialized
3131 variable warnings.
3132
3133 3. Added the "runtest" target to Makefile.
3134
3135 4. Set default compiler flag to -O2 rather than just -O.
3136
3137
3138 Version 1.01 19-Nov-97
3139 ----------------------
3140
3141 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
3142 like /([ab]*)*/, that is, for classes with more than one character in them.
3143
3144 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
3145 as /((?>a*))*/ (a PCRE_EXTRA facility).
3146
3147
3148 Version 1.00 18-Nov-97
3149 ----------------------
3150
3151 1. Added compile-time macros to support systems such as SunOS4 which don't have
3152 memmove() or strerror() but have other things that can be used instead.
3153
3154 2. Arranged that "make clean" removes the executables.
3155
3156
3157 Version 0.99 27-Oct-97
3158 ----------------------
3159
3160 1. Fixed bug in code for optimizing classes with only one character. It was
3161 initializing a 32-byte map regardless, which could cause it to run off the end
3162 of the memory it had got.
3163
3164 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
3165
3166
3167 Version 0.98 22-Oct-97
3168 ----------------------
3169
3170 1. Fixed bug in code for handling temporary memory usage when there are more
3171 back references than supplied space in the ovector. This could cause segfaults.
3172
3173
3174 Version 0.97 21-Oct-97
3175 ----------------------
3176
3177 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
3178
3179 2. Optimized negated single characters not to use a bit map.
3180
3181 3. Brought error texts together as macro definitions; clarified some of them;
3182 fixed one that was wrong - it said "range out of order" when it meant "invalid
3183 escape sequence".
3184
3185 4. Changed some char * arguments to const char *.
3186
3187 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
3188
3189 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
3190 pcretest.
3191
3192
3193 Version 0.96 16-Oct-97
3194 ----------------------
3195
3196 1. Added a simple "pgrep" utility to the distribution.
3197
3198 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
3199 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
3200 where "ddd" means "one or more decimal digits".
3201
3202 3. Fixed serious bug. If a pattern had a back reference, but the call to
3203 pcre_exec() didn't supply a large enough ovector to record the related
3204 identifying subpattern, the match always failed. PCRE now remembers the number
3205 of the largest back reference, and gets some temporary memory in which to save
3206 the offsets during matching if necessary, in order to ensure that
3207 backreferences always work.
3208
3209 4. Increased the compatibility with Perl in a number of ways:
3210
3211 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
3212 to request this handling. The option can be set at compile or exec time.
3213
3214 (b) $ matches before a terminating newline by default; an option
3215 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
3216 mode). The option can be set at compile or exec time.
3217
3218 (c) The handling of \ followed by a digit other than 0 is now supposed to be
3219 the same as Perl's. If the decimal number it represents is less than 10
3220 or there aren't that many previous left capturing parentheses, an octal
3221 escape is read. Inside a character class, it's always an octal escape,
3222 even if it is a single digit.
3223
3224 (d) An escaped but undefined alphabetic character is taken as a literal,
3225 unless PCRE_EXTRA is set. Currently this just reserves the remaining
3226 escapes.
3227
3228 (e) {0} is now permitted. (The previous item is removed from the compiled
3229 pattern).
3230
3231 5. Changed all the names of code files so that the basic parts are no longer
3232 than 10 characters, and abolished the teeny "globals.c" file.
3233
3234 6. Changed the handling of character classes; they are now done with a 32-byte
3235 bit map always.
3236
3237 7. Added the -d and /D options to pcretest to make it possible to look at the
3238 internals of compilation without having to recompile pcre.
3239
3240
3241 Version 0.95 23-Sep-97
3242 ----------------------
3243
3244 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
3245 \x20 at the start of a run of normal characters. These were being treated as
3246 real characters, instead of the source characters being re-checked.
3247
3248
3249 Version 0.94 18-Sep-97
3250 ----------------------
3251
3252 1. The functions are now thread-safe, with the caveat that the global variables
3253 containing pointers to malloc() and free() or alternative functions are the
3254 same for all threads.
3255
3256 2. Get pcre_study() to generate a bitmap of initial characters for non-
3257 anchored patterns when this is possible, and use it if passed to pcre_exec().
3258
3259
3260 Version 0.93 15-Sep-97
3261 ----------------------
3262
3263 1. /(b)|(:+)/ was computing an incorrect first character.
3264
3265 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
3266 but not actually doing anything yet.
3267
3268 3. Treat "-" characters in classes that cannot be part of ranges as literals,
3269 as Perl does (e.g. [-az] or [az-]).
3270
3271 4. Set the anchored flag if a branch starts with .* or .*? because that tests
3272 all possible positions.
3273
3274 5. Split up into different modules to avoid including unneeded functions in a
3275 compiled binary. However, compile and exec are still in one module. The "study"
3276 function is split off.
3277
3278 6. The character tables are now in a separate module whose source is generated
3279 by an auxiliary program - but can then be edited by hand if required. There are
3280 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
3281 toupper() in the code.
3282
3283 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
3284 make them global. Abolish the function for setting them, as the caller can now
3285 set them directly.
3286
3287
3288 Version 0.92 11-Sep-97
3289 ----------------------
3290
3291 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
3292 (e.g. /a{1,3}/) was broken (I mis-optimized it).
3293
3294 2. Caseless matching was not working in character classes if the characters in
3295 the pattern were in upper case.
3296
3297 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
3298
3299 4. Make PCRE_ANCHORED public and accept as a compile option.
3300
3301 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
3302 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
3303 pass them.
3304
3305 6. Give an error if bad option bits passed at compile or run time.
3306
3307 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
3308 pcretest to cause it to pass that flag.
3309
3310 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
3311 options, and the first character, if set.
3312
3313 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
3314
3315
3316 Version 0.91 10-Sep-97
3317 ----------------------
3318
3319 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
3320 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
3321
3322 2. PCRE was looping on encountering an indefinitely repeated back reference to
3323 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
3324 Perl does - treats the match as successful.
3325
3326 ****

Properties

Name Value
svn:eol-style native
svn:keywords "Author Date Id Revision Url"

  ViewVC Help
Powered by ViewVC 1.1.5