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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 359 - (hide annotations) (download)
Wed Jul 9 16:20:19 2008 UTC (6 years ago) by ph10
File size: 158007 byte(s)
Add PCRE_CALL_CONVENTION for MSVC users; add some missing PCRE_EXP_DEFNs.

1 nigel 3 ChangeLog for PCRE
2     ------------------
3    
4 ph10 349 Version 8.0 02 Jul-08
5     ---------------------
6    
7     1. Replaced UCP searching code with optimized version as implemented for Ad
8     Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
9 ph10 351 stage table and inline lookup instead of a function, giving speed ups of 2
10     to 5 times on some simple patterns that I tested. Permission was given to
11     distribute the MultiStage2.py script that generates the tables (it's not in
12     the tarball, but is in the Subversion repository).
13 ph10 349
14 ph10 351 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
15     scripts.
16 ph10 358
17 ph10 353 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
18     a group with a zero qualifier. The result of the study could be incorrect,
19 ph10 358 or the function might crash, depending on the pattern.
20    
21     4. Caseless matching was not working for non-ASCII characters in back
22 ph10 354 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
23 ph10 358 It now works when Unicode Property Support is available.
24    
25 ph10 355 5. In pcretest, an escape such as \x{de} in the data was always generating
26     a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
27     non-UTF-8 mode. If the value is greater than 255, it gives a warning about
28 ph10 358 truncation.
29 ph10 349
30 ph10 358 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
31    
32 ph10 357 7. Added two (int) casts to pcregrep when printing the difference of two
33     pointers, in case they are 64-bit values.
34 ph10 358
35     8. Added comments about Mac OS X stack usage to the pcrestack man page and to
36     test 2 if it fails.
37 ph10 359
38     9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
39     and a #define of that name to empty if it is not externally set. This is to
40     allow users of MSVC to set it if necessary.
41    
42     10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
43     the convenience functions in the pcre_get.c source file.
44 ph10 351
45 ph10 357
46 ph10 347 Version 7.7 07-May-08
47 ph10 321 ---------------------
48    
49     1. Applied Craig's patch to sort out a long long problem: "If we can't convert
50 ph10 345 a string to a long long, pretend we don't even have a long long." This is
51 ph10 321 done by checking for the strtoq, strtoll, and _strtoi64 functions.
52 ph10 345
53 ph10 322 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
54     pre-7.6 versions, which defined a global no_arg variable instead of putting
55 ph10 332 it in the RE class. (See also #8 below.)
56 ph10 345
57     3. Remove a line of dead code, identified by coverity and reported by Nuno
58     Lopes.
59    
60 ph10 324 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
61 ph10 321
62 ph10 324 (1) The include/exclude patterns were being applied to the whole pathnames
63 ph10 345 of files, instead of just to the final components.
64    
65 ph10 324 (2) If there was more than one level of directory, the subdirectories were
66     skipped unless they satisfied the include/exclude conditions. This is
67     inconsistent with GNU grep (and could even be seen as contrary to the
68     pcregrep specification - which I improved to make it absolutely clear).
69     The action now is always to scan all levels of directory, and just
70     apply the include/exclude patterns to regular files.
71 ph10 345
72 ph10 325 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
73 ph10 345 --exclude_dir in the tests to avoid scanning .svn directories.
74    
75 ph10 326 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
76 ph10 345 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
77 ph10 326 doesn't support NULs in patterns.
78 ph10 345
79     7. Added some missing "const"s to declarations of static tables in
80     pcre_compile.c and pcre_dfa_exec.c.
81    
82 ph10 328 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
83 ph10 329 caused by fix #2 above. (Subsequently also a second patch to fix the
84 ph10 345 first patch. And a third patch - this was a messy problem.)
85 ph10 321
86 ph10 345 9. Applied Craig's patch to remove the use of push_back().
87    
88 ph10 332 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
89     matching function regexec().
90 ph10 345
91     11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
92     which, however, unlike Perl's \g{...}, are subroutine calls, not back
93 ph10 333 references. PCRE supports relative numbers with this syntax (I don't think
94     Oniguruma does).
95 ph10 345
96     12. Previously, a group with a zero repeat such as (...){0} was completely
97 ph10 335 omitted from the compiled regex. However, this means that if the group
98     was called as a subroutine from elsewhere in the pattern, things went wrong
99 ph10 345 (an internal error was given). Such groups are now left in the compiled
100     pattern, with a new opcode that causes them to be skipped at execution
101 ph10 335 time.
102 ph10 345
103 ph10 341 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
104     to the way PCRE behaves:
105 ph10 345
106 ph10 336 (a) A lone ] character is dis-allowed (Perl treats it as data).
107 ph10 345
108     (b) A back reference to an unmatched subpattern matches an empty string
109 ph10 336 (Perl fails the current match path).
110 ph10 345
111 ph10 341 (c) A data ] in a character class must be notated as \] because if the
112 ph10 345 first data character in a class is ], it defines an empty class. (In
113 ph10 341 Perl it is not possible to have an empty class.) The empty class []
114     never matches; it forces failure and is equivalent to (*FAIL) or (?!).
115     The negative empty class [^] matches any one character, independently
116     of the DOTALL setting.
117 ph10 345
118     14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
119 ph10 340 non-existent subpattern following a character class starting with ']' and
120     containing () gave an internal compiling error instead of "reference to
121     non-existent subpattern". Fortunately, when the pattern did exist, the
122 ph10 345 compiled code was correct. (When scanning forwards to check for the
123 ph10 340 existencd of the subpattern, it was treating the data ']' as terminating
124 ph10 345 the class, so got the count wrong. When actually compiling, the reference
125 ph10 340 was subsequently set up correctly.)
126 ph10 345
127 ph10 341 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
128 ph10 345 it was being rejected as not supported by pcre_dfa_exec(), even though
129     other assertions are supported. I have made pcre_dfa_exec() support
130     (*FAIL).
131    
132 ph10 342 16. The implementation of 13c above involved the invention of a new opcode,
133 ph10 345 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
134     cannot be changed at match time, I realized I could make a small
135     improvement to matching performance by compiling OP_ALLANY instead of
136     OP_ANY for "." when DOTALL was set, and then removing the runtime tests
137     on the OP_ANY path.
138    
139     17. Compiling pcretest on Windows with readline support failed without the
140 ph10 343 following two fixes: (1) Make the unistd.h include conditional on
141 ph10 345 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
142    
143 ph10 344 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
144 ph10 345 ncurses library to be included for pcretest when ReadLine support is
145     requested, but also to allow for it to be overridden. This patch came from
146     Daniel Bergström.
147 ph10 347
148 ph10 346 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
149 ph10 347 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
150     any errors with the current Unicode tables. Thanks to Peter Kankowski for
151     spotting this.
152 ph10 324
153 ph10 332
154 ph10 319 Version 7.6 28-Jan-08
155 ph10 300 ---------------------
156    
157 ph10 302 1. A character class containing a very large number of characters with
158     codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
159     overflow.
160 ph10 309
161     2. Patch to cut out the "long long" test in pcrecpp_unittest when
162     HAVE_LONG_LONG is not defined.
163    
164 ph10 303 3. Applied Christian Ehrlicher's patch to update the CMake build files to
165 ph10 304 bring them up to date and include new features. This patch includes:
166 ph10 309
167 ph10 304 - Fixed PH's badly added libz and libbz2 support.
168     - Fixed a problem with static linking.
169 ph10 312 - Added pcredemo. [But later removed - see 7 below.]
170 ph10 304 - Fixed dftables problem and added an option.
171     - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
172     HAVE_LONG_LONG.
173     - Added readline support for pcretest.
174 ph10 309 - Added an listing of the option settings after cmake has run.
175    
176 ph10 314 4. A user submitted a patch to Makefile that makes it easy to create
177     "pcre.dll" under mingw when using Configure/Make. I added stuff to
178     Makefile.am that cause it to include this special target, without
179     affecting anything else. Note that the same mingw target plus all
180     the other distribution libraries and programs are now supported
181     when configuring with CMake (see 6 below) instead of with
182     Configure/Make.
183 ph10 309
184 ph10 308 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
185     This is an attempt to solve the reported problem "pcrecpp::no_arg is not
186 ph10 309 exported in the Windows port". It has not yet been confirmed that the patch
187     solves the problem, but it does no harm.
188 ph10 313
189 ph10 311 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
190 ph10 319 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
191 ph10 314 with CMake, and also correct the comment about stack recursion.
192 ph10 313
193 ph10 312 7. Remove the automatic building of pcredemo from the ./configure system and
194 ph10 313 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
195     of a program that users should build themselves after PCRE is installed, so
196     building it automatically is not really right. What is more, it gave
197 ph10 312 trouble in some build environments.
198 ph10 300
199 ph10 319 8. Further tidies to CMakeLists.txt from Sheri and Christian.
200 ph10 308
201 ph10 319
202 ph10 298 Version 7.5 10-Jan-08
203 ph10 263 ---------------------
204    
205     1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
206     values in parens when parsing an RE using the C++ wrapper."
207 ph10 286
208 ph10 264 2. Negative specials like \S did not work in character classes in UTF-8 mode.
209     Characters greater than 255 were excluded from the class instead of being
210     included.
211 ph10 286
212     3. The same bug as (2) above applied to negated POSIX classes such as
213 ph10 265 [:^space:].
214 ph10 286
215 ph10 267 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
216     defined or documented. It seems to have been a typo for PCRE_STATIC, so
217 ph10 286 I have changed it.
218    
219 ph10 268 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
220 ph10 286 first named subpattern) and a construct such as (?&a) would reference the
221     first named subpattern whose name started with "a" (in other words, the
222 ph10 272 length check was missing). Both these problems are fixed. "Subpattern name
223     expected" is now given for (?&) (a zero-length name), and this patch also
224     makes it give the same error for \k'' (previously it complained that that
225     was a reference to a non-existent subpattern).
226 ph10 286
227 ph10 269 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
228 ph10 286 this is right because (?- can be followed by option settings as well as by
229 ph10 269 digits. I have, however, made the messages clearer.
230 ph10 286
231 ph10 270 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
232     than the number used in the conditional) now cause a compile-time error.
233     This is actually not compatible with Perl, which accepts such patterns, but
234     treats the conditional as always being FALSE (as PCRE used to), but it
235     seems to me that giving a diagnostic is better.
236 ph10 286
237 ph10 274 8. Change "alphameric" to the more common word "alphanumeric" in comments
238 ph10 275 and messages.
239 ph10 286
240     9. Fix two occurrences of "backslash" in comments that should have been
241     "backspace".
242    
243     10. Remove two redundant lines of code that can never be obeyed (their function
244     was moved elsewhere).
245    
246     11. The program that makes PCRE's Unicode character property table had a bug
247     which caused it to generate incorrect table entries for sequences of
248 ph10 277 characters that have the same character type, but are in different scripts.
