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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 85 - (show annotations) (download)
Sat Feb 24 21:41:13 2007 UTC (7 years, 1 month ago) by nigel
File size: 84049 byte(s)
Load pcre-6.4 into code/trunk.

1 ChangeLog for PCRE
2 ------------------
3
4 Version 6.4 05-Sep-05
5 ---------------------
6
7 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
8 "--" to be printed when multiple files were scanned, even when none of the
9 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
10 consider it to be a bug, and have restored the previous behaviour.
11
12 2. A couple of code tidies to get rid of compiler warnings.
13
14 3. The pcretest program used to cheat by referring to symbols in the library
15 whose names begin with _pcre_. These are internal symbols that are not
16 really supposed to be visible externally, and in some environments it is
17 possible to suppress them. The cheating is now confined to including
18 certain files from the library's source, which is a bit cleaner.
19
20 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
21 file's purpose clearer.
22
23 5. Reorganized pcre_ucp_findchar().
24
25
26 Version 6.3 15-Aug-05
27 ---------------------
28
29 1. The file libpcre.pc.in did not have general read permission in the tarball.
30
31 2. There were some problems when building without C++ support:
32
33 (a) If C++ support was not built, "make install" and "make test" still
34 tried to test it.
35
36 (b) There were problems when the value of CXX was explicitly set. Some
37 changes have been made to try to fix these, and ...
38
39 (c) --disable-cpp can now be used to explicitly disable C++ support.
40
41 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
42 backslash in a target when C++ was disabled. This confuses some
43 versions of "make", apparently. Using an intermediate variable solves
44 this. (Same for CPP_LOBJ.)
45
46 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
47 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
48 necessary on certain architectures.
49
50 4. Added a setting of -export-symbols-regex to the link command to remove
51 those symbols that are exported in the C sense, but actually are local
52 within the library, and not documented. Their names all begin with
53 "_pcre_". This is not a perfect job, because (a) we have to except some
54 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
55 available (and never for static libraries). I have made a note to try to
56 find a way round (a) in the future.
57
58
59 Version 6.2 01-Aug-05
60 ---------------------
61
62 1. There was no test for integer overflow of quantifier values. A construction
63 such as {1111111111111111} would give undefined results. What is worse, if
64 a minimum quantifier for a parenthesized subpattern overflowed and became
65 negative, the calculation of the memory size went wrong. This could have
66 led to memory overwriting.
67
68 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
69
70 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
71 operating environments where this matters.
72
73 4. Applied Giuseppe Maxia's patch to add additional features for controlling
74 PCRE options from within the C++ wrapper.
75
76 5. Named capturing subpatterns were not being correctly counted when a pattern
77 was compiled. This caused two problems: (a) If there were more than 100
78 such subpatterns, the calculation of the memory needed for the whole
79 compiled pattern went wrong, leading to an overflow error. (b) Numerical
80 back references of the form \12, where the number was greater than 9, were
81 not recognized as back references, even though there were sufficient
82 previous subpatterns.
83
84 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
85 versions of gcc, e.g. 2.95.4.
86
87
88 Version 6.1 21-Jun-05
89 ---------------------
90
91 1. There was one reference to the variable "posix" in pcretest.c that was not
92 surrounded by "#if !defined NOPOSIX".
93
94 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
95 the cross-check on the old pcre_info() function, for the benefit of the
96 cut-down version of PCRE that is currently imported into Exim.
97
98 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
99 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
100 patterns that start (?i) but I don't think that matters. The use of (?i) is
101 just an example; this all applies to the other options as well.
102
103 4. Since libtool seems to echo the compile commands it is issuing, the output
104 from "make" can be reduced a bit by putting "@" in front of each libtool
105 compile command.
106
107 5. Patch from the folks at Google for configure.in to be a bit more thorough
108 in checking for a suitable C++ installation before trying to compile the
109 C++ stuff. This should fix a reported problem when a compiler was present,
110 but no suitable headers.
111
112 6. The man pages all had just "PCRE" as their title. I have changed them to
113 be the relevant file name. I have also arranged that these names are
114 retained in the file doc/pcre.txt, which is a concatenation in text format
115 of all the man pages except the little individual ones for each function.
116
117 7. The NON-UNIX-USE file had not been updated for the different set of source
118 files that come with release 6. I also added a few comments about the C++
119 wrapper.
120
121
122 Version 6.0 07-Jun-05
123 ---------------------
124
125 1. Some minor internal re-organization to help with my DFA experiments.
126
127 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
128 didn't matter for the library itself when fully configured, but did matter
129 when compiling without UCP support, or within Exim, where the ucp files are
130 not imported.
131
132 3. Refactoring of the library code to split up the various functions into
133 different source modules. The addition of the new DFA matching code (see
134 below) to a single monolithic source would have made it really too
135 unwieldy, quite apart from causing all the code to be include in a
136 statically linked application, when only some functions are used. This is
137 relevant even without the DFA addition now that patterns can be compiled in
138 one application and matched in another.
139
140 The downside of splitting up is that there have to be some external
141 functions and data tables that are used internally in different modules of
142 the library but which are not part of the API. These have all had their
143 names changed to start with "_pcre_" so that they are unlikely to clash
144 with other external names.
145
146 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
147 a different (DFA) algorithm. Although it is slower than the original
148 function, it does have some advantages for certain types of matching
149 problem.
150
151 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
152 including restarting after a partial match.
153
154 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
155 defined when compiling for Windows was sent to me. I have put it into the
156 code, though I have no means of testing or verifying it.
157
158 7. Added the pcre_refcount() auxiliary function.
159
160 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
161 match before or at the first newline in the subject string. In pcretest,
162 the /f option on a pattern can be used to set this.
163
164 9. A repeated \w when used in UTF-8 mode with characters greater than 256
165 would behave wrongly. This has been present in PCRE since release 4.0.
166
167 10. A number of changes to the pcregrep command:
168
169 (a) Refactored how -x works; insert ^(...)$ instead of setting
170 PCRE_ANCHORED and checking the length, in preparation for adding
171 something similar for -w.
172
173 (b) Added the -w (match as a word) option.
174
175 (c) Refactored the way lines are read and buffered so as to have more
176 than one at a time available.
177
178 (d) Implemented a pcregrep test script.
179
180 (e) Added the -M (multiline match) option. This allows patterns to match
181 over several lines of the subject. The buffering ensures that at least
182 8K, or the rest of the document (whichever is the shorter) is available
183 for matching (and similarly the previous 8K for lookbehind assertions).
184
185 (f) Changed the --help output so that it now says
186
187 -w, --word-regex(p)
188
189 instead of two lines, one with "regex" and the other with "regexp"
190 because that confused at least one person since the short forms are the
191 same. (This required a bit of code, as the output is generated
192 automatically from a table. It wasn't just a text change.)
193
194 (g) -- can be used to terminate pcregrep options if the next thing isn't an
195 option but starts with a hyphen. Could be a pattern or a path name
196 starting with a hyphen, for instance.
197
198 (h) "-" can be given as a file name to represent stdin.
199
200 (i) When file names are being printed, "(standard input)" is used for
201 the standard input, for compatibility with GNU grep. Previously
202 "<stdin>" was used.
203
204 (j) The option --label=xxx can be used to supply a name to be used for
205 stdin when file names are being printed. There is no short form.
206
207 (k) Re-factored the options decoding logic because we are going to add
208 two more options that take data. Such options can now be given in four
209 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
210
211 (l) Added the -A, -B, and -C options for requesting that lines of context
212 around matches be printed.
213
214 (m) Added the -L option to print the names of files that do not contain
215 any matching lines, that is, the complement of -l.
216
217 (n) The return code is 2 if any file cannot be opened, but pcregrep does
218 continue to scan other files.
219
220 (o) The -s option was incorrectly implemented. For compatibility with other
221 greps, it now suppresses the error message for a non-existent or non-
222 accessible file (but not the return code). There is a new option called
223 -q that suppresses the output of matching lines, which was what -s was
224 previously doing.
225
226 (p) Added --include and --exclude options to specify files for inclusion
227 and exclusion when recursing.
