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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 380 - (show annotations) (download)
Tue Mar 3 12:32:47 2009 UTC (5 years, 1 month ago) by ph10
File size: 161808 byte(s)
The pattern (?(?=.*b)b|^) was incorrectly processed as requiring a match at the 
start of the subject or after a newline.

1 ChangeLog for PCRE
2 ------------------
3
4 Version 7.9 xx-xxx-09
5 ---------------------
6
7 1. When building with support for bzlib/zlib (pcregrep) and/or readline
8 (pcretest), all targets were linked against these libraries. This included
9 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
10 libraries. This caused unwanted dependencies to be created. This problem
11 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
12 pcretest is linked with readline.
13
14 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
15 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
16 moved outside it again, because FALSE and TRUE are already defined in AIX,
17 but BOOL is not.
18
19 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
20 PCRE_MATCH_LIMIT_RETURSION values as ints, when they should be long ints.
21
22 4. The pcregrep documentation said spaces were inserted as well as colons (or
23 hyphens) following file names and line numbers when outputting matching
24 lines. This is not true; no spaces are inserted. I have also clarified the
25 wording for the --colour (or --color) option.
26
27 5. In pcregrep, when --colour was used with -o, the list of matching strings
28 was not coloured; this is different to GNU grep, so I have changed it to be
29 the same.
30
31 6. When --colo(u)r was used in pcregrep, only the first matching substring in
32 each matching line was coloured. Now it goes on to look for further matches
33 of any of the test patterns, which is the same behaviour as GNU grep.
34
35 7. A pattern that could match an empty string could cause pcregrep to loop; it
36 doesn't make sense to accept an empty string match in pcregrep, so I have
37 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
38 seems to be how GNU grep behaves.
39
40 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
41 start or after a newline", because the conditional assertion was not being
42 skipped when checking for this condition.
43
44
45 Version 7.8 05-Sep-08
46 ---------------------
47
48 1. Replaced UCP searching code with optimized version as implemented for Ad
49 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
50 stage table and inline lookup instead of a function, giving speed ups of 2
51 to 5 times on some simple patterns that I tested. Permission was given to
52 distribute the MultiStage2.py script that generates the tables (it's not in
53 the tarball, but is in the Subversion repository).
54
55 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
56 scripts.
57
58 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
59 a group with a zero qualifier. The result of the study could be incorrect,
60 or the function might crash, depending on the pattern.
61
62 4. Caseless matching was not working for non-ASCII characters in back
63 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
64 It now works when Unicode Property Support is available.
65
66 5. In pcretest, an escape such as \x{de} in the data was always generating
67 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
68 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
69 truncation.
70
71 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
72
73 7. Added two (int) casts to pcregrep when printing the difference of two
74 pointers, in case they are 64-bit values.
75
76 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
77 test 2 if it fails.
78
79 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
80 and a #define of that name to empty if it is not externally set. This is to
81 allow users of MSVC to set it if necessary.
82
83 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
84 the convenience functions in the pcre_get.c source file.
85
86 11. An option change at the start of a pattern that had top-level alternatives
87 could cause overwriting and/or a crash. This command provoked a crash in
88 some environments:
89
90 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
91
92 This potential security problem was recorded as CVE-2008-2371.
93
94 12. For a pattern where the match had to start at the beginning or immediately
95 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
96 pcre_dfa_exec() could read past the end of the passed subject if there was
97 no match. To help with detecting such bugs (e.g. with valgrind), I modified
98 pcretest so that it places the subject at the end of its malloc-ed buffer.
99
100 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
101 exec() might read past the end of the data buffer in UTF-8 mode.
102
103 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
104 the data contained the byte 0x85 as part of a UTF-8 character within its
105 first line. This applied both to normal and DFA matching.
106
107 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
108 /^[^d]*?$/8 failed to match "abc".
109
110 16. Added a missing copyright notice to pcrecpp_internal.h.
111
112 17. Make it more clear in the documentation that values returned from
113 pcre_exec() in ovector are byte offsets, not character counts.
114
115 18. Tidied a few places to stop certain compilers from issuing warnings.
116
117 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
118 supplied by Stefan Weber. I made a further small update for 7.8 because
119 there is a change of source arrangements: the pcre_searchfuncs.c module is
120 replaced by pcre_ucd.c.
121
122
123 Version 7.7 07-May-08
124 ---------------------
125
126 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
127 a string to a long long, pretend we don't even have a long long." This is
128 done by checking for the strtoq, strtoll, and _strtoi64 functions.
129
130 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
131 pre-7.6 versions, which defined a global no_arg variable instead of putting
132 it in the RE class. (See also #8 below.)
133
134 3. Remove a line of dead code, identified by coverity and reported by Nuno
135 Lopes.
136
137 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
138
139 (1) The include/exclude patterns were being applied to the whole pathnames
140 of files, instead of just to the final components.
141
142 (2) If there was more than one level of directory, the subdirectories were
143 skipped unless they satisfied the include/exclude conditions. This is
144 inconsistent with GNU grep (and could even be seen as contrary to the
145 pcregrep specification - which I improved to make it absolutely clear).
146 The action now is always to scan all levels of directory, and just
147 apply the include/exclude patterns to regular files.
148
149 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
150 --exclude_dir in the tests to avoid scanning .svn directories.
151
152 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
153 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
154 doesn't support NULs in patterns.
155
156 7. Added some missing "const"s to declarations of static tables in
157 pcre_compile.c and pcre_dfa_exec.c.
158
159 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
160 caused by fix #2 above. (Subsequently also a second patch to fix the
161 first patch. And a third patch - this was a messy problem.)
162
163 9. Applied Craig's patch to remove the use of push_back().
164
165 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
166 matching function regexec().
167
168 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
169 which, however, unlike Perl's \g{...}, are subroutine calls, not back
170 references. PCRE supports relative numbers with this syntax (I don't think
171 Oniguruma does).
172
173 12. Previously, a group with a zero repeat such as (...){0} was completely
174 omitted from the compiled regex. However, this means that if the group
175 was called as a subroutine from elsewhere in the pattern, things went wrong
176 (an internal error was given). Such groups are now left in the compiled
177 pattern, with a new opcode that causes them to be skipped at execution
178 time.
179
180 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
181 to the way PCRE behaves:
182
183 (a) A lone ] character is dis-allowed (Perl treats it as data).
184
185 (b) A back reference to an unmatched subpattern matches an empty string
186 (Perl fails the current match path).
187
188 (c) A data ] in a character class must be notated as \] because if the
189 first data character in a class is ], it defines an empty class. (In
190 Perl it is not possible to have an empty class.) The empty class []
191 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
192 The negative empty class [^] matches any one character, independently
193 of the DOTALL setting.
194
195 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
196 non-existent subpattern following a character class starting with ']' and
197 containing () gave an internal compiling error instead of "reference to
198 non-existent subpattern". Fortunately, when the pattern did exist, the
199 compiled code was correct. (When scanning forwards to check for the
200 existencd of the subpattern, it was treating the data ']' as terminating
201 the class, so got the count wrong. When actually compiling, the reference
202 was subsequently set up correctly.)
203
204 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
205 it was being rejected as not supported by pcre_dfa_exec(), even though
206 other assertions are supported. I have made pcre_dfa_exec() support
207 (*FAIL).
208
209 16. The implementation of 13c above involved the invention of a new opcode,
210 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
211 cannot be changed at match time, I realized I could make a small
212 improvement to matching performance by compiling OP_ALLANY instead of
213 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
214 on the OP_ANY path.
215
216 17. Compiling pcretest on Windows with readline support failed without the
217 following two fixes: (1) Make the unistd.h include conditional on
218 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
219
220 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
221 ncurses library to be included for pcretest when ReadLine support is
222 requested, but also to allow for it to be overridden. This patch came from
223 Daniel Bergström.
224
225 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
226 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
227 any errors with the current Unicode tables. Thanks to Peter Kankowski for
228 spotting this.
229
230
231 Version 7.6 28-Jan-08
232 ---------------------
233
234 1. A character class containing a very large number of characters with
235 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
236 overflow.
237
238 2. Patch to cut out the "long long" test in pcrecpp_unittest when
239 HAVE_LONG_LONG is not defined.
240
241 3. Applied Christian Ehrlicher's patch to update the CMake build files to
242 bring them up to date and include new features. This patch includes:
243
244 - Fixed PH's badly added libz and libbz2 support.
245 - Fixed a problem with static linking.
246 - Added pcredemo. [But later removed - see 7 below.]
247 - Fixed dftables problem and added an option.
248 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
249 HAVE_LONG_LONG.
250 - Added readline support for pcretest.
251 - Added an listing of the option settings after cmake has run.
252
253 4. A user submitted a patch to Makefile that makes it easy to create
254 "pcre.dll" under mingw when using Configure/Make. I added stuff to
255 Makefile.am that cause it to include this special target, without
256 affecting anything else. Note that the same mingw target plus all
257 the other distribution libraries and programs are now supported
258 when configuring with CMake (see 6 below) instead of with
259 Configure/Make.
260
261 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
262 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
263 exported in the Windows port". It has not yet been confirmed that the patch
264 solves the problem, but it does no harm.
265
266 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
267 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
268 with CMake, and also correct the comment about stack recursion.
269
270 7. Remove the automatic building of pcredemo from the ./configure system and
271 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
272 of a program that users should build themselves after PCRE is installed, so
273 building it automatically is not really right. What is more, it gave
274 trouble in some build environments.
275
276 8. Further tidies to CMakeLists.txt from Sheri and Christian.
277
278
279 Version 7.5 10-Jan-08
280 ---------------------
281
282 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
283 values in parens when parsing an RE using the C++ wrapper."
284
285 2. Negative specials like \S did not work in character classes in UTF-8 mode.
286 Characters greater than 255 were excluded from the class instead of being
287 included.
288
289 3. The same bug as (2) above applied to negated POSIX classes such as
290 [:^space:].
291
292 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
293 defined or documented. It seems to have been a typo for PCRE_STATIC, so
294 I have changed it.
295
296 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
297 first named subpattern) and a construct such as (?&a) would reference the
298 first named subpattern whose name started with "a" (in other words, the
299 length check was missing). Both these problems are fixed. "Subpattern name
300 expected" is now given for (?&) (a zero-length name), and this patch also
301 makes it give the same error for \k'' (previously it complained that that
302 was a reference to a non-existent subpattern).
303
304 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
305 this is right because (?- can be followed by option settings as well as by
306 digits. I have, however, made the messages clearer.
307
308 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
309 than the number used in the conditional) now cause a compile-time error.
310 This is actually not compatible with Perl, which accepts such patterns, but
311 treats the conditional as always being FALSE (as PCRE used to), but it
312 seems to me that giving a diagnostic is better.
313
314 8. Change "alphameric" to the more common word "alphanumeric" in comments
315 and messages.
316
317 9. Fix two occurrences of "backslash" in comments that should have been
318 "backspace".
319
320 10. Remove two redundant lines of code that can never be obeyed (their function
321 was moved elsewhere).
322
323 11. The program that makes PCRE's Unicode character property table had a bug
324 which caused it to generate incorrect table entries for sequences of
325 characters that have the same character type, but are in different scripts.
326 It amalgamated them into a single range, with the script of the first of
327 them. In other words, some characters were in the wrong script. There were
328 thirteen such cases, affecting characters in the following ranges:
329
330 U+002b0 - U+002c1
331 U+0060c - U+0060d
332 U+0061e - U+00612
333 U+0064b - U+0065e
334 U+0074d - U+0076d
335 U+01800 - U+01805
336 U+01d00 - U+01d77
337 U+01d9b - U+01dbf
338 U+0200b - U+0200f
339 U+030fc - U+030fe
340 U+03260 - U+0327f
341 U+0fb46 - U+0fbb1
342 U+10450 - U+1049d
343
344 12. The -o option (show only the matching part of a line) for pcregrep was not
345 compatible with GNU grep in that, if there was more than one match in a
346 line, it showed only the first of them. It now behaves in the same way as
347 GNU grep.
348
349 13. If the -o and -v options were combined for pcregrep, it printed a blank
350 line for every non-matching line. GNU grep prints nothing, and pcregrep now
351 does the same. The return code can be used to tell if there were any
352 non-matching lines.
353
354 14. Added --file-offsets and --line-offsets to pcregrep.
355
356 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
357 infinitely looping recursion. The bug was that positive lookaheads were not
358 being skipped when checking for a possible empty match (negative lookaheads
359 and both kinds of lookbehind were skipped).
360
361 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
362 inclusion of <windows.h> to before rather than after the definition of
363 INVALID_FILE_ATTRIBUTES (patch from David Byron).
364
365 17. Specifying a possessive quantifier with a specific limit for a Unicode
366 character property caused pcre_compile() to compile bad code, which led at
367 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
368 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
369 caused the error; without that there was no problem.
370
371 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
372
373 19. Added --enable-pcretest-libreadline.
374
375 20. In pcrecpp.cc, the variable 'count' was incremented twice in
376 RE::GlobalReplace(). As a result, the number of replacements returned was
377 double what it should be. I removed one of the increments, but Craig sent a
378 later patch that removed the other one (the right fix) and added unit tests
379 that check the return values (which was not done before).
380
381 21. Several CMake things:
382
383 (1) Arranged that, when cmake is used on Unix, the libraries end up with
384 the names libpcre and libpcreposix, not just pcre and pcreposix.
385
386 (2) The above change means that pcretest and pcregrep are now correctly
387 linked with the newly-built libraries, not previously installed ones.
388
389 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
390
391 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
392 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
393 UTF-8 newline character). The key issue is that the pattern starts .*;
394 this means that the match must be either at the beginning, or after a
395 newline. The bug was in the code for advancing after a failed match and
396 checking that the new position followed a newline. It was not taking
397 account of UTF-8 characters correctly.
398
399 23. PCRE was behaving differently from Perl in the way it recognized POSIX
400 character classes. PCRE was not treating the sequence [:...:] as a
401 character class unless the ... were all letters. Perl, however, seems to
402 allow any characters between [: and :], though of course it rejects as
403 unknown any "names" that contain non-letters, because all the known class
404 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
405 for example, whereas PCRE did not - it did not recognize a POSIX character
406 class. This seemed a bit dangerous, so the code has been changed to be
407 closer to Perl. The behaviour is not identical to Perl, because PCRE will
408 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
409 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
410 Perl does, and where it didn't before.
411
412 24. Rewrite so as to remove the single use of %n from pcregrep because in some
413 Windows environments %n is disabled by default.
414
415
416 Version 7.4 21-Sep-07
417 ---------------------
418
419 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
420 means that a class such as [\s] counted as "explicit reference to CR or
421 LF". That isn't really right - the whole point of the change was to try to
422 help when there was an actual mention of one of the two characters. So now
423 the change happens only if \r or \n (or a literal CR or LF) character is
424 encountered.
425
426 2. The 32-bit options word was also used for 6 internal flags, but the numbers
427 of both had grown to the point where there were only 3 bits left.
428 Fortunately, there was spare space in the data structure, and so I have
429 moved the internal flags into a new 16-bit field to free up more option
430 bits.
431
432 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
433 but did not set the internal JCHANGED flag - either of these is enough to
434 control the way the "get" function works - but the PCRE_INFO_JCHANGED
435 facility is supposed to tell if (?J) was ever used, so now (?J) at the
436 start sets both bits.