249 ph10 286 It amalgamated them into a single range, with the script of the first of
250 ph10 277 them. In other words, some characters were in the wrong script. There were
251     thirteen such cases, affecting characters in the following ranges:
252 ph10 286
253 ph10 277 U+002b0 - U+002c1
254     U+0060c - U+0060d
255 ph10 286 U+0061e - U+00612
256 ph10 277 U+0064b - U+0065e
257     U+0074d - U+0076d
258     U+01800 - U+01805
259     U+01d00 - U+01d77
260     U+01d9b - U+01dbf
261     U+0200b - U+0200f
262     U+030fc - U+030fe
263     U+03260 - U+0327f
264     U+0fb46 - U+0fbb1
265     U+10450 - U+1049d
266 ph10 286
267 ph10 279 12. The -o option (show only the matching part of a line) for pcregrep was not
268     compatible with GNU grep in that, if there was more than one match in a
269     line, it showed only the first of them. It now behaves in the same way as
270     GNU grep.
271 ph10 286
272     13. If the -o and -v options were combined for pcregrep, it printed a blank
273     line for every non-matching line. GNU grep prints nothing, and pcregrep now
274     does the same. The return code can be used to tell if there were any
275     non-matching lines.
276    
277 ph10 289 14. Added --file-offsets and --line-offsets to pcregrep.
278    
279     15. The pattern (?=something)(?R) was not being diagnosed as a potentially
280 ph10 282 infinitely looping recursion. The bug was that positive lookaheads were not
281 ph10 286 being skipped when checking for a possible empty match (negative lookaheads
282     and both kinds of lookbehind were skipped).
283    
284 ph10 289 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
285 ph10 284 inclusion of <windows.h> to before rather than after the definition of
286     INVALID_FILE_ATTRIBUTES (patch from David Byron).
287 ph10 263
288 ph10 289 17. Specifying a possessive quantifier with a specific limit for a Unicode
289 ph10 286 character property caused pcre_compile() to compile bad code, which led at
290     runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
291     are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
292     caused the error; without that there was no problem.
293 ph10 263
294 ph10 289 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
295 ph10 286
296 ph10 289 19. Added --enable-pcretest-libreadline.
297 ph10 286
298 ph10 289 20. In pcrecpp.cc, the variable 'count' was incremented twice in
299 ph10 288 RE::GlobalReplace(). As a result, the number of replacements returned was
300 ph10 298 double what it should be. I removed one of the increments, but Craig sent a
301     later patch that removed the other one (the right fix) and added unit tests
302     that check the return values (which was not done before).
303 ph10 292
304 ph10 291 21. Several CMake things:
305 ph10 287
306 ph10 291 (1) Arranged that, when cmake is used on Unix, the libraries end up with
307 ph10 292 the names libpcre and libpcreposix, not just pcre and pcreposix.
308 ph10 288
309 ph10 292 (2) The above change means that pcretest and pcregrep are now correctly
310     linked with the newly-built libraries, not previously installed ones.
311 ph10 291
312 ph10 292 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
313 ph10 298
314     22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
315     crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
316     UTF-8 newline character). The key issue is that the pattern starts .*;
317     this means that the match must be either at the beginning, or after a
318     newline. The bug was in the code for advancing after a failed match and
319     checking that the new position followed a newline. It was not taking
320 ph10 294 account of UTF-8 characters correctly.
321 ph10 298
322     23. PCRE was behaving differently from Perl in the way it recognized POSIX
323     character classes. PCRE was not treating the sequence [:...:] as a
324     character class unless the ... were all letters. Perl, however, seems to
325     allow any characters between [: and :], though of course it rejects as
326     unknown any "names" that contain non-letters, because all the known class
327     names consist only of letters. Thus, Perl gives an error for [[:1234:]],
328     for example, whereas PCRE did not - it did not recognize a POSIX character
329     class. This seemed a bit dangerous, so the code has been changed to be
330     closer to Perl. The behaviour is not identical to Perl, because PCRE will
331     diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
332     treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
333     Perl does, and where it didn't before.
334    
335 ph10 296 24. Rewrite so as to remove the single use of %n from pcregrep because in some
336     Windows environments %n is disabled by default.
337 ph10 292
338    
339 ph10 260 Version 7.4 21-Sep-07
340 ph10 230 ---------------------
341    
342     1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
343 ph10 231 means that a class such as [\s] counted as "explicit reference to CR or
344     LF". That isn't really right - the whole point of the change was to try to
345     help when there was an actual mention of one of the two characters. So now
346     the change happens only if \r or \n (or a literal CR or LF) character is
347 ph10 230 encountered.
348 ph10 231
349     2. The 32-bit options word was also used for 6 internal flags, but the numbers
350     of both had grown to the point where there were only 3 bits left.
351     Fortunately, there was spare space in the data structure, and so I have
352     moved the internal flags into a new 16-bit field to free up more option
353 ph10 230 bits.
354 ph10 231
355     3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
356     but did not set the internal JCHANGED flag - either of these is enough to
357     control the way the "get" function works - but the PCRE_INFO_JCHANGED
358     facility is supposed to tell if (?J) was ever used, so now (?J) at the
359 ph10 230 start sets both bits.
360    
361 ph10 231 4. Added options (at build time, compile time, exec time) to change \R from
362     matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
363 ph10 230
364 ph10 243 5. doc/pcresyntax.html was missing from the distribution.
365    
366     6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
367 ph10 233 compatibility, even though it is no longer used.
368 ph10 243
369     7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
370 ph10 254 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
371 ph10 259 windows.h file is present (where different names are used). [This was
372 ph10 257 reversed later after testing - see 16 below.]
373 ph10 243
374     8. Changed all #include <config.h> to #include "config.h". There were also
375 ph10 236 some further <pcre.h> cases that I changed to "pcre.h".
376 ph10 243
377 ph10 239 9. When pcregrep was used with the --colour option, it missed the line ending
378 ph10 243 sequence off the lines that it output.
379    
380     10. It was pointed out to me that arrays of string pointers cause lots of
381     relocations when a shared library is dynamically loaded. A technique of
382     using a single long string with a table of offsets can drastically reduce
383     these. I have refactored PCRE in four places to do this. The result is
384 ph10 240 dramatic:
385 ph10 243
386 ph10 240 Originally: 290
387     After changing UCP table: 187
388 ph10 243 After changing error message table: 43
389 ph10 240 After changing table of "verbs" 36
390     After changing table of Posix names 22
391 ph10 243
392 ph10 240 Thanks to the folks working on Gregex for glib for this insight.
393 ph10 247
394 ph10 244 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
395 ph10 247 unicode-properties was also set.
396    
397 ph10 248 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
398 ph10 231
399 ph10 253 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
400     checked only for CRLF.
401 ph10 233
402 ph10 259 14. Added casts to pcretest.c to avoid compiler warnings.
403    
404 ph10 256 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
405 ph10 253
406 ph10 257 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
407 ph10 259 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
408     entirely. This removes changes made in 7 above.
409 ph10 256
410 ph10 261 17. The CMake files have been updated, and there is now more information about
411     building with CMake in the NON-UNIX-USE document.
412 ph10 257
413 ph10 261
414 ph10 228 Version 7.3 28-Aug-07
415 ph10 157 ---------------------
416    
417 ph10 189 1. In the rejigging of the build system that eventually resulted in 7.1, the
418     line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
419     brackets there is not right, since it causes compilers to look for an
420     installed pcre.h, not the version that is in the source that is being
421     compiled (which of course may be different). I have changed it back to:
422 ph10 197
423 ph10 189 #include "pcre.h"
424 ph10 197
425     I have a vague recollection that the change was concerned with compiling in
426     different directories, but in the new build system, that is taken care of
427     by the VPATH setting the Makefile.
428    
429 ph10 190 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
430     when the subject happened to end in the byte 0x85 (e.g. if the last
431     character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
432     characters but of course it shouldn't be taken as a newline when it is part
433     of another character. The bug was that, for an unlimited repeat of . in
434     not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
435     characters when looking for a newline.
436 ph10 197
437     3. A small performance improvement in the DOTALL UTF-8 mode .* case.
438    
439     4. Debugging: adjusted the names of opcodes for different kinds of parentheses
440     in debug output.
441    
442 ph10 201 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
443     long printing in the pcrecpp unittest when running under MinGW.
444 ph10 189
445 ph10 197 6. ESC_K was left out of the EBCDIC table.
446 ph10 189
447 ph10 197 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
448     parentheses; I made it 1000, which seemed large enough. Unfortunately, the
449     limit also applies to "virtual nesting" when a pattern is recursive, and in
450     this case 1000 isn't so big. I have been able to remove this limit at the
451     expense of backing off one optimization in certain circumstances. Normally,
452     when pcre_exec() would call its internal match() function recursively and
453     immediately return the result unconditionally, it uses a "tail recursion"
454     feature to save stack. However, when a subpattern that can match an empty
455     string has an unlimited repetition quantifier, it no longer makes this
456     optimization. That gives it a stack frame in which to save the data for
457     checking that an empty string has been matched. Previously this was taken
458     from the 1000-entry workspace that had been reserved. So now there is no
459     explicit limit, but more stack is used.
460 ph10 189
461 ph10 200 8. Applied Daniel's patches to solve problems with the import/export magic
462     syntax that is required for Windows, and which was going wrong for the
463     pcreposix and pcrecpp parts of the library. These were overlooked when this
464     problem was solved for the main library.
465 ph10 197
466 ph10 202 9. There were some crude static tests to avoid integer overflow when computing
467     the size of patterns that contain repeated groups with explicit upper
468     limits. As the maximum quantifier is 65535, the maximum group length was
469     set at 30,000 so that the product of these two numbers did not overflow a
470     32-bit integer. However, it turns out that people want to use groups that
471     are longer than 30,000 bytes (though not repeat them that many times).
472     Change 7.0/17 (the refactoring of the way the pattern size is computed) has
473     made it possible to implement the integer overflow checks in a much more
474     dynamic way, which I have now done. The artificial limitation on group
475     length has been removed - we now have only the limit on the total length of
476     the compiled pattern, which depends on the LINK_SIZE setting.
477 ph10 208
478     10. Fixed a bug in the documentation for get/copy named substring when
479     duplicate names are permitted. If none of the named substrings are set, the
480     functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
481     empty string.
482    
483     11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
484     instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
485     because the ] is interpreted as the first data character and the
486     terminating ] is not found. PCRE has been made compatible with Perl in this
487     regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
488     cause memory overwriting.
489    
490 ph10 206 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
491     string has been matched (to stop an infinite loop). It was not recognizing
492 ph10 208 a conditional subpattern that could match an empty string if that
493 ph10 206 subpattern was within another subpattern. For example, it looped when
494 ph10 208 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
495 ph10 206 condition was not nested. This bug has been fixed.