228
229 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
230 Hopefully, it now does.
231
232 12. Missing cast in pcre_study().
233
234 13. Added an "uninstall" target to the makefile.
235
236 14. Replaced "extern" in the function prototypes in Makefile.in with
237 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
238 world, but is set differently for Windows.
239
240 15. Added a second compiling function called pcre_compile2(). The only
241 difference is that it has an extra argument, which is a pointer to an
242 integer error code. When there is a compile-time failure, this is set
243 non-zero, in addition to the error test pointer being set to point to an
244 error message. The new argument may be NULL if no error number is required
245 (but then you may as well call pcre_compile(), which is now just a
246 wrapper). This facility is provided because some applications need a
247 numeric error indication, but it has also enabled me to tidy up the way
248 compile-time errors are handled in the POSIX wrapper.
249
250 16. Added VPATH=.libs to the makefile; this should help when building with one
251 prefix path and installing with another. (Or so I'm told by someone who
252 knows more about this stuff than I do.)
253
254 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
255 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
256 match everything, including newlines. This is not POSIX-compatible, but
257 somebody wanted the feature. From pcretest it can be activated by using
258 both the P and the s flags.
259
260 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
261
262 19. libpcre.pc was being incorrectly installed as executable.
263
264 20. A couple of places in pcretest check for end-of-line by looking for '\n';
265 it now also looks for '\r' so that it will work unmodified on Windows.
266
267 21. Added Google's contributed C++ wrapper to the distribution.
268
269 22. Added some untidy missing memory free() calls in pcretest, to keep
270 Electric Fence happy when testing.
271
272
273
274 Version 5.0 13-Sep-04
275 ---------------------
276
277 1. Internal change: literal characters are no longer packed up into items
278 containing multiple characters in a single byte-string. Each character
279 is now matched using a separate opcode. However, there may be more than one
280 byte in the character in UTF-8 mode.
281
282 2. The pcre_callout_block structure has two new fields: pattern_position and
283 next_item_length. These contain the offset in the pattern to the next match
284 item, and its length, respectively.
285
286 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
287 insertion of callouts before each pattern item. Added the /C option to
288 pcretest to make use of this.
289
290 4. On the advice of a Windows user, the lines
291
292 #if defined(_WIN32) || defined(WIN32)
293 _setmode( _fileno( stdout ), 0x8000 );
294 #endif /* defined(_WIN32) || defined(WIN32) */
295
296 have been added to the source of pcretest. This apparently does useful
297 magic in relation to line terminators.
298
299 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
300 for the benefit of those environments where the "b" makes a difference.
301
302 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
303 to know about it. I have put a hack into configure.in that adds in code
304 to set GCC=yes if CC=icc. This seems to end up at a point in the
305 generated configure script that is early enough to affect the setting of
306 compiler options, which is what is needed, but I have no means of testing
307 whether it really works. (The user who reported this had patched the
308 generated configure script, which of course I cannot do.)
309
310 LATER: After change 22 below (new libtool files), the configure script
311 seems to know about icc (and also ecc). Therefore, I have commented out
312 this hack in configure.in.
313
314 7. Added support for pkg-config (2 patches were sent in).
315
316 8. Negated POSIX character classes that used a combination of internal tables
317 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
318 [[:^ascii]]. Typically, they would match almost any characters. The other
319 POSIX classes were not broken in this way.
320
321 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
322 to find the match, as PCRE was deluded into thinking that the match had to
323 start at the start point or following a newline. The same bug applied to
324 patterns with negative forward assertions or any backward assertions
325 preceding ".*" at the start, unless the pattern required a fixed first
326 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
327
328 10. In UTF-8 mode, when moving forwards in the subject after a failed match
329 starting at the last subject character, bytes beyond the end of the subject
330 string were read.
331
332 11. Renamed the variable "class" as "classbits" to make life easier for C++
333 users. (Previously there was a macro definition, but it apparently wasn't
334 enough.)
335
336 12. Added the new field "tables" to the extra data so that tables can be passed
337 in at exec time, or the internal tables can be re-selected. This allows
338 a compiled regex to be saved and re-used at a later time by a different
339 program that might have everything at different addresses.
340
341 13. Modified the pcre-config script so that, when run on Solaris, it shows a
342 -R library as well as a -L library.
343
344 14. The debugging options of pcretest (-d on the command line or D on a
345 pattern) showed incorrect output for anything following an extended class
346 that contained multibyte characters and which was followed by a quantifier.
347
348 15. Added optional support for general category Unicode character properties
349 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
350 support. It adds about 90K to the size of the library. The meanings of the
351 inbuilt class escapes such as \d and \s have NOT been changed.
352
353 16. Updated pcredemo.c to include calls to free() to release the memory for the
354 compiled pattern.
355
356 17. The generated file chartables.c was being created in the source directory
357 instead of in the building directory. This caused the build to fail if the
358 source directory was different from the building directory, and was
359 read-only.
360
361 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
362 file. No doubt somebody will tell me if they don't make sense... Also added
363 Dan Mooney's comments about building on OpenVMS.
364
365 19. Added support for partial matching via the PCRE_PARTIAL option for
366 pcre_exec() and the \P data escape in pcretest.
367
368 20. Extended pcretest with 3 new pattern features:
369
370 (i) A pattern option of the form ">rest-of-line" causes pcretest to
371 write the compiled pattern to the file whose name is "rest-of-line".
372 This is a straight binary dump of the data, with the saved pointer to
373 the character tables forced to be NULL. The study data, if any, is
374 written too. After writing, pcretest reads a new pattern.
375
376 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
377 compiled pattern from the given file. There must not be any
378 occurrences of "<" in the file name (pretty unlikely); if there are,
379 pcretest will instead treat the initial "<" as a pattern delimiter.
380 After reading in the pattern, pcretest goes on to read data lines as
381 usual.
382
383 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
384 and 16-bit fields in a compiled pattern, to simulate a pattern that
385 was compiled on a host of opposite endianness.
386
387 21. The pcre-exec() function can now cope with patterns that were compiled on
388 hosts of opposite endianness, with this restriction:
389
390 As for any compiled expression that is saved and used later, the tables
391 pointer field cannot be preserved; the extra_data field in the arguments
392 to pcre_exec() should be used to pass in a tables address if a value
393 other than the default internal tables were used at compile time.
394
395 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
396 now diagnosed as an error. Previously, most of the time, a negative number
397 would have been treated as zero, but if in addition "ovector" was passed as
398 NULL, a crash could occur.
399
400 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
401 new versions from the libtool 1.5 distribution (the last one is a copy of
402 a file called libtool.m4). This seems to have fixed the need to patch
403 "configure" to support Darwin 1.3 (which I used to do). However, I still
404 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
405 workstation).
406
407 24. Changed the PCRE licence to be the more standard "BSD" licence.
408
409
410 Version 4.5 01-Dec-03
411 ---------------------
412
413 1. There has been some re-arrangement of the code for the match() function so
414 that it can be compiled in a version that does not call itself recursively.
415 Instead, it keeps those local variables that need separate instances for
416 each "recursion" in a frame on the heap, and gets/frees frames whenever it
417 needs to "recurse". Keeping track of where control must go is done by means
418 of setjmp/longjmp. The whole thing is implemented by a set of macros that
419 hide most of the details from the main code, and operates only if
420 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
421 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
422 operating.
423
424 To make it easier for callers to provide specially tailored get/free
425 functions for this usage, two new functions, pcre_stack_malloc, and
426 pcre_stack_free, are used. They are always called in strict stacking order,
427 and the size of block requested is always the same.
428
429 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
430 PCRE has been compiled to use the stack or the heap for recursion. The
431 -C option of pcretest uses this to show which version is compiled.
432
433 A new data escape \S, is added to pcretest; it causes the amounts of store
434 obtained and freed by both kinds of malloc/free at match time to be added
435 to the output.
436
437 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
438 what's available on my current Linux desktop machine.
439
440 3. When matching a UTF-8 string, the test for a valid string at the start has
441 been extended. If start_offset is not zero, PCRE now checks that it points
442 to a byte that is the start of a UTF-8 character. If not, it returns
443 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
444 this is necessary because there may be backward assertions in the pattern.