437
438 4. Added options (at build time, compile time, exec time) to change \R from
439 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
440
441 5. doc/pcresyntax.html was missing from the distribution.
442
443 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
444 compatibility, even though it is no longer used.
445
446 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
447 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
448 windows.h file is present (where different names are used). [This was
449 reversed later after testing - see 16 below.]
450
451 8. Changed all #include <config.h> to #include "config.h". There were also
452 some further <pcre.h> cases that I changed to "pcre.h".
453
454 9. When pcregrep was used with the --colour option, it missed the line ending
455 sequence off the lines that it output.
456
457 10. It was pointed out to me that arrays of string pointers cause lots of
458 relocations when a shared library is dynamically loaded. A technique of
459 using a single long string with a table of offsets can drastically reduce
460 these. I have refactored PCRE in four places to do this. The result is
461 dramatic:
462
463 Originally: 290
464 After changing UCP table: 187
465 After changing error message table: 43
466 After changing table of "verbs" 36
467 After changing table of Posix names 22
468
469 Thanks to the folks working on Gregex for glib for this insight.
470
471 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
472 unicode-properties was also set.
473
474 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
475
476 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
477 checked only for CRLF.
478
479 14. Added casts to pcretest.c to avoid compiler warnings.
480
481 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
482
483 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
484 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
485 entirely. This removes changes made in 7 above.
486
487 17. The CMake files have been updated, and there is now more information about
488 building with CMake in the NON-UNIX-USE document.
489
490
491 Version 7.3 28-Aug-07
492 ---------------------
493
494 1. In the rejigging of the build system that eventually resulted in 7.1, the
495 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
496 brackets there is not right, since it causes compilers to look for an
497 installed pcre.h, not the version that is in the source that is being
498 compiled (which of course may be different). I have changed it back to:
499
500 #include "pcre.h"
501
502 I have a vague recollection that the change was concerned with compiling in
503 different directories, but in the new build system, that is taken care of
504 by the VPATH setting the Makefile.
505
506 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
507 when the subject happened to end in the byte 0x85 (e.g. if the last
508 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
509 characters but of course it shouldn't be taken as a newline when it is part
510 of another character. The bug was that, for an unlimited repeat of . in
511 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
512 characters when looking for a newline.
513
514 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
515
516 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
517 in debug output.
518
519 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
520 long printing in the pcrecpp unittest when running under MinGW.
521
522 6. ESC_K was left out of the EBCDIC table.
523
524 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
525 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
526 limit also applies to "virtual nesting" when a pattern is recursive, and in
527 this case 1000 isn't so big. I have been able to remove this limit at the
528 expense of backing off one optimization in certain circumstances. Normally,
529 when pcre_exec() would call its internal match() function recursively and
530 immediately return the result unconditionally, it uses a "tail recursion"
531 feature to save stack. However, when a subpattern that can match an empty
532 string has an unlimited repetition quantifier, it no longer makes this
533 optimization. That gives it a stack frame in which to save the data for
534 checking that an empty string has been matched. Previously this was taken
535 from the 1000-entry workspace that had been reserved. So now there is no
536 explicit limit, but more stack is used.
537
538 8. Applied Daniel's patches to solve problems with the import/export magic
539 syntax that is required for Windows, and which was going wrong for the
540 pcreposix and pcrecpp parts of the library. These were overlooked when this
541 problem was solved for the main library.
542
543 9. There were some crude static tests to avoid integer overflow when computing
544 the size of patterns that contain repeated groups with explicit upper
545 limits. As the maximum quantifier is 65535, the maximum group length was
546 set at 30,000 so that the product of these two numbers did not overflow a
547 32-bit integer. However, it turns out that people want to use groups that
548 are longer than 30,000 bytes (though not repeat them that many times).
549 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
550 made it possible to implement the integer overflow checks in a much more
551 dynamic way, which I have now done. The artificial limitation on group
552 length has been removed - we now have only the limit on the total length of
553 the compiled pattern, which depends on the LINK_SIZE setting.
554
555 10. Fixed a bug in the documentation for get/copy named substring when
556 duplicate names are permitted. If none of the named substrings are set, the
557 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
558 empty string.
559
560 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
561 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
562 because the ] is interpreted as the first data character and the
563 terminating ] is not found. PCRE has been made compatible with Perl in this
564 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
565 cause memory overwriting.
566
567 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
568 string has been matched (to stop an infinite loop). It was not recognizing
569 a conditional subpattern that could match an empty string if that
570 subpattern was within another subpattern. For example, it looped when
571 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
572 condition was not nested. This bug has been fixed.
573
574 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
575 past the start of the subject in the presence of bytes with the top bit
576 set, for example "\x8aBCD".
577
578 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
579 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
580
581 14. Optimized (?!) to (*FAIL).
582
583 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
584 This restricts code points to be within the range 0 to 0x10FFFF, excluding
585 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
586 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
587 does: it's just the validity check that is more restrictive.
588
589 16. Inserted checks for integer overflows during escape sequence (backslash)
590 processing, and also fixed erroneous offset values for syntax errors during
591 backslash processing.
592
593 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
594 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
595
596 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
597 caused an overrun.
598
599 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
600 something other than just ASCII characters) inside a group that had an
601 unlimited repeat caused a loop at compile time (while checking to see
602 whether the group could match an empty string).
603
604 20. Debugging a pattern containing \p or \P could cause a crash. For example,
605 [\P{Any}] did so. (Error in the code for printing property names.)
606
607 21. An orphan \E inside a character class could cause a crash.
608
609 22. A repeated capturing bracket such as (A)? could cause a wild memory
610 reference during compilation.
611
612 23. There are several functions in pcre_compile() that scan along a compiled
613 expression for various reasons (e.g. to see if it's fixed length for look
614 behind). There were bugs in these functions when a repeated \p or \P was
615 present in the pattern. These operators have additional parameters compared
616 with \d, etc, and these were not being taken into account when moving along
617 the compiled data. Specifically:
618
619 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
620 length.
621
622 (b) An item such as \pL+ within a repeated group could cause crashes or
623 loops.
624
625 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
626 "reference to non-existent subpattern" error.
627
628 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
629
630 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
631 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
632
633 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
634
635 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
636 character were causing crashes (broken optimization).
637
638 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
639 \p or \P) caused a compile-time loop.
640
641 28. More problems have arisen in unanchored patterns when CRLF is a valid line
642 break. For example, the unstudied pattern [\r\n]A does not match the string
643 "\r\nA" because change 7.0/46 below moves the current point on by two
644 characters after failing to match at the start. However, the pattern \nA
645 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
646 the same is true. There doesn't seem any very clean way out of this, but
647 what I have chosen to do makes the common cases work: PCRE now takes note
648 of whether there can be an explicit match for \r or \n anywhere in the
649 pattern, and if so, 7.0/46 no longer applies. As part of this change,
650 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
651 pattern has explicit CR or LF references.
652
653 29. Added (*CR) etc for changing newline setting at start of pattern.
654
655
656 Version 7.2 19-Jun-07
657 ---------------------
658
659 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
660 which is apparently normally available under Windows.
661
662 2. Re-jig the pcregrep tests with different newline settings in an attempt
663 to make them independent of the local environment's newline setting.
664
665 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
666
667 4. Some of the "internals" tests were previously cut out when the link size
668 was not 2, because the output contained actual offsets. The recent new
669 "Z" feature of pcretest means that these can be cut out, making the tests
670 usable with all link sizes.
671
672 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
673 stack recursion. This gives a massive performance boost under BSD, but just
674 a small improvement under Linux. However, it saves one field in the frame
675 in all cases.
676
677 6. Added more features from the forthcoming Perl 5.10:
678
679 (a) (?-n) (where n is a string of digits) is a relative subroutine or
680 recursion call. It refers to the nth most recently opened parentheses.
681
682 (b) (?+n) is also a relative subroutine call; it refers to the nth next
683 to be opened parentheses.
684
685 (c) Conditions that refer to capturing parentheses can be specified
686 relatively, for example, (?(-2)... or (?(+3)...
687
688 (d) \K resets the start of the current match so that everything before
689 is not part of it.
690
691 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
692
693 (f) \g{name} is another synonym - part of Perl 5.10's unification of
694 reference syntax.
695
696 (g) (?| introduces a group in which the numbering of parentheses in each
697 alternative starts with the same number.
698
699 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
700
701 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
702 PCRE_INFO_JCHANGED.
703
704 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
705 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
706 for detecting groups that can match an empty string.
707
708 9. A pattern with a very large number of alternatives (more than several
709 hundred) was running out of internal workspace during the pre-compile
710 phase, where pcre_compile() figures out how much memory will be needed. A
711 bit of new cunning has reduced the workspace needed for groups with
712 alternatives. The 1000-alternative test pattern now uses 12 bytes of
713 workspace instead of running out of the 4096 that are available.
714
715 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
716
717 11. Applied patch from Google to remove an optimization that didn't quite work.
718 The report of the bug said:
719
720 pcrecpp::RE("a*").FullMatch("aaa") matches, while
721 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
722 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
723
724 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
725 it matched the wrong number of bytes.
726
727
728 Version 7.1 24-Apr-07
729 ---------------------
730
731 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
732 that is more "standard", making use of automake and other Autotools. There
733 is some re-arrangement of the files and adjustment of comments consequent
734 on this.
735
736 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
737 for recursive directory scanning broke on some systems because the files
738 are not scanned in any specific order and on different systems the order
739 was different. A call to "sort" has been inserted into RunGrepTest for the
740 approprate test as a short-term fix. In the longer term there may be an
741 alternative.
742
743 3. I had an email from Eric Raymond about problems translating some of PCRE's
744 man pages to HTML (despite the fact that I distribute HTML pages, some
745 people do their own conversions for various reasons). The problems
746 concerned the use of low-level troff macros .br and .in. I have therefore
747 removed all such uses from the man pages (some were redundant, some could
748 be replaced by .nf/.fi pairs). The 132html script that I use to generate
749 HTML has been updated to handle .nf/.fi and to complain if it encounters
750 .br or .in.
751
752 4. Updated comments in configure.ac that get placed in config.h.in and also
753 arranged for config.h to be included in the distribution, with the name
754 config.h.generic, for the benefit of those who have to compile without
755 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
756
757 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
758 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
759 makevp.bat for the current PCRE, using the additional files
760 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
761
762 6. A Windows user reported a minor discrepancy with test 2, which turned out
763 to be caused by a trailing space on an input line that had got lost in his
764 copy. The trailing space was an accident, so I've just removed it.
765
766 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
767 that is needed.
768
769 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
770 as "const" (a) because they are and (b) because it helps the PHP
771 maintainers who have recently made a script to detect big data structures
772 in the php code that should be moved to the .rodata section. I remembered
773 to update Builducptable as well, so it won't revert if ucptable.h is ever
774 re-created.
775
776 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
777 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
778 order to be able to cut out the UTF-8 tables in the latter when UTF-8
779 support is not required. This saves 1.5-2K of code, which is important in
780 some applications.
781
782 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
783 so as not to refer to the tables, even though these functions will never be
784 called when UTF-8 support is disabled. Otherwise there are problems with a
785 shared library.
786
787 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
788
789 (a) It was defining its arguments as char * instead of void *.
790
791 (b) It was assuming that all moves were upwards in memory; this was true
792 a long time ago when I wrote it, but is no longer the case.
793
794 The emulated memove() is provided for those environments that have neither
795 memmove() nor bcopy(). I didn't think anyone used it these days, but that
796 is clearly not the case, as these two bugs were recently reported.
797
798 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
799 and Detrail to create the HTML documentation, the .txt form of the man
800 pages, and it removes trailing spaces from listed files. It also creates
801 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
802 case, it wraps all the #defines with #ifndefs. This script should be run
803 before "make dist".
804
805 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
806 with Unicode property support.
807
808 (a) For a maximizing quantifier, if the two different cases of the
809 character were of different lengths in their UTF-8 codings (there are
810 some cases like this - I found 11), and the matching function had to
811 back up over a mixture of the two cases, it incorrectly assumed they
812 were both the same length.
813
814 (b) When PCRE was configured to use the heap rather than the stack for
815 recursion during matching, it was not correctly preserving the data for
816 the other case of a UTF-8 character when checking ahead for a match
817 while processing a minimizing repeat. If the check also involved
818 matching a wide character, but failed, corruption could cause an
819 erroneous result when trying to check for a repeat of the original
820 character.
821
822 13. Some tidying changes to the testing mechanism:
823
824 (a) The RunTest script now detects the internal link size and whether there
825 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
826 values substituted by "configure". (The RunGrepTest script already did
827 this for UTF-8.) The configure.ac script no longer substitutes the
828 relevant variables.
829
830 (b) The debugging options /B and /D in pcretest show the compiled bytecode
831 with length and offset values. This means that the output is different
832 for different internal link sizes. Test 2 is skipped for link sizes
833 other than 2 because of this, bypassing the problem. Unfortunately,
834 there was also a test in test 3 (the locale tests) that used /B and
835 failed for link sizes other than 2. Rather than cut the whole test out,
836 I have added a new /Z option to pcretest that replaces the length and
837 offset values with spaces. This is now used to make test 3 independent
838 of link size. (Test 2 will be tidied up later.)
839
840 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
841 segmentation fault instead of returning the appropriate error message.
842
843 15. In multiline mode when the newline sequence was set to "any", the pattern
844 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
845 This doesn't seem right; it now treats the CRLF combination as the line
846 ending, and so does not match in that case. It's only a pattern such as ^$
847 that would hit this one: something like ^ABC$ would have failed after \r
848 and then tried again after \r\n.
849
850 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
851 in an attempt to make files that differ only in their line terminators
852 compare equal. This works on Linux.
853
854 17. Under certain error circumstances pcregrep might try to free random memory
855 as it exited. This is now fixed, thanks to valgrind.
856
857 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
858 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
859 was because its rules for how to advance for /g after matching an empty
860 string at the end of a line did not allow for this case. They now check for
861 it specially.
862
863 20. pcretest is supposed to handle patterns and data of any length, by
864 extending its buffers when necessary. It was getting this wrong when the
865 buffer for a data line had to be extended.
866
867 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
868 CRLF as a newline sequence.
869
870 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
871 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
872 I have nevertheless tidied it up.
873
874 23. Added some casts to kill warnings from HP-UX ia64 compiler.
875
876 24. Added a man page for pcre-config.
877
878
879 Version 7.0 19-Dec-06
880 ---------------------
881
882 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
883 moving to gcc 4.1.1.
884
885 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
886 sys/time.h, which is documented as needed for this function. It doesn't
887 seem to matter on Linux, but it showed up on some releases of OS X.
888
889 3. It seems that there are systems where bytes whose values are greater than
890 127 match isprint() in the "C" locale. The "C" locale should be the
891 default when a C program starts up. In most systems, only ASCII printing
892 characters match isprint(). This difference caused the output from pcretest
893 to vary, making some of the tests fail. I have changed pcretest so that:
894
895 (a) When it is outputting text in the compiled version of a pattern, bytes
896 other than 32-126 are always shown as hex escapes.
897
898 (b) When it is outputting text that is a matched part of a subject string,
899 it does the same, unless a different locale has been set for the match
900 (using the /L modifier). In this case, it uses isprint() to decide.