496 ph10 208
497 ph10 207 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
498     past the start of the subject in the presence of bytes with the top bit
499     set, for example "\x8aBCD".
500 ph10 211
501 ph10 210 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
502     (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
503 ph10 200
504 ph10 211 14. Optimized (?!) to (*FAIL).
505 ph10 202
506 ph10 212 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
507     This restricts code points to be within the range 0 to 0x10FFFF, excluding
508     the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
509     full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
510     does: it's just the validity check that is more restrictive.
511 ph10 220
512     16. Inserted checks for integer overflows during escape sequence (backslash)
513     processing, and also fixed erroneous offset values for syntax errors during
514     backslash processing.
515    
516 ph10 214 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
517 ph10 220 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
518    
519 ph10 215 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
520     caused an overrun.
521 ph10 220
522     19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
523     something other than just ASCII characters) inside a group that had an
524     unlimited repeat caused a loop at compile time (while checking to see
525     whether the group could match an empty string).
526    
527 ph10 217 20. Debugging a pattern containing \p or \P could cause a crash. For example,
528     [\P{Any}] did so. (Error in the code for printing property names.)
529 ph10 210
530 ph10 220 21. An orphan \E inside a character class could cause a crash.
531    
532     22. A repeated capturing bracket such as (A)? could cause a wild memory
533 ph10 218 reference during compilation.
534 ph10 220
535     23. There are several functions in pcre_compile() that scan along a compiled
536     expression for various reasons (e.g. to see if it's fixed length for look
537 ph10 218 behind). There were bugs in these functions when a repeated \p or \P was
538 ph10 220 present in the pattern. These operators have additional parameters compared
539     with \d, etc, and these were not being taken into account when moving along
540 ph10 218 the compiled data. Specifically:
541 ph10 220
542     (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
543     length.
544    
545     (b) An item such as \pL+ within a repeated group could cause crashes or
546 ph10 218 loops.
547 ph10 220
548 ph10 218 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
549 ph10 220 "reference to non-existent subpattern" error.
550    
551 ph10 221 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
552    
553 ph10 219 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
554 ph10 220 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
555 ph10 211
556 ph10 222 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
557 ph10 218
558 ph10 227 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
559 ph10 223 character were causing crashes (broken optimization).
560 ph10 227
561 ph10 224 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
562 ph10 227 \p or \P) caused a compile-time loop.
563    
564 ph10 226 28. More problems have arisen in unanchored patterns when CRLF is a valid line
565     break. For example, the unstudied pattern [\r\n]A does not match the string
566     "\r\nA" because change 7.0/46 below moves the current point on by two
567     characters after failing to match at the start. However, the pattern \nA
568 ph10 227 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
569     the same is true. There doesn't seem any very clean way out of this, but
570     what I have chosen to do makes the common cases work: PCRE now takes note
571     of whether there can be an explicit match for \r or \n anywhere in the
572     pattern, and if so, 7.0/46 no longer applies. As part of this change,
573     there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
574     pattern has explicit CR or LF references.
575 ph10 222
576 ph10 227 29. Added (*CR) etc for changing newline setting at start of pattern.
577 ph10 223
578 ph10 227
579 ph10 189 Version 7.2 19-Jun-07
580     ---------------------
581    
582 ph10 157 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
583     which is apparently normally available under Windows.
584 ph10 159
585     2. Re-jig the pcregrep tests with different newline settings in an attempt
586     to make them independent of the local environment's newline setting.
587 ph10 157
588 ph10 160 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
589 ph10 165
590 ph10 161 4. Some of the "internals" tests were previously cut out when the link size
591     was not 2, because the output contained actual offsets. The recent new
592 ph10 165 "Z" feature of pcretest means that these can be cut out, making the tests
593     usable with all link sizes.
594    
595 ph10 164 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
596     stack recursion. This gives a massive performance boost under BSD, but just
597 ph10 165 a small improvement under Linux. However, it saves one field in the frame
598 ph10 164 in all cases.
599 ph10 172
600 ph10 166 6. Added more features from the forthcoming Perl 5.10:
601 ph10 172
602 ph10 166 (a) (?-n) (where n is a string of digits) is a relative subroutine or
603     recursion call. It refers to the nth most recently opened parentheses.
604 ph10 172
605 ph10 166 (b) (?+n) is also a relative subroutine call; it refers to the nth next
606 ph10 172 to be opened parentheses.
607    
608     (c) Conditions that refer to capturing parentheses can be specified
609 ph10 167 relatively, for example, (?(-2)... or (?(+3)...
610 ph10 172
611 ph10 168 (d) \K resets the start of the current match so that everything before
612 ph10 172 is not part of it.
613    
614 ph10 171 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
615 ph10 172
616 ph10 171 (f) \g{name} is another synonym - part of Perl 5.10's unification of
617 ph10 172 reference syntax.
618 ph10 182
619 ph10 175 (g) (?| introduces a group in which the numbering of parentheses in each
620 ph10 182 alternative starts with the same number.
621 ph10 172
622 ph10 182 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
623    
624 ph10 172 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
625     PCRE_INFO_JCHANGED.
626    
627     8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
628     terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
629 ph10 170 for detecting groups that can match an empty string.
630 ph10 159
631 ph10 172 9. A pattern with a very large number of alternatives (more than several
632     hundred) was running out of internal workspace during the pre-compile
633     phase, where pcre_compile() figures out how much memory will be needed. A
634     bit of new cunning has reduced the workspace needed for groups with
635     alternatives. The 1000-alternative test pattern now uses 12 bytes of
636     workspace instead of running out of the 4096 that are available.
637 ph10 182
638 ph10 176 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
639 ph10 172
640 ph10 179 11. Applied patch from Google to remove an optimization that didn't quite work.
641     The report of the bug said:
642 ph10 182
643 ph10 179 pcrecpp::RE("a*").FullMatch("aaa") matches, while
644     pcrecpp::RE("a*?").FullMatch("aaa") does not, and
645 ph10 182 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
646 ph10 185
647 ph10 184 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
648 ph10 185 it matched the wrong number of bytes.
649 ph10 172
650 ph10 179
651 ph10 155 Version 7.1 24-Apr-07
652 ph10 98 ---------------------
653    
654 ph10 111 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
655 ph10 122 that is more "standard", making use of automake and other Autotools. There
656 ph10 99 is some re-arrangement of the files and adjustment of comments consequent
657     on this.
658 ph10 111
659     2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
660     for recursive directory scanning broke on some systems because the files
661     are not scanned in any specific order and on different systems the order
662     was different. A call to "sort" has been inserted into RunGrepTest for the
663     approprate test as a short-term fix. In the longer term there may be an
664 ph10 100 alternative.
665 ph10 111
666 ph10 100 3. I had an email from Eric Raymond about problems translating some of PCRE's
667 ph10 111 man pages to HTML (despite the fact that I distribute HTML pages, some
668     people do their own conversions for various reasons). The problems
669     concerned the use of low-level troff macros .br and .in. I have therefore
670     removed all such uses from the man pages (some were redundant, some could
671 ph10 113 be replaced by .nf/.fi pairs). The 132html script that I use to generate
672     HTML has been updated to handle .nf/.fi and to complain if it encounters
673     .br or .in.
674 ph10 111
675 ph10 100 4. Updated comments in configure.ac that get placed in config.h.in and also
676 ph10 123 arranged for config.h to be included in the distribution, with the name
677 ph10 111 config.h.generic, for the benefit of those who have to compile without
678     Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
679    
680     5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
681     Weber: (1) pcre_internal.h was missing some function renames; (2) updated
682 ph10 127 makevp.bat for the current PCRE, using the additional files
683 ph10 135 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
684 ph10 111
685     6. A Windows user reported a minor discrepancy with test 2, which turned out
686     to be caused by a trailing space on an input line that had got lost in his
687 ph10 102 copy. The trailing space was an accident, so I've just removed it.
688 ph10 111
689 ph10 104 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
690 ph10 111 that is needed.
691    
692 ph10 105 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
693     as "const" (a) because they are and (b) because it helps the PHP
694     maintainers who have recently made a script to detect big data structures
695 ph10 111 in the php code that should be moved to the .rodata section. I remembered
696     to update Builducptable as well, so it won't revert if ucptable.h is ever
697 ph10 105 re-created.
698 ph10 111
699     9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
700 ph10 107 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
701 ph10 111 order to be able to cut out the UTF-8 tables in the latter when UTF-8
702     support is not required. This saves 1.5-2K of code, which is important in
703 ph10 107 some applications.
704 ph10 111
705 ph10 110 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
706 ph10 111 so as not to refer to the tables, even though these functions will never be
707     called when UTF-8 support is disabled. Otherwise there are problems with a
708     shared library.
709    
710 ph10 118 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
711    
712     (a) It was defining its arguments as char * instead of void *.
713    
714 ph10 123 (b) It was assuming that all moves were upwards in memory; this was true
715     a long time ago when I wrote it, but is no longer the case.
716    
717 ph10 118 The emulated memove() is provided for those environments that have neither
718 ph10 123 memmove() nor bcopy(). I didn't think anyone used it these days, but that
719 ph10 118 is clearly not the case, as these two bugs were recently reported.
720 ph10 123
721 ph10 111 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
722 ph10 123 and Detrail to create the HTML documentation, the .txt form of the man
723     pages, and it removes trailing spaces from listed files. It also creates
724     pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
725     case, it wraps all the #defines with #ifndefs. This script should be run
726 ph10 111 before "make dist".
727 ph10 123
728 ph10 115 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
729     with Unicode property support.
730 ph10 123
731     (a) For a maximizing quantifier, if the two different cases of the
732     character were of different lengths in their UTF-8 codings (there are
733     some cases like this - I found 11), and the matching function had to
734 ph10 115 back up over a mixture of the two cases, it incorrectly assumed they
735     were both the same length.
736 ph10 123
737     (b) When PCRE was configured to use the heap rather than the stack for
738     recursion during matching, it was not correctly preserving the data for
739     the other case of a UTF-8 character when checking ahead for a match
740     while processing a minimizing repeat. If the check also involved
741 ph10 115 matching a wide character, but failed, corruption could cause an
742     erroneous result when trying to check for a repeat of the original
743     character.
744 ph10 123
745 ph10 116 13. Some tidying changes to the testing mechanism:
746 ph10 98
747 ph10 116 (a) The RunTest script now detects the internal link size and whether there
748     is UTF-8 and UCP support by running ./pcretest -C instead of relying on
749 ph10 123 values substituted by "configure". (The RunGrepTest script already did
750     this for UTF-8.) The configure.ac script no longer substitutes the
751     relevant variables.
