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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 67 - (show annotations) (download)
Sat Feb 24 21:40:13 2007 UTC (7 years, 5 months ago) by nigel
File size: 54877 byte(s)
Load pcre-4.2 into code/trunk.

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

webmaster@exim.org
ViewVC Help
Powered by ViewVC 1.1.12