901
902 4. Fixed a major bug that caused incorrect computation of the amount of memory
903 required for a compiled pattern when options that changed within the
904 pattern affected the logic of the preliminary scan that determines the
905 length. The relevant options are -x, and -i in UTF-8 mode. The result was
906 that the computed length was too small. The symptoms of this bug were
907 either the PCRE error "internal error: code overflow" from pcre_compile(),
908 or a glibc crash with a message such as "pcretest: free(): invalid next
909 size (fast)". Examples of patterns that provoked this bug (shown in
910 pcretest format) are:
911
912 /(?-x: )/x
913 /(?x)(?-x: \s*#\s*)/
914 /((?i)[\x{c0}])/8
915 /(?i:[\x{c0}])/8
916
917 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
918 is now done differently.
919
920 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
921 wrapper classes; (b) implement a new function in the C++ scanner that is
922 more efficient than the old way of doing things because it avoids levels of
923 recursion in the regex matching; (c) add a paragraph to the documentation
924 for the FullMatch() function.
925
926 6. The escape sequence \n was being treated as whatever was defined as
927 "newline". Not only was this contrary to the documentation, which states
928 that \n is character 10 (hex 0A), but it also went horribly wrong when
929 "newline" was defined as CRLF. This has been fixed.
930
931 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
932 was being set to -1 for the "end of line" case (supposedly a value that no
933 character can have). Though this value is never used (the check for end of
934 line is "zero bytes in current character"), it caused compiler complaints.
935 I've changed it to 0xffffffff.
936
937 8. In pcre_version.c, the version string was being built by a sequence of
938 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
939 string (as it is for production releases) called a macro with an empty
940 argument. The C standard says the result of this is undefined. The gcc
941 compiler treats it as an empty string (which was what was wanted) but it is
942 reported that Visual C gives an error. The source has been hacked around to
943 avoid this problem.
944
945 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
946 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
947 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
948 of them did).
949
950 10. Originally, pcretest opened its input and output without "b"; then I was
951 told that "b" was needed in some environments, so it was added for release
952 5.0 to both the input and output. (It makes no difference on Unix-like
953 systems.) Later I was told that it is wrong for the input on Windows. I've
954 now abstracted the modes into two macros, to make it easier to fiddle with
955 them, and removed "b" from the input mode under Windows.
956
957 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
958
959 12. Added -help and --help to pcretest as an official way of being reminded
960 of the options.
961
962 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
963 and pcrecpp.cc because they annoy compilers at high warning levels.
964
965 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
966
967 15. Fixed an occurrence of == in configure.ac that should have been = (shell
968 scripts are not C programs :-) and which was not noticed because it works
969 on Linux.
970
971 16. pcretest is supposed to handle any length of pattern and data line (as one
972 line or as a continued sequence of lines) by extending its input buffer if
973 necessary. This feature was broken for very long pattern lines, leading to
974 a string of junk being passed to pcre_compile() if the pattern was longer
975 than about 50K.
976
977 17. I have done a major re-factoring of the way pcre_compile() computes the
978 amount of memory needed for a compiled pattern. Previously, there was code
979 that made a preliminary scan of the pattern in order to do this. That was
980 OK when PCRE was new, but as the facilities have expanded, it has become
981 harder and harder to keep it in step with the real compile phase, and there
982 have been a number of bugs (see for example, 4 above). I have now found a
983 cunning way of running the real compile function in a "fake" mode that
984 enables it to compute how much memory it would need, while actually only
985 ever using a few hundred bytes of working memory and without too many
986 tests of the mode. This should make future maintenance and development
987 easier. A side effect of this work is that the limit of 200 on the nesting
988 depth of parentheses has been removed (though this was never a serious
989 limitation, I suspect). However, there is a downside: pcre_compile() now
990 runs more slowly than before (30% or more, depending on the pattern). I
991 hope this isn't a big issue. There is no effect on runtime performance.
992
993 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
994 newline (only possible for the last line of a file) and it was a
995 pattern that set a locale (followed by /Lsomething), pcretest crashed.
996
997 19. Added additional timing features to pcretest. (1) The -tm option now times
998 matching only, not compiling. (2) Both -t and -tm can be followed, as a
999 separate command line item, by a number that specifies the number of
1000 repeats to use when timing. The default is 50000; this gives better
1001 precision, but takes uncomfortably long for very large patterns.
1002
1003 20. Extended pcre_study() to be more clever in cases where a branch of a
1004 subpattern has no definite first character. For example, (a*|b*)[cd] would
1005 previously give no result from pcre_study(). Now it recognizes that the
1006 first character must be a, b, c, or d.
1007
1008 21. There was an incorrect error "recursive call could loop indefinitely" if
1009 a subpattern (or the entire pattern) that was being tested for matching an
1010 empty string contained only one non-empty item after a nested subpattern.
1011 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
1012 incorrectly, because the \d was being skipped in the check.
1013
1014 22. The pcretest program now has a new pattern option /B and a command line
1015 option -b, which is equivalent to adding /B to every pattern. This causes
1016 it to show the compiled bytecode, without the additional information that
1017 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
1018 is the same as /B/I).
1019
1020 23. A new optimization is now able automatically to treat some sequences such
1021 as a*b as a*+b. More specifically, if something simple (such as a character
1022 or a simple class like \d) has an unlimited quantifier, and is followed by
1023 something that cannot possibly match the quantified thing, the quantifier
1024 is automatically "possessified".
1025
1026 24. A recursive reference to a subpattern whose number was greater than 39
1027 went wrong under certain circumstances in UTF-8 mode. This bug could also
1028 have affected the operation of pcre_study().
1029
1030 25. Realized that a little bit of performance could be had by replacing
1031 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
1032
1033 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
1034
1035 27. Possessive quantifiers such as a++ were previously implemented by turning
1036 them into atomic groups such as ($>a+). Now they have their own opcodes,
1037 which improves performance. This includes the automatically created ones
1038 from 23 above.
1039
1040 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
1041 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
1042 the first matched character to be a colon. This applied both to named and
1043 numbered groups.
1044
1045 29. The ucpinternal.h header file was missing its idempotency #ifdef.
1046
1047 30. I was sent a "project" file called libpcre.a.dev which I understand makes
1048 building PCRE on Windows easier, so I have included it in the distribution.
1049
1050 31. There is now a check in pcretest against a ridiculously large number being
1051 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
1052 loop, the loop is abandoned.
1053
1054 32. Forward references to subpatterns in conditions such as (?(2)...) where
1055 subpattern 2 is defined later cause pcre_compile() to search forwards in
1056 the pattern for the relevant set of parentheses. This search went wrong
1057 when there were unescaped parentheses in a character class, parentheses
1058 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
1059
1060 33. "Subroutine" calls and backreferences were previously restricted to
1061 referencing subpatterns earlier in the regex. This restriction has now
1062 been removed.
1063
1064 34. Added a number of extra features that are going to be in Perl 5.10. On the
1065 whole, these are just syntactic alternatives for features that PCRE had
1066 previously implemented using the Python syntax or my own invention. The
1067 other formats are all retained for compatibility.
1068
1069 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
1070 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
1071 also .NET compatible.
1072
1073 (b) A recursion or subroutine call to a named group can now be defined as
1074 (?&name) as well as (?P>name).
1075
1076 (c) A backreference to a named group can now be defined as \k<name> or
1077 \k'name' as well as (?P=name). The new forms, as well as being in Perl
1078 5.10, are also .NET compatible.
1079
1080 (d) A conditional reference to a named group can now use the syntax
1081 (?(<name>) or (?('name') as well as (?(name).
1082
1083 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
1084 groups (named and numbered) that are never evaluated inline, but can be
1085 called as "subroutines" from elsewhere. In effect, the DEFINE condition
1086 is always false. There may be only one alternative in such a group.
1087
1088 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
1089 as the simple (?(R). The condition is true only if the most recent
1090 recursion is that of the given number or name. It does not search out
1091 through the entire recursion stack.
1092
1093 (g) The escape \gN or \g{N} has been added, where N is a positive or
1094 negative number, specifying an absolute or relative reference.
1095
1096 35. Tidied to get rid of some further signed/unsigned compiler warnings and
1097 some "unreachable code" warnings.
1098
1099 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
1100 things, this adds five new scripts.
1101
1102 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
1103 There were also incompatibilities regarding the handling of \Q..\E inside
1104 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
1105 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
1106
1107 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
1108 matches an empty string, and forcibly breaks the loop. There were bugs in
1109 this code in non-simple cases. For a pattern such as ^(a()*)* matched
1110 against aaaa the result was just "a" rather than "aaaa", for example. Two
1111 separate and independent bugs (that affected different cases) have been
1112 fixed.
1113
1114 39. Refactored the code to abolish the use of different opcodes for small
1115 capturing bracket numbers. This is a tidy that I avoided doing when I
1116 removed the limit on the number of capturing brackets for 3.5 back in 2001.
1117 The new approach is not only tidier, it makes it possible to reduce the
1118 memory needed to fix the previous bug (38).
1119
1120 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
1121 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
1122 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
1123 mode.
1124
1125 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
1126 report.
1127
1128 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
1129 copy construction and assignment in the C++ wrapper.
1130
1131 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
1132 couple of bugs that could have given wrong results in the "--newline=crlf"
1133 case.
1134
1135 44. Added a number of casts and did some reorganization of signed/unsigned int
1136 variables following suggestions from Dair Grant. Also renamed the variable
1137 "this" as "item" because it is a C++ keyword.
1138
1139 45. Arranged for dftables to add
1140
1141 #include "pcre_internal.h"
1142
1143 to pcre_chartables.c because without it, gcc 4.x may remove the array
1144 definition from the final binary if PCRE is built into a static library and
1145 dead code stripping is activated.
1146
1147 46. For an unanchored pattern, if a match attempt fails at the start of a
1148 newline sequence, and the newline setting is CRLF or ANY, and the next two
1149 characters are CRLF, advance by two characters instead of one.
1150
1151
1152 Version 6.7 04-Jul-06
1153 ---------------------
1154
1155 1. In order to handle tests when input lines are enormously long, pcretest has
1156 been re-factored so that it automatically extends its buffers when
1157 necessary. The code is crude, but this _is_ just a test program. The
1158 default size has been increased from 32K to 50K.
1159
1160 2. The code in pcre_study() was using the value of the re argument before
1161 testing it for NULL. (Of course, in any sensible call of the function, it
1162 won't be NULL.)
1163
1164 3. The memmove() emulation function in pcre_internal.h, which is used on
1165 systems that lack both memmove() and bcopy() - that is, hardly ever -
1166 was missing a "static" storage class specifier.
1167
1168 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
1169 containing an extended class (one that cannot be represented by a bitmap
1170 because it contains high-valued characters or Unicode property items, e.g.
1171 [\pZ]). Almost always one would set UTF-8 mode when processing such a
1172 pattern, but PCRE should not loop if you do not (it no longer does).
1173 [Detail: two cases were found: (a) a repeated subpattern containing an
1174 extended class; (b) a recursive reference to a subpattern that followed a
1175 previous extended class. It wasn't skipping over the extended class
1176 correctly when UTF-8 mode was not set.]
1177
1178 5. A negated single-character class was not being recognized as fixed-length
1179 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
1180 compile error "lookbehind assertion is not fixed length".
1181
1182 6. The RunPerlTest auxiliary script was showing an unexpected difference
1183 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
1184 write a Perl script that can interpret lines of an input file either as
1185 byte characters or as UTF-8, which is what "perltest" was being required to
1186 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
1187 can't do is switch easily at run time between having the "use utf8;" pragma
1188 or not. In the end, I fudged it by using the RunPerlTest script to insert
1189 "use utf8;" explicitly for the UTF-8 tests.
1190
1191 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
1192 the end of the subject string, contrary to the documentation and to what
1193 Perl does. This was true of both matching functions. Now it matches only at
1194 the start of the subject and immediately after *internal* newlines.
1195
1196 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
1197 a pointer to an int instead of a pointer to an unsigned long int. This
1198 caused problems on 64-bit systems.
1199
1200 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
1201 instance of the 'standard' template library not being so standard".
1202
1203 10. There was no check on the number of named subpatterns nor the maximum
1204 length of a subpattern name. The product of these values is used to compute
1205 the size of the memory block for a compiled pattern. By supplying a very
1206 long subpattern name and a large number of named subpatterns, the size
1207 computation could be caused to overflow. This is now prevented by limiting
1208 the length of names to 32 characters, and the number of named subpatterns
1209 to 10,000.
1210
1211 11. Subpatterns that are repeated with specific counts have to be replicated in
1212 the compiled pattern. The size of memory for this was computed from the
1213 length of the subpattern and the repeat count. The latter is limited to
1214 65535, but there was no limit on the former, meaning that integer overflow
1215 could in principle occur. The compiled length of a repeated subpattern is
1216 now limited to 30,000 bytes in order to prevent this.
1217
1218 12. Added the optional facility to have named substrings with the same name.
1219
1220 13. Added the ability to use a named substring as a condition, using the
1221 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
1222 are numbers (not recommended). Forward references are permitted.
1223
1224 14. Added forward references in named backreferences (if you see what I mean).
1225
1226 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
1227 pattern could run off the end of the subject. For example, the pattern
1228 "(?s)(.{1,5})"8 did this with the subject "ab".
1229
1230 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
1231 PCRE_CASELESS was set when matching characters that were quantified with ?
1232 or *.
1233
1234 17. A character class other than a single negated character that had a minimum
1235 but no maximum quantifier - for example [ab]{6,} - was not handled
1236 correctly by pce_dfa_exec(). It would match only one character.
1237
1238 18. A valid (though odd) pattern that looked like a POSIX character
1239 class but used an invalid character after [ (for example [[,abc,]]) caused
1240 pcre_compile() to give the error "Failed: internal error: code overflow" or
1241 in some cases to crash with a glibc free() error. This could even happen if
1242 the pattern terminated after [[ but there just happened to be a sequence of
1243 letters, a binary zero, and a closing ] in the memory that followed.
1244
1245 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
1246 over the years. Originally (before any Unicode support), just the bottom 8
1247 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
1248 output from "man perlunicode" includes this:
1249
1250 The regular expression compiler produces polymorphic opcodes. That
1251 is, the pattern adapts to the data and automatically switches to
1252 the Unicode character scheme when presented with Unicode data--or
1253 instead uses a traditional byte scheme when presented with byte
1254 data.
1255
1256 Sadly, a wide octal escape does not cause a switch, and in a string with
1257 no other multibyte characters, these octal escapes are treated as before.
1258 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
1259 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
1260 Unicode string.
1261
1262 I have not perpetrated such confusion in PCRE. Up till now, it took just
1263 the bottom 8 bits, as in old Perl. I have now made octal escapes with
1264 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
1265 translate to the appropriate multibyte character.
1266
1267 29. Applied some refactoring to reduce the number of warnings from Microsoft
1268 and Borland compilers. This has included removing the fudge introduced
1269 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
1270 a warning about an unused variable.
1271
1272 21. PCRE has not included VT (character 0x0b) in the set of whitespace
1273 characters since release 4.0, because Perl (from release 5.004) does not.
1274 [Or at least, is documented not to: some releases seem to be in conflict
1275 with the documentation.] However, when a pattern was studied with
1276 pcre_study() and all its branches started with \s, PCRE still included VT
1277 as a possible starting character. Of course, this did no harm; it just
1278 caused an unnecessary match attempt.