752    
753 ph10 116 (b) The debugging options /B and /D in pcretest show the compiled bytecode
754     with length and offset values. This means that the output is different
755     for different internal link sizes. Test 2 is skipped for link sizes
756     other than 2 because of this, bypassing the problem. Unfortunately,
757     there was also a test in test 3 (the locale tests) that used /B and
758 ph10 123 failed for link sizes other than 2. Rather than cut the whole test out,
759     I have added a new /Z option to pcretest that replaces the length and
760     offset values with spaces. This is now used to make test 3 independent
761 ph10 122 of link size. (Test 2 will be tidied up later.)
762 ph10 123
763     14. If erroroffset was passed as NULL to pcre_compile, it provoked a
764 ph10 122 segmentation fault instead of returning the appropriate error message.
765 ph10 134
766 ph10 131 15. In multiline mode when the newline sequence was set to "any", the pattern
767 ph10 134 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
768     This doesn't seem right; it now treats the CRLF combination as the line
769     ending, and so does not match in that case. It's only a pattern such as ^$
770     that would hit this one: something like ^ABC$ would have failed after \r
771     and then tried again after \r\n.
772    
773 ph10 131 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
774 ph10 134 in an attempt to make files that differ only in their line terminators
775     compare equal. This works on Linux.
776 ph10 142
777 ph10 141 17. Under certain error circumstances pcregrep might try to free random memory
778     as it exited. This is now fixed, thanks to valgrind.
779 ph10 142
780 ph10 141 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
781 ph10 142 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
782     was because its rules for how to advance for /g after matching an empty
783 ph10 143 string at the end of a line did not allow for this case. They now check for
784     it specially.
785 ph10 150
786     20. pcretest is supposed to handle patterns and data of any length, by
787     extending its buffers when necessary. It was getting this wrong when the
788 ph10 147 buffer for a data line had to be extended.
789 ph10 150
790 ph10 149 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
791 ph10 150 CRLF as a newline sequence.
792 ph10 152
793 ph10 151 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
794 ph10 152 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
795     I have nevertheless tidied it up.
796 ph10 111
797 ph10 152 23. Added some casts to kill warnings from HP-UX ia64 compiler.
798 ph10 123
799 ph10 153 24. Added a man page for pcre-config.
800 ph10 152
801 ph10 153
802 nigel 93 Version 7.0 19-Dec-06
803     ---------------------
804    
805     1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
806     moving to gcc 4.1.1.
807    
808     2. The -S option for pcretest uses setrlimit(); I had omitted to #include
809     sys/time.h, which is documented as needed for this function. It doesn't
810     seem to matter on Linux, but it showed up on some releases of OS X.
811    
812     3. It seems that there are systems where bytes whose values are greater than
813     127 match isprint() in the "C" locale. The "C" locale should be the
814     default when a C program starts up. In most systems, only ASCII printing
815     characters match isprint(). This difference caused the output from pcretest
816     to vary, making some of the tests fail. I have changed pcretest so that:
817    
818     (a) When it is outputting text in the compiled version of a pattern, bytes
819     other than 32-126 are always shown as hex escapes.
820    
821     (b) When it is outputting text that is a matched part of a subject string,
822     it does the same, unless a different locale has been set for the match
823     (using the /L modifier). In this case, it uses isprint() to decide.
824    
825     4. Fixed a major bug that caused incorrect computation of the amount of memory
826     required for a compiled pattern when options that changed within the
827     pattern affected the logic of the preliminary scan that determines the
828     length. The relevant options are -x, and -i in UTF-8 mode. The result was
829     that the computed length was too small. The symptoms of this bug were
830     either the PCRE error "internal error: code overflow" from pcre_compile(),
831     or a glibc crash with a message such as "pcretest: free(): invalid next
832     size (fast)". Examples of patterns that provoked this bug (shown in
833     pcretest format) are:
834    
835     /(?-x: )/x
836     /(?x)(?-x: \s*#\s*)/
837     /((?i)[\x{c0}])/8
838     /(?i:[\x{c0}])/8
839    
840     HOWEVER: Change 17 below makes this fix obsolete as the memory computation
841     is now done differently.
842    
843     5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
844     wrapper classes; (b) implement a new function in the C++ scanner that is
845     more efficient than the old way of doing things because it avoids levels of
846     recursion in the regex matching; (c) add a paragraph to the documentation
847     for the FullMatch() function.
848    
849     6. The escape sequence \n was being treated as whatever was defined as
850     "newline". Not only was this contrary to the documentation, which states
851     that \n is character 10 (hex 0A), but it also went horribly wrong when
852     "newline" was defined as CRLF. This has been fixed.
853    
854     7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
855     was being set to -1 for the "end of line" case (supposedly a value that no
856     character can have). Though this value is never used (the check for end of
857     line is "zero bytes in current character"), it caused compiler complaints.
858     I've changed it to 0xffffffff.
859    
860     8. In pcre_version.c, the version string was being built by a sequence of
861     C macros that, in the event of PCRE_PRERELEASE being defined as an empty
862     string (as it is for production releases) called a macro with an empty
863     argument. The C standard says the result of this is undefined. The gcc
864     compiler treats it as an empty string (which was what was wanted) but it is
865     reported that Visual C gives an error. The source has been hacked around to
866     avoid this problem.
867    
868     9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
869     builds of pcretest, and changed the call to _setmode() to use _O_BINARY
870     instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
871     of them did).
872    
873     10. Originally, pcretest opened its input and output without "b"; then I was
874     told that "b" was needed in some environments, so it was added for release
875     5.0 to both the input and output. (It makes no difference on Unix-like
876     systems.) Later I was told that it is wrong for the input on Windows. I've
877     now abstracted the modes into two macros, to make it easier to fiddle with
878     them, and removed "b" from the input mode under Windows.
879    
880     11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
881    
882     12. Added -help and --help to pcretest as an official way of being reminded
883     of the options.
884    
885     13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
886     and pcrecpp.cc because they annoy compilers at high warning levels.
887    
888     14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
889    
890     15. Fixed an occurrence of == in configure.ac that should have been = (shell
891     scripts are not C programs :-) and which was not noticed because it works
892     on Linux.
893    
894     16. pcretest is supposed to handle any length of pattern and data line (as one
895     line or as a continued sequence of lines) by extending its input buffer if
896     necessary. This feature was broken for very long pattern lines, leading to
897     a string of junk being passed to pcre_compile() if the pattern was longer
898     than about 50K.
899    
900     17. I have done a major re-factoring of the way pcre_compile() computes the
901     amount of memory needed for a compiled pattern. Previously, there was code
902     that made a preliminary scan of the pattern in order to do this. That was
903     OK when PCRE was new, but as the facilities have expanded, it has become
904     harder and harder to keep it in step with the real compile phase, and there
905     have been a number of bugs (see for example, 4 above). I have now found a
906     cunning way of running the real compile function in a "fake" mode that
907     enables it to compute how much memory it would need, while actually only
908     ever using a few hundred bytes of working memory and without too many
909     tests of the mode. This should make future maintenance and development
910     easier. A side effect of this work is that the limit of 200 on the nesting
911     depth of parentheses has been removed (though this was never a serious
912     limitation, I suspect). However, there is a downside: pcre_compile() now
913     runs more slowly than before (30% or more, depending on the pattern). I
914     hope this isn't a big issue. There is no effect on runtime performance.
915    
916     18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
917     newline (only possible for the last line of a file) and it was a
918     pattern that set a locale (followed by /Lsomething), pcretest crashed.
919    
920     19. Added additional timing features to pcretest. (1) The -tm option now times
921     matching only, not compiling. (2) Both -t and -tm can be followed, as a
922     separate command line item, by a number that specifies the number of
923     repeats to use when timing. The default is 50000; this gives better
924     precision, but takes uncomfortably long for very large patterns.
925    
926     20. Extended pcre_study() to be more clever in cases where a branch of a
927     subpattern has no definite first character. For example, (a*|b*)[cd] would
928     previously give no result from pcre_study(). Now it recognizes that the
929     first character must be a, b, c, or d.
930    
931     21. There was an incorrect error "recursive call could loop indefinitely" if
932     a subpattern (or the entire pattern) that was being tested for matching an
933     empty string contained only one non-empty item after a nested subpattern.
934     For example, the pattern (?>\x{100}*)\d(?R) provoked this error
935     incorrectly, because the \d was being skipped in the check.
936    
937     22. The pcretest program now has a new pattern option /B and a command line
938     option -b, which is equivalent to adding /B to every pattern. This causes
939     it to show the compiled bytecode, without the additional information that
940     -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
941     is the same as /B/I).
942    
943     23. A new optimization is now able automatically to treat some sequences such
944     as a*b as a*+b. More specifically, if something simple (such as a character
945     or a simple class like \d) has an unlimited quantifier, and is followed by
946     something that cannot possibly match the quantified thing, the quantifier
947     is automatically "possessified".
948    
949     24. A recursive reference to a subpattern whose number was greater than 39
950     went wrong under certain circumstances in UTF-8 mode. This bug could also
951     have affected the operation of pcre_study().
952    
953     25. Realized that a little bit of performance could be had by replacing
954     (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
955    
956     26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
957    
958     27. Possessive quantifiers such as a++ were previously implemented by turning
959     them into atomic groups such as ($>a+). Now they have their own opcodes,
960     which improves performance. This includes the automatically created ones
961     from 23 above.
962    
963     28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
964     lookahead was broken if it was not anchored. PCRE was mistakenly expecting
965     the first matched character to be a colon. This applied both to named and
966     numbered groups.
967    
968     29. The ucpinternal.h header file was missing its idempotency #ifdef.
969    
970     30. I was sent a "project" file called libpcre.a.dev which I understand makes
971     building PCRE on Windows easier, so I have included it in the distribution.
972    
973     31. There is now a check in pcretest against a ridiculously large number being
974     returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
975     loop, the loop is abandoned.
976    
977     32. Forward references to subpatterns in conditions such as (?(2)...) where
978     subpattern 2 is defined later cause pcre_compile() to search forwards in
979     the pattern for the relevant set of parentheses. This search went wrong
980     when there were unescaped parentheses in a character class, parentheses
981     escaped with \Q...\E, or parentheses in a #-comment in /x mode.
982    
983     33. "Subroutine" calls and backreferences were previously restricted to
984     referencing subpatterns earlier in the regex. This restriction has now
985     been removed.
986    
987     34. Added a number of extra features that are going to be in Perl 5.10. On the
988     whole, these are just syntactic alternatives for features that PCRE had
989     previously implemented using the Python syntax or my own invention. The
990     other formats are all retained for compatibility.
991    
992     (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
993     as (?P<name>...). The new forms, as well as being in Perl 5.10, are
994     also .NET compatible.
995    
996     (b) A recursion or subroutine call to a named group can now be defined as
997     (?&name) as well as (?P>name).
