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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 69 - (show annotations) (download)
Sat Feb 24 21:40:18 2007 UTC (7 years, 7 months ago) by nigel
File size: 57033 byte(s)
Load pcre-4.3 into code/trunk.

1 ChangeLog for PCRE
2 ------------------
3
4 Version 4.3 21-May-03
5 ---------------------
6
7 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
8 Makefile.
9
10 2. Some refactoring to improve the quality of the code:
11
12 (i) The utf8_table... variables are now declared "const".
13
14 (ii) The code for \cx, which used the "case flipping" table to upper case
15 lower case letters, now just substracts 32. This is ASCII-specific,
16 but the whole concept of \cx is ASCII-specific, so it seems
17 reasonable.
18
19 (iii) PCRE was using its character types table to recognize decimal and
20 hexadecimal digits in the pattern. This is silly, because it handles
21 only 0-9, a-f, and A-F, but the character types table is locale-
22 specific, which means strange things might happen. A private
23 table is now used for this - though it costs 256 bytes, a table is
24 much faster than multiple explicit tests. Of course, the standard
25 character types table is still used for matching digits in subject
26 strings against \d.
27
28 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
29 ending in _t are). So I've renamed it as ESC_tee.
30
31 3. The first argument for regexec() in the POSIX wrapper should have been
32 defined as "const".
33
34 4. Changed pcretest to use malloc() for its buffers so that they can be
35 Electric Fenced for debugging.
36
37 5. There were several places in the code where, in UTF-8 mode, PCRE would try
38 to read one or more bytes before the start of the subject string. Often this
39 had no effect on PCRE's behaviour, but in some circumstances it could
40 provoke a segmentation fault.
41
42 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
43 to try to read one or more bytes before the start of the subject string.
44
45 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
46 UTF-8 support could misbehave in various ways if the subject string
47 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
48 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
49 back over UTF-8 characters.)
50
51
52 Version 4.2 14-Apr-03
53 ---------------------
54
55 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
56
57 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
58 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
59 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
60 [WIN_PREFIX]: new variable, "cyg" for Cygwin
61 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
62 and BUILD_EXEEXT
63 Note: automatic setting of the BUILD variables is not yet working
64 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
65 compile-time but not at link-time
66 [LINK]: use for linking executables only
67 make different versions for Windows and non-Windows
68 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
69 libraries
70 [LINK_FOR_BUILD]: new variable
71 [OBJEXT]: use throughout
72 [EXEEXT]: use throughout
73 <winshared>: new target
74 <wininstall>: new target
75 <dftables.o>: use native compiler
76 <dftables>: use native linker
77 <install>: handle Windows platform correctly
78 <clean>: ditto
79 <check>: ditto
80 copy DLL to top builddir before testing
81
82 As part of these changes, -no-undefined was removed again. This was reported
83 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
84 in any case.
85
86 3. Some tidies to get rid of compiler warnings:
87
88 . In the match_data structure, match_limit was an unsigned long int, whereas
89 match_call_count was an int. I've made them both unsigned long ints.
90
91 . In pcretest the fact that a const uschar * doesn't automatically cast to
92 a void * provoked a warning.
93
94 . Turning on some more compiler warnings threw up some "shadow" variables
95 and a few more missing casts.
96
97 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
98 option, a class that contained a single character with a value between 128
99 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
100
101 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
102 option, a class that contained several characters, but with at least one
103 whose value was between 128 and 255 caused PCRE to crash.
104
105
106 Version 4.1 12-Mar-03
107 ---------------------
108
109 1. Compiling with gcc -pedantic found a couple of places where casts were
110 needed, and a string in dftables.c that was longer than standard compilers are
111 required to support.
112
113 2. Compiling with Sun's compiler found a few more places where the code could
114 be tidied up in order to avoid warnings.
115
116 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
117 first of these names is deprecated in the latest Autoconf in favour of the name
118 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
119 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
120 analogy I have changed it to CFLAGS_FOR_BUILD.
121
122 4. Added -no-undefined to the linking command in the Makefile, because this is
123 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
124 linking step for the pcreposix library.
125
126 5. PCRE was failing to diagnose the case of two named groups with the same
127 name.
128
129 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
130 literal character that is needed in the subject for a match, and scans along to
131 ensure that it is present before embarking on the full matching process. This
132 saves time in cases of nested unlimited repeats that are never going to match.
133 Problem: the scan can take a lot of time if the subject is very long (e.g.
134 megabytes), thus penalizing straightforward matches. It is now done only if the
135 amount of subject to be scanned is less than 1000 bytes.
136
137 7. A lesser problem with the same optimization is that it was recording the
138 first character of an anchored pattern as "needed", thus provoking a search
139 right along the subject, even when the first match of the pattern was going to
140 fail. The "needed" character is now not set for anchored patterns, unless it
141 follows something in the pattern that is of non-fixed length. Thus, it still
142 fulfils its original purpose of finding quick non-matches in cases of nested
143 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
144
145
146 Version 4.0 17-Feb-03
147 ---------------------
148
149 1. If a comment in an extended regex that started immediately after a meta-item
150 extended to the end of string, PCRE compiled incorrect data. This could lead to
151 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
152
153 2. Moved to autoconf 2.53 and libtool 1.4.2.
154
155 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
156 the special perltest8 script is no longer needed - all the tests can be run
157 from a single perltest script.