1279
1280 22. Removed a now-redundant internal flag bit that recorded the fact that case
1281 dependency changed within the pattern. This was once needed for "required
1282 byte" processing, but is no longer used. This recovers a now-scarce options
1283 bit. Also moved the least significant internal flag bit to the most-
1284 significant bit of the word, which was not previously used (hangover from
1285 the days when it was an int rather than a uint) to free up another bit for
1286 the future.
1287
1288 23. Added support for CRLF line endings as well as CR and LF. As well as the
1289 default being selectable at build time, it can now be changed at runtime
1290 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
1291 specify that it is scanning data with non-default line endings.
1292
1293 24. Changed the definition of CXXLINK to make it agree with the definition of
1294 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
1295
1296 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
1297 recursions. This makes a big different to stack usage for some patterns.
1298
1299 26. If a subpattern containing a named recursion or subroutine reference such
1300 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
1301 the space required for the compiled pattern went wrong and gave too small a
1302 value. Depending on the environment, this could lead to "Failed: internal
1303 error: code overflow at offset 49" or "glibc detected double free or
1304 corruption" errors.
1305
1306 27. Applied patches from Google (a) to support the new newline modes and (b) to
1307 advance over multibyte UTF-8 characters in GlobalReplace.
1308
1309 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
1310 difference for some implementation of PCRE in some Windows version.
1311
1312 29. Added some extra testing facilities to pcretest:
1313
1314 \q<number> in a data line sets the "match limit" value
1315 \Q<number> in a data line sets the "match recursion limt" value
1316 -S <number> sets the stack size, where <number> is in megabytes
1317
1318 The -S option isn't available for Windows.
1319
1320
1321 Version 6.6 06-Feb-06
1322 ---------------------
1323
1324 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
1325 in pcreposix.h. I have copied the definition from pcre.h.
1326
1327 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
1328 because pcre.h is no longer a built file.
1329
1330 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
1331 not normally included in the compiled code.
1332
1333
1334 Version 6.5 01-Feb-06
1335 ---------------------
1336
1337 1. When using the partial match feature with pcre_dfa_exec(), it was not
1338 anchoring the second and subsequent partial matches at the new starting
1339 point. This could lead to incorrect results. For example, with the pattern
1340 /1234/, partially matching against "123" and then "a4" gave a match.
1341
1342 2. Changes to pcregrep:
1343
1344 (a) All non-match returns from pcre_exec() were being treated as failures
1345 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
1346 error message is output. Some extra information is given for the
1347 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
1348 probably the only errors that are likely to be caused by users (by
1349 specifying a regex that has nested indefinite repeats, for instance).
1350 If there are more than 20 of these errors, pcregrep is abandoned.
1351
1352 (b) A binary zero was treated as data while matching, but terminated the
1353 output line if it was written out. This has been fixed: binary zeroes
1354 are now no different to any other data bytes.
1355
1356 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
1357 used to set a locale for matching. The --locale=xxxx long option has
1358 been added (no short equivalent) to specify a locale explicitly on the
1359 pcregrep command, overriding the environment variables.
1360
1361 (d) When -B was used with -n, some line numbers in the output were one less
1362 than they should have been.
1363
1364 (e) Added the -o (--only-matching) option.
1365
1366 (f) If -A or -C was used with -c (count only), some lines of context were
1367 accidentally printed for the final match.
1368
1369 (g) Added the -H (--with-filename) option.
1370
1371 (h) The combination of options -rh failed to suppress file names for files
1372 that were found from directory arguments.
1373
1374 (i) Added the -D (--devices) and -d (--directories) options.
1375
1376 (j) Added the -F (--fixed-strings) option.
1377
1378 (k) Allow "-" to be used as a file name for -f as well as for a data file.
1379
1380 (l) Added the --colo(u)r option.
1381
1382 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
1383 is not present by default.
1384
1385 3. A nasty bug was discovered in the handling of recursive patterns, that is,
1386 items such as (?R) or (?1), when the recursion could match a number of
1387 alternatives. If it matched one of the alternatives, but subsequently,
1388 outside the recursion, there was a failure, the code tried to back up into
1389 the recursion. However, because of the way PCRE is implemented, this is not
1390 possible, and the result was an incorrect result from the match.
1391
1392 In order to prevent this happening, the specification of recursion has
1393 been changed so that all such subpatterns are automatically treated as
1394 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
1395
1396 4. I had overlooked the fact that, in some locales, there are characters for
1397 which isalpha() is true but neither isupper() nor islower() are true. In
1398 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
1399 and ordfeminine) are like this. This affected the treatment of \w and \W
1400 when they appeared in character classes, but not when they appeared outside
1401 a character class. The bit map for "word" characters is now created
1402 separately from the results of isalnum() instead of just taking it from the
1403 upper, lower, and digit maps. (Plus the underscore character, of course.)
1404
1405 5. The above bug also affected the handling of POSIX character classes such as
1406 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
1407 permanent tables. Instead, the bit maps for such a class were previously
1408 created as the appropriate unions of the upper, lower, and digit bitmaps.
1409 Now they are created by subtraction from the [[:word:]] class, which has
1410 its own bitmap.
1411
1412 6. The [[:blank:]] character class matches horizontal, but not vertical space.
1413 It is created by subtracting the vertical space characters (\x09, \x0a,
1414 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
1415 subtraction was done in the overall bitmap for a character class, meaning
1416 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
1417 be recognized. This bug has been fixed.
1418
1419 7. Patches from the folks at Google:
1420
1421 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
1422 real life, but is still worth protecting against".
1423
1424 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
1425 regular expressions".
1426
1427 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
1428 have it.
1429
1430 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
1431 "configure" and the latter not, in order to fix a problem somebody had
1432 with compiling the Arg class on HP-UX.
1433
1434 (e) Improve the error-handling of the C++ wrapper a little bit.
1435
1436 (f) New tests for checking recursion limiting.
1437
1438 8. The pcre_memmove() function, which is used only if the environment does not
1439 have a standard memmove() function (and is therefore rarely compiled),
1440 contained two bugs: (a) use of int instead of size_t, and (b) it was not
1441 returning a result (though PCRE never actually uses the result).
1442
1443 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
1444 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
1445 returned instead of calling malloc() with an overflowing number that would
1446 most likely cause subsequent chaos.
1447
1448 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
1449
1450 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
1451 with this option is matched, the nmatch and pmatch options of regexec() are
1452 ignored.
1453
1454 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
1455 provided in case anyone wants to the the POSIX interface with UTF-8
1456 strings.
1457
1458 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
1459 C++ linking (needed for some HP-UX environments).
1460
1461 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
1462 (unused parameter) and in the pcre_printint() function (omitted "default"
1463 switch label when the default is to do nothing).
1464
1465 15. Added some code to make it possible, when PCRE is compiled as a C++
1466 library, to replace subject pointers for pcre_exec() with a smart pointer
1467 class, thus making it possible to process discontinuous strings.
1468
1469 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
1470 much the same function. They were added by different people who were trying
1471 to make PCRE easy to compile on non-Unix systems. It has been suggested
1472 that PCRE_EXPORT be abolished now that there is more automatic apparatus
1473 for compiling on Windows systems. I have therefore replaced it with
1474 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
1475 defaults to "extern" for C or "extern C" for C++, which works fine on
1476 Unix-like systems. It is now possible to override the value of PCRE_DATA_
1477 SCOPE with something explicit in config.h. In addition:
1478
1479 (a) pcreposix.h still had just "extern" instead of either of these macros;
1480 I have replaced it with PCRE_DATA_SCOPE.
1481
1482 (b) Functions such as _pcre_xclass(), which are internal to the library,
1483 but external in the C sense, all had PCRE_EXPORT in their definitions.
1484 This is apparently wrong for the Windows case, so I have removed it.
1485 (It makes no difference on Unix-like systems.)
1486
1487 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
1488 of recursive calls to match(). This is different to MATCH_LIMIT because
1489 that limits the total number of calls to match(), not all of which increase
1490 the depth of recursion. Limiting the recursion depth limits the amount of
1491 stack (or heap if NO_RECURSE is set) that is used. The default can be set
1492 when PCRE is compiled, and changed at run time. A patch from Google adds
1493 this functionality to the C++ interface.
1494
1495 18. Changes to the handling of Unicode character properties:
1496
1497 (a) Updated the table to Unicode 4.1.0.
1498
1499 (b) Recognize characters that are not in the table as "Cn" (undefined).
1500
1501 (c) I revised the way the table is implemented to a much improved format
1502 which includes recognition of ranges. It now supports the ranges that
1503 are defined in UnicodeData.txt, and it also amalgamates other
1504 characters into ranges. This has reduced the number of entries in the
1505 table from around 16,000 to around 3,000, thus reducing its size
1506 considerably. I realized I did not need to use a tree structure after
1507 all - a binary chop search is just as efficient. Having reduced the
1508 number of entries, I extended their size from 6 bytes to 8 bytes to
1509 allow for more data.
1510
1511 (d) Added support for Unicode script names via properties such as \p{Han}.
1512
1513 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
1514 matching that character.
1515
1516 20. When matching a repeated Unicode property with a minimum greater than zero,
1517 (for example \pL{2,}), PCRE could look past the end of the subject if it
1518 reached it while seeking the minimum number of characters. This could
1519 happen only if some of the characters were more than one byte long, because
1520 there is a check for at least the minimum number of bytes.
1521
1522 21. Refactored the implementation of \p and \P so as to be more general, to
1523 allow for more different types of property in future. This has changed the
1524 compiled form incompatibly. Anybody with saved compiled patterns that use
1525 \p or \P will have to recompile them.
1526
1527 22. Added "Any" and "L&" to the supported property types.
1528
1529 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
1530 but give a compile time error if the value is greater than 0xff.
1531
1532 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
1533 accidentally not being installed or uninstalled.
1534
1535 25. The pcre.h file was built from pcre.h.in, but the only changes that were
1536 made were to insert the current release number. This seemed silly, because
1537 it made things harder for people building PCRE on systems that don't run
1538 "configure". I have turned pcre.h into a distributed file, no longer built
1539 by "configure", with the version identification directly included. There is
1540 no longer a pcre.h.in file.
1541
1542 However, this change necessitated a change to the pcre-config script as
1543 well. It is built from pcre-config.in, and one of the substitutions was the
1544 release number. I have updated configure.ac so that ./configure now finds
1545 the release number by grepping pcre.h.
1546
1547 26. Added the ability to run the tests under valgrind.
1548
1549
1550 Version 6.4 05-Sep-05
1551 ---------------------
1552
1553 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
1554 "--" to be printed when multiple files were scanned, even when none of the
1555 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
1556 consider it to be a bug, and have restored the previous behaviour.
1557
1558 2. A couple of code tidies to get rid of compiler warnings.
1559
1560 3. The pcretest program used to cheat by referring to symbols in the library
1561 whose names begin with _pcre_. These are internal symbols that are not
1562 really supposed to be visible externally, and in some environments it is
1563 possible to suppress them. The cheating is now confined to including
1564 certain files from the library's source, which is a bit cleaner.
1565
1566 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
1567 file's purpose clearer.
1568
1569 5. Reorganized pcre_ucp_findchar().
1570
1571
1572 Version 6.3 15-Aug-05
1573 ---------------------
1574
1575 1. The file libpcre.pc.in did not have general read permission in the tarball.
1576
1577 2. There were some problems when building without C++ support:
1578
1579 (a) If C++ support was not built, "make install" and "make test" still
1580 tried to test it.
1581
1582 (b) There were problems when the value of CXX was explicitly set. Some
1583 changes have been made to try to fix these, and ...
1584
1585 (c) --disable-cpp can now be used to explicitly disable C++ support.
1586
1587 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
1588 backslash in a target when C++ was disabled. This confuses some
1589 versions of "make", apparently. Using an intermediate variable solves
1590 this. (Same for CPP_LOBJ.)
1591
1592 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
1593 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
1594 necessary on certain architectures.
1595
1596 4. Added a setting of -export-symbols-regex to the link command to remove
1597 those symbols that are exported in the C sense, but actually are local
1598 within the library, and not documented. Their names all begin with
1599 "_pcre_". This is not a perfect job, because (a) we have to except some
1600 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
1601 available (and never for static libraries). I have made a note to try to
1602 find a way round (a) in the future.
1603
1604
1605 Version 6.2 01-Aug-05
1606 ---------------------
1607
1608 1. There was no test for integer overflow of quantifier values. A construction
1609 such as {1111111111111111} would give undefined results. What is worse, if
1610 a minimum quantifier for a parenthesized subpattern overflowed and became
1611 negative, the calculation of the memory size went wrong. This could have
1612 led to memory overwriting.
1613
1614 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
1615
1616 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
1617 operating environments where this matters.
1618
1619 4. Applied Giuseppe Maxia's patch to add additional features for controlling
1620 PCRE options from within the C++ wrapper.
1621
1622 5. Named capturing subpatterns were not being correctly counted when a pattern
1623 was compiled. This caused two problems: (a) If there were more than 100
1624 such subpatterns, the calculation of the memory needed for the whole
1625 compiled pattern went wrong, leading to an overflow error. (b) Numerical
1626 back references of the form \12, where the number was greater than 9, were
1627 not recognized as back references, even though there were sufficient
1628 previous subpatterns.
1629
1630 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
1631 versions of gcc, e.g. 2.95.4.
1632
1633
1634 Version 6.1 21-Jun-05
1635 ---------------------
1636
1637 1. There was one reference to the variable "posix" in pcretest.c that was not
1638 surrounded by "#if !defined NOPOSIX".
1639
1640 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
1641 the cross-check on the old pcre_info() function, for the benefit of the
1642 cut-down version of PCRE that is currently imported into Exim.
1643
1644 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
1645 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
1646 patterns that start (?i) but I don't think that matters. The use of (?i) is
1647 just an example; this all applies to the other options as well.
1648
1649 4. Since libtool seems to echo the compile commands it is issuing, the output
1650 from "make" can be reduced a bit by putting "@" in front of each libtool
1651 compile command.
1652
1653 5. Patch from the folks at Google for configure.in to be a bit more thorough
1654 in checking for a suitable C++ installation before trying to compile the
1655 C++ stuff. This should fix a reported problem when a compiler was present,
1656 but no suitable headers.
1657
1658 6. The man pages all had just "PCRE" as their title. I have changed them to
1659 be the relevant file name. I have also arranged that these names are
1660 retained in the file doc/pcre.txt, which is a concatenation in text format
1661 of all the man pages except the little individual ones for each function.
1662
1663 7. The NON-UNIX-USE file had not been updated for the different set of source
1664 files that come with release 6. I also added a few comments about the C++
1665 wrapper.
1666
1667
1668 Version 6.0 07-Jun-05
1669 ---------------------
1670
1671 1. Some minor internal re-organization to help with my DFA experiments.
1672
1673 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
1674 didn't matter for the library itself when fully configured, but did matter
1675 when compiling without UCP support, or within Exim, where the ucp files are
1676 not imported.
1677
1678 3. Refactoring of the library code to split up the various functions into
1679 different source modules. The addition of the new DFA matching code (see
1680 below) to a single monolithic source would have made it really too
1681 unwieldy, quite apart from causing all the code to be include in a
1682 statically linked application, when only some functions are used. This is
1683 relevant even without the DFA addition now that patterns can be compiled in
1684 one application and matched in another.
