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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 324 - (show annotations) (download)
Fri Mar 7 19:48:32 2008 UTC (6 years, 7 months ago) by ph10
File size: 151973 byte(s)
Fix bugs with --include and --exclude in pcregrep.

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

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