158
159 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
160 by \s. It has now been removed in PCRE. This means it isn't recognized as
161 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
162 class [:space:] *does* include VT, thereby creating a mess.
163
164 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
165 space and tab.
166
167 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
168 its new features into the main test script, reducing the number of scripts.
169
170 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
171 were backward compatible, and made the (?i) apply to the whole pattern, as if
172 /i were given. Now it behaves more logically, and applies the option setting
173 only to what follows. PCRE has been changed to follow suit. However, if it
174 finds options settings right at the start of the pattern, it extracts them into
175 the global options, as before. Thus, they show up in the info data.
176
177 8. Added support for the \Q...\E escape sequence. Characters in between are
178 treated as literals. This is slightly different from Perl in that $ and @ are
179 also handled as literals inside the quotes. In Perl, they will cause variable
180 interpolation. Note the following examples:
181
182 Pattern PCRE matches Perl matches
183
184 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
185 \Qabc\$xyz\E abc\$xyz abc\$xyz
186 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
187
188 For compatibility with Perl, \Q...\E sequences are recognized inside character
189 classes as well as outside them.
190
191 9. Re-organized 3 code statements in pcretest to avoid "overflow in
192 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
193 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
194 signed/unsigned warnings.
195
196 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
197 option for pcretest, so I've replaced it by a simple function that does just
198 that job.
199
200 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
201 "pcregrep -".
202
203 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
204 Java package. This provides some syntactic sugar for simple cases of what my
205 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
206 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
207 item, you can use this simplified notation. Note that only makes sense with
208 greedy quantifiers. Consequently, the use of the possessive quantifier forces
209 greediness, whatever the setting of the PCRE_UNGREEDY option.
210
211 13. A change of greediness default within a pattern was not taking effect at
212 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
213 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
214 was abstracted outside.
215
216 14. PCRE now supports the \G assertion. It is true when the current matching
217 position is at the start point of the match. This differs from \A when the
218 starting offset is non-zero. Used with the /g option of pcretest (or similar
219 code), it works in the same way as it does for Perl's /g option. If all
220 alternatives of a regex begin with \G, the expression is anchored to the start
221 match position, and the "anchored" flag is set in the compiled expression.
222
223 15. Some bugs concerning the handling of certain option changes within patterns
224 have been fixed. These applied to options other than (?ims). For example,
225 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
226 been the other way round. Some of this was related to change 7 above.
227
228 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
229 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
230 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
231 POSIX classes only within a class (e.g. /[[:alpha:]]/).
232
233 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
234 mode. Unlike ".", it always matches newline, whatever the setting of
235 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
236 assertions. Perl allows it, but it doesn't (in general) work because it can't
237 calculate the length of the lookbehind. At least, that's the case for Perl
238 5.8.0 - I've been told they are going to document that it doesn't work in
239 future.
240
241 18. Added an error diagnosis for escapes that PCRE does not support: these are
242 \L, \l, \N, \P, \p, \U, \u, and \X.
243
244 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
245 reading past the end of the pattern in cases such as /[abcd/.
246
247 20. PCRE was getting more memory than necessary for patterns with classes that
248 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
249
250 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
251 compiling PCRE for use with Virtual Pascal.
252
253 22. Small fix to the Makefile to make it work properly if the build is done
254 outside the source tree.
255
256 23. Added a new extension: a condition to go with recursion. If a conditional
257 subpattern starts with (?(R) the "true" branch is used if recursion has
258 happened, whereas the "false" branch is used only at the top level.
259
260 24. When there was a very long string of literal characters (over 255 bytes
261 without UTF support, over 250 bytes with UTF support), the computation of how
262 much memory was required could be incorrect, leading to segfaults or other
263 strange effects.
264
265 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
266 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
267 there was a subsequent back reference to those brackets. This meant that, for
268 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
269 possible to check for precisely this case. All we can do is abandon the
270 optimization if .* occurs inside capturing brackets when there are any back
271 references whatsoever. (See below for a better fix that came later.)
272
273 26. The handling of the optimization for finding the first character of a
274 non-anchored pattern, and for finding a character that is required later in the
275 match were failing in some cases. This didn't break the matching; it just
276 failed to optimize when it could. The way this is done has been re-implemented.
277
278 27. Fixed typo in error message for invalid (?R item (it said "(?p").
279
280 28. Added a new feature that provides some of the functionality that Perl
281 provides with (?{...}). The facility is termed a "callout". The way it is done
282 in PCRE is for the caller to provide an optional function, by setting
283 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
284 global variable. By default it is unset, which disables all calling out. To get
285 the function called, the regex must include (?C) at appropriate points. This
286 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
287 This provides a means of identifying different callout points. When PCRE
288 reaches such a point in the regex, if pcre_callout has been set, the external
289 function is called. It is provided with data in a structure called
290 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
291 matching continues; if it returns a non-zero value, the match at the current
292 point fails. However, backtracking will occur if possible. [This was changed
293 later and other features added - see item 49 below.]
294
295 29. pcretest is upgraded to test the callout functionality. It provides a
296 callout function that displays information. By default, it shows the start of
297 the match and the current position in the text. There are some new data escapes
298 to vary what happens:
299
300 \C+ in addition, show current contents of captured substrings
301 \C- do not supply a callout function
302 \C!n return 1 when callout number n is reached
303 \C!n!m return 1 when callout number n is reached for the mth time
304
305 30. If pcregrep was called with the -l option and just a single file name, it
306 output "<stdin>" if a match was found, instead of the file name.