1685
1686 The downside of splitting up is that there have to be some external
1687 functions and data tables that are used internally in different modules of
1688 the library but which are not part of the API. These have all had their
1689 names changed to start with "_pcre_" so that they are unlikely to clash
1690 with other external names.
1691
1692 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
1693 a different (DFA) algorithm. Although it is slower than the original
1694 function, it does have some advantages for certain types of matching
1695 problem.
1696
1697 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
1698 including restarting after a partial match.
1699
1700 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
1701 defined when compiling for Windows was sent to me. I have put it into the
1702 code, though I have no means of testing or verifying it.
1703
1704 7. Added the pcre_refcount() auxiliary function.
1705
1706 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
1707 match before or at the first newline in the subject string. In pcretest,
1708 the /f option on a pattern can be used to set this.
1709
1710 9. A repeated \w when used in UTF-8 mode with characters greater than 256
1711 would behave wrongly. This has been present in PCRE since release 4.0.
1712
1713 10. A number of changes to the pcregrep command:
1714
1715 (a) Refactored how -x works; insert ^(...)$ instead of setting
1716 PCRE_ANCHORED and checking the length, in preparation for adding
1717 something similar for -w.
1718
1719 (b) Added the -w (match as a word) option.
1720
1721 (c) Refactored the way lines are read and buffered so as to have more
1722 than one at a time available.
1723
1724 (d) Implemented a pcregrep test script.
1725
1726 (e) Added the -M (multiline match) option. This allows patterns to match
1727 over several lines of the subject. The buffering ensures that at least
1728 8K, or the rest of the document (whichever is the shorter) is available
1729 for matching (and similarly the previous 8K for lookbehind assertions).
1730
1731 (f) Changed the --help output so that it now says
1732
1733 -w, --word-regex(p)
1734
1735 instead of two lines, one with "regex" and the other with "regexp"
1736 because that confused at least one person since the short forms are the
1737 same. (This required a bit of code, as the output is generated
1738 automatically from a table. It wasn't just a text change.)
1739
1740 (g) -- can be used to terminate pcregrep options if the next thing isn't an
1741 option but starts with a hyphen. Could be a pattern or a path name
1742 starting with a hyphen, for instance.
1743
1744 (h) "-" can be given as a file name to represent stdin.
1745
1746 (i) When file names are being printed, "(standard input)" is used for
1747 the standard input, for compatibility with GNU grep. Previously
1748 "<stdin>" was used.
1749
1750 (j) The option --label=xxx can be used to supply a name to be used for
1751 stdin when file names are being printed. There is no short form.
1752
1753 (k) Re-factored the options decoding logic because we are going to add
1754 two more options that take data. Such options can now be given in four
1755 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
1756
1757 (l) Added the -A, -B, and -C options for requesting that lines of context
1758 around matches be printed.
1759
1760 (m) Added the -L option to print the names of files that do not contain
1761 any matching lines, that is, the complement of -l.
1762
1763 (n) The return code is 2 if any file cannot be opened, but pcregrep does
1764 continue to scan other files.
1765
1766 (o) The -s option was incorrectly implemented. For compatibility with other
1767 greps, it now suppresses the error message for a non-existent or non-
1768 accessible file (but not the return code). There is a new option called
1769 -q that suppresses the output of matching lines, which was what -s was
1770 previously doing.
1771
1772 (p) Added --include and --exclude options to specify files for inclusion
1773 and exclusion when recursing.
1774
1775 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
1776 Hopefully, it now does.
1777
1778 12. Missing cast in pcre_study().
1779
1780 13. Added an "uninstall" target to the makefile.
1781
1782 14. Replaced "extern" in the function prototypes in Makefile.in with
1783 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
1784 world, but is set differently for Windows.
1785
1786 15. Added a second compiling function called pcre_compile2(). The only
1787 difference is that it has an extra argument, which is a pointer to an
1788 integer error code. When there is a compile-time failure, this is set
1789 non-zero, in addition to the error test pointer being set to point to an
1790 error message. The new argument may be NULL if no error number is required
1791 (but then you may as well call pcre_compile(), which is now just a
1792 wrapper). This facility is provided because some applications need a
1793 numeric error indication, but it has also enabled me to tidy up the way
1794 compile-time errors are handled in the POSIX wrapper.
1795
1796 16. Added VPATH=.libs to the makefile; this should help when building with one
1797 prefix path and installing with another. (Or so I'm told by someone who
1798 knows more about this stuff than I do.)
1799
1800 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
1801 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
1802 match everything, including newlines. This is not POSIX-compatible, but
1803 somebody wanted the feature. From pcretest it can be activated by using
1804 both the P and the s flags.
1805
1806 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
1807
1808 19. libpcre.pc was being incorrectly installed as executable.
1809
1810 20. A couple of places in pcretest check for end-of-line by looking for '\n';
1811 it now also looks for '\r' so that it will work unmodified on Windows.
1812
1813 21. Added Google's contributed C++ wrapper to the distribution.
1814
1815 22. Added some untidy missing memory free() calls in pcretest, to keep
1816 Electric Fence happy when testing.
1817
1818
1819
1820 Version 5.0 13-Sep-04
1821 ---------------------
1822
1823 1. Internal change: literal characters are no longer packed up into items
1824 containing multiple characters in a single byte-string. Each character
1825 is now matched using a separate opcode. However, there may be more than one
1826 byte in the character in UTF-8 mode.
1827
1828 2. The pcre_callout_block structure has two new fields: pattern_position and
1829 next_item_length. These contain the offset in the pattern to the next match
1830 item, and its length, respectively.
1831
1832 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
1833 insertion of callouts before each pattern item. Added the /C option to
1834 pcretest to make use of this.
1835
1836 4. On the advice of a Windows user, the lines
1837
1838 #if defined(_WIN32) || defined(WIN32)
1839 _setmode( _fileno( stdout ), 0x8000 );
1840 #endif /* defined(_WIN32) || defined(WIN32) */
1841
1842 have been added to the source of pcretest. This apparently does useful
1843 magic in relation to line terminators.
1844
1845 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
1846 for the benefit of those environments where the "b" makes a difference.
1847
1848 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
1849 to know about it. I have put a hack into configure.in that adds in code
1850 to set GCC=yes if CC=icc. This seems to end up at a point in the
1851 generated configure script that is early enough to affect the setting of
1852 compiler options, which is what is needed, but I have no means of testing
1853 whether it really works. (The user who reported this had patched the
1854 generated configure script, which of course I cannot do.)
1855
1856 LATER: After change 22 below (new libtool files), the configure script
1857 seems to know about icc (and also ecc). Therefore, I have commented out
1858 this hack in configure.in.
1859
1860 7. Added support for pkg-config (2 patches were sent in).
1861
1862 8. Negated POSIX character classes that used a combination of internal tables
1863 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
1864 [[:^ascii]]. Typically, they would match almost any characters. The other
1865 POSIX classes were not broken in this way.
1866
1867 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
1868 to find the match, as PCRE was deluded into thinking that the match had to
1869 start at the start point or following a newline. The same bug applied to
1870 patterns with negative forward assertions or any backward assertions
1871 preceding ".*" at the start, unless the pattern required a fixed first
1872 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
1873
1874 10. In UTF-8 mode, when moving forwards in the subject after a failed match
1875 starting at the last subject character, bytes beyond the end of the subject
1876 string were read.
1877
1878 11. Renamed the variable "class" as "classbits" to make life easier for C++
1879 users. (Previously there was a macro definition, but it apparently wasn't
1880 enough.)
1881
1882 12. Added the new field "tables" to the extra data so that tables can be passed
1883 in at exec time, or the internal tables can be re-selected. This allows
1884 a compiled regex to be saved and re-used at a later time by a different
1885 program that might have everything at different addresses.
1886
1887 13. Modified the pcre-config script so that, when run on Solaris, it shows a
1888 -R library as well as a -L library.
1889
1890 14. The debugging options of pcretest (-d on the command line or D on a
1891 pattern) showed incorrect output for anything following an extended class
1892 that contained multibyte characters and which was followed by a quantifier.
1893
1894 15. Added optional support for general category Unicode character properties
1895 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
1896 support. It adds about 90K to the size of the library. The meanings of the
1897 inbuilt class escapes such as \d and \s have NOT been changed.
1898
1899 16. Updated pcredemo.c to include calls to free() to release the memory for the
1900 compiled pattern.
1901
1902 17. The generated file chartables.c was being created in the source directory
1903 instead of in the building directory. This caused the build to fail if the
1904 source directory was different from the building directory, and was
1905 read-only.
1906
1907 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
1908 file. No doubt somebody will tell me if they don't make sense... Also added
1909 Dan Mooney's comments about building on OpenVMS.
1910
1911 19. Added support for partial matching via the PCRE_PARTIAL option for
1912 pcre_exec() and the \P data escape in pcretest.
1913
1914 20. Extended pcretest with 3 new pattern features:
1915
1916 (i) A pattern option of the form ">rest-of-line" causes pcretest to
1917 write the compiled pattern to the file whose name is "rest-of-line".
1918 This is a straight binary dump of the data, with the saved pointer to
1919 the character tables forced to be NULL. The study data, if any, is
1920 written too. After writing, pcretest reads a new pattern.
1921
1922 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
1923 compiled pattern from the given file. There must not be any
1924 occurrences of "<" in the file name (pretty unlikely); if there are,
1925 pcretest will instead treat the initial "<" as a pattern delimiter.
1926 After reading in the pattern, pcretest goes on to read data lines as
1927 usual.
1928
1929 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
1930 and 16-bit fields in a compiled pattern, to simulate a pattern that
1931 was compiled on a host of opposite endianness.
1932
1933 21. The pcre-exec() function can now cope with patterns that were compiled on
1934 hosts of opposite endianness, with this restriction:
1935
1936 As for any compiled expression that is saved and used later, the tables
1937 pointer field cannot be preserved; the extra_data field in the arguments
1938 to pcre_exec() should be used to pass in a tables address if a value
1939 other than the default internal tables were used at compile time.
1940
1941 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
1942 now diagnosed as an error. Previously, most of the time, a negative number
1943 would have been treated as zero, but if in addition "ovector" was passed as
1944 NULL, a crash could occur.
1945
1946 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
1947 new versions from the libtool 1.5 distribution (the last one is a copy of
1948 a file called libtool.m4). This seems to have fixed the need to patch
1949 "configure" to support Darwin 1.3 (which I used to do). However, I still
1950 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
1951 workstation).
1952
1953 24. Changed the PCRE licence to be the more standard "BSD" licence.
1954
1955
1956 Version 4.5 01-Dec-03
1957 ---------------------
1958
1959 1. There has been some re-arrangement of the code for the match() function so
1960 that it can be compiled in a version that does not call itself recursively.
1961 Instead, it keeps those local variables that need separate instances for
1962 each "recursion" in a frame on the heap, and gets/frees frames whenever it
1963 needs to "recurse". Keeping track of where control must go is done by means
1964 of setjmp/longjmp. The whole thing is implemented by a set of macros that
1965 hide most of the details from the main code, and operates only if
1966 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
1967 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
1968 operating.
1969
1970 To make it easier for callers to provide specially tailored get/free
1971 functions for this usage, two new functions, pcre_stack_malloc, and
1972 pcre_stack_free, are used. They are always called in strict stacking order,
1973 and the size of block requested is always the same.
1974
1975 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
1976 PCRE has been compiled to use the stack or the heap for recursion. The
1977 -C option of pcretest uses this to show which version is compiled.
1978
1979 A new data escape \S, is added to pcretest; it causes the amounts of store
1980 obtained and freed by both kinds of malloc/free at match time to be added
1981 to the output.
1982
1983 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
1984 what's available on my current Linux desktop machine.
1985
1986 3. When matching a UTF-8 string, the test for a valid string at the start has
1987 been extended. If start_offset is not zero, PCRE now checks that it points
1988 to a byte that is the start of a UTF-8 character. If not, it returns
1989 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
1990 this is necessary because there may be backward assertions in the pattern.
1991 When matching the same subject several times, it may save resources to use
1992 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
1993
1994 4. The code for checking the validity of UTF-8 strings has been tightened so
1995 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
1996 containing "overlong sequences".
1997
1998 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
1999 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
2000 should have been "&", but it just so happened that all the cases this let
2001 through by mistake were picked up later in the function.
2002
2003 6. I had used a variable called "isblank" - this is a C99 function, causing
2004 some compilers to warn. To avoid this, I renamed it (as "blankclass").
2005
2006 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
2007 prompting; (b) run "./pcretest /dev/null" at the start of the test script
2008 so the version is shown; (c) stop "make test" echoing "./RunTest".
2009
2010 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
2011
2012 9. The prototype for memmove() for systems that don't have it was using
2013 size_t, but the inclusion of the header that defines size_t was later. I've
2014 moved the #includes for the C headers earlier to avoid this.
2015
2016 10. Added some adjustments to the code to make it easier to compiler on certain
2017 special systems:
2018
2019 (a) Some "const" qualifiers were missing.
2020 (b) Added the macro EXPORT before all exported functions; by default this
2021 is defined to be empty.
2022 (c) Changed the dftables auxiliary program (that builds chartables.c) so
2023 that it reads its output file name as an argument instead of writing
2024 to the standard output and assuming this can be redirected.
2025
2026 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
2027 class containing characters with values greater than 255, PCRE compilation
2028 went into a loop.
2029
2030 12. A recursive reference to a subpattern that was within another subpattern
2031 that had a minimum quantifier of zero caused PCRE to crash. For example,
2032 (x(y(?2))z)? provoked this bug with a subject that got as far as the
2033 recursion. If the recursively-called subpattern itself had a zero repeat,
2034 that was OK.
2035
2036 13. In pcretest, the buffer for reading a data line was set at 30K, but the
2037 buffer into which it was copied (for escape processing) was still set at
2038 1024, so long lines caused crashes.
2039
2040 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
2041 "internal error: code overflow...". This applied to any character class
2042 that was followed by a possessive quantifier.
2043
2044 15. Modified the Makefile to add libpcre.la as a prerequisite for
2045 libpcreposix.la because I was told this is needed for a parallel build to
2046 work.
2047
2048 16. If a pattern that contained .* following optional items at the start was
2049 studied, the wrong optimizing data was generated, leading to matching
2050 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
2051 matching string must start with a or b or c. The correct conclusion for
2052 this pattern is that a match can start with any character.
2053
2054
2055 Version 4.4 13-Aug-03
2056 ---------------------
2057
2058 1. In UTF-8 mode, a character class containing characters with values between
2059 127 and 255 was not handled correctly if the compiled pattern was studied.
2060 In fixing this, I have also improved the studying algorithm for such
2061 classes (slightly).
2062
2063 2. Three internal functions had redundant arguments passed to them. Removal
2064 might give a very teeny performance improvement.
2065
2066 3. Documentation bug: the value of the capture_top field in a callout is *one
2067 more than* the number of the hightest numbered captured substring.
2068
2069 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
2070 in incorrectly linking with a previously installed version. They now link
2071 explicitly with libpcre.la.
2072
2073 5. configure.in no longer needs to recognize Cygwin specially.
2074
2075 6. A problem in pcre.in for Windows platforms is fixed.