445 When matching the same subject several times, it may save resources to use
446 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
447
448 4. The code for checking the validity of UTF-8 strings has been tightened so
449 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
450 containing "overlong sequences".
451
452 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
453 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
454 should have been "&", but it just so happened that all the cases this let
455 through by mistake were picked up later in the function.
456
457 6. I had used a variable called "isblank" - this is a C99 function, causing
458 some compilers to warn. To avoid this, I renamed it (as "blankclass").
459
460 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
461 prompting; (b) run "./pcretest /dev/null" at the start of the test script
462 so the version is shown; (c) stop "make test" echoing "./RunTest".
463
464 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
465
466 9. The prototype for memmove() for systems that don't have it was using
467 size_t, but the inclusion of the header that defines size_t was later. I've
468 moved the #includes for the C headers earlier to avoid this.
469
470 10. Added some adjustments to the code to make it easier to compiler on certain
471 special systems:
472
473 (a) Some "const" qualifiers were missing.
474 (b) Added the macro EXPORT before all exported functions; by default this
475 is defined to be empty.
476 (c) Changed the dftables auxiliary program (that builds chartables.c) so
477 that it reads its output file name as an argument instead of writing
478 to the standard output and assuming this can be redirected.
479
480 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
481 class containing characters with values greater than 255, PCRE compilation
482 went into a loop.
483
484 12. A recursive reference to a subpattern that was within another subpattern
485 that had a minimum quantifier of zero caused PCRE to crash. For example,
486 (x(y(?2))z)? provoked this bug with a subject that got as far as the
487 recursion. If the recursively-called subpattern itself had a zero repeat,
488 that was OK.
489
490 13. In pcretest, the buffer for reading a data line was set at 30K, but the
491 buffer into which it was copied (for escape processing) was still set at
492 1024, so long lines caused crashes.
493
494 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
495 "internal error: code overflow...". This applied to any character class
496 that was followed by a possessive quantifier.
497
498 15. Modified the Makefile to add libpcre.la as a prerequisite for
499 libpcreposix.la because I was told this is needed for a parallel build to
500 work.
501
502 16. If a pattern that contained .* following optional items at the start was
503 studied, the wrong optimizing data was generated, leading to matching
504 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
505 matching string must start with a or b or c. The correct conclusion for
506 this pattern is that a match can start with any character.
507
508
509 Version 4.4 13-Aug-03
510 ---------------------
511
512 1. In UTF-8 mode, a character class containing characters with values between
513 127 and 255 was not handled correctly if the compiled pattern was studied.
514 In fixing this, I have also improved the studying algorithm for such
515 classes (slightly).
516
517 2. Three internal functions had redundant arguments passed to them. Removal
518 might give a very teeny performance improvement.
519
520 3. Documentation bug: the value of the capture_top field in a callout is *one
521 more than* the number of the hightest numbered captured substring.
522
523 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
524 in incorrectly linking with a previously installed version. They now link
525 explicitly with libpcre.la.
526
527 5. configure.in no longer needs to recognize Cygwin specially.
528
529 6. A problem in pcre.in for Windows platforms is fixed.
530
531 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
532 pcretest, it used to include the size of the study block as part of its
533 output. Unfortunately, the structure contains a field that has a different
534 size on different hardware architectures. This meant that the tests that
535 showed this size failed. As the block is currently always of a fixed size,
536 this information isn't actually particularly useful in pcretest output, so
537 I have just removed it.
538
539 8. Three pre-processor statements accidentally did not start in column 1.
540 Sadly, there are *still* compilers around that complain, even though
541 standard C has not required this for well over a decade. Sigh.
542
543 9. In pcretest, the code for checking callouts passed small integers in the
544 callout_data field, which is a void * field. However, some picky compilers
545 complained about the casts involved for this on 64-bit systems. Now
546 pcretest passes the address of the small integer instead, which should get
547 rid of the warnings.
548
549 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
550 both compile and run time, and gives an error if an invalid UTF-8 sequence
551 is found. There is a option for disabling this check in cases where the
552 string is known to be correct and/or the maximum performance is wanted.
553
554 11. In response to a bug report, I changed one line in Makefile.in from
555
556 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
557 to
558 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
559
560 to look similar to other lines, but I have no way of telling whether this
561 is the right thing to do, as I do not use Windows. No doubt I'll get told
562 if it's wrong...
563
564
565 Version 4.3 21-May-03
566 ---------------------
567
568 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
569 Makefile.
570
571 2. Some refactoring to improve the quality of the code:
572
573 (i) The utf8_table... variables are now declared "const".
574
575 (ii) The code for \cx, which used the "case flipping" table to upper case
576 lower case letters, now just substracts 32. This is ASCII-specific,
577 but the whole concept of \cx is ASCII-specific, so it seems
578 reasonable.
579
580 (iii) PCRE was using its character types table to recognize decimal and
581 hexadecimal digits in the pattern. This is silly, because it handles
582 only 0-9, a-f, and A-F, but the character types table is locale-
583 specific, which means strange things might happen. A private
584 table is now used for this - though it costs 256 bytes, a table is
585 much faster than multiple explicit tests. Of course, the standard
586 character types table is still used for matching digits in subject
587 strings against \d.
588
589 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
590 ending in _t are). So I've renamed it as ESC_tee.
591
592 3. The first argument for regexec() in the POSIX wrapper should have been
593 defined as "const".
594
595 4. Changed pcretest to use malloc() for its buffers so that they can be
596 Electric Fenced for debugging.
597
598 5. There were several places in the code where, in UTF-8 mode, PCRE would try
599 to read one or more bytes before the start of the subject string. Often this
600 had no effect on PCRE's behaviour, but in some circumstances it could
601 provoke a segmentation fault.
602
603 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
604 to try to read one or more bytes before the start of the subject string.
605
606 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
607 UTF-8 support could misbehave in various ways if the subject string
608 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
609 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
610 back over UTF-8 characters.)
611
612
613 Version 4.2 14-Apr-03
614 ---------------------
615
616 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
617
618 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
619 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
620 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
621 [WIN_PREFIX]: new variable, "cyg" for Cygwin
622 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
623 and BUILD_EXEEXT
624 Note: automatic setting of the BUILD variables is not yet working
625 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
626 compile-time but not at link-time
627 [LINK]: use for linking executables only
628 make different versions for Windows and non-Windows
629 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
630 libraries
631 [LINK_FOR_BUILD]: new variable
632 [OBJEXT]: use throughout
633 [EXEEXT]: use throughout
634 <winshared>: new target
635 <wininstall>: new target
636 <dftables.o>: use native compiler
637 <dftables>: use native linker
638 <install>: handle Windows platform correctly
639 <clean>: ditto
640 <check>: ditto
641 copy DLL to top builddir before testing
642
643 As part of these changes, -no-undefined was removed again. This was reported
644 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
645 in any case.
646
647 3. Some tidies to get rid of compiler warnings:
648
649 . In the match_data structure, match_limit was an unsigned long int, whereas
650 match_call_count was an int. I've made them both unsigned long ints.
651
652 . In pcretest the fact that a const uschar * doesn't automatically cast to
653 a void * provoked a warning.
654
655 . Turning on some more compiler warnings threw up some "shadow" variables
656 and a few more missing casts.
657
658 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
659 option, a class that contained a single character with a value between 128
660 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
661
662 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
663 option, a class that contained several characters, but with at least one
664 whose value was between 128 and 255 caused PCRE to crash.
665
666
667 Version 4.1 12-Mar-03
668 ---------------------
669
670 1. Compiling with gcc -pedantic found a couple of places where casts were
671 needed, and a string in dftables.c that was longer than standard compilers are
672 required to support.
673
674 2. Compiling with Sun's compiler found a few more places where the code could
675 be tidied up in order to avoid warnings.
676
677 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
678 first of these names is deprecated in the latest Autoconf in favour of the name
679 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
680 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
681 analogy I have changed it to CFLAGS_FOR_BUILD.
682
683 4. Added -no-undefined to the linking command in the Makefile, because this is
684 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
685 linking step for the pcreposix library.