307
308 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
309 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
310 pcre_exec(). This saves a malloc/free per call. The default value of
311 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
312 when configuring.
313
314 32. The default maximum size of a compiled pattern is 64K. There have been a
315 few cases of people hitting this limit. The code now uses macros to handle the
316 storing of links as offsets within the compiled pattern. It defaults to 2-byte
317 links, but this can be changed to 3 or 4 bytes by --with-link-size when
318 configuring. Tests 2 and 5 work only with 2-byte links because they output
319 debugging information about compiled patterns.
320
321 33. Internal code re-arrangements:
322
323 (a) Moved the debugging function for printing out a compiled regex into
324 its own source file (printint.c) and used #include to pull it into
325 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
326 separate copies.
327
328 (b) Defined the list of op-code names for debugging as a macro in
329 internal.h so that it is next to the definition of the opcodes.
330
331 (c) Defined a table of op-code lengths for simpler skipping along compiled
332 code. This is again a macro in internal.h so that it is next to the
333 definition of the opcodes.
334
335 34. Added support for recursive calls to individual subpatterns, along the
336 lines of Robin Houston's patch (but implemented somewhat differently).
337
338 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
339 allow it to read and process whole directories in Win32. This code was
340 contributed by Lionel Fourquaux; it has not been tested by me.
341
342 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
343 used to name a group. Names consist of alphanumerics and underscores, and must
344 be unique. Back references use the syntax (?P=name) and recursive calls use
345 (?P>name) which is a PCRE extension to the Python extension. Groups still have
346 numbers. The function pcre_fullinfo() can be used after compilation to extract
347 a name/number map. There are three relevant calls:
348
349 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
350 PCRE_INFO_NAMECOUNT yields the number of entries
351 PCRE_INFO_NAMETABLE yields a pointer to the map.
352
353 The map is a vector of fixed-size entries. The size of each entry depends on
354 the length of the longest name used. The first two bytes of each entry are the
355 group number, most significant byte first. There follows the corresponding
356 name, zero terminated. The names are in alphabetical order.
357
358 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
359 case instead of 255. Making it the same both with and without UTF-8 support
360 means that the same test output works with both.
361
362 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
363 calling malloc() with a zero argument.
364
365 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
366 optimization. I've improved things by keeping a bitmap of backreferences with
367 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
368 fact referenced, the optimization can be applied. It is unlikely that a
369 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
370 the match to follow \n) will appear inside brackets with a number greater than
371 31, but if it does, any back reference > 31 suppresses the optimization.
372
373 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
374 of disabling numbered capturing parentheses. Any opening parenthesis that is
375 not followed by ? behaves as if it were followed by ?: but named parentheses
376 can still be used for capturing (and they will acquire numbers in the usual
377 way).
378
379 41. Redesigned the return codes from the match() function into yes/no/error so
380 that errors can be passed back from deep inside the nested calls. A malloc
381 failure while inside a recursive subpattern call now causes the
382 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
383
384 42. It is now possible to set a limit on the number of times the match()
385 function is called in a call to pcre_exec(). This facility makes it possible to
386 limit the amount of recursion and backtracking, though not in a directly
387 obvious way, because the match() function is used in a number of different
388 circumstances. The count starts from zero for each position in the subject
389 string (for non-anchored patterns). The default limit is, for compatibility, a
390 large number, namely 10 000 000. You can change this in two ways:
391
392 (a) When configuring PCRE before making, you can use --with-match-limit=n
393 to set a default value for the compiled library.
394
395 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
396 a different value is set. See 45 below.
397
398 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
399
400 43. Added a new function pcre_config(int, void *) to enable run-time extraction
401 of things that can be changed at compile time. The first argument specifies
402 what is wanted and the second points to where the information is to be placed.
403 The current list of available information is:
404
405 PCRE_CONFIG_UTF8
406
407 The output is an integer that is set to one if UTF-8 support is available;
408 otherwise it is set to zero.
409
410 PCRE_CONFIG_NEWLINE
411
412 The output is an integer that it set to the value of the code that is used for
413 newline. It is either LF (10) or CR (13).
414
415 PCRE_CONFIG_LINK_SIZE
416
417 The output is an integer that contains the number of bytes used for internal
418 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
419
420 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
421
422 The output is an integer that contains the threshold above which the POSIX
423 interface uses malloc() for output vectors. See item 31 above.
424
425 PCRE_CONFIG_MATCH_LIMIT
426
427 The output is an unsigned integer that contains the default limit of the number
428 of match() calls in a pcre_exec() execution. See 42 above.
429
430 44. pcretest has been upgraded by the addition of the -C option. This causes it
431 to extract all the available output from the new pcre_config() function, and to
432 output it. The program then exits immediately.
433
434 45. A need has arisen to pass over additional data with calls to pcre_exec() in
435 order to support additional features. One way would have been to define
436 pcre_exec2() (for example) with extra arguments, but this would not have been
437 extensible, and would also have required all calls to the original function to
438 be mapped to the new one. Instead, I have chosen to extend the mechanism that
439 is used for passing in "extra" data from pcre_study().