2076
2077 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
2078 pcretest, it used to include the size of the study block as part of its
2079 output. Unfortunately, the structure contains a field that has a different
2080 size on different hardware architectures. This meant that the tests that
2081 showed this size failed. As the block is currently always of a fixed size,
2082 this information isn't actually particularly useful in pcretest output, so
2083 I have just removed it.
2084
2085 8. Three pre-processor statements accidentally did not start in column 1.
2086 Sadly, there are *still* compilers around that complain, even though
2087 standard C has not required this for well over a decade. Sigh.
2088
2089 9. In pcretest, the code for checking callouts passed small integers in the
2090 callout_data field, which is a void * field. However, some picky compilers
2091 complained about the casts involved for this on 64-bit systems. Now
2092 pcretest passes the address of the small integer instead, which should get
2093 rid of the warnings.
2094
2095 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
2096 both compile and run time, and gives an error if an invalid UTF-8 sequence
2097 is found. There is a option for disabling this check in cases where the
2098 string is known to be correct and/or the maximum performance is wanted.
2099
2100 11. In response to a bug report, I changed one line in Makefile.in from
2101
2102 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
2103 to
2104 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
2105
2106 to look similar to other lines, but I have no way of telling whether this
2107 is the right thing to do, as I do not use Windows. No doubt I'll get told
2108 if it's wrong...
2109
2110
2111 Version 4.3 21-May-03
2112 ---------------------
2113
2114 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
2115 Makefile.
2116
2117 2. Some refactoring to improve the quality of the code:
2118
2119 (i) The utf8_table... variables are now declared "const".
2120
2121 (ii) The code for \cx, which used the "case flipping" table to upper case
2122 lower case letters, now just substracts 32. This is ASCII-specific,
2123 but the whole concept of \cx is ASCII-specific, so it seems
2124 reasonable.
2125
2126 (iii) PCRE was using its character types table to recognize decimal and
2127 hexadecimal digits in the pattern. This is silly, because it handles
2128 only 0-9, a-f, and A-F, but the character types table is locale-
2129 specific, which means strange things might happen. A private
2130 table is now used for this - though it costs 256 bytes, a table is
2131 much faster than multiple explicit tests. Of course, the standard
2132 character types table is still used for matching digits in subject
2133 strings against \d.
2134
2135 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
2136 ending in _t are). So I've renamed it as ESC_tee.
2137
2138 3. The first argument for regexec() in the POSIX wrapper should have been
2139 defined as "const".
2140
2141 4. Changed pcretest to use malloc() for its buffers so that they can be
2142 Electric Fenced for debugging.
2143
2144 5. There were several places in the code where, in UTF-8 mode, PCRE would try
2145 to read one or more bytes before the start of the subject string. Often this
2146 had no effect on PCRE's behaviour, but in some circumstances it could
2147 provoke a segmentation fault.
2148
2149 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
2150 to try to read one or more bytes before the start of the subject string.
2151
2152 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
2153 UTF-8 support could misbehave in various ways if the subject string
2154 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
2155 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
2156 back over UTF-8 characters.)
2157
2158
2159 Version 4.2 14-Apr-03
2160 ---------------------
2161
2162 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
2163
2164 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
2165 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
2166 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
2167 [WIN_PREFIX]: new variable, "cyg" for Cygwin
2168 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
2169 and BUILD_EXEEXT
2170 Note: automatic setting of the BUILD variables is not yet working
2171 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
2172 compile-time but not at link-time
2173 [LINK]: use for linking executables only
2174 make different versions for Windows and non-Windows
2175 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
2176 libraries
2177 [LINK_FOR_BUILD]: new variable
2178 [OBJEXT]: use throughout
2179 [EXEEXT]: use throughout
2180 <winshared>: new target
2181 <wininstall>: new target
2182 <dftables.o>: use native compiler
2183 <dftables>: use native linker
2184 <install>: handle Windows platform correctly
2185 <clean>: ditto
2186 <check>: ditto
2187 copy DLL to top builddir before testing
2188
2189 As part of these changes, -no-undefined was removed again. This was reported
2190 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
2191 in any case.
2192
2193 3. Some tidies to get rid of compiler warnings:
2194
2195 . In the match_data structure, match_limit was an unsigned long int, whereas
2196 match_call_count was an int. I've made them both unsigned long ints.
2197
2198 . In pcretest the fact that a const uschar * doesn't automatically cast to
2199 a void * provoked a warning.
2200
2201 . Turning on some more compiler warnings threw up some "shadow" variables
2202 and a few more missing casts.
2203
2204 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
2205 option, a class that contained a single character with a value between 128
2206 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
2207
2208 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
2209 option, a class that contained several characters, but with at least one
2210 whose value was between 128 and 255 caused PCRE to crash.
2211
2212
2213 Version 4.1 12-Mar-03
2214 ---------------------
2215
2216 1. Compiling with gcc -pedantic found a couple of places where casts were
2217 needed, and a string in dftables.c that was longer than standard compilers are
2218 required to support.
2219
2220 2. Compiling with Sun's compiler found a few more places where the code could
2221 be tidied up in order to avoid warnings.
2222
2223 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
2224 first of these names is deprecated in the latest Autoconf in favour of the name
2225 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
2226 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
2227 analogy I have changed it to CFLAGS_FOR_BUILD.
2228
2229 4. Added -no-undefined to the linking command in the Makefile, because this is
2230 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
2231 linking step for the pcreposix library.
2232
2233 5. PCRE was failing to diagnose the case of two named groups with the same
2234 name.
2235
2236 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
2237 literal character that is needed in the subject for a match, and scans along to
2238 ensure that it is present before embarking on the full matching process. This
2239 saves time in cases of nested unlimited repeats that are never going to match.
2240 Problem: the scan can take a lot of time if the subject is very long (e.g.
2241 megabytes), thus penalizing straightforward matches. It is now done only if the
2242 amount of subject to be scanned is less than 1000 bytes.
2243
2244 7. A lesser problem with the same optimization is that it was recording the
2245 first character of an anchored pattern as "needed", thus provoking a search
2246 right along the subject, even when the first match of the pattern was going to
2247 fail. The "needed" character is now not set for anchored patterns, unless it
2248 follows something in the pattern that is of non-fixed length. Thus, it still
2249 fulfils its original purpose of finding quick non-matches in cases of nested
2250 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
2251
2252
2253 Version 4.0 17-Feb-03
2254 ---------------------
2255
2256 1. If a comment in an extended regex that started immediately after a meta-item
2257 extended to the end of string, PCRE compiled incorrect data. This could lead to
2258 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
2259
2260 2. Moved to autoconf 2.53 and libtool 1.4.2.
2261
2262 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
2263 the special perltest8 script is no longer needed - all the tests can be run
2264 from a single perltest script.
2265
2266 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
2267 by \s. It has now been removed in PCRE. This means it isn't recognized as
2268 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
2269 class [:space:] *does* include VT, thereby creating a mess.
2270
2271 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
2272 space and tab.
2273
2274 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
2275 its new features into the main test script, reducing the number of scripts.
2276
2277 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
2278 were backward compatible, and made the (?i) apply to the whole pattern, as if
2279 /i were given. Now it behaves more logically, and applies the option setting
2280 only to what follows. PCRE has been changed to follow suit. However, if it
2281 finds options settings right at the start of the pattern, it extracts them into
2282 the global options, as before. Thus, they show up in the info data.
2283
2284 8. Added support for the \Q...\E escape sequence. Characters in between are
2285 treated as literals. This is slightly different from Perl in that $ and @ are
2286 also handled as literals inside the quotes. In Perl, they will cause variable
2287 interpolation. Note the following examples:
2288
2289 Pattern PCRE matches Perl matches
2290
2291 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
2292 \Qabc\$xyz\E abc\$xyz abc\$xyz
2293 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
2294
2295 For compatibility with Perl, \Q...\E sequences are recognized inside character
2296 classes as well as outside them.
2297
2298 9. Re-organized 3 code statements in pcretest to avoid "overflow in
2299 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
2300 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
2301 signed/unsigned warnings.
2302
2303 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
2304 option for pcretest, so I've replaced it by a simple function that does just
2305 that job.
2306
2307 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
2308 "pcregrep -".
2309
2310 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
2311 Java package. This provides some syntactic sugar for simple cases of what my
2312 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
2313 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
2314 item, you can use this simplified notation. Note that only makes sense with
2315 greedy quantifiers. Consequently, the use of the possessive quantifier forces
2316 greediness, whatever the setting of the PCRE_UNGREEDY option.
2317
2318 13. A change of greediness default within a pattern was not taking effect at
2319 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
2320 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
2321 was abstracted outside.
2322
2323 14. PCRE now supports the \G assertion. It is true when the current matching
2324 position is at the start point of the match. This differs from \A when the
2325 starting offset is non-zero. Used with the /g option of pcretest (or similar
2326 code), it works in the same way as it does for Perl's /g option. If all
2327 alternatives of a regex begin with \G, the expression is anchored to the start
2328 match position, and the "anchored" flag is set in the compiled expression.
2329
2330 15. Some bugs concerning the handling of certain option changes within patterns
2331 have been fixed. These applied to options other than (?ims). For example,
2332 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
2333 been the other way round. Some of this was related to change 7 above.
2334
2335 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
2336 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
2337 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
2338 POSIX classes only within a class (e.g. /[[:alpha:]]/).
2339
2340 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
2341 mode. Unlike ".", it always matches newline, whatever the setting of
2342 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
2343 assertions. Perl allows it, but it doesn't (in general) work because it can't
2344 calculate the length of the lookbehind. At least, that's the case for Perl
2345 5.8.0 - I've been told they are going to document that it doesn't work in
2346 future.
2347
2348 18. Added an error diagnosis for escapes that PCRE does not support: these are
2349 \L, \l, \N, \P, \p, \U, \u, and \X.
2350
2351 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
2352 reading past the end of the pattern in cases such as /[abcd/.
2353
2354 20. PCRE was getting more memory than necessary for patterns with classes that
2355 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
2356
2357 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
2358 compiling PCRE for use with Virtual Pascal.
2359
2360 22. Small fix to the Makefile to make it work properly if the build is done
2361 outside the source tree.
2362
2363 23. Added a new extension: a condition to go with recursion. If a conditional
2364 subpattern starts with (?(R) the "true" branch is used if recursion has
2365 happened, whereas the "false" branch is used only at the top level.
2366
2367 24. When there was a very long string of literal characters (over 255 bytes
2368 without UTF support, over 250 bytes with UTF support), the computation of how
2369 much memory was required could be incorrect, leading to segfaults or other
2370 strange effects.
2371
2372 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
2373 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
2374 there was a subsequent back reference to those brackets. This meant that, for
2375 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
2376 possible to check for precisely this case. All we can do is abandon the
2377 optimization if .* occurs inside capturing brackets when there are any back
2378 references whatsoever. (See below for a better fix that came later.)
2379
2380 26. The handling of the optimization for finding the first character of a
2381 non-anchored pattern, and for finding a character that is required later in the
2382 match were failing in some cases. This didn't break the matching; it just
2383 failed to optimize when it could. The way this is done has been re-implemented.
2384
2385 27. Fixed typo in error message for invalid (?R item (it said "(?p").
2386
2387 28. Added a new feature that provides some of the functionality that Perl
2388 provides with (?{...}). The facility is termed a "callout". The way it is done
2389 in PCRE is for the caller to provide an optional function, by setting
2390 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
2391 global variable. By default it is unset, which disables all calling out. To get
2392 the function called, the regex must include (?C) at appropriate points. This
2393 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
2394 This provides a means of identifying different callout points. When PCRE
2395 reaches such a point in the regex, if pcre_callout has been set, the external
2396 function is called. It is provided with data in a structure called
2397 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
2398 matching continues; if it returns a non-zero value, the match at the current
2399 point fails. However, backtracking will occur if possible. [This was changed
2400 later and other features added - see item 49 below.]
2401
2402 29. pcretest is upgraded to test the callout functionality. It provides a
2403 callout function that displays information. By default, it shows the start of
2404 the match and the current position in the text. There are some new data escapes
2405 to vary what happens:
2406
2407 \C+ in addition, show current contents of captured substrings
2408 \C- do not supply a callout function
2409 \C!n return 1 when callout number n is reached
2410 \C!n!m return 1 when callout number n is reached for the mth time
2411
2412 30. If pcregrep was called with the -l option and just a single file name, it
2413 output "<stdin>" if a match was found, instead of the file name.
2414
2415 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
2416 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
2417 pcre_exec(). This saves a malloc/free per call. The default value of
2418 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
2419 when configuring.
2420
2421 32. The default maximum size of a compiled pattern is 64K. There have been a
2422 few cases of people hitting this limit. The code now uses macros to handle the
2423 storing of links as offsets within the compiled pattern. It defaults to 2-byte
2424 links, but this can be changed to 3 or 4 bytes by --with-link-size when
2425 configuring. Tests 2 and 5 work only with 2-byte links because they output
2426 debugging information about compiled patterns.
2427
2428 33. Internal code re-arrangements:
2429
2430 (a) Moved the debugging function for printing out a compiled regex into
2431 its own source file (printint.c) and used #include to pull it into
2432 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
2433 separate copies.
2434
2435 (b) Defined the list of op-code names for debugging as a macro in
2436 internal.h so that it is next to the definition of the opcodes.
2437
2438 (c) Defined a table of op-code lengths for simpler skipping along compiled
2439 code. This is again a macro in internal.h so that it is next to the
2440 definition of the opcodes.
2441
2442 34. Added support for recursive calls to individual subpatterns, along the
2443 lines of Robin Houston's patch (but implemented somewhat differently).
2444
2445 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
2446 allow it to read and process whole directories in Win32. This code was
2447 contributed by Lionel Fourquaux; it has not been tested by me.
2448
2449 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
2450 used to name a group. Names consist of alphanumerics and underscores, and must
2451 be unique. Back references use the syntax (?P=name) and recursive calls use
2452 (?P>name) which is a PCRE extension to the Python extension. Groups still have
2453 numbers. The function pcre_fullinfo() can be used after compilation to extract
2454 a name/number map. There are three relevant calls:
2455
2456 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
2457 PCRE_INFO_NAMECOUNT yields the number of entries
2458 PCRE_INFO_NAMETABLE yields a pointer to the map.
2459
2460 The map is a vector of fixed-size entries. The size of each entry depends on
2461 the length of the longest name used. The first two bytes of each entry are the
2462 group number, most significant byte first. There follows the corresponding
2463 name, zero terminated. The names are in alphabetical order.
2464
2465 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
2466 case instead of 255. Making it the same both with and without UTF-8 support
2467 means that the same test output works with both.
2468
2469 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
2470 calling malloc() with a zero argument.
2471
2472 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
2473 optimization. I've improved things by keeping a bitmap of backreferences with
2474 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
2475 fact referenced, the optimization can be applied. It is unlikely that a
2476 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
2477 the match to follow \n) will appear inside brackets with a number greater than
2478 31, but if it does, any back reference > 31 suppresses the optimization.
2479
2480 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
2481 of disabling numbered capturing parentheses. Any opening parenthesis that is
2482 not followed by ? behaves as if it were followed by ?: but named parentheses
2483 can still be used for capturing (and they will acquire numbers in the usual
2484 way).