686
687 5. PCRE was failing to diagnose the case of two named groups with the same
688 name.
689
690 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
691 literal character that is needed in the subject for a match, and scans along to
692 ensure that it is present before embarking on the full matching process. This
693 saves time in cases of nested unlimited repeats that are never going to match.
694 Problem: the scan can take a lot of time if the subject is very long (e.g.
695 megabytes), thus penalizing straightforward matches. It is now done only if the
696 amount of subject to be scanned is less than 1000 bytes.
697
698 7. A lesser problem with the same optimization is that it was recording the
699 first character of an anchored pattern as "needed", thus provoking a search
700 right along the subject, even when the first match of the pattern was going to
701 fail. The "needed" character is now not set for anchored patterns, unless it
702 follows something in the pattern that is of non-fixed length. Thus, it still
703 fulfils its original purpose of finding quick non-matches in cases of nested
704 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
705
706
707 Version 4.0 17-Feb-03
708 ---------------------
709
710 1. If a comment in an extended regex that started immediately after a meta-item
711 extended to the end of string, PCRE compiled incorrect data. This could lead to
712 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
713
714 2. Moved to autoconf 2.53 and libtool 1.4.2.
715
716 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
717 the special perltest8 script is no longer needed - all the tests can be run
718 from a single perltest script.
719
720 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
721 by \s. It has now been removed in PCRE. This means it isn't recognized as
722 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
723 class [:space:] *does* include VT, thereby creating a mess.
724
725 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
726 space and tab.
727
728 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
729 its new features into the main test script, reducing the number of scripts.
730
731 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
732 were backward compatible, and made the (?i) apply to the whole pattern, as if
733 /i were given. Now it behaves more logically, and applies the option setting
734 only to what follows. PCRE has been changed to follow suit. However, if it
735 finds options settings right at the start of the pattern, it extracts them into
736 the global options, as before. Thus, they show up in the info data.
737
738 8. Added support for the \Q...\E escape sequence. Characters in between are
739 treated as literals. This is slightly different from Perl in that $ and @ are
740 also handled as literals inside the quotes. In Perl, they will cause variable
741 interpolation. Note the following examples:
742
743 Pattern PCRE matches Perl matches
744
745 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
746 \Qabc\$xyz\E abc\$xyz abc\$xyz
747 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
748
749 For compatibility with Perl, \Q...\E sequences are recognized inside character
750 classes as well as outside them.
751
752 9. Re-organized 3 code statements in pcretest to avoid "overflow in
753 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
754 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
755 signed/unsigned warnings.
756
757 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
758 option for pcretest, so I've replaced it by a simple function that does just
759 that job.
760
761 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
762 "pcregrep -".
763
764 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
765 Java package. This provides some syntactic sugar for simple cases of what my
766 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
767 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
768 item, you can use this simplified notation. Note that only makes sense with
769 greedy quantifiers. Consequently, the use of the possessive quantifier forces
770 greediness, whatever the setting of the PCRE_UNGREEDY option.
771
772 13. A change of greediness default within a pattern was not taking effect at
773 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
774 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
775 was abstracted outside.
776
777 14. PCRE now supports the \G assertion. It is true when the current matching
778 position is at the start point of the match. This differs from \A when the
779 starting offset is non-zero. Used with the /g option of pcretest (or similar
780 code), it works in the same way as it does for Perl's /g option. If all
781 alternatives of a regex begin with \G, the expression is anchored to the start
782 match position, and the "anchored" flag is set in the compiled expression.
783
784 15. Some bugs concerning the handling of certain option changes within patterns
785 have been fixed. These applied to options other than (?ims). For example,
786 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
787 been the other way round. Some of this was related to change 7 above.
788
789 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
790 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
791 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
792 POSIX classes only within a class (e.g. /[[:alpha:]]/).
793
794 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
795 mode. Unlike ".", it always matches newline, whatever the setting of
796 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
797 assertions. Perl allows it, but it doesn't (in general) work because it can't
798 calculate the length of the lookbehind. At least, that's the case for Perl
799 5.8.0 - I've been told they are going to document that it doesn't work in
800 future.
801
802 18. Added an error diagnosis for escapes that PCRE does not support: these are
803 \L, \l, \N, \P, \p, \U, \u, and \X.
804
805 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
806 reading past the end of the pattern in cases such as /[abcd/.
807
808 20. PCRE was getting more memory than necessary for patterns with classes that
809 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
810
811 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
812 compiling PCRE for use with Virtual Pascal.
813
814 22. Small fix to the Makefile to make it work properly if the build is done
815 outside the source tree.
816
817 23. Added a new extension: a condition to go with recursion. If a conditional
818 subpattern starts with (?(R) the "true" branch is used if recursion has
819 happened, whereas the "false" branch is used only at the top level.
820
821 24. When there was a very long string of literal characters (over 255 bytes
822 without UTF support, over 250 bytes with UTF support), the computation of how
823 much memory was required could be incorrect, leading to segfaults or other
824 strange effects.
825
826 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
827 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
828 there was a subsequent back reference to those brackets. This meant that, for
829 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
830 possible to check for precisely this case. All we can do is abandon the
831 optimization if .* occurs inside capturing brackets when there are any back
832 references whatsoever. (See below for a better fix that came later.)
833
834 26. The handling of the optimization for finding the first character of a
835 non-anchored pattern, and for finding a character that is required later in the
836 match were failing in some cases. This didn't break the matching; it just
837 failed to optimize when it could. The way this is done has been re-implemented.
838
839 27. Fixed typo in error message for invalid (?R item (it said "(?p").
840
841 28. Added a new feature that provides some of the functionality that Perl
842 provides with (?{...}). The facility is termed a "callout". The way it is done
843 in PCRE is for the caller to provide an optional function, by setting
844 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
845 global variable. By default it is unset, which disables all calling out. To get
846 the function called, the regex must include (?C) at appropriate points. This
847 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
848 This provides a means of identifying different callout points. When PCRE
849 reaches such a point in the regex, if pcre_callout has been set, the external
850 function is called. It is provided with data in a structure called
851 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
852 matching continues; if it returns a non-zero value, the match at the current
853 point fails. However, backtracking will occur if possible. [This was changed
854 later and other features added - see item 49 below.]
855
856 29. pcretest is upgraded to test the callout functionality. It provides a
857 callout function that displays information. By default, it shows the start of
858 the match and the current position in the text. There are some new data escapes
859 to vary what happens:
860
861 \C+ in addition, show current contents of captured substrings
862 \C- do not supply a callout function
863 \C!n return 1 when callout number n is reached
864 \C!n!m return 1 when callout number n is reached for the mth time
865
866 30. If pcregrep was called with the -l option and just a single file name, it
867 output "<stdin>" if a match was found, instead of the file name.
868
869 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
870 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
871 pcre_exec(). This saves a malloc/free per call. The default value of
872 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
873 when configuring.
874
875 32. The default maximum size of a compiled pattern is 64K. There have been a
876 few cases of people hitting this limit. The code now uses macros to handle the
877 storing of links as offsets within the compiled pattern. It defaults to 2-byte
878 links, but this can be changed to 3 or 4 bytes by --with-link-size when
879 configuring. Tests 2 and 5 work only with 2-byte links because they output
880 debugging information about compiled patterns.
881
882 33. Internal code re-arrangements:
883
884 (a) Moved the debugging function for printing out a compiled regex into
885 its own source file (printint.c) and used #include to pull it into
886 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
887 separate copies.
888
889 (b) Defined the list of op-code names for debugging as a macro in
890 internal.h so that it is next to the definition of the opcodes.
891
892 (c) Defined a table of op-code lengths for simpler skipping along compiled
893 code. This is again a macro in internal.h so that it is next to the
894 definition of the opcodes.
895
896 34. Added support for recursive calls to individual subpatterns, along the
897 lines of Robin Houston's patch (but implemented somewhat differently).
898
899 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
900 allow it to read and process whole directories in Win32. This code was
901 contributed by Lionel Fourquaux; it has not been tested by me.