440
441 The pcre_extra structure is now exposed and defined in pcre.h. It currently
442 contains the following fields:
443
444 flags a bitmap indicating which of the following fields are set
445 study_data opaque data from pcre_study()
446 match_limit a way of specifying a limit on match() calls for a specific
447 call to pcre_exec()
448 callout_data data for callouts (see 49 below)
449
450 The flag bits are also defined in pcre.h, and are
451
452 PCRE_EXTRA_STUDY_DATA
453 PCRE_EXTRA_MATCH_LIMIT
454 PCRE_EXTRA_CALLOUT_DATA
455
456 The pcre_study() function now returns one of these new pcre_extra blocks, with
457 the actual study data pointed to by the study_data field, and the
458 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
459 before. That is, this change is entirely upwards-compatible and requires no
460 change to existing code.
461
462 If you want to pass in additional data to pcre_exec(), you can either place it
463 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
464 block.
465
466 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
467 data string contains the escape sequence \M, pcretest calls pcre_exec() several
468 times with different match limits, until it finds the minimum value needed for
469 pcre_exec() to complete. The value is then output. This can be instructive; for
470 most simple matches the number is quite small, but for pathological cases it
471 gets very large very quickly.
472
473 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
474 returns the size of the data block pointed to by the study_data field in a
475 pcre_extra block, that is, the value that was passed as the argument to
476 pcre_malloc() when PCRE was getting memory in which to place the information
477 created by pcre_study(). The fourth argument should point to a size_t variable.
478 pcretest has been extended so that this information is shown after a successful
479 pcre_study() call when information about the compiled regex is being displayed.
480
481 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
482 because what follows is always an absolute path. (Later: it turns out that this
483 is more than cosmetic for MinGW, because it doesn't like empty path
484 components.)
485
486 49. Some changes have been made to the callout feature (see 28 above):
487
488 (i) A callout function now has three choices for what it returns:
489
490 0 => success, carry on matching
491 > 0 => failure at this point, but backtrack if possible
492 < 0 => serious error, return this value from pcre_exec()
493
494 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
495 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
496 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
497 use by callout functions. It will never be used by PCRE itself.
498
499 (ii) The pcre_extra structure (see 45 above) has a void * field called
500 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
501 pcre_callout_block structure has a field of the same name. The contents of
502 the field passed in the pcre_extra structure are passed to the callout
503 function in the corresponding field in the callout block. This makes it
504 easier to use the same callout-containing regex from multiple threads. For
505 testing, the pcretest program has a new data escape
506
507 \C*n pass the number n (may be negative) as callout_data
508
509 If the callout function in pcretest receives a non-zero value as
510 callout_data, it returns that value.
511
512 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
513 there were some redundant $(CFLAGS) in commands that are now specified as
514 $(LINK), which already includes $(CFLAGS).
515
516 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
517 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
518 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
519 one-byte characters throughout. Note that case-insensitive matching applies
520 only to characters whose values are less than 256. PCRE doesn't support the
521 notion of cases for higher-valued characters.
522
523 (i) A character class whose characters are all within 0-255 is handled as
524 a bit map, and the map is inverted for negative classes. Previously, a
525 character > 255 always failed to match such a class; however it should
526 match if the class was a negative one (e.g. [^ab]). This has been fixed.
527
528 (ii) A negated character class with a single character < 255 is coded as
529 "not this character" (OP_NOT). This wasn't working properly when the test
530 character was multibyte, either singly or repeated.
531
532 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
533 mode, for example: \x{100}{2,3}.
534
535 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
536 singly or repeated) now correctly test multibyte characters. However,
537 PCRE doesn't recognize any characters with values greater than 255 as
538 digits, spaces, or word characters. Such characters always match \D, \S,
539 and \W, and never match \d, \s, or \w.
540
541 (v) Classes may now contain characters and character ranges with values
542 greater than 255. For example: [ab\x{100}-\x{400}].
543
544 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
545 PCRE in UTF-8 mode.
546
547 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
548 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
549 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
550 value.)
551
552 53. The single man page has become too large. I have therefore split it up into
553 a number of separate man pages. These also give rise to individual HTML pages;
554 these are now put in a separate directory, and there is an index.html page that
555 lists them all. Some hyperlinking between the pages has been installed.
556
557 54. Added convenience functions for handling named capturing parentheses.
558
559 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
560 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
561 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
562 are faulted.
563
564 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
565 calling configure. These values are used when compiling the dftables.c program
566 which is run to generate the source of the default character tables. They
567 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
568 you will need to set these values.
569
570 57. Updated the building process for Windows DLL, as provided by Fred Cox.
571
572
573 Version 3.9 02-Jan-02
574 ---------------------
575
576 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
577
578 2. If --disable-static was given, the building process failed when trying to
579 build pcretest and pcregrep. (For some reason it was using libtool to compile
580 them, which is not right, as they aren't part of the library.)
581
582
583 Version 3.8 18-Dec-01
584 ---------------------
585
586 1. The experimental UTF-8 code was completely screwed up. It was packing the
587 bytes in the wrong order. How dumb can you get?
588
589
590 Version 3.7 29-Oct-01
591 ---------------------
592
593 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
594 This caused pcretest, when used on the test data, to segfault. Unfortunately,
595 this didn't happen under Solaris 8, where I normally test things.
596
597 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
598 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
599 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
600 this fix an hour or so after the initial 3.7 release.)
601
602
603 Version 3.6 23-Oct-01
604 ---------------------
605
606 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
607 offsets passed as NULL with zero offset count.
608
609 2. The config.guess and config.sub files had not been updated when I moved to
610 the latest autoconf.
611
612
613 Version 3.5 15-Aug-01
614 ---------------------
615
616 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
617 had been forgotten.