998    
999     (c) A backreference to a named group can now be defined as \k<name> or
1000     \k'name' as well as (?P=name). The new forms, as well as being in Perl
1001     5.10, are also .NET compatible.
1002    
1003     (d) A conditional reference to a named group can now use the syntax
1004     (?(<name>) or (?('name') as well as (?(name).
1005    
1006     (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
1007     groups (named and numbered) that are never evaluated inline, but can be
1008     called as "subroutines" from elsewhere. In effect, the DEFINE condition
1009     is always false. There may be only one alternative in such a group.
1010    
1011     (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
1012     as the simple (?(R). The condition is true only if the most recent
1013     recursion is that of the given number or name. It does not search out
1014     through the entire recursion stack.
1015    
1016     (g) The escape \gN or \g{N} has been added, where N is a positive or
1017     negative number, specifying an absolute or relative reference.
1018    
1019     35. Tidied to get rid of some further signed/unsigned compiler warnings and
1020     some "unreachable code" warnings.
1021    
1022     36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
1023     things, this adds five new scripts.
1024    
1025     37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
1026     There were also incompatibilities regarding the handling of \Q..\E inside
1027     character classes, for example with patterns like [\Qa\E-\Qz\E] where the
1028     hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
1029    
1030     38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
1031     matches an empty string, and forcibly breaks the loop. There were bugs in
1032     this code in non-simple cases. For a pattern such as ^(a()*)* matched
1033     against aaaa the result was just "a" rather than "aaaa", for example. Two
1034     separate and independent bugs (that affected different cases) have been
1035     fixed.
1036    
1037     39. Refactored the code to abolish the use of different opcodes for small
1038     capturing bracket numbers. This is a tidy that I avoided doing when I
1039     removed the limit on the number of capturing brackets for 3.5 back in 2001.
1040     The new approach is not only tidier, it makes it possible to reduce the
1041     memory needed to fix the previous bug (38).
1042    
1043     40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
1044     sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
1045     processing dot, circumflex, or dollar metacharacters, or #-comments in /x
1046     mode.
1047    
1048     41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
1049     report.
1050    
1051     42. Applied patch, originally from Ari Pollak, modified by Google, to allow
1052     copy construction and assignment in the C++ wrapper.
1053    
1054     43. Updated pcregrep to support "--newline=any". In the process, I fixed a
1055     couple of bugs that could have given wrong results in the "--newline=crlf"
1056     case.
1057    
1058     44. Added a number of casts and did some reorganization of signed/unsigned int
1059     variables following suggestions from Dair Grant. Also renamed the variable
1060     "this" as "item" because it is a C++ keyword.
1061    
1062     45. Arranged for dftables to add
1063    
1064     #include "pcre_internal.h"
1065    
1066     to pcre_chartables.c because without it, gcc 4.x may remove the array
1067     definition from the final binary if PCRE is built into a static library and
1068     dead code stripping is activated.
1069    
1070     46. For an unanchored pattern, if a match attempt fails at the start of a
1071     newline sequence, and the newline setting is CRLF or ANY, and the next two
1072     characters are CRLF, advance by two characters instead of one.
1073    
1074    
1075 nigel 91 Version 6.7 04-Jul-06
1076     ---------------------
1077    
1078     1. In order to handle tests when input lines are enormously long, pcretest has
1079     been re-factored so that it automatically extends its buffers when
1080     necessary. The code is crude, but this _is_ just a test program. The
1081     default size has been increased from 32K to 50K.
1082    
1083     2. The code in pcre_study() was using the value of the re argument before
1084     testing it for NULL. (Of course, in any sensible call of the function, it
1085     won't be NULL.)
1086    
1087     3. The memmove() emulation function in pcre_internal.h, which is used on
1088     systems that lack both memmove() and bcopy() - that is, hardly ever -
1089     was missing a "static" storage class specifier.
1090    
1091     4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
1092     containing an extended class (one that cannot be represented by a bitmap
1093     because it contains high-valued characters or Unicode property items, e.g.
1094     [\pZ]). Almost always one would set UTF-8 mode when processing such a
1095     pattern, but PCRE should not loop if you do not (it no longer does).
1096     [Detail: two cases were found: (a) a repeated subpattern containing an
1097     extended class; (b) a recursive reference to a subpattern that followed a
1098     previous extended class. It wasn't skipping over the extended class
1099     correctly when UTF-8 mode was not set.]
1100    
1101     5. A negated single-character class was not being recognized as fixed-length
1102     in lookbehind assertions such as (?<=[^f]), leading to an incorrect
1103     compile error "lookbehind assertion is not fixed length".
1104    
1105     6. The RunPerlTest auxiliary script was showing an unexpected difference
1106     between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
1107     write a Perl script that can interpret lines of an input file either as
1108     byte characters or as UTF-8, which is what "perltest" was being required to
1109     do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
1110     can't do is switch easily at run time between having the "use utf8;" pragma
1111     or not. In the end, I fudged it by using the RunPerlTest script to insert
1112     "use utf8;" explicitly for the UTF-8 tests.
1113    
1114     7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
1115     the end of the subject string, contrary to the documentation and to what
1116     Perl does. This was true of both matching functions. Now it matches only at
1117     the start of the subject and immediately after *internal* newlines.
1118    
1119     8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
1120     a pointer to an int instead of a pointer to an unsigned long int. This
1121     caused problems on 64-bit systems.
1122    
1123     9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
1124     instance of the 'standard' template library not being so standard".
1125    
1126     10. There was no check on the number of named subpatterns nor the maximum
1127     length of a subpattern name. The product of these values is used to compute
1128     the size of the memory block for a compiled pattern. By supplying a very
1129     long subpattern name and a large number of named subpatterns, the size
1130     computation could be caused to overflow. This is now prevented by limiting
1131     the length of names to 32 characters, and the number of named subpatterns
1132     to 10,000.
1133    
1134     11. Subpatterns that are repeated with specific counts have to be replicated in
1135     the compiled pattern. The size of memory for this was computed from the
1136     length of the subpattern and the repeat count. The latter is limited to
1137     65535, but there was no limit on the former, meaning that integer overflow
1138     could in principle occur. The compiled length of a repeated subpattern is
1139     now limited to 30,000 bytes in order to prevent this.
1140    
1141     12. Added the optional facility to have named substrings with the same name.
1142    
1143     13. Added the ability to use a named substring as a condition, using the
1144     Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
1145     are numbers (not recommended). Forward references are permitted.
1146    
1147     14. Added forward references in named backreferences (if you see what I mean).
1148    
1149     15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
1150     pattern could run off the end of the subject. For example, the pattern
1151     "(?s)(.{1,5})"8 did this with the subject "ab".
1152    
1153     16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
1154     PCRE_CASELESS was set when matching characters that were quantified with ?
1155     or *.
1156    
1157     17. A character class other than a single negated character that had a minimum
1158     but no maximum quantifier - for example [ab]{6,} - was not handled
1159     correctly by pce_dfa_exec(). It would match only one character.
1160    
1161     18. A valid (though odd) pattern that looked like a POSIX character
1162     class but used an invalid character after [ (for example [[,abc,]]) caused
1163     pcre_compile() to give the error "Failed: internal error: code overflow" or
1164     in some cases to crash with a glibc free() error. This could even happen if
1165     the pattern terminated after [[ but there just happened to be a sequence of
1166     letters, a binary zero, and a closing ] in the memory that followed.
1167    
1168     19. Perl's treatment of octal escapes in the range \400 to \777 has changed
1169     over the years. Originally (before any Unicode support), just the bottom 8
1170     bits were taken. Thus, for example, \500 really meant \100. Nowadays the
1171     output from "man perlunicode" includes this:
1172    
1173     The regular expression compiler produces polymorphic opcodes. That
1174     is, the pattern adapts to the data and automatically switches to
1175     the Unicode character scheme when presented with Unicode data--or
1176     instead uses a traditional byte scheme when presented with byte
1177     data.
1178    
1179     Sadly, a wide octal escape does not cause a switch, and in a string with
1180     no other multibyte characters, these octal escapes are treated as before.
1181     Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
1182     /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
1183     Unicode string.
1184    
1185     I have not perpetrated such confusion in PCRE. Up till now, it took just
1186     the bottom 8 bits, as in old Perl. I have now made octal escapes with
1187     values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
1188     translate to the appropriate multibyte character.
1189    
1190     29. Applied some refactoring to reduce the number of warnings from Microsoft
1191     and Borland compilers. This has included removing the fudge introduced
1192     seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
1193     a warning about an unused variable.
1194    
1195     21. PCRE has not included VT (character 0x0b) in the set of whitespace
1196     characters since release 4.0, because Perl (from release 5.004) does not.
1197     [Or at least, is documented not to: some releases seem to be in conflict
1198     with the documentation.] However, when a pattern was studied with
1199     pcre_study() and all its branches started with \s, PCRE still included VT
1200     as a possible starting character. Of course, this did no harm; it just
1201     caused an unnecessary match attempt.
1202    
1203     22. Removed a now-redundant internal flag bit that recorded the fact that case
1204     dependency changed within the pattern. This was once needed for "required
1205     byte" processing, but is no longer used. This recovers a now-scarce options
1206     bit. Also moved the least significant internal flag bit to the most-
1207     significant bit of the word, which was not previously used (hangover from
1208     the days when it was an int rather than a uint) to free up another bit for
1209     the future.
1210    
1211     23. Added support for CRLF line endings as well as CR and LF. As well as the
1212     default being selectable at build time, it can now be changed at runtime
1213     via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
1214     specify that it is scanning data with non-default line endings.
1215    
1216     24. Changed the definition of CXXLINK to make it agree with the definition of
1217     LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
1218    
1219     25. Applied Ian Taylor's patches to avoid using another stack frame for tail
1220     recursions. This makes a big different to stack usage for some patterns.
1221    
1222     26. If a subpattern containing a named recursion or subroutine reference such
1223     as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
1224     the space required for the compiled pattern went wrong and gave too small a
1225     value. Depending on the environment, this could lead to "Failed: internal
1226     error: code overflow at offset 49" or "glibc detected double free or
1227     corruption" errors.
1228    
1229     27. Applied patches from Google (a) to support the new newline modes and (b) to
1230     advance over multibyte UTF-8 characters in GlobalReplace.
1231    
1232     28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
1233     difference for some implementation of PCRE in some Windows version.
1234    
1235     29. Added some extra testing facilities to pcretest:
1236    
1237     \q<number> in a data line sets the "match limit" value
1238     \Q<number> in a data line sets the "match recursion limt" value
1239     -S <number> sets the stack size, where <number> is in megabytes
1240    
1241     The -S option isn't available for Windows.
1242    
1243    
1244 nigel 89 Version 6.6 06-Feb-06
1245     ---------------------
1246    
1247     1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
1248     in pcreposix.h. I have copied the definition from pcre.h.