2485
2486 41. Redesigned the return codes from the match() function into yes/no/error so
2487 that errors can be passed back from deep inside the nested calls. A malloc
2488 failure while inside a recursive subpattern call now causes the
2489 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
2490
2491 42. It is now possible to set a limit on the number of times the match()
2492 function is called in a call to pcre_exec(). This facility makes it possible to
2493 limit the amount of recursion and backtracking, though not in a directly
2494 obvious way, because the match() function is used in a number of different
2495 circumstances. The count starts from zero for each position in the subject
2496 string (for non-anchored patterns). The default limit is, for compatibility, a
2497 large number, namely 10 000 000. You can change this in two ways:
2498
2499 (a) When configuring PCRE before making, you can use --with-match-limit=n
2500 to set a default value for the compiled library.
2501
2502 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
2503 a different value is set. See 45 below.
2504
2505 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
2506
2507 43. Added a new function pcre_config(int, void *) to enable run-time extraction
2508 of things that can be changed at compile time. The first argument specifies
2509 what is wanted and the second points to where the information is to be placed.
2510 The current list of available information is:
2511
2512 PCRE_CONFIG_UTF8
2513
2514 The output is an integer that is set to one if UTF-8 support is available;
2515 otherwise it is set to zero.
2516
2517 PCRE_CONFIG_NEWLINE
2518
2519 The output is an integer that it set to the value of the code that is used for
2520 newline. It is either LF (10) or CR (13).
2521
2522 PCRE_CONFIG_LINK_SIZE
2523
2524 The output is an integer that contains the number of bytes used for internal
2525 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
2526
2527 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
2528
2529 The output is an integer that contains the threshold above which the POSIX
2530 interface uses malloc() for output vectors. See item 31 above.
2531
2532 PCRE_CONFIG_MATCH_LIMIT
2533
2534 The output is an unsigned integer that contains the default limit of the number
2535 of match() calls in a pcre_exec() execution. See 42 above.
2536
2537 44. pcretest has been upgraded by the addition of the -C option. This causes it
2538 to extract all the available output from the new pcre_config() function, and to
2539 output it. The program then exits immediately.
2540
2541 45. A need has arisen to pass over additional data with calls to pcre_exec() in
2542 order to support additional features. One way would have been to define
2543 pcre_exec2() (for example) with extra arguments, but this would not have been
2544 extensible, and would also have required all calls to the original function to
2545 be mapped to the new one. Instead, I have chosen to extend the mechanism that
2546 is used for passing in "extra" data from pcre_study().
2547
2548 The pcre_extra structure is now exposed and defined in pcre.h. It currently
2549 contains the following fields:
2550
2551 flags a bitmap indicating which of the following fields are set
2552 study_data opaque data from pcre_study()
2553 match_limit a way of specifying a limit on match() calls for a specific
2554 call to pcre_exec()
2555 callout_data data for callouts (see 49 below)
2556
2557 The flag bits are also defined in pcre.h, and are
2558
2559 PCRE_EXTRA_STUDY_DATA
2560 PCRE_EXTRA_MATCH_LIMIT
2561 PCRE_EXTRA_CALLOUT_DATA
2562
2563 The pcre_study() function now returns one of these new pcre_extra blocks, with
2564 the actual study data pointed to by the study_data field, and the
2565 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
2566 before. That is, this change is entirely upwards-compatible and requires no
2567 change to existing code.
2568
2569 If you want to pass in additional data to pcre_exec(), you can either place it
2570 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
2571 block.
2572
2573 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
2574 data string contains the escape sequence \M, pcretest calls pcre_exec() several
2575 times with different match limits, until it finds the minimum value needed for
2576 pcre_exec() to complete. The value is then output. This can be instructive; for
2577 most simple matches the number is quite small, but for pathological cases it
2578 gets very large very quickly.
2579
2580 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
2581 returns the size of the data block pointed to by the study_data field in a
2582 pcre_extra block, that is, the value that was passed as the argument to
2583 pcre_malloc() when PCRE was getting memory in which to place the information
2584 created by pcre_study(). The fourth argument should point to a size_t variable.
2585 pcretest has been extended so that this information is shown after a successful
2586 pcre_study() call when information about the compiled regex is being displayed.
2587
2588 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
2589 because what follows is always an absolute path. (Later: it turns out that this
2590 is more than cosmetic for MinGW, because it doesn't like empty path
2591 components.)
2592
2593 49. Some changes have been made to the callout feature (see 28 above):
2594
2595 (i) A callout function now has three choices for what it returns:
2596
2597 0 => success, carry on matching
2598 > 0 => failure at this point, but backtrack if possible
2599 < 0 => serious error, return this value from pcre_exec()
2600
2601 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
2602 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
2603 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
2604 use by callout functions. It will never be used by PCRE itself.
2605
2606 (ii) The pcre_extra structure (see 45 above) has a void * field called
2607 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
2608 pcre_callout_block structure has a field of the same name. The contents of
2609 the field passed in the pcre_extra structure are passed to the callout
2610 function in the corresponding field in the callout block. This makes it
2611 easier to use the same callout-containing regex from multiple threads. For
2612 testing, the pcretest program has a new data escape
2613
2614 \C*n pass the number n (may be negative) as callout_data
2615
2616 If the callout function in pcretest receives a non-zero value as
2617 callout_data, it returns that value.
2618
2619 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
2620 there were some redundant $(CFLAGS) in commands that are now specified as
2621 $(LINK), which already includes $(CFLAGS).
2622
2623 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
2624 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
2625 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
2626 one-byte characters throughout. Note that case-insensitive matching applies
2627 only to characters whose values are less than 256. PCRE doesn't support the
2628 notion of cases for higher-valued characters.
2629
2630 (i) A character class whose characters are all within 0-255 is handled as
2631 a bit map, and the map is inverted for negative classes. Previously, a
2632 character > 255 always failed to match such a class; however it should
2633 match if the class was a negative one (e.g. [^ab]). This has been fixed.
2634
2635 (ii) A negated character class with a single character < 255 is coded as
2636 "not this character" (OP_NOT). This wasn't working properly when the test
2637 character was multibyte, either singly or repeated.
2638
2639 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
2640 mode, for example: \x{100}{2,3}.
2641
2642 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
2643 singly or repeated) now correctly test multibyte characters. However,
2644 PCRE doesn't recognize any characters with values greater than 255 as
2645 digits, spaces, or word characters. Such characters always match \D, \S,
2646 and \W, and never match \d, \s, or \w.
2647
2648 (v) Classes may now contain characters and character ranges with values
2649 greater than 255. For example: [ab\x{100}-\x{400}].
2650
2651 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
2652 PCRE in UTF-8 mode.
2653
2654 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
2655 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
2656 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
2657 value.)
2658
2659 53. The single man page has become too large. I have therefore split it up into
2660 a number of separate man pages. These also give rise to individual HTML pages;
2661 these are now put in a separate directory, and there is an index.html page that
2662 lists them all. Some hyperlinking between the pages has been installed.
2663
2664 54. Added convenience functions for handling named capturing parentheses.
2665
2666 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
2667 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
2668 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
2669 are faulted.
2670
2671 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
2672 calling configure. These values are used when compiling the dftables.c program
2673 which is run to generate the source of the default character tables. They
2674 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
2675 you will need to set these values.
2676
2677 57. Updated the building process for Windows DLL, as provided by Fred Cox.
2678
2679
2680 Version 3.9 02-Jan-02
2681 ---------------------
2682
2683 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
2684
2685 2. If --disable-static was given, the building process failed when trying to
2686 build pcretest and pcregrep. (For some reason it was using libtool to compile
2687 them, which is not right, as they aren't part of the library.)
2688
2689
2690 Version 3.8 18-Dec-01
2691 ---------------------
2692
2693 1. The experimental UTF-8 code was completely screwed up. It was packing the
2694 bytes in the wrong order. How dumb can you get?
2695
2696
2697 Version 3.7 29-Oct-01
2698 ---------------------
2699
2700 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
2701 This caused pcretest, when used on the test data, to segfault. Unfortunately,
2702 this didn't happen under Solaris 8, where I normally test things.
2703
2704 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
2705 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
2706 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
2707 this fix an hour or so after the initial 3.7 release.)
2708
2709
2710 Version 3.6 23-Oct-01
2711 ---------------------
2712
2713 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
2714 offsets passed as NULL with zero offset count.
2715
2716 2. The config.guess and config.sub files had not been updated when I moved to
2717 the latest autoconf.
2718
2719
2720 Version 3.5 15-Aug-01
2721 ---------------------
2722
2723 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
2724 had been forgotten.
2725
2726 2. By using declared but undefined structures, we can avoid using "void"
2727 definitions in pcre.h while keeping the internal definitions of the structures
2728 private.
2729
2730 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
2731 user point of view, this means that both static and shared libraries are built
2732 by default, but this can be individually controlled. More of the work of
2733 handling this static/shared cases is now inside libtool instead of PCRE's make
2734 file.
2735
2736 4. The pcretest utility is now installed along with pcregrep because it is
2737 useful for users (to test regexs) and by doing this, it automatically gets
2738 relinked by libtool. The documentation has been turned into a man page, so
2739 there are now .1, .txt, and .html versions in /doc.
2740
2741 5. Upgrades to pcregrep:
2742 (i) Added long-form option names like gnu grep.
2743 (ii) Added --help to list all options with an explanatory phrase.
2744 (iii) Added -r, --recursive to recurse into sub-directories.
2745 (iv) Added -f, --file to read patterns from a file.
2746
2747 6. pcre_exec() was referring to its "code" argument before testing that
2748 argument for NULL (and giving an error if it was NULL).
2749
2750 7. Upgraded Makefile.in to allow for compiling in a different directory from
2751 the source directory.
2752
2753 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
2754 options bits, the pointer it was passed was to an int instead of to an unsigned
2755 long int. This mattered only on 64-bit systems.
2756
2757 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
2758 generated) instead of pcre.in, which it its source. Also made the same change
2759 in several of the .c files.
2760
2761 10. A new release of gcc defines printf() as a macro, which broke pcretest
2762 because it had an ifdef in the middle of a string argument for printf(). Fixed
2763 by using separate calls to printf().
2764
2765 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
2766 script, to force use of CR or LF instead of \n in the source. On non-Unix
2767 systems, the value can be set in config.h.
2768
2769 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
2770 absolute limit. Changed the text of the error message to make this clear, and
2771 likewise updated the man page.
2772
2773 13. The limit of 99 on the number of capturing subpatterns has been removed.
2774 The new limit is 65535, which I hope will not be a "real" limit.
2775
2776
2777 Version 3.4 22-Aug-00
2778 ---------------------
2779
2780 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
2781
2782 2. Diagnose condition (?(0) as an error instead of crashing on matching.
2783
2784
2785 Version 3.3 01-Aug-00
2786 ---------------------
2787
2788 1. If an octal character was given, but the value was greater than \377, it
2789 was not getting masked to the least significant bits, as documented. This could
2790 lead to crashes in some systems.
2791
2792 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
2793 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
2794
2795 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
2796 These just pass their arguments on to (pcre_free)(), but they are provided
2797 because some uses of PCRE bind it to non-C systems that can call its functions,
2798 but cannot call free() or pcre_free() directly.
2799
2800 4. Add "make test" as a synonym for "make check". Corrected some comments in
2801 the Makefile.
2802
2803 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
2804 Makefile.
2805
2806 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
2807 command called pgrep for grepping around the active processes.
2808
2809 7. Added the beginnings of support for UTF-8 character strings.
2810
2811 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
2812 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
2813 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
2814 out for the ar command.)
2815
2816
2817 Version 3.2 12-May-00
2818 ---------------------
2819
2820 This is purely a bug fixing release.
2821
2822 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
2823 of ZA. This was just one example of several cases that could provoke this bug,
2824 which was introduced by change 9 of version 2.00. The code for breaking
2825 infinite loops after an iteration that matches an empty string was't working
2826 correctly.
2827
2828 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
2829 when matched against abbab (for example). After matching an empty string, it
2830 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
2831 caused it to match further down the string than it should.
2832
2833 3. The code contained an inclusion of sys/types.h. It isn't clear why this
2834 was there because it doesn't seem to be needed, and it causes trouble on some
2835 systems, as it is not a Standard C header. It has been removed.
2836
2837 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
2838 were reported on the Macintosh. The changes were from
2839
2840 while ((c = *(++ptr)) != 0 && c != '\n');
2841 to
2842 while ((c = *(++ptr)) != 0 && c != '\n') ;
2843
2844 Totally extraordinary, but if that's what it takes...
2845
2846 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
2847 available. Added HAVE_BCOPY and an autoconf test for it; if neither
2848 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
2849 assumes the way PCRE uses memmove() (always moving upwards).
2850
2851 6. PCRE is being used in one environment where strchr() is not available. There
2852 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
2853 faster code anyway.
2854
2855
2856 Version 3.1 09-Feb-00
2857 ---------------------
2858
2859 The only change in this release is the fixing of some bugs in Makefile.in for
2860 the "install" target:
2861
2862 (1) It was failing to install pcreposix.h.
2863
2864 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
2865
2866
2867 Version 3.0 01-Feb-00
2868 ---------------------
2869
2870 1. Add support for the /+ modifier to perltest (to output $` like it does in
2871 pcretest).
2872
2873 2. Add support for the /g modifier to perltest.
2874
2875 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
2876 matches null strings.
2877
2878 4. Fix perltest so that it doesn't do unwanted things when fed an empty
2879 pattern. Perl treats empty patterns specially - it reuses the most recent
2880 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
2881 effect.
2882
2883 5. The POSIX interface was broken in that it was just handing over the POSIX
2884 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
2885 required a bigger vector, with some working space on the end. This means that
2886 the POSIX wrapper now has to get and free some memory, and copy the results.
2887
2888 6. Added some simple autoconf support, placing the test data and the
2889 documentation in separate directories, re-organizing some of the
2890 information files, and making it build pcre-config (a GNU standard). Also added
2891 libtool support for building PCRE as a shared library, which is now the
2892 default.
2893
2894 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
2895 09 are not valid octal constants. Single digits will be used for minor values
2896 less than 10.
2897
2898 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
2899 existing programs that set these in the POSIX interface can use PCRE without
2900 modification.
2901
2902 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
2903 return all that pcre_info() returns, plus additional data. The pcre_info()
2904 function is retained for compatibility, but is considered to be obsolete.
2905
2906 10. Added experimental recursion feature (?R) to handle one common case that
2907 Perl 5.6 will be able to do with (?p{...}).
2908
2909 11. Added support for POSIX character classes like [:alpha:], which Perl is
2910 adopting.
2911
2912
2913 Version 2.08 31-Aug-99
2914 ----------------------
2915
2916 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
2917 trying to match at the startoffset position, but instead was moving forward to
2918 the next newline as if a previous match had failed.
2919
2920 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
2921 and could get into a loop if a null string was matched other than at the start
2922 of the subject.
2923
2924 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
2925 be distinguished at compile time, and for completeness also added PCRE_DATE.
2926
2927 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
2928 in GnuWin32 environments.
2929
2930
2931 Version 2.07 29-Jul-99
2932 ----------------------
2933
2934 1. The documentation is now supplied in plain text form and HTML as well as in
2935 the form of man page sources.
2936
2937 2. C++ compilers don't like assigning (void *) values to other pointer types.
2938 In particular this affects malloc(). Although there is no problem in Standard
2939 C, I've put in casts to keep C++ compilers happy.
2940
2941 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
2942 should be (const char *).