618
619 2. By using declared but undefined structures, we can avoid using "void"
620 definitions in pcre.h while keeping the internal definitions of the structures
621 private.
622
623 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
624 user point of view, this means that both static and shared libraries are built
625 by default, but this can be individually controlled. More of the work of
626 handling this static/shared cases is now inside libtool instead of PCRE's make
627 file.
628
629 4. The pcretest utility is now installed along with pcregrep because it is
630 useful for users (to test regexs) and by doing this, it automatically gets
631 relinked by libtool. The documentation has been turned into a man page, so
632 there are now .1, .txt, and .html versions in /doc.
633
634 5. Upgrades to pcregrep:
635 (i) Added long-form option names like gnu grep.
636 (ii) Added --help to list all options with an explanatory phrase.
637 (iii) Added -r, --recursive to recurse into sub-directories.
638 (iv) Added -f, --file to read patterns from a file.
639
640 6. pcre_exec() was referring to its "code" argument before testing that
641 argument for NULL (and giving an error if it was NULL).
642
643 7. Upgraded Makefile.in to allow for compiling in a different directory from
644 the source directory.
645
646 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
647 options bits, the pointer it was passed was to an int instead of to an unsigned
648 long int. This mattered only on 64-bit systems.
649
650 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
651 generated) instead of pcre.in, which it its source. Also made the same change
652 in several of the .c files.
653
654 10. A new release of gcc defines printf() as a macro, which broke pcretest
655 because it had an ifdef in the middle of a string argument for printf(). Fixed
656 by using separate calls to printf().
657
658 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
659 script, to force use of CR or LF instead of \n in the source. On non-Unix
660 systems, the value can be set in config.h.
661
662 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
663 absolute limit. Changed the text of the error message to make this clear, and
664 likewise updated the man page.
665
666 13. The limit of 99 on the number of capturing subpatterns has been removed.
667 The new limit is 65535, which I hope will not be a "real" limit.
668
669
670 Version 3.4 22-Aug-00
671 ---------------------
672
673 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
674
675 2. Diagnose condition (?(0) as an error instead of crashing on matching.
676
677
678 Version 3.3 01-Aug-00
679 ---------------------
680
681 1. If an octal character was given, but the value was greater than \377, it
682 was not getting masked to the least significant bits, as documented. This could
683 lead to crashes in some systems.
684
685 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
686 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
687
688 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
689 These just pass their arguments on to (pcre_free)(), but they are provided
690 because some uses of PCRE bind it to non-C systems that can call its functions,
691 but cannot call free() or pcre_free() directly.
692
693 4. Add "make test" as a synonym for "make check". Corrected some comments in
694 the Makefile.
695
696 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
697 Makefile.
698
699 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
700 command called pgrep for grepping around the active processes.
701
702 7. Added the beginnings of support for UTF-8 character strings.
703
704 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
705 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
706 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
707 out for the ar command.)
708
709
710 Version 3.2 12-May-00
711 ---------------------
712
713 This is purely a bug fixing release.
714
715 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
716 of ZA. This was just one example of several cases that could provoke this bug,
717 which was introduced by change 9 of version 2.00. The code for breaking
718 infinite loops after an iteration that matches an empty string was't working
719 correctly.
720
721 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
722 when matched against abbab (for example). After matching an empty string, it
723 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
724 caused it to match further down the string than it should.
725
726 3. The code contained an inclusion of sys/types.h. It isn't clear why this
727 was there because it doesn't seem to be needed, and it causes trouble on some
728 systems, as it is not a Standard C header. It has been removed.
729
730 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
731 were reported on the Macintosh. The changes were from
732
733 while ((c = *(++ptr)) != 0 && c != '\n');
734 to
735 while ((c = *(++ptr)) != 0 && c != '\n') ;
736
737 Totally extraordinary, but if that's what it takes...
738
739 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
740 available. Added HAVE_BCOPY and an autoconf test for it; if neither
741 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
742 assumes the way PCRE uses memmove() (always moving upwards).
743
744 6. PCRE is being used in one environment where strchr() is not available. There
745 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
746 faster code anyway.
747
748
749 Version 3.1 09-Feb-00
750 ---------------------
751
752 The only change in this release is the fixing of some bugs in Makefile.in for
753 the "install" target:
754
755 (1) It was failing to install pcreposix.h.
756
757 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
758
759
760 Version 3.0 01-Feb-00
761 ---------------------
762
763 1. Add support for the /+ modifier to perltest (to output $` like it does in
764 pcretest).
765
766 2. Add support for the /g modifier to perltest.
767
768 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
769 matches null strings.
770
771 4. Fix perltest so that it doesn't do unwanted things when fed an empty
772 pattern. Perl treats empty patterns specially - it reuses the most recent
773 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
774 effect.
775
776 5. The POSIX interface was broken in that it was just handing over the POSIX
777 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
778 required a bigger vector, with some working space on the end. This means that
779 the POSIX wrapper now has to get and free some memory, and copy the results.
780
781 6. Added some simple autoconf support, placing the test data and the
782 documentation in separate directories, re-organizing some of the
783 information files, and making it build pcre-config (a GNU standard). Also added
784 libtool support for building PCRE as a shared library, which is now the
785 default.
786
787 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
788 09 are not valid octal constants. Single digits will be used for minor values
789 less than 10.
790
791 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
792 existing programs that set these in the POSIX interface can use PCRE without
793 modification.
794
795 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
796 return all that pcre_info() returns, plus additional data. The pcre_info()
797 function is retained for compatibility, but is considered to be obsolete.