902
903 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
904 used to name a group. Names consist of alphanumerics and underscores, and must
905 be unique. Back references use the syntax (?P=name) and recursive calls use
906 (?P>name) which is a PCRE extension to the Python extension. Groups still have
907 numbers. The function pcre_fullinfo() can be used after compilation to extract
908 a name/number map. There are three relevant calls:
909
910 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
911 PCRE_INFO_NAMECOUNT yields the number of entries
912 PCRE_INFO_NAMETABLE yields a pointer to the map.
913
914 The map is a vector of fixed-size entries. The size of each entry depends on
915 the length of the longest name used. The first two bytes of each entry are the
916 group number, most significant byte first. There follows the corresponding
917 name, zero terminated. The names are in alphabetical order.
918
919 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
920 case instead of 255. Making it the same both with and without UTF-8 support
921 means that the same test output works with both.
922
923 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
924 calling malloc() with a zero argument.
925
926 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
927 optimization. I've improved things by keeping a bitmap of backreferences with
928 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
929 fact referenced, the optimization can be applied. It is unlikely that a
930 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
931 the match to follow \n) will appear inside brackets with a number greater than
932 31, but if it does, any back reference > 31 suppresses the optimization.
933
934 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
935 of disabling numbered capturing parentheses. Any opening parenthesis that is
936 not followed by ? behaves as if it were followed by ?: but named parentheses
937 can still be used for capturing (and they will acquire numbers in the usual
938 way).
939
940 41. Redesigned the return codes from the match() function into yes/no/error so
941 that errors can be passed back from deep inside the nested calls. A malloc
942 failure while inside a recursive subpattern call now causes the
943 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
944
945 42. It is now possible to set a limit on the number of times the match()
946 function is called in a call to pcre_exec(). This facility makes it possible to
947 limit the amount of recursion and backtracking, though not in a directly
948 obvious way, because the match() function is used in a number of different
949 circumstances. The count starts from zero for each position in the subject
950 string (for non-anchored patterns). The default limit is, for compatibility, a
951 large number, namely 10 000 000. You can change this in two ways:
952
953 (a) When configuring PCRE before making, you can use --with-match-limit=n
954 to set a default value for the compiled library.
955
956 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
957 a different value is set. See 45 below.
958
959 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
960
961 43. Added a new function pcre_config(int, void *) to enable run-time extraction
962 of things that can be changed at compile time. The first argument specifies
963 what is wanted and the second points to where the information is to be placed.
964 The current list of available information is:
965
966 PCRE_CONFIG_UTF8
967
968 The output is an integer that is set to one if UTF-8 support is available;
969 otherwise it is set to zero.
970
971 PCRE_CONFIG_NEWLINE
972
973 The output is an integer that it set to the value of the code that is used for
974 newline. It is either LF (10) or CR (13).
975
976 PCRE_CONFIG_LINK_SIZE
977
978 The output is an integer that contains the number of bytes used for internal
979 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
980
981 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
982
983 The output is an integer that contains the threshold above which the POSIX
984 interface uses malloc() for output vectors. See item 31 above.
985
986 PCRE_CONFIG_MATCH_LIMIT
987
988 The output is an unsigned integer that contains the default limit of the number
989 of match() calls in a pcre_exec() execution. See 42 above.
990
991 44. pcretest has been upgraded by the addition of the -C option. This causes it
992 to extract all the available output from the new pcre_config() function, and to
993 output it. The program then exits immediately.
994
995 45. A need has arisen to pass over additional data with calls to pcre_exec() in
996 order to support additional features. One way would have been to define
997 pcre_exec2() (for example) with extra arguments, but this would not have been
998 extensible, and would also have required all calls to the original function to
999 be mapped to the new one. Instead, I have chosen to extend the mechanism that
1000 is used for passing in "extra" data from pcre_study().
1001
1002 The pcre_extra structure is now exposed and defined in pcre.h. It currently
1003 contains the following fields:
1004
1005 flags a bitmap indicating which of the following fields are set
1006 study_data opaque data from pcre_study()
1007 match_limit a way of specifying a limit on match() calls for a specific
1008 call to pcre_exec()
1009 callout_data data for callouts (see 49 below)
1010
1011 The flag bits are also defined in pcre.h, and are
1012
1013 PCRE_EXTRA_STUDY_DATA
1014 PCRE_EXTRA_MATCH_LIMIT
1015 PCRE_EXTRA_CALLOUT_DATA
1016
1017 The pcre_study() function now returns one of these new pcre_extra blocks, with
1018 the actual study data pointed to by the study_data field, and the
1019 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
1020 before. That is, this change is entirely upwards-compatible and requires no
1021 change to existing code.
1022
1023 If you want to pass in additional data to pcre_exec(), you can either place it
1024 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
1025 block.
1026
1027 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
1028 data string contains the escape sequence \M, pcretest calls pcre_exec() several
1029 times with different match limits, until it finds the minimum value needed for
1030 pcre_exec() to complete. The value is then output. This can be instructive; for
1031 most simple matches the number is quite small, but for pathological cases it
1032 gets very large very quickly.
1033
1034 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
1035 returns the size of the data block pointed to by the study_data field in a
1036 pcre_extra block, that is, the value that was passed as the argument to
1037 pcre_malloc() when PCRE was getting memory in which to place the information
1038 created by pcre_study(). The fourth argument should point to a size_t variable.
1039 pcretest has been extended so that this information is shown after a successful
1040 pcre_study() call when information about the compiled regex is being displayed.
1041
1042 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
1043 because what follows is always an absolute path. (Later: it turns out that this
1044 is more than cosmetic for MinGW, because it doesn't like empty path
1045 components.)
1046
1047 49. Some changes have been made to the callout feature (see 28 above):
1048
1049 (i) A callout function now has three choices for what it returns:
1050
1051 0 => success, carry on matching
1052 > 0 => failure at this point, but backtrack if possible
1053 < 0 => serious error, return this value from pcre_exec()
1054
1055 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
1056 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
1057 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
1058 use by callout functions. It will never be used by PCRE itself.
1059
1060 (ii) The pcre_extra structure (see 45 above) has a void * field called
1061 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
1062 pcre_callout_block structure has a field of the same name. The contents of
1063 the field passed in the pcre_extra structure are passed to the callout
1064 function in the corresponding field in the callout block. This makes it
1065 easier to use the same callout-containing regex from multiple threads. For
1066 testing, the pcretest program has a new data escape
1067
1068 \C*n pass the number n (may be negative) as callout_data
1069
1070 If the callout function in pcretest receives a non-zero value as
1071 callout_data, it returns that value.
1072
1073 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
1074 there were some redundant $(CFLAGS) in commands that are now specified as
1075 $(LINK), which already includes $(CFLAGS).
1076
1077 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
1078 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
1079 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
1080 one-byte characters throughout. Note that case-insensitive matching applies
1081 only to characters whose values are less than 256. PCRE doesn't support the
1082 notion of cases for higher-valued characters.
1083
1084 (i) A character class whose characters are all within 0-255 is handled as
1085 a bit map, and the map is inverted for negative classes. Previously, a
1086 character > 255 always failed to match such a class; however it should
1087 match if the class was a negative one (e.g. [^ab]). This has been fixed.
1088
1089 (ii) A negated character class with a single character < 255 is coded as
1090 "not this character" (OP_NOT). This wasn't working properly when the test
1091 character was multibyte, either singly or repeated.
1092
1093 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
1094 mode, for example: \x{100}{2,3}.
1095
1096 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
1097 singly or repeated) now correctly test multibyte characters. However,
1098 PCRE doesn't recognize any characters with values greater than 255 as
1099 digits, spaces, or word characters. Such characters always match \D, \S,
1100 and \W, and never match \d, \s, or \w.
1101
1102 (v) Classes may now contain characters and character ranges with values
1103 greater than 255. For example: [ab\x{100}-\x{400}].
1104
1105 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
1106 PCRE in UTF-8 mode.
1107
1108 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
1109 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
1110 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
1111 value.)
1112
1113 53. The single man page has become too large. I have therefore split it up into
1114 a number of separate man pages. These also give rise to individual HTML pages;
1115 these are now put in a separate directory, and there is an index.html page that
1116 lists them all. Some hyperlinking between the pages has been installed.