2943
2944 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
2945 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
2946 However, I haven't made this a standard facility. The documentation doesn't
2947 mention it, and the Makefile doesn't support it.
2948
2949 5. The Makefile now contains an "install" target, with editable destinations at
2950 the top of the file. The pcretest program is not installed.
2951
2952 6. pgrep -V now gives the PCRE version number and date.
2953
2954 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
2955 causing the entire string to be ignored, instead of just the last character.
2956
2957 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
2958 non-matching string, it can take a very, very long time, even for strings of
2959 quite modest length, because of the nested recursion. PCRE now does better in
2960 some of these cases. It does this by remembering the last required literal
2961 character in the pattern, and pre-searching the subject to ensure it is present
2962 before running the real match. In other words, it applies a heuristic to detect
2963 some types of certain failure quickly, and in the above example, if presented
2964 with a string that has no trailing " it gives "no match" very quickly.
2965
2966 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
2967 other alternatives are tried instead.
2968
2969
2970 Version 2.06 09-Jun-99
2971 ----------------------
2972
2973 1. Change pcretest's output for amount of store used to show just the code
2974 space, because the remainder (the data block) varies in size between 32-bit and
2975 64-bit systems.
2976
2977 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
2978 start matching at. This allows lookbehinds to work when searching for multiple
2979 occurrences in a string.
2980
2981 3. Added additional options to pcretest for testing multiple occurrences:
2982
2983 /+ outputs the rest of the string that follows a match
2984 /g loops for multiple occurrences, using the new startoffset argument
2985 /G loops for multiple occurrences by passing an incremented pointer
2986
2987 4. PCRE wasn't doing the "first character" optimization for patterns starting
2988 with \b or \B, though it was doing it for other lookbehind assertions. That is,
2989 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
2990 the letter 'x'. On long subject strings, this gives a significant speed-up.
2991
2992
2993 Version 2.05 21-Apr-99
2994 ----------------------
2995
2996 1. Changed the type of magic_number from int to long int so that it works
2997 properly on 16-bit systems.
2998
2999 2. Fixed a bug which caused patterns starting with .* not to work correctly
3000 when the subject string contained newline characters. PCRE was assuming
3001 anchoring for such patterns in all cases, which is not correct because .* will
3002 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
3003 DOTALL is set at top level; otherwise it knows that patterns starting with .*
3004 must be retried after every newline in the subject.
3005
3006
3007 Version 2.04 18-Feb-99
3008 ----------------------
3009
3010 1. For parenthesized subpatterns with repeats whose minimum was zero, the
3011 computation of the store needed to hold the pattern was incorrect (too large).
3012 If such patterns were nested a few deep, this could multiply and become a real
3013 problem.
3014
3015 2. Added /M option to pcretest to show the memory requirement of a specific
3016 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
3017
3018 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
3019 compiled in such a way that the backtracking after subsequent failure was
3020 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
3021 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
3022
3023
3024 Version 2.03 02-Feb-99
3025 ----------------------
3026
3027 1. Fixed typo and small mistake in man page.
3028
3029 2. Added 4th condition (GPL supersedes if conflict) and created separate
3030 LICENCE file containing the conditions.
3031
3032 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
3033 Perl, that is the internal \ allows the delimiter to be included in the
3034 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
3035 the final delimiter, add \ to the end of the pattern (to test the error).
3036
3037 4. Added the convenience functions for extracting substrings after a successful
3038 match. Updated pcretest to make it able to test these functions.
3039
3040
3041 Version 2.02 14-Jan-99
3042 ----------------------
3043
3044 1. Initialized the working variables associated with each extraction so that
3045 their saving and restoring doesn't refer to uninitialized store.
3046
3047 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
3048 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
3049 fix the problem.
3050
3051 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
3052 calls, and wasn't printing the correct value for compiling calls. Increased the
3053 default value of LOOPREPEAT, and the number of significant figures in the
3054 times.
3055
3056 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
3057
3058 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
3059 a building problem on Windows NT with a FAT file system.
3060
3061
3062 Version 2.01 21-Oct-98
3063 ----------------------
3064
3065 1. Changed the API for pcre_compile() to allow for the provision of a pointer
3066 to character tables built by pcre_maketables() in the current locale. If NULL
3067 is passed, the default tables are used.
3068
3069
3070 Version 2.00 24-Sep-98
3071 ----------------------
3072
3073 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
3074 it any more.
3075
3076 2. Allow quantification of (?>) groups, and make it work correctly.
3077
3078 3. The first character computation wasn't working for (?>) groups.
3079
3080 4. Correct the implementation of \Z (it is permitted to match on the \n at the
3081 end of the subject) and add 5.005's \z, which really does match only at the
3082 very end of the subject.
3083
3084 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
3085
3086 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
3087 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
3088 localized options. All options to pcre_study() were also removed.
3089
3090 7. Add other new features from 5.005:
3091
3092 $(?<= positive lookbehind
3093 $(?<! negative lookbehind
3094 (?imsx-imsx) added the unsetting capability
3095 such a setting is global if at outer level; local otherwise
3096 (?imsx-imsx:) non-capturing groups with option setting
3097 (?(cond)re|re) conditional pattern matching
3098
3099 A backreference to itself in a repeated group matches the previous
3100 captured string.
3101
3102 8. General tidying up of studying (both automatic and via "study")
3103 consequential on the addition of new assertions.
3104
3105 9. As in 5.005, unlimited repeated groups that could match an empty substring
3106 are no longer faulted at compile time. Instead, the loop is forcibly broken at
3107 runtime if any iteration does actually match an empty substring.
3108
3109 10. Include the RunTest script in the distribution.
3110
3111 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
3112 discrepancies, some of which were old and were also with respect to 5.004. They
3113 have now been fixed.
3114
3115
3116 Version 1.09 28-Apr-98
3117 ----------------------
3118
3119 1. A negated single character class followed by a quantifier with a minimum
3120 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
3121 program crashes, or just wrong answers. This did not apply to negated classes
3122 containing more than one character, or to minima other than one.
3123
3124
3125 Version 1.08 27-Mar-98
3126 ----------------------
3127
3128 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
3129
3130 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
3131 latter must appear before anything that relies on it in the pattern.
3132
3133
3134 Version 1.07 16-Feb-98
3135 ----------------------
3136
3137 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
3138 repeat of a potentially empty string).
3139
3140
3141 Version 1.06 23-Jan-98
3142 ----------------------
3143
3144 1. Added Markus Oberhumer's little patches for C++.
3145
3146 2. Literal strings longer than 255 characters were broken.
3147
3148
3149 Version 1.05 23-Dec-97
3150 ----------------------
3151
3152 1. Negated character classes containing more than one character were failing if
3153 PCRE_CASELESS was set at run time.
3154
3155
3156 Version 1.04 19-Dec-97
3157 ----------------------
3158
3159 1. Corrected the man page, where some "const" qualifiers had been omitted.
3160
3161 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
3162 input syntax.
3163
3164 3. Fixed memory leak which occurred when a regex with back references was
3165 matched with an offsets vector that wasn't big enough. The temporary memory
3166 that is used in this case wasn't being freed if the match failed.
3167
3168 4. Tidied pcretest to ensure it frees memory that it gets.
3169
3170 5. Temporary memory was being obtained in the case where the passed offsets
3171 vector was exactly big enough.
3172
3173 6. Corrected definition of offsetof() from change 5 below.
3174
3175 7. I had screwed up change 6 below and broken the rules for the use of
3176 setjmp(). Now fixed.
3177
3178
3179 Version 1.03 18-Dec-97
3180 ----------------------
3181
3182 1. A erroneous regex with a missing opening parenthesis was correctly
3183 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
3184 on some systems.
3185
3186 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
3187 it was reported that one broken compiler failed on the former because "code" is
3188 also an independent variable.
3189
3190 3. The erroneous regex a[]b caused an array overrun reference.
3191
3192 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
3193 fail on data ending with that character. (It was going on too far, and checking
3194 the next character, typically a binary zero.) This was specific to the
3195 optimized code for single-character negative classes.
3196
3197 5. Added a contributed patch from the TIN world which does the following:
3198
3199 + Add an undef for memmove, in case the the system defines a macro for it.
3200
3201 + Add a definition of offsetof(), in case there isn't one. (I don't know
3202 the reason behind this - offsetof() is part of the ANSI standard - but
3203 it does no harm).
3204
3205 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
3206 most of the places where whitespace preceded '#'. I have given up and
3207 allowed the remaining 2 cases to be at the margin.
3208
3209 + Rename some variables in pcre to eliminate shadowing. This seems very
3210 pedantic, but does no harm, of course.
3211
3212 6. Moved the call to setjmp() into its own function, to get rid of warnings
3213 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
3214
3215 7. Constructs such as \d{8,} were compiling into the equivalent of
3216 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
3217 outcome, but in this particular case used more store than had been allocated,
3218 which caused the bug to be discovered because it threw up an internal error.
3219
3220 8. The debugging code in both pcre and pcretest for outputting the compiled
3221 form of a regex was going wrong in the case of back references followed by
3222 curly-bracketed repeats.
3223
3224
3225 Version 1.02 12-Dec-97
3226 ----------------------
3227
3228 1. Typos in pcre.3 and comments in the source fixed.
3229
3230 2. Applied a contributed patch to get rid of places where it used to remove
3231 'const' from variables, and fixed some signed/unsigned and uninitialized
3232 variable warnings.
3233
3234 3. Added the "runtest" target to Makefile.
3235
3236 4. Set default compiler flag to -O2 rather than just -O.
3237
3238
3239 Version 1.01 19-Nov-97
3240 ----------------------
3241
3242 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
3243 like /([ab]*)*/, that is, for classes with more than one character in them.
3244
3245 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
3246 as /((?>a*))*/ (a PCRE_EXTRA facility).
3247
3248
3249 Version 1.00 18-Nov-97
3250 ----------------------
3251
3252 1. Added compile-time macros to support systems such as SunOS4 which don't have
3253 memmove() or strerror() but have other things that can be used instead.
3254
3255 2. Arranged that "make clean" removes the executables.
3256
3257
3258 Version 0.99 27-Oct-97
3259 ----------------------
3260
3261 1. Fixed bug in code for optimizing classes with only one character. It was
3262 initializing a 32-byte map regardless, which could cause it to run off the end
3263 of the memory it had got.
3264
3265 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
3266
3267
3268 Version 0.98 22-Oct-97
3269 ----------------------
3270
3271 1. Fixed bug in code for handling temporary memory usage when there are more
3272 back references than supplied space in the ovector. This could cause segfaults.
3273
3274
3275 Version 0.97 21-Oct-97
3276 ----------------------
3277
3278 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
3279
3280 2. Optimized negated single characters not to use a bit map.
3281
3282 3. Brought error texts together as macro definitions; clarified some of them;
3283 fixed one that was wrong - it said "range out of order" when it meant "invalid
3284 escape sequence".
3285
3286 4. Changed some char * arguments to const char *.
3287
3288 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
3289
3290 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
3291 pcretest.
3292
3293
3294 Version 0.96 16-Oct-97
3295 ----------------------
3296
3297 1. Added a simple "pgrep" utility to the distribution.
3298
3299 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
3300 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
3301 where "ddd" means "one or more decimal digits".
3302
3303 3. Fixed serious bug. If a pattern had a back reference, but the call to
3304 pcre_exec() didn't supply a large enough ovector to record the related
3305 identifying subpattern, the match always failed. PCRE now remembers the number
3306 of the largest back reference, and gets some temporary memory in which to save
3307 the offsets during matching if necessary, in order to ensure that
3308 backreferences always work.
3309
3310 4. Increased the compatibility with Perl in a number of ways:
3311
3312 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
3313 to request this handling. The option can be set at compile or exec time.
3314
3315 (b) $ matches before a terminating newline by default; an option
3316 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
3317 mode). The option can be set at compile or exec time.
3318
3319 (c) The handling of \ followed by a digit other than 0 is now supposed to be
3320 the same as Perl's. If the decimal number it represents is less than 10
3321 or there aren't that many previous left capturing parentheses, an octal
3322 escape is read. Inside a character class, it's always an octal escape,
3323 even if it is a single digit.
3324
3325 (d) An escaped but undefined alphabetic character is taken as a literal,
3326 unless PCRE_EXTRA is set. Currently this just reserves the remaining
3327 escapes.
3328
3329 (e) {0} is now permitted. (The previous item is removed from the compiled
3330 pattern).
3331
3332 5. Changed all the names of code files so that the basic parts are no longer
3333 than 10 characters, and abolished the teeny "globals.c" file.
3334
3335 6. Changed the handling of character classes; they are now done with a 32-byte
3336 bit map always.
3337
3338 7. Added the -d and /D options to pcretest to make it possible to look at the
3339 internals of compilation without having to recompile pcre.
3340
3341
3342 Version 0.95 23-Sep-97
3343 ----------------------
3344
3345 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
3346 \x20 at the start of a run of normal characters. These were being treated as
3347 real characters, instead of the source characters being re-checked.
3348
3349
3350 Version 0.94 18-Sep-97
3351 ----------------------
3352
3353 1. The functions are now thread-safe, with the caveat that the global variables
3354 containing pointers to malloc() and free() or alternative functions are the
3355 same for all threads.
3356
3357 2. Get pcre_study() to generate a bitmap of initial characters for non-
3358 anchored patterns when this is possible, and use it if passed to pcre_exec().
3359
3360
3361 Version 0.93 15-Sep-97
3362 ----------------------
3363
3364 1. /(b)|(:+)/ was computing an incorrect first character.
3365
3366 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
3367 but not actually doing anything yet.
3368
3369 3. Treat "-" characters in classes that cannot be part of ranges as literals,
3370 as Perl does (e.g. [-az] or [az-]).
3371
3372 4. Set the anchored flag if a branch starts with .* or .*? because that tests
3373 all possible positions.
3374
3375 5. Split up into different modules to avoid including unneeded functions in a
3376 compiled binary. However, compile and exec are still in one module. The "study"
3377 function is split off.
3378
3379 6. The character tables are now in a separate module whose source is generated
3380 by an auxiliary program - but can then be edited by hand if required. There are
3381 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
3382 toupper() in the code.
3383
3384 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
3385 make them global. Abolish the function for setting them, as the caller can now
3386 set them directly.
3387
3388
3389 Version 0.92 11-Sep-97
3390 ----------------------
3391
3392 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
3393 (e.g. /a{1,3}/) was broken (I mis-optimized it).
3394
3395 2. Caseless matching was not working in character classes if the characters in
3396 the pattern were in upper case.
3397
3398 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
3399
3400 4. Make PCRE_ANCHORED public and accept as a compile option.
3401
3402 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
3403 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
3404 pass them.
3405
3406 6. Give an error if bad option bits passed at compile or run time.
3407
3408 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
3409 pcretest to cause it to pass that flag.
3410
3411 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
3412 options, and the first character, if set.
3413
3414 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
3415
3416
3417 Version 0.91 10-Sep-97
3418 ----------------------
3419
3420 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
3421 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
3422
3423 2. PCRE was looping on encountering an indefinitely repeated back reference to
3424 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
3425 Perl does - treats the match as successful.
3426
3427 ****

Properties

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

webmaster@exim.org
ViewVC Help
Powered by ViewVC 1.1.12