798
799 10. Added experimental recursion feature (?R) to handle one common case that
800 Perl 5.6 will be able to do with (?p{...}).
801
802 11. Added support for POSIX character classes like [:alpha:], which Perl is
803 adopting.
804
805
806 Version 2.08 31-Aug-99
807 ----------------------
808
809 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
810 trying to match at the startoffset position, but instead was moving forward to
811 the next newline as if a previous match had failed.
812
813 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
814 and could get into a loop if a null string was matched other than at the start
815 of the subject.
816
817 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
818 be distinguished at compile time, and for completeness also added PCRE_DATE.
819
820 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
821 in GnuWin32 environments.
822
823
824 Version 2.07 29-Jul-99
825 ----------------------
826
827 1. The documentation is now supplied in plain text form and HTML as well as in
828 the form of man page sources.
829
830 2. C++ compilers don't like assigning (void *) values to other pointer types.
831 In particular this affects malloc(). Although there is no problem in Standard
832 C, I've put in casts to keep C++ compilers happy.
833
834 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
835 should be (const char *).
836
837 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
838 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
839 However, I haven't made this a standard facility. The documentation doesn't
840 mention it, and the Makefile doesn't support it.
841
842 5. The Makefile now contains an "install" target, with editable destinations at
843 the top of the file. The pcretest program is not installed.
844
845 6. pgrep -V now gives the PCRE version number and date.
846
847 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
848 causing the entire string to be ignored, instead of just the last character.
849
850 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
851 non-matching string, it can take a very, very long time, even for strings of
852 quite modest length, because of the nested recursion. PCRE now does better in
853 some of these cases. It does this by remembering the last required literal
854 character in the pattern, and pre-searching the subject to ensure it is present
855 before running the real match. In other words, it applies a heuristic to detect
856 some types of certain failure quickly, and in the above example, if presented
857 with a string that has no trailing " it gives "no match" very quickly.
858
859 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
860 other alternatives are tried instead.
861
862
863 Version 2.06 09-Jun-99
864 ----------------------
865
866 1. Change pcretest's output for amount of store used to show just the code
867 space, because the remainder (the data block) varies in size between 32-bit and
868 64-bit systems.
869
870 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
871 start matching at. This allows lookbehinds to work when searching for multiple
872 occurrences in a string.
873
874 3. Added additional options to pcretest for testing multiple occurrences:
875
876 /+ outputs the rest of the string that follows a match
877 /g loops for multiple occurrences, using the new startoffset argument
878 /G loops for multiple occurrences by passing an incremented pointer
879
880 4. PCRE wasn't doing the "first character" optimization for patterns starting
881 with \b or \B, though it was doing it for other lookbehind assertions. That is,
882 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
883 the letter 'x'. On long subject strings, this gives a significant speed-up.
884
885
886 Version 2.05 21-Apr-99
887 ----------------------
888
889 1. Changed the type of magic_number from int to long int so that it works
890 properly on 16-bit systems.
891
892 2. Fixed a bug which caused patterns starting with .* not to work correctly
893 when the subject string contained newline characters. PCRE was assuming
894 anchoring for such patterns in all cases, which is not correct because .* will
895 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
896 DOTALL is set at top level; otherwise it knows that patterns starting with .*
897 must be retried after every newline in the subject.
898
899
900 Version 2.04 18-Feb-99
901 ----------------------
902
903 1. For parenthesized subpatterns with repeats whose minimum was zero, the
904 computation of the store needed to hold the pattern was incorrect (too large).
905 If such patterns were nested a few deep, this could multiply and become a real
906 problem.
907
908 2. Added /M option to pcretest to show the memory requirement of a specific
909 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
910
911 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
912 compiled in such a way that the backtracking after subsequent failure was
913 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
914 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
915
916
917 Version 2.03 02-Feb-99
918 ----------------------
919
920 1. Fixed typo and small mistake in man page.
921
922 2. Added 4th condition (GPL supersedes if conflict) and created separate
923 LICENCE file containing the conditions.
924
925 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
926 Perl, that is the internal \ allows the delimiter to be included in the
927 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
928 the final delimiter, add \ to the end of the pattern (to test the error).
929
930 4. Added the convenience functions for extracting substrings after a successful
931 match. Updated pcretest to make it able to test these functions.
932
933
934 Version 2.02 14-Jan-99
935 ----------------------
936
937 1. Initialized the working variables associated with each extraction so that
938 their saving and restoring doesn't refer to uninitialized store.
939
940 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
941 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
942 fix the problem.
943
944 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
945 calls, and wasn't printing the correct value for compiling calls. Increased the
946 default value of LOOPREPEAT, and the number of significant figures in the
947 times.
948
949 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
950
951 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
952 a building problem on Windows NT with a FAT file system.
953
954
955 Version 2.01 21-Oct-98
956 ----------------------
957
958 1. Changed the API for pcre_compile() to allow for the provision of a pointer
959 to character tables built by pcre_maketables() in the current locale. If NULL
960 is passed, the default tables are used.
961
962
963 Version 2.00 24-Sep-98
964 ----------------------
965
966 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
967 it any more.
968
969 2. Allow quantification of (?>) groups, and make it work correctly.
970
971 3. The first character computation wasn't working for (?>) groups.
972
973 4. Correct the implementation of \Z (it is permitted to match on the \n at the
974 end of the subject) and add 5.005's \z, which really does match only at the
975 very end of the subject.