1249    
1250     2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
1251     because pcre.h is no longer a built file.
1252    
1253     3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
1254     not normally included in the compiled code.
1255    
1256    
1257 nigel 87 Version 6.5 01-Feb-06
1258     ---------------------
1259    
1260     1. When using the partial match feature with pcre_dfa_exec(), it was not
1261     anchoring the second and subsequent partial matches at the new starting
1262     point. This could lead to incorrect results. For example, with the pattern
1263     /1234/, partially matching against "123" and then "a4" gave a match.
1264    
1265     2. Changes to pcregrep:
1266    
1267     (a) All non-match returns from pcre_exec() were being treated as failures
1268     to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
1269     error message is output. Some extra information is given for the
1270     PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
1271     probably the only errors that are likely to be caused by users (by
1272     specifying a regex that has nested indefinite repeats, for instance).
1273     If there are more than 20 of these errors, pcregrep is abandoned.
1274    
1275     (b) A binary zero was treated as data while matching, but terminated the
1276     output line if it was written out. This has been fixed: binary zeroes
1277     are now no different to any other data bytes.
1278    
1279     (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
1280     used to set a locale for matching. The --locale=xxxx long option has
1281     been added (no short equivalent) to specify a locale explicitly on the
1282     pcregrep command, overriding the environment variables.
1283    
1284     (d) When -B was used with -n, some line numbers in the output were one less
1285     than they should have been.
1286    
1287     (e) Added the -o (--only-matching) option.
1288    
1289     (f) If -A or -C was used with -c (count only), some lines of context were
1290     accidentally printed for the final match.
1291    
1292     (g) Added the -H (--with-filename) option.
1293    
1294     (h) The combination of options -rh failed to suppress file names for files
1295     that were found from directory arguments.
1296    
1297     (i) Added the -D (--devices) and -d (--directories) options.
1298    
1299     (j) Added the -F (--fixed-strings) option.
1300    
1301     (k) Allow "-" to be used as a file name for -f as well as for a data file.
1302    
1303     (l) Added the --colo(u)r option.
1304    
1305     (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
1306     is not present by default.
1307    
1308     3. A nasty bug was discovered in the handling of recursive patterns, that is,
1309     items such as (?R) or (?1), when the recursion could match a number of
1310     alternatives. If it matched one of the alternatives, but subsequently,
1311     outside the recursion, there was a failure, the code tried to back up into
1312     the recursion. However, because of the way PCRE is implemented, this is not
1313     possible, and the result was an incorrect result from the match.
1314    
1315     In order to prevent this happening, the specification of recursion has
1316     been changed so that all such subpatterns are automatically treated as
1317     atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
1318    
1319     4. I had overlooked the fact that, in some locales, there are characters for
1320     which isalpha() is true but neither isupper() nor islower() are true. In
1321     the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
1322     and ordfeminine) are like this. This affected the treatment of \w and \W
1323     when they appeared in character classes, but not when they appeared outside
1324     a character class. The bit map for "word" characters is now created
1325     separately from the results of isalnum() instead of just taking it from the
1326     upper, lower, and digit maps. (Plus the underscore character, of course.)
1327    
1328     5. The above bug also affected the handling of POSIX character classes such as
1329     [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
1330     permanent tables. Instead, the bit maps for such a class were previously
1331     created as the appropriate unions of the upper, lower, and digit bitmaps.
1332     Now they are created by subtraction from the [[:word:]] class, which has
1333     its own bitmap.
1334    
1335     6. The [[:blank:]] character class matches horizontal, but not vertical space.
1336     It is created by subtracting the vertical space characters (\x09, \x0a,
1337     \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
1338     subtraction was done in the overall bitmap for a character class, meaning
1339     that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
1340     be recognized. This bug has been fixed.
1341    
1342     7. Patches from the folks at Google:
1343    
1344     (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
1345     real life, but is still worth protecting against".
1346    
1347     (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
1348     regular expressions".
1349    
1350     (c) pcre_scanner.cc: avoid use of std::count() because not all systems
1351     have it.
1352    
1353     (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
1354     "configure" and the latter not, in order to fix a problem somebody had
1355     with compiling the Arg class on HP-UX.
1356    
1357     (e) Improve the error-handling of the C++ wrapper a little bit.
1358    
1359     (f) New tests for checking recursion limiting.
1360    
1361     8. The pcre_memmove() function, which is used only if the environment does not
1362     have a standard memmove() function (and is therefore rarely compiled),
1363     contained two bugs: (a) use of int instead of size_t, and (b) it was not
1364     returning a result (though PCRE never actually uses the result).
1365    
1366     9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
1367     large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
1368     returned instead of calling malloc() with an overflowing number that would
1369     most likely cause subsequent chaos.
1370    
1371     10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
1372    
1373     11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
1374     with this option is matched, the nmatch and pmatch options of regexec() are
1375     ignored.
1376    
1377     12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
1378     provided in case anyone wants to the the POSIX interface with UTF-8
1379     strings.
1380    
1381     13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
1382     C++ linking (needed for some HP-UX environments).
1383    
1384     14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
1385     (unused parameter) and in the pcre_printint() function (omitted "default"
1386     switch label when the default is to do nothing).
1387    
1388     15. Added some code to make it possible, when PCRE is compiled as a C++
1389     library, to replace subject pointers for pcre_exec() with a smart pointer
1390     class, thus making it possible to process discontinuous strings.
1391    
1392     16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
1393     much the same function. They were added by different people who were trying
1394     to make PCRE easy to compile on non-Unix systems. It has been suggested
1395     that PCRE_EXPORT be abolished now that there is more automatic apparatus
1396     for compiling on Windows systems. I have therefore replaced it with
1397     PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
1398     defaults to "extern" for C or "extern C" for C++, which works fine on
1399     Unix-like systems. It is now possible to override the value of PCRE_DATA_
1400     SCOPE with something explicit in config.h. In addition:
1401    
1402     (a) pcreposix.h still had just "extern" instead of either of these macros;
1403     I have replaced it with PCRE_DATA_SCOPE.
1404    
1405     (b) Functions such as _pcre_xclass(), which are internal to the library,
1406     but external in the C sense, all had PCRE_EXPORT in their definitions.
1407     This is apparently wrong for the Windows case, so I have removed it.
1408     (It makes no difference on Unix-like systems.)
1409    
1410     17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
1411     of recursive calls to match(). This is different to MATCH_LIMIT because
1412     that limits the total number of calls to match(), not all of which increase
1413     the depth of recursion. Limiting the recursion depth limits the amount of
1414     stack (or heap if NO_RECURSE is set) that is used. The default can be set
1415     when PCRE is compiled, and changed at run time. A patch from Google adds
1416     this functionality to the C++ interface.
1417    
1418     18. Changes to the handling of Unicode character properties:
1419    
1420     (a) Updated the table to Unicode 4.1.0.
1421    
1422     (b) Recognize characters that are not in the table as "Cn" (undefined).
1423    
1424     (c) I revised the way the table is implemented to a much improved format
1425     which includes recognition of ranges. It now supports the ranges that
1426     are defined in UnicodeData.txt, and it also amalgamates other
1427     characters into ranges. This has reduced the number of entries in the
1428     table from around 16,000 to around 3,000, thus reducing its size
1429     considerably. I realized I did not need to use a tree structure after
1430     all - a binary chop search is just as efficient. Having reduced the
1431     number of entries, I extended their size from 6 bytes to 8 bytes to
1432     allow for more data.
1433    
1434     (d) Added support for Unicode script names via properties such as \p{Han}.
1435    
1436     19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
1437     matching that character.
1438    
1439     20. When matching a repeated Unicode property with a minimum greater than zero,
1440     (for example \pL{2,}), PCRE could look past the end of the subject if it
1441     reached it while seeking the minimum number of characters. This could
1442     happen only if some of the characters were more than one byte long, because
1443     there is a check for at least the minimum number of bytes.
1444    
1445     21. Refactored the implementation of \p and \P so as to be more general, to
1446     allow for more different types of property in future. This has changed the
1447     compiled form incompatibly. Anybody with saved compiled patterns that use
1448     \p or \P will have to recompile them.
1449    
1450     22. Added "Any" and "L&" to the supported property types.
1451    
1452     23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
1453     but give a compile time error if the value is greater than 0xff.
1454    
1455     24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
1456     accidentally not being installed or uninstalled.
1457    
1458     25. The pcre.h file was built from pcre.h.in, but the only changes that were
1459     made were to insert the current release number. This seemed silly, because
1460     it made things harder for people building PCRE on systems that don't run
1461     "configure". I have turned pcre.h into a distributed file, no longer built
1462     by "configure", with the version identification directly included. There is
1463     no longer a pcre.h.in file.
1464    
1465     However, this change necessitated a change to the pcre-config script as
1466     well. It is built from pcre-config.in, and one of the substitutions was the
1467     release number. I have updated configure.ac so that ./configure now finds
1468     the release number by grepping pcre.h.
1469    
1470     26. Added the ability to run the tests under valgrind.
1471    
1472    
1473 nigel 85 Version 6.4 05-Sep-05
1474     ---------------------
1475    
1476     1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
1477     "--" to be printed when multiple files were scanned, even when none of the
1478     -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
1479     consider it to be a bug, and have restored the previous behaviour.
1480    
1481     2. A couple of code tidies to get rid of compiler warnings.
1482    
1483     3. The pcretest program used to cheat by referring to symbols in the library
1484     whose names begin with _pcre_. These are internal symbols that are not
1485     really supposed to be visible externally, and in some environments it is
1486     possible to suppress them. The cheating is now confined to including
1487     certain files from the library's source, which is a bit cleaner.
1488    
1489     4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
1490     file's purpose clearer.
1491    
1492     5. Reorganized pcre_ucp_findchar().
1493    
1494    
1495 nigel 83 Version 6.3 15-Aug-05
1496     ---------------------
1497    
1498     1. The file libpcre.pc.in did not have general read permission in the tarball.
1499    
1500     2. There were some problems when building without C++ support:
1501    
1502     (a) If C++ support was not built, "make install" and "make test" still
1503     tried to test it.
1504    
1505     (b) There were problems when the value of CXX was explicitly set. Some
1506     changes have been made to try to fix these, and ...
1507    
1508     (c) --disable-cpp can now be used to explicitly disable C++ support.
1509    
1510     (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
1511     backslash in a target when C++ was disabled. This confuses some
1512     versions of "make", apparently. Using an intermediate variable solves
1513     this. (Same for CPP_LOBJ.)
1514    
1515     3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
1516     (non-Windows) now includes $(CFLAGS) because these flags are sometimes
1517     necessary on certain architectures.