1117
1118 54. Added convenience functions for handling named capturing parentheses.
1119
1120 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
1121 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
1122 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
1123 are faulted.
1124
1125 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
1126 calling configure. These values are used when compiling the dftables.c program
1127 which is run to generate the source of the default character tables. They
1128 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
1129 you will need to set these values.
1130
1131 57. Updated the building process for Windows DLL, as provided by Fred Cox.
1132
1133
1134 Version 3.9 02-Jan-02
1135 ---------------------
1136
1137 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
1138
1139 2. If --disable-static was given, the building process failed when trying to
1140 build pcretest and pcregrep. (For some reason it was using libtool to compile
1141 them, which is not right, as they aren't part of the library.)
1142
1143
1144 Version 3.8 18-Dec-01
1145 ---------------------
1146
1147 1. The experimental UTF-8 code was completely screwed up. It was packing the
1148 bytes in the wrong order. How dumb can you get?
1149
1150
1151 Version 3.7 29-Oct-01
1152 ---------------------
1153
1154 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
1155 This caused pcretest, when used on the test data, to segfault. Unfortunately,
1156 this didn't happen under Solaris 8, where I normally test things.
1157
1158 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
1159 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
1160 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
1161 this fix an hour or so after the initial 3.7 release.)
1162
1163
1164 Version 3.6 23-Oct-01
1165 ---------------------
1166
1167 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
1168 offsets passed as NULL with zero offset count.
1169
1170 2. The config.guess and config.sub files had not been updated when I moved to
1171 the latest autoconf.
1172
1173
1174 Version 3.5 15-Aug-01
1175 ---------------------
1176
1177 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
1178 had been forgotten.
1179
1180 2. By using declared but undefined structures, we can avoid using "void"
1181 definitions in pcre.h while keeping the internal definitions of the structures
1182 private.
1183
1184 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
1185 user point of view, this means that both static and shared libraries are built
1186 by default, but this can be individually controlled. More of the work of
1187 handling this static/shared cases is now inside libtool instead of PCRE's make
1188 file.
1189
1190 4. The pcretest utility is now installed along with pcregrep because it is
1191 useful for users (to test regexs) and by doing this, it automatically gets
1192 relinked by libtool. The documentation has been turned into a man page, so
1193 there are now .1, .txt, and .html versions in /doc.
1194
1195 5. Upgrades to pcregrep:
1196 (i) Added long-form option names like gnu grep.
1197 (ii) Added --help to list all options with an explanatory phrase.
1198 (iii) Added -r, --recursive to recurse into sub-directories.
1199 (iv) Added -f, --file to read patterns from a file.
1200
1201 6. pcre_exec() was referring to its "code" argument before testing that
1202 argument for NULL (and giving an error if it was NULL).
1203
1204 7. Upgraded Makefile.in to allow for compiling in a different directory from
1205 the source directory.
1206
1207 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
1208 options bits, the pointer it was passed was to an int instead of to an unsigned
1209 long int. This mattered only on 64-bit systems.
1210
1211 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
1212 generated) instead of pcre.in, which it its source. Also made the same change
1213 in several of the .c files.
1214
1215 10. A new release of gcc defines printf() as a macro, which broke pcretest
1216 because it had an ifdef in the middle of a string argument for printf(). Fixed
1217 by using separate calls to printf().
1218
1219 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
1220 script, to force use of CR or LF instead of \n in the source. On non-Unix
1221 systems, the value can be set in config.h.
1222
1223 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
1224 absolute limit. Changed the text of the error message to make this clear, and
1225 likewise updated the man page.
1226
1227 13. The limit of 99 on the number of capturing subpatterns has been removed.
1228 The new limit is 65535, which I hope will not be a "real" limit.
1229
1230
1231 Version 3.4 22-Aug-00
1232 ---------------------
1233
1234 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
1235
1236 2. Diagnose condition (?(0) as an error instead of crashing on matching.
1237
1238
1239 Version 3.3 01-Aug-00
1240 ---------------------
1241
1242 1. If an octal character was given, but the value was greater than \377, it
1243 was not getting masked to the least significant bits, as documented. This could
1244 lead to crashes in some systems.
1245
1246 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
1247 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
1248
1249 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
1250 These just pass their arguments on to (pcre_free)(), but they are provided
1251 because some uses of PCRE bind it to non-C systems that can call its functions,
1252 but cannot call free() or pcre_free() directly.
1253
1254 4. Add "make test" as a synonym for "make check". Corrected some comments in
1255 the Makefile.
1256
1257 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
1258 Makefile.
1259
1260 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
1261 command called pgrep for grepping around the active processes.
1262
1263 7. Added the beginnings of support for UTF-8 character strings.
1264
1265 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
1266 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
1267 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
1268 out for the ar command.)
1269
1270
1271 Version 3.2 12-May-00
1272 ---------------------
1273
1274 This is purely a bug fixing release.
1275
1276 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
1277 of ZA. This was just one example of several cases that could provoke this bug,
1278 which was introduced by change 9 of version 2.00. The code for breaking
1279 infinite loops after an iteration that matches an empty string was't working
1280 correctly.
1281
1282 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
1283 when matched against abbab (for example). After matching an empty string, it
1284 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
1285 caused it to match further down the string than it should.
1286
1287 3. The code contained an inclusion of sys/types.h. It isn't clear why this
1288 was there because it doesn't seem to be needed, and it causes trouble on some
1289 systems, as it is not a Standard C header. It has been removed.
1290
1291 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
1292 were reported on the Macintosh. The changes were from
1293
1294 while ((c = *(++ptr)) != 0 && c != '\n');
1295 to
1296 while ((c = *(++ptr)) != 0 && c != '\n') ;
1297
1298 Totally extraordinary, but if that's what it takes...
1299
1300 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
1301 available. Added HAVE_BCOPY and an autoconf test for it; if neither
1302 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
1303 assumes the way PCRE uses memmove() (always moving upwards).
1304
1305 6. PCRE is being used in one environment where strchr() is not available. There
1306 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
1307 faster code anyway.
1308
1309
1310 Version 3.1 09-Feb-00
1311 ---------------------
1312
1313 The only change in this release is the fixing of some bugs in Makefile.in for
1314 the "install" target:
1315
1316 (1) It was failing to install pcreposix.h.
1317
1318 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
1319
1320
1321 Version 3.0 01-Feb-00
1322 ---------------------
1323
1324 1. Add support for the /+ modifier to perltest (to output $` like it does in
1325 pcretest).
1326
1327 2. Add support for the /g modifier to perltest.
1328
1329 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
1330 matches null strings.
1331
1332 4. Fix perltest so that it doesn't do unwanted things when fed an empty
1333 pattern. Perl treats empty patterns specially - it reuses the most recent
1334 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
1335 effect.
1336
1337 5. The POSIX interface was broken in that it was just handing over the POSIX
1338 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
1339 required a bigger vector, with some working space on the end. This means that
1340 the POSIX wrapper now has to get and free some memory, and copy the results.
1341
1342 6. Added some simple autoconf support, placing the test data and the
1343 documentation in separate directories, re-organizing some of the
1344 information files, and making it build pcre-config (a GNU standard). Also added
1345 libtool support for building PCRE as a shared library, which is now the
1346 default.
1347
1348 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
1349 09 are not valid octal constants. Single digits will be used for minor values
1350 less than 10.
1351
1352 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
1353 existing programs that set these in the POSIX interface can use PCRE without
1354 modification.
1355
1356 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
1357 return all that pcre_info() returns, plus additional data. The pcre_info()
1358 function is retained for compatibility, but is considered to be obsolete.
1359
1360 10. Added experimental recursion feature (?R) to handle one common case that
1361 Perl 5.6 will be able to do with (?p{...}).
1362
1363 11. Added support for POSIX character classes like [:alpha:], which Perl is
1364 adopting.
1365
1366
1367 Version 2.08 31-Aug-99
1368 ----------------------
1369
1370 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
1371 trying to match at the startoffset position, but instead was moving forward to
1372 the next newline as if a previous match had failed.