976
977 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
978
979 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
980 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
981 localized options. All options to pcre_study() were also removed.
982
983 7. Add other new features from 5.005:
984
985 $(?<= positive lookbehind
986 $(?<! negative lookbehind
987 (?imsx-imsx) added the unsetting capability
988 such a setting is global if at outer level; local otherwise
989 (?imsx-imsx:) non-capturing groups with option setting
990 (?(cond)re|re) conditional pattern matching
991
992 A backreference to itself in a repeated group matches the previous
993 captured string.
994
995 8. General tidying up of studying (both automatic and via "study")
996 consequential on the addition of new assertions.
997
998 9. As in 5.005, unlimited repeated groups that could match an empty substring
999 are no longer faulted at compile time. Instead, the loop is forcibly broken at
1000 runtime if any iteration does actually match an empty substring.
1001
1002 10. Include the RunTest script in the distribution.
1003
1004 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
1005 discrepancies, some of which were old and were also with respect to 5.004. They
1006 have now been fixed.
1007
1008
1009 Version 1.09 28-Apr-98
1010 ----------------------
1011
1012 1. A negated single character class followed by a quantifier with a minimum
1013 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
1014 program crashes, or just wrong answers. This did not apply to negated classes
1015 containing more than one character, or to minima other than one.
1016
1017
1018 Version 1.08 27-Mar-98
1019 ----------------------
1020
1021 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
1022
1023 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
1024 latter must appear before anything that relies on it in the pattern.
1025
1026
1027 Version 1.07 16-Feb-98
1028 ----------------------
1029
1030 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
1031 repeat of a potentially empty string).
1032
1033
1034 Version 1.06 23-Jan-98
1035 ----------------------
1036
1037 1. Added Markus Oberhumer's little patches for C++.
1038
1039 2. Literal strings longer than 255 characters were broken.
1040
1041
1042 Version 1.05 23-Dec-97
1043 ----------------------
1044
1045 1. Negated character classes containing more than one character were failing if
1046 PCRE_CASELESS was set at run time.
1047
1048
1049 Version 1.04 19-Dec-97
1050 ----------------------
1051
1052 1. Corrected the man page, where some "const" qualifiers had been omitted.
1053
1054 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
1055 input syntax.
1056
1057 3. Fixed memory leak which occurred when a regex with back references was
1058 matched with an offsets vector that wasn't big enough. The temporary memory
1059 that is used in this case wasn't being freed if the match failed.
1060
1061 4. Tidied pcretest to ensure it frees memory that it gets.
1062
1063 5. Temporary memory was being obtained in the case where the passed offsets
1064 vector was exactly big enough.
1065
1066 6. Corrected definition of offsetof() from change 5 below.
1067
1068 7. I had screwed up change 6 below and broken the rules for the use of
1069 setjmp(). Now fixed.
1070
1071
1072 Version 1.03 18-Dec-97
1073 ----------------------
1074
1075 1. A erroneous regex with a missing opening parenthesis was correctly
1076 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
1077 on some systems.
1078
1079 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
1080 it was reported that one broken compiler failed on the former because "code" is
1081 also an independent variable.
1082
1083 3. The erroneous regex a[]b caused an array overrun reference.
1084
1085 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
1086 fail on data ending with that character. (It was going on too far, and checking
1087 the next character, typically a binary zero.) This was specific to the
1088 optimized code for single-character negative classes.
1089
1090 5. Added a contributed patch from the TIN world which does the following:
1091
1092 + Add an undef for memmove, in case the the system defines a macro for it.
1093
1094 + Add a definition of offsetof(), in case there isn't one. (I don't know
1095 the reason behind this - offsetof() is part of the ANSI standard - but
1096 it does no harm).
1097
1098 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
1099 most of the places where whitespace preceded '#'. I have given up and
1100 allowed the remaining 2 cases to be at the margin.
1101
1102 + Rename some variables in pcre to eliminate shadowing. This seems very
1103 pedantic, but does no harm, of course.
1104
1105 6. Moved the call to setjmp() into its own function, to get rid of warnings
1106 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
1107
1108 7. Constructs such as \d{8,} were compiling into the equivalent of
1109 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
1110 outcome, but in this particular case used more store than had been allocated,
1111 which caused the bug to be discovered because it threw up an internal error.
1112
1113 8. The debugging code in both pcre and pcretest for outputting the compiled
1114 form of a regex was going wrong in the case of back references followed by
1115 curly-bracketed repeats.
1116
1117
1118 Version 1.02 12-Dec-97
1119 ----------------------
1120
1121 1. Typos in pcre.3 and comments in the source fixed.
1122
1123 2. Applied a contributed patch to get rid of places where it used to remove
1124 'const' from variables, and fixed some signed/unsigned and uninitialized
1125 variable warnings.
1126
1127 3. Added the "runtest" target to Makefile.
1128
1129 4. Set default compiler flag to -O2 rather than just -O.
1130
1131
1132 Version 1.01 19-Nov-97
1133 ----------------------
1134
1135 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
1136 like /([ab]*)*/, that is, for classes with more than one character in them.
1137
1138 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
1139 as /((?>a*))*/ (a PCRE_EXTRA facility).
1140
1141
1142 Version 1.00 18-Nov-97
1143 ----------------------
1144
1145 1. Added compile-time macros to support systems such as SunOS4 which don't have
1146 memmove() or strerror() but have other things that can be used instead.