1518    
1519     4. Added a setting of -export-symbols-regex to the link command to remove
1520     those symbols that are exported in the C sense, but actually are local
1521     within the library, and not documented. Their names all begin with
1522     "_pcre_". This is not a perfect job, because (a) we have to except some
1523     symbols that pcretest ("illegally") uses, and (b) the facility isn't always
1524     available (and never for static libraries). I have made a note to try to
1525     find a way round (a) in the future.
1526    
1527    
1528 nigel 81 Version 6.2 01-Aug-05
1529     ---------------------
1530    
1531     1. There was no test for integer overflow of quantifier values. A construction
1532     such as {1111111111111111} would give undefined results. What is worse, if
1533     a minimum quantifier for a parenthesized subpattern overflowed and became
1534     negative, the calculation of the memory size went wrong. This could have
1535     led to memory overwriting.
1536    
1537     2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
1538    
1539     3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
1540     operating environments where this matters.
1541    
1542     4. Applied Giuseppe Maxia's patch to add additional features for controlling
1543     PCRE options from within the C++ wrapper.
1544    
1545     5. Named capturing subpatterns were not being correctly counted when a pattern
1546     was compiled. This caused two problems: (a) If there were more than 100
1547     such subpatterns, the calculation of the memory needed for the whole
1548     compiled pattern went wrong, leading to an overflow error. (b) Numerical
1549     back references of the form \12, where the number was greater than 9, were
1550     not recognized as back references, even though there were sufficient
1551     previous subpatterns.
1552    
1553     6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
1554     versions of gcc, e.g. 2.95.4.
1555    
1556    
1557 nigel 79 Version 6.1 21-Jun-05
1558     ---------------------
1559    
1560     1. There was one reference to the variable "posix" in pcretest.c that was not
1561     surrounded by "#if !defined NOPOSIX".
1562    
1563     2. Make it possible to compile pcretest without DFA support, UTF8 support, or
1564     the cross-check on the old pcre_info() function, for the benefit of the
1565     cut-down version of PCRE that is currently imported into Exim.
1566    
1567     3. A (silly) pattern starting with (?i)(?-i) caused an internal space
1568     allocation error. I've done the easy fix, which wastes 2 bytes for sensible
1569     patterns that start (?i) but I don't think that matters. The use of (?i) is
1570     just an example; this all applies to the other options as well.
1571    
1572     4. Since libtool seems to echo the compile commands it is issuing, the output
1573     from "make" can be reduced a bit by putting "@" in front of each libtool
1574     compile command.
1575    
1576     5. Patch from the folks at Google for configure.in to be a bit more thorough
1577     in checking for a suitable C++ installation before trying to compile the
1578     C++ stuff. This should fix a reported problem when a compiler was present,
1579     but no suitable headers.
1580    
1581     6. The man pages all had just "PCRE" as their title. I have changed them to
1582     be the relevant file name. I have also arranged that these names are
1583     retained in the file doc/pcre.txt, which is a concatenation in text format
1584     of all the man pages except the little individual ones for each function.
1585    
1586     7. The NON-UNIX-USE file had not been updated for the different set of source
1587     files that come with release 6. I also added a few comments about the C++
1588     wrapper.
1589    
1590    
1591 nigel 77 Version 6.0 07-Jun-05
1592     ---------------------
1593    
1594     1. Some minor internal re-organization to help with my DFA experiments.
1595    
1596     2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
1597     didn't matter for the library itself when fully configured, but did matter
1598     when compiling without UCP support, or within Exim, where the ucp files are
1599     not imported.
1600    
1601     3. Refactoring of the library code to split up the various functions into
1602     different source modules. The addition of the new DFA matching code (see
1603     below) to a single monolithic source would have made it really too
1604     unwieldy, quite apart from causing all the code to be include in a
1605     statically linked application, when only some functions are used. This is
1606     relevant even without the DFA addition now that patterns can be compiled in
1607     one application and matched in another.
1608    
1609     The downside of splitting up is that there have to be some external
1610     functions and data tables that are used internally in different modules of
1611     the library but which are not part of the API. These have all had their
1612     names changed to start with "_pcre_" so that they are unlikely to clash
1613     with other external names.
1614    
1615     4. Added an alternate matching function, pcre_dfa_exec(), which matches using
1616     a different (DFA) algorithm. Although it is slower than the original
1617     function, it does have some advantages for certain types of matching
1618     problem.
1619    
1620     5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
1621     including restarting after a partial match.
1622    
1623     6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
1624     defined when compiling for Windows was sent to me. I have put it into the
1625     code, though I have no means of testing or verifying it.
1626    
1627     7. Added the pcre_refcount() auxiliary function.
1628    
1629     8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
1630     match before or at the first newline in the subject string. In pcretest,
1631     the /f option on a pattern can be used to set this.
1632    
1633     9. A repeated \w when used in UTF-8 mode with characters greater than 256
1634     would behave wrongly. This has been present in PCRE since release 4.0.
1635    
1636     10. A number of changes to the pcregrep command:
1637    
1638     (a) Refactored how -x works; insert ^(...)$ instead of setting
1639     PCRE_ANCHORED and checking the length, in preparation for adding
1640     something similar for -w.
1641    
1642     (b) Added the -w (match as a word) option.
1643    
1644     (c) Refactored the way lines are read and buffered so as to have more
1645     than one at a time available.
1646    
1647     (d) Implemented a pcregrep test script.
1648    
1649     (e) Added the -M (multiline match) option. This allows patterns to match
1650     over several lines of the subject. The buffering ensures that at least
1651     8K, or the rest of the document (whichever is the shorter) is available
1652     for matching (and similarly the previous 8K for lookbehind assertions).
1653    
1654     (f) Changed the --help output so that it now says
1655    
1656     -w, --word-regex(p)
1657    
1658     instead of two lines, one with "regex" and the other with "regexp"
1659     because that confused at least one person since the short forms are the
1660     same. (This required a bit of code, as the output is generated
1661     automatically from a table. It wasn't just a text change.)
1662    
1663     (g) -- can be used to terminate pcregrep options if the next thing isn't an
1664     option but starts with a hyphen. Could be a pattern or a path name
1665     starting with a hyphen, for instance.
1666    
1667     (h) "-" can be given as a file name to represent stdin.
1668    
1669     (i) When file names are being printed, "(standard input)" is used for
1670     the standard input, for compatibility with GNU grep. Previously
1671     "<stdin>" was used.
1672    
1673     (j) The option --label=xxx can be used to supply a name to be used for
1674     stdin when file names are being printed. There is no short form.
1675    
1676     (k) Re-factored the options decoding logic because we are going to add
1677     two more options that take data. Such options can now be given in four
1678     different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
1679    
1680     (l) Added the -A, -B, and -C options for requesting that lines of context
1681     around matches be printed.
1682    
1683     (m) Added the -L option to print the names of files that do not contain
1684     any matching lines, that is, the complement of -l.
1685    
1686     (n) The return code is 2 if any file cannot be opened, but pcregrep does
1687     continue to scan other files.
1688    
1689     (o) The -s option was incorrectly implemented. For compatibility with other
1690     greps, it now suppresses the error message for a non-existent or non-
1691     accessible file (but not the return code). There is a new option called
1692     -q that suppresses the output of matching lines, which was what -s was
1693     previously doing.
1694    
1695     (p) Added --include and --exclude options to specify files for inclusion
1696     and exclusion when recursing.
1697    
1698     11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
1699     Hopefully, it now does.
1700    
1701     12. Missing cast in pcre_study().
1702    
1703     13. Added an "uninstall" target to the makefile.
1704    
1705     14. Replaced "extern" in the function prototypes in Makefile.in with
1706     "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
1707     world, but is set differently for Windows.
1708    
1709     15. Added a second compiling function called pcre_compile2(). The only
1710     difference is that it has an extra argument, which is a pointer to an
1711     integer error code. When there is a compile-time failure, this is set
1712     non-zero, in addition to the error test pointer being set to point to an
1713     error message. The new argument may be NULL if no error number is required
1714     (but then you may as well call pcre_compile(), which is now just a
1715     wrapper). This facility is provided because some applications need a
1716     numeric error indication, but it has also enabled me to tidy up the way
1717     compile-time errors are handled in the POSIX wrapper.
1718    
1719     16. Added VPATH=.libs to the makefile; this should help when building with one
1720     prefix path and installing with another. (Or so I'm told by someone who
1721     knows more about this stuff than I do.)
1722    
1723     17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
1724     passes PCRE_DOTALL to the pcre_compile() function, making the "." character
1725     match everything, including newlines. This is not POSIX-compatible, but
1726     somebody wanted the feature. From pcretest it can be activated by using
1727     both the P and the s flags.
1728    
1729     18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
1730    
1731     19. libpcre.pc was being incorrectly installed as executable.
1732    
1733     20. A couple of places in pcretest check for end-of-line by looking for '\n';
1734     it now also looks for '\r' so that it will work unmodified on Windows.
1735    
1736     21. Added Google's contributed C++ wrapper to the distribution.
1737    
1738     22. Added some untidy missing memory free() calls in pcretest, to keep
1739     Electric Fence happy when testing.
1740    
1741    
1742    
1743 nigel 75 Version 5.0 13-Sep-04
1744     ---------------------
1745    
1746     1. Internal change: literal characters are no longer packed up into items
1747     containing multiple characters in a single byte-string. Each character
1748     is now matched using a separate opcode. However, there may be more than one
1749     byte in the character in UTF-8 mode.
1750    
1751     2. The pcre_callout_block structure has two new fields: pattern_position and
1752     next_item_length. These contain the offset in the pattern to the next match
1753     item, and its length, respectively.
1754    
1755     3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
1756     insertion of callouts before each pattern item. Added the /C option to
1757     pcretest to make use of this.
1758    
1759     4. On the advice of a Windows user, the lines
1760    
1761     #if defined(_WIN32) || defined(WIN32)
1762     _setmode( _fileno( stdout ), 0x8000 );
1763     #endif /* defined(_WIN32) || defined(WIN32) */
1764    
1765     have been added to the source of pcretest. This apparently does useful
1766     magic in relation to line terminators.
1767    
1768     5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
1769     for the benefit of those environments where the "b" makes a difference.
1770    
1771     6. The icc compiler has the same options as gcc, but "configure" doesn't seem
1772     to know about it. I have put a hack into configure.in that adds in code
1773     to set GCC=yes if CC=icc. This seems to end up at a point in the
1774     generated configure script that is early enough to affect the setting of
1775     compiler options, which is what is needed, but I have no means of testing
1776     whether it really works. (The user who reported this had patched the
1777     generated configure script, which of course I cannot do.)
1778    
1779     LATER: After change 22 below (new libtool files), the configure script
1780     seems to know about icc (and also ecc). Therefore, I have commented out
1781     this hack in configure.in.