1373
1374 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
1375 and could get into a loop if a null string was matched other than at the start
1376 of the subject.
1377
1378 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
1379 be distinguished at compile time, and for completeness also added PCRE_DATE.
1380
1381 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
1382 in GnuWin32 environments.
1383
1384
1385 Version 2.07 29-Jul-99
1386 ----------------------
1387
1388 1. The documentation is now supplied in plain text form and HTML as well as in
1389 the form of man page sources.
1390
1391 2. C++ compilers don't like assigning (void *) values to other pointer types.
1392 In particular this affects malloc(). Although there is no problem in Standard
1393 C, I've put in casts to keep C++ compilers happy.
1394
1395 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
1396 should be (const char *).
1397
1398 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
1399 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
1400 However, I haven't made this a standard facility. The documentation doesn't
1401 mention it, and the Makefile doesn't support it.
1402
1403 5. The Makefile now contains an "install" target, with editable destinations at
1404 the top of the file. The pcretest program is not installed.
1405
1406 6. pgrep -V now gives the PCRE version number and date.
1407
1408 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
1409 causing the entire string to be ignored, instead of just the last character.
1410
1411 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
1412 non-matching string, it can take a very, very long time, even for strings of
1413 quite modest length, because of the nested recursion. PCRE now does better in
1414 some of these cases. It does this by remembering the last required literal
1415 character in the pattern, and pre-searching the subject to ensure it is present
1416 before running the real match. In other words, it applies a heuristic to detect
1417 some types of certain failure quickly, and in the above example, if presented
1418 with a string that has no trailing " it gives "no match" very quickly.
1419
1420 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
1421 other alternatives are tried instead.
1422
1423
1424 Version 2.06 09-Jun-99
1425 ----------------------
1426
1427 1. Change pcretest's output for amount of store used to show just the code
1428 space, because the remainder (the data block) varies in size between 32-bit and
1429 64-bit systems.
1430
1431 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
1432 start matching at. This allows lookbehinds to work when searching for multiple
1433 occurrences in a string.
1434
1435 3. Added additional options to pcretest for testing multiple occurrences:
1436
1437 /+ outputs the rest of the string that follows a match
1438 /g loops for multiple occurrences, using the new startoffset argument
1439 /G loops for multiple occurrences by passing an incremented pointer
1440
1441 4. PCRE wasn't doing the "first character" optimization for patterns starting
1442 with \b or \B, though it was doing it for other lookbehind assertions. That is,
1443 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
1444 the letter 'x'. On long subject strings, this gives a significant speed-up.
1445
1446
1447 Version 2.05 21-Apr-99
1448 ----------------------
1449
1450 1. Changed the type of magic_number from int to long int so that it works
1451 properly on 16-bit systems.
1452
1453 2. Fixed a bug which caused patterns starting with .* not to work correctly
1454 when the subject string contained newline characters. PCRE was assuming
1455 anchoring for such patterns in all cases, which is not correct because .* will
1456 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
1457 DOTALL is set at top level; otherwise it knows that patterns starting with .*
1458 must be retried after every newline in the subject.
1459
1460
1461 Version 2.04 18-Feb-99
1462 ----------------------
1463
1464 1. For parenthesized subpatterns with repeats whose minimum was zero, the
1465 computation of the store needed to hold the pattern was incorrect (too large).
1466 If such patterns were nested a few deep, this could multiply and become a real
1467 problem.
1468
1469 2. Added /M option to pcretest to show the memory requirement of a specific
1470 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
1471
1472 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
1473 compiled in such a way that the backtracking after subsequent failure was
1474 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
1475 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
1476
1477
1478 Version 2.03 02-Feb-99
1479 ----------------------
1480
1481 1. Fixed typo and small mistake in man page.
1482
1483 2. Added 4th condition (GPL supersedes if conflict) and created separate
1484 LICENCE file containing the conditions.
1485
1486 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
1487 Perl, that is the internal \ allows the delimiter to be included in the
1488 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
1489 the final delimiter, add \ to the end of the pattern (to test the error).
1490
1491 4. Added the convenience functions for extracting substrings after a successful
1492 match. Updated pcretest to make it able to test these functions.
1493
1494
1495 Version 2.02 14-Jan-99
1496 ----------------------
1497
1498 1. Initialized the working variables associated with each extraction so that
1499 their saving and restoring doesn't refer to uninitialized store.
1500
1501 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
1502 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
1503 fix the problem.
1504
1505 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
1506 calls, and wasn't printing the correct value for compiling calls. Increased the
1507 default value of LOOPREPEAT, and the number of significant figures in the
1508 times.
1509
1510 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
1511
1512 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
1513 a building problem on Windows NT with a FAT file system.
1514
1515
1516 Version 2.01 21-Oct-98
1517 ----------------------
1518
1519 1. Changed the API for pcre_compile() to allow for the provision of a pointer
1520 to character tables built by pcre_maketables() in the current locale. If NULL
1521 is passed, the default tables are used.
1522
1523
1524 Version 2.00 24-Sep-98
1525 ----------------------
1526
1527 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
1528 it any more.
1529
1530 2. Allow quantification of (?>) groups, and make it work correctly.
1531
1532 3. The first character computation wasn't working for (?>) groups.
1533
1534 4. Correct the implementation of \Z (it is permitted to match on the \n at the
1535 end of the subject) and add 5.005's \z, which really does match only at the
1536 very end of the subject.
1537
1538 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
1539
1540 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
1541 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
1542 localized options. All options to pcre_study() were also removed.
1543
1544 7. Add other new features from 5.005:
1545
1546 $(?<= positive lookbehind
1547 $(?<! negative lookbehind
1548 (?imsx-imsx) added the unsetting capability
1549 such a setting is global if at outer level; local otherwise
1550 (?imsx-imsx:) non-capturing groups with option setting
1551 (?(cond)re|re) conditional pattern matching
1552
1553 A backreference to itself in a repeated group matches the previous
1554 captured string.
1555
1556 8. General tidying up of studying (both automatic and via "study")
1557 consequential on the addition of new assertions.
1558
1559 9. As in 5.005, unlimited repeated groups that could match an empty substring
1560 are no longer faulted at compile time. Instead, the loop is forcibly broken at
1561 runtime if any iteration does actually match an empty substring.
1562
1563 10. Include the RunTest script in the distribution.
1564
1565 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
1566 discrepancies, some of which were old and were also with respect to 5.004. They
1567 have now been fixed.
1568
1569
1570 Version 1.09 28-Apr-98
1571 ----------------------
1572
1573 1. A negated single character class followed by a quantifier with a minimum
1574 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
1575 program crashes, or just wrong answers. This did not apply to negated classes
1576 containing more than one character, or to minima other than one.
1577
1578
1579 Version 1.08 27-Mar-98
1580 ----------------------
1581
1582 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
1583
1584 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
1585 latter must appear before anything that relies on it in the pattern.
1586
1587
1588 Version 1.07 16-Feb-98
1589 ----------------------
1590
1591 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
1592 repeat of a potentially empty string).
1593
1594
1595 Version 1.06 23-Jan-98
1596 ----------------------
1597
1598 1. Added Markus Oberhumer's little patches for C++.
1599
1600 2. Literal strings longer than 255 characters were broken.
1601
1602
1603 Version 1.05 23-Dec-97
1604 ----------------------
1605
1606 1. Negated character classes containing more than one character were failing if
1607 PCRE_CASELESS was set at run time.
1608
1609
1610 Version 1.04 19-Dec-97
1611 ----------------------
1612
1613 1. Corrected the man page, where some "const" qualifiers had been omitted.
1614
1615 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
1616 input syntax.
1617
1618 3. Fixed memory leak which occurred when a regex with back references was
1619 matched with an offsets vector that wasn't big enough. The temporary memory
1620 that is used in this case wasn't being freed if the match failed.
1621
1622 4. Tidied pcretest to ensure it frees memory that it gets.
1623
1624 5. Temporary memory was being obtained in the case where the passed offsets
1625 vector was exactly big enough.
1626
1627 6. Corrected definition of offsetof() from change 5 below.
