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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 330 - (show annotations) (download)
Sat Apr 5 15:31:38 2008 UTC (6 years, 4 months ago) by ph10
File size: 152656 byte(s)
Craig's further amended patch for OS X.

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

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