1147
1148 2. Arranged that "make clean" removes the executables.
1149
1150
1151 Version 0.99 27-Oct-97
1152 ----------------------
1153
1154 1. Fixed bug in code for optimizing classes with only one character. It was
1155 initializing a 32-byte map regardless, which could cause it to run off the end
1156 of the memory it had got.
1157
1158 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
1159
1160
1161 Version 0.98 22-Oct-97
1162 ----------------------
1163
1164 1. Fixed bug in code for handling temporary memory usage when there are more
1165 back references than supplied space in the ovector. This could cause segfaults.
1166
1167
1168 Version 0.97 21-Oct-97
1169 ----------------------
1170
1171 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
1172
1173 2. Optimized negated single characters not to use a bit map.
1174
1175 3. Brought error texts together as macro definitions; clarified some of them;
1176 fixed one that was wrong - it said "range out of order" when it meant "invalid
1177 escape sequence".
1178
1179 4. Changed some char * arguments to const char *.
1180
1181 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
1182
1183 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
1184 pcretest.
1185
1186
1187 Version 0.96 16-Oct-97
1188 ----------------------
1189
1190 1. Added a simple "pgrep" utility to the distribution.
1191
1192 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
1193 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
1194 where "ddd" means "one or more decimal digits".
1195
1196 3. Fixed serious bug. If a pattern had a back reference, but the call to
1197 pcre_exec() didn't supply a large enough ovector to record the related
1198 identifying subpattern, the match always failed. PCRE now remembers the number
1199 of the largest back reference, and gets some temporary memory in which to save
1200 the offsets during matching if necessary, in order to ensure that
1201 backreferences always work.
1202
1203 4. Increased the compatibility with Perl in a number of ways:
1204
1205 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
1206 to request this handling. The option can be set at compile or exec time.
1207
1208 (b) $ matches before a terminating newline by default; an option
1209 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
1210 mode). The option can be set at compile or exec time.
1211
1212 (c) The handling of \ followed by a digit other than 0 is now supposed to be
1213 the same as Perl's. If the decimal number it represents is less than 10
1214 or there aren't that many previous left capturing parentheses, an octal
1215 escape is read. Inside a character class, it's always an octal escape,
1216 even if it is a single digit.
1217
1218 (d) An escaped but undefined alphabetic character is taken as a literal,
1219 unless PCRE_EXTRA is set. Currently this just reserves the remaining
1220 escapes.
1221
1222 (e) {0} is now permitted. (The previous item is removed from the compiled
1223 pattern).
1224
1225 5. Changed all the names of code files so that the basic parts are no longer
1226 than 10 characters, and abolished the teeny "globals.c" file.
1227
1228 6. Changed the handling of character classes; they are now done with a 32-byte
1229 bit map always.
1230
1231 7. Added the -d and /D options to pcretest to make it possible to look at the
1232 internals of compilation without having to recompile pcre.
1233
1234
1235 Version 0.95 23-Sep-97
1236 ----------------------
1237
1238 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
1239 \x20 at the start of a run of normal characters. These were being treated as
1240 real characters, instead of the source characters being re-checked.
1241
1242
1243 Version 0.94 18-Sep-97
1244 ----------------------
1245
1246 1. The functions are now thread-safe, with the caveat that the global variables
1247 containing pointers to malloc() and free() or alternative functions are the
1248 same for all threads.
1249
1250 2. Get pcre_study() to generate a bitmap of initial characters for non-
1251 anchored patterns when this is possible, and use it if passed to pcre_exec().
1252
1253
1254 Version 0.93 15-Sep-97
1255 ----------------------
1256
1257 1. /(b)|(:+)/ was computing an incorrect first character.
1258
1259 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
1260 but not actually doing anything yet.
1261
1262 3. Treat "-" characters in classes that cannot be part of ranges as literals,
1263 as Perl does (e.g. [-az] or [az-]).
1264
1265 4. Set the anchored flag if a branch starts with .* or .*? because that tests
1266 all possible positions.
1267
1268 5. Split up into different modules to avoid including unneeded functions in a
1269 compiled binary. However, compile and exec are still in one module. The "study"
1270 function is split off.
1271
1272 6. The character tables are now in a separate module whose source is generated
1273 by an auxiliary program - but can then be edited by hand if required. There are
1274 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
1275 toupper() in the code.
1276
1277 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
1278 make them global. Abolish the function for setting them, as the caller can now
1279 set them directly.
1280
1281
1282 Version 0.92 11-Sep-97
1283 ----------------------
1284
1285 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
1286 (e.g. /a{1,3}/) was broken (I mis-optimized it).
1287
1288 2. Caseless matching was not working in character classes if the characters in
1289 the pattern were in upper case.
1290
1291 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
1292
1293 4. Make PCRE_ANCHORED public and accept as a compile option.
1294
1295 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
1296 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
1297 pass them.
1298
1299 6. Give an error if bad option bits passed at compile or run time.
1300
1301 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
1302 pcretest to cause it to pass that flag.
1303
1304 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
1305 options, and the first character, if set.
1306
1307 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
1308
1309
1310 Version 0.91 10-Sep-97
1311 ----------------------
1312
1313 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
1314 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
1315
1316 2. PCRE was looping on encountering an indefinitely repeated back reference to
1317 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
1318 Perl does - treats the match as successful.
1319
1320 ****

webmaster@exim.org
ViewVC Help
Powered by ViewVC 1.1.12