1628
1629 7. I had screwed up change 6 below and broken the rules for the use of
1630 setjmp(). Now fixed.
1631
1632
1633 Version 1.03 18-Dec-97
1634 ----------------------
1635
1636 1. A erroneous regex with a missing opening parenthesis was correctly
1637 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
1638 on some systems.
1639
1640 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
1641 it was reported that one broken compiler failed on the former because "code" is
1642 also an independent variable.
1643
1644 3. The erroneous regex a[]b caused an array overrun reference.
1645
1646 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
1647 fail on data ending with that character. (It was going on too far, and checking
1648 the next character, typically a binary zero.) This was specific to the
1649 optimized code for single-character negative classes.
1650
1651 5. Added a contributed patch from the TIN world which does the following:
1652
1653 + Add an undef for memmove, in case the the system defines a macro for it.
1654
1655 + Add a definition of offsetof(), in case there isn't one. (I don't know
1656 the reason behind this - offsetof() is part of the ANSI standard - but
1657 it does no harm).
1658
1659 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
1660 most of the places where whitespace preceded '#'. I have given up and
1661 allowed the remaining 2 cases to be at the margin.
1662
1663 + Rename some variables in pcre to eliminate shadowing. This seems very
1664 pedantic, but does no harm, of course.
1665
1666 6. Moved the call to setjmp() into its own function, to get rid of warnings
1667 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
1668
1669 7. Constructs such as \d{8,} were compiling into the equivalent of
1670 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
1671 outcome, but in this particular case used more store than had been allocated,
1672 which caused the bug to be discovered because it threw up an internal error.
1673
1674 8. The debugging code in both pcre and pcretest for outputting the compiled
1675 form of a regex was going wrong in the case of back references followed by
1676 curly-bracketed repeats.
1677
1678
1679 Version 1.02 12-Dec-97
1680 ----------------------
1681
1682 1. Typos in pcre.3 and comments in the source fixed.
1683
1684 2. Applied a contributed patch to get rid of places where it used to remove
1685 'const' from variables, and fixed some signed/unsigned and uninitialized
1686 variable warnings.
1687
1688 3. Added the "runtest" target to Makefile.
1689
1690 4. Set default compiler flag to -O2 rather than just -O.
1691
1692
1693 Version 1.01 19-Nov-97
1694 ----------------------
1695
1696 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
1697 like /([ab]*)*/, that is, for classes with more than one character in them.
1698
1699 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
1700 as /((?>a*))*/ (a PCRE_EXTRA facility).
1701
1702
1703 Version 1.00 18-Nov-97
1704 ----------------------
1705
1706 1. Added compile-time macros to support systems such as SunOS4 which don't have
1707 memmove() or strerror() but have other things that can be used instead.
1708
1709 2. Arranged that "make clean" removes the executables.
1710
1711
1712 Version 0.99 27-Oct-97
1713 ----------------------
1714
1715 1. Fixed bug in code for optimizing classes with only one character. It was
1716 initializing a 32-byte map regardless, which could cause it to run off the end
1717 of the memory it had got.
1718
1719 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
1720
1721
1722 Version 0.98 22-Oct-97
1723 ----------------------
1724
1725 1. Fixed bug in code for handling temporary memory usage when there are more
1726 back references than supplied space in the ovector. This could cause segfaults.
1727
1728
1729 Version 0.97 21-Oct-97
1730 ----------------------
1731
1732 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
1733
1734 2. Optimized negated single characters not to use a bit map.
1735
1736 3. Brought error texts together as macro definitions; clarified some of them;
1737 fixed one that was wrong - it said "range out of order" when it meant "invalid
1738 escape sequence".
1739
1740 4. Changed some char * arguments to const char *.
1741
1742 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
1743
1744 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
1745 pcretest.
1746
1747
1748 Version 0.96 16-Oct-97
1749 ----------------------
1750
1751 1. Added a simple "pgrep" utility to the distribution.
1752
1753 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
1754 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
1755 where "ddd" means "one or more decimal digits".
1756
1757 3. Fixed serious bug. If a pattern had a back reference, but the call to
1758 pcre_exec() didn't supply a large enough ovector to record the related
1759 identifying subpattern, the match always failed. PCRE now remembers the number
1760 of the largest back reference, and gets some temporary memory in which to save
1761 the offsets during matching if necessary, in order to ensure that
1762 backreferences always work.
1763
1764 4. Increased the compatibility with Perl in a number of ways:
1765
1766 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
1767 to request this handling. The option can be set at compile or exec time.
1768
1769 (b) $ matches before a terminating newline by default; an option
1770 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
1771 mode). The option can be set at compile or exec time.
1772
1773 (c) The handling of \ followed by a digit other than 0 is now supposed to be
1774 the same as Perl's. If the decimal number it represents is less than 10
1775 or there aren't that many previous left capturing parentheses, an octal
1776 escape is read. Inside a character class, it's always an octal escape,
1777 even if it is a single digit.
1778
1779 (d) An escaped but undefined alphabetic character is taken as a literal,
1780 unless PCRE_EXTRA is set. Currently this just reserves the remaining
1781 escapes.
1782
1783 (e) {0} is now permitted. (The previous item is removed from the compiled
1784 pattern).
1785
1786 5. Changed all the names of code files so that the basic parts are no longer
1787 than 10 characters, and abolished the teeny "globals.c" file.
1788
1789 6. Changed the handling of character classes; they are now done with a 32-byte
1790 bit map always.
1791
1792 7. Added the -d and /D options to pcretest to make it possible to look at the
1793 internals of compilation without having to recompile pcre.
1794
1795
1796 Version 0.95 23-Sep-97
1797 ----------------------
1798
1799 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
1800 \x20 at the start of a run of normal characters. These were being treated as
1801 real characters, instead of the source characters being re-checked.
1802
1803
1804 Version 0.94 18-Sep-97
1805 ----------------------
1806
1807 1. The functions are now thread-safe, with the caveat that the global variables
1808 containing pointers to malloc() and free() or alternative functions are the
1809 same for all threads.
1810
1811 2. Get pcre_study() to generate a bitmap of initial characters for non-
1812 anchored patterns when this is possible, and use it if passed to pcre_exec().
1813
1814
1815 Version 0.93 15-Sep-97
1816 ----------------------
1817
1818 1. /(b)|(:+)/ was computing an incorrect first character.
1819
1820 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
1821 but not actually doing anything yet.
1822
1823 3. Treat "-" characters in classes that cannot be part of ranges as literals,
1824 as Perl does (e.g. [-az] or [az-]).
1825
1826 4. Set the anchored flag if a branch starts with .* or .*? because that tests
1827 all possible positions.
1828
1829 5. Split up into different modules to avoid including unneeded functions in a
1830 compiled binary. However, compile and exec are still in one module. The "study"
1831 function is split off.
1832
1833 6. The character tables are now in a separate module whose source is generated
1834 by an auxiliary program - but can then be edited by hand if required. There are
1835 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
1836 toupper() in the code.
1837
1838 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
1839 make them global. Abolish the function for setting them, as the caller can now
1840 set them directly.
1841
1842
1843 Version 0.92 11-Sep-97
1844 ----------------------
1845
1846 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
1847 (e.g. /a{1,3}/) was broken (I mis-optimized it).
1848
1849 2. Caseless matching was not working in character classes if the characters in
1850 the pattern were in upper case.
1851
1852 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
1853
1854 4. Make PCRE_ANCHORED public and accept as a compile option.
1855
1856 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
1857 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
1858 pass them.
1859
1860 6. Give an error if bad option bits passed at compile or run time.
1861
1862 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
1863 pcretest to cause it to pass that flag.
1864
1865 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
1866 options, and the first character, if set.
1867
1868 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
1869
1870
1871 Version 0.91 10-Sep-97
1872 ----------------------
1873
1874 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
1875 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
1876
1877 2. PCRE was looping on encountering an indefinitely repeated back reference to
1878 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
1879 Perl does - treats the match as successful.
1880
1881 ****

webmaster@exim.org
ViewVC Help
Powered by ViewVC 1.1.12