1782    
1783     7. Added support for pkg-config (2 patches were sent in).
1784    
1785     8. Negated POSIX character classes that used a combination of internal tables
1786     were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
1787     [[:^ascii]]. Typically, they would match almost any characters. The other
1788     POSIX classes were not broken in this way.
1789    
1790     9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
1791     to find the match, as PCRE was deluded into thinking that the match had to
1792     start at the start point or following a newline. The same bug applied to
1793     patterns with negative forward assertions or any backward assertions
1794     preceding ".*" at the start, unless the pattern required a fixed first
1795     character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
1796    
1797     10. In UTF-8 mode, when moving forwards in the subject after a failed match
1798     starting at the last subject character, bytes beyond the end of the subject
1799     string were read.
1800    
1801     11. Renamed the variable "class" as "classbits" to make life easier for C++
1802     users. (Previously there was a macro definition, but it apparently wasn't
1803     enough.)
1804    
1805     12. Added the new field "tables" to the extra data so that tables can be passed
1806     in at exec time, or the internal tables can be re-selected. This allows
1807     a compiled regex to be saved and re-used at a later time by a different
1808     program that might have everything at different addresses.
1809    
1810     13. Modified the pcre-config script so that, when run on Solaris, it shows a
1811     -R library as well as a -L library.
1812    
1813     14. The debugging options of pcretest (-d on the command line or D on a
1814     pattern) showed incorrect output for anything following an extended class
1815     that contained multibyte characters and which was followed by a quantifier.
1816    
1817     15. Added optional support for general category Unicode character properties
1818     via the \p, \P, and \X escapes. Unicode property support implies UTF-8
1819     support. It adds about 90K to the size of the library. The meanings of the
1820     inbuilt class escapes such as \d and \s have NOT been changed.
1821    
1822     16. Updated pcredemo.c to include calls to free() to release the memory for the
1823     compiled pattern.
1824    
1825     17. The generated file chartables.c was being created in the source directory
1826     instead of in the building directory. This caused the build to fail if the
1827     source directory was different from the building directory, and was
1828     read-only.
1829    
1830     18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
1831     file. No doubt somebody will tell me if they don't make sense... Also added
1832     Dan Mooney's comments about building on OpenVMS.
1833    
1834     19. Added support for partial matching via the PCRE_PARTIAL option for
1835     pcre_exec() and the \P data escape in pcretest.
1836    
1837     20. Extended pcretest with 3 new pattern features:
1838    
1839     (i) A pattern option of the form ">rest-of-line" causes pcretest to
1840     write the compiled pattern to the file whose name is "rest-of-line".
1841     This is a straight binary dump of the data, with the saved pointer to
1842     the character tables forced to be NULL. The study data, if any, is
1843     written too. After writing, pcretest reads a new pattern.
1844    
1845     (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
1846     compiled pattern from the given file. There must not be any
1847     occurrences of "<" in the file name (pretty unlikely); if there are,
1848     pcretest will instead treat the initial "<" as a pattern delimiter.
1849     After reading in the pattern, pcretest goes on to read data lines as
1850     usual.
1851    
1852     (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
1853     and 16-bit fields in a compiled pattern, to simulate a pattern that
1854     was compiled on a host of opposite endianness.
1855    
1856     21. The pcre-exec() function can now cope with patterns that were compiled on
1857     hosts of opposite endianness, with this restriction:
1858    
1859     As for any compiled expression that is saved and used later, the tables
1860     pointer field cannot be preserved; the extra_data field in the arguments
1861     to pcre_exec() should be used to pass in a tables address if a value
1862     other than the default internal tables were used at compile time.
1863    
1864     22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
1865     now diagnosed as an error. Previously, most of the time, a negative number
1866     would have been treated as zero, but if in addition "ovector" was passed as
1867     NULL, a crash could occur.
1868    
1869     23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
1870     new versions from the libtool 1.5 distribution (the last one is a copy of
1871     a file called libtool.m4). This seems to have fixed the need to patch
1872     "configure" to support Darwin 1.3 (which I used to do). However, I still
1873     had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
1874     workstation).
1875    
1876     24. Changed the PCRE licence to be the more standard "BSD" licence.
1877    
1878    
1879 nigel 73 Version 4.5 01-Dec-03
1880     ---------------------
1881    
1882     1. There has been some re-arrangement of the code for the match() function so
1883     that it can be compiled in a version that does not call itself recursively.
1884     Instead, it keeps those local variables that need separate instances for
1885     each "recursion" in a frame on the heap, and gets/frees frames whenever it
1886     needs to "recurse". Keeping track of where control must go is done by means
1887     of setjmp/longjmp. The whole thing is implemented by a set of macros that
1888     hide most of the details from the main code, and operates only if
1889     NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
1890     "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
1891     operating.
1892    
1893     To make it easier for callers to provide specially tailored get/free
1894     functions for this usage, two new functions, pcre_stack_malloc, and
1895     pcre_stack_free, are used. They are always called in strict stacking order,
1896     and the size of block requested is always the same.
1897    
1898     The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
1899     PCRE has been compiled to use the stack or the heap for recursion. The
1900     -C option of pcretest uses this to show which version is compiled.
1901    
1902     A new data escape \S, is added to pcretest; it causes the amounts of store
1903     obtained and freed by both kinds of malloc/free at match time to be added
1904     to the output.
1905    
1906     2. Changed the locale test to use "fr_FR" instead of "fr" because that's
1907     what's available on my current Linux desktop machine.
1908    
1909     3. When matching a UTF-8 string, the test for a valid string at the start has
1910     been extended. If start_offset is not zero, PCRE now checks that it points
1911     to a byte that is the start of a UTF-8 character. If not, it returns
1912     PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
1913     this is necessary because there may be backward assertions in the pattern.
1914     When matching the same subject several times, it may save resources to use
1915     PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
1916    
1917     4. The code for checking the validity of UTF-8 strings has been tightened so
1918     that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
1919     containing "overlong sequences".
1920    
1921     5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
1922     I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
1923     should have been "&", but it just so happened that all the cases this let
1924     through by mistake were picked up later in the function.
1925    
1926     6. I had used a variable called "isblank" - this is a C99 function, causing
1927     some compilers to warn. To avoid this, I renamed it (as "blankclass").
1928    
1929     7. Cosmetic: (a) only output another newline at the end of pcretest if it is
1930     prompting; (b) run "./pcretest /dev/null" at the start of the test script
1931     so the version is shown; (c) stop "make test" echoing "./RunTest".
1932    
1933     8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
1934    
1935     9. The prototype for memmove() for systems that don't have it was using
1936     size_t, but the inclusion of the header that defines size_t was later. I've
1937     moved the #includes for the C headers earlier to avoid this.
1938    
1939     10. Added some adjustments to the code to make it easier to compiler on certain
1940     special systems:
1941    
1942     (a) Some "const" qualifiers were missing.
1943     (b) Added the macro EXPORT before all exported functions; by default this
1944     is defined to be empty.
1945     (c) Changed the dftables auxiliary program (that builds chartables.c) so
1946     that it reads its output file name as an argument instead of writing
1947     to the standard output and assuming this can be redirected.
1948    
1949     11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
1950     class containing characters with values greater than 255, PCRE compilation
1951     went into a loop.
1952    
1953     12. A recursive reference to a subpattern that was within another subpattern
1954     that had a minimum quantifier of zero caused PCRE to crash. For example,
1955     (x(y(?2))z)? provoked this bug with a subject that got as far as the
1956     recursion. If the recursively-called subpattern itself had a zero repeat,
1957     that was OK.
1958    
1959     13. In pcretest, the buffer for reading a data line was set at 30K, but the
1960     buffer into which it was copied (for escape processing) was still set at
1961     1024, so long lines caused crashes.
1962    
1963     14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
1964     "internal error: code overflow...". This applied to any character class
1965     that was followed by a possessive quantifier.
1966    
1967     15. Modified the Makefile to add libpcre.la as a prerequisite for
1968     libpcreposix.la because I was told this is needed for a parallel build to
1969     work.
1970    
1971     16. If a pattern that contained .* following optional items at the start was
1972     studied, the wrong optimizing data was generated, leading to matching
1973     errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
1974     matching string must start with a or b or c. The correct conclusion for
1975     this pattern is that a match can start with any character.
1976    
1977    
1978 nigel 71 Version 4.4 13-Aug-03
1979     ---------------------
1980    
1981     1. In UTF-8 mode, a character class containing characters with values between
1982     127 and 255 was not handled correctly if the compiled pattern was studied.
1983     In fixing this, I have also improved the studying algorithm for such
1984     classes (slightly).
1985    
1986     2. Three internal functions had redundant arguments passed to them. Removal
1987     might give a very teeny performance improvement.
1988    
1989     3. Documentation bug: the value of the capture_top field in a callout is *one
1990     more than* the number of the hightest numbered captured substring.
1991    
1992     4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
1993     in incorrectly linking with a previously installed version. They now link
1994     explicitly with libpcre.la.
1995    
1996     5. configure.in no longer needs to recognize Cygwin specially.
1997    
1998     6. A problem in pcre.in for Windows platforms is fixed.
1999    
2000     7. If a pattern was successfully studied, and the -d (or /D) flag was given to
2001     pcretest, it used to include the size of the study block as part of its
2002     output. Unfortunately, the structure contains a field that has a different
2003     size on different hardware architectures. This meant that the tests that
2004     showed this size failed. As the block is currently always of a fixed size,
2005     this information isn't actually particularly useful in pcretest output, so
2006     I have just removed it.
2007    
2008     8. Three pre-processor statements accidentally did not start in column 1.
2009     Sadly, there are *still* compilers around that complain, even though
2010     standard C has not required this for well over a decade. Sigh.
2011    
2012     9. In pcretest, the code for checking callouts passed small integers in the
2013     callout_data field, which is a void * field. However, some picky compilers
2014     complained about the casts involved for this on 64-bit systems. Now
2015     pcretest passes the address of the small integer instead, which should get
2016     rid of the warnings.
2017    
2018     10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
2019     both compile and run time, and gives an error if an invalid UTF-8 sequence
2020     is found. There is a option for disabling this check in cases where the
2021     string is known to be correct and/or the maximum performance is wanted.
2022    
2023     11. In response to a bug report, I changed one line in Makefile.in from
2024    
2025     -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
2026     to
2027     -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
2028    
2029     to look similar to other lines, but I have no way of telling whether this
2030     is the right thing to do, as I do not use Windows. No doubt I'll get told
2031     if it's wrong...
2032    
2033    
2034 nigel 69 Version 4.3 21-May-03
2035     ---------------------
2036    
2037     1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
2038     Makefile.
2039    
2040     2. Some refactoring to improve the quality of the code:
2041    
2042     (i) The utf8_table... variables are now declared "const".
2043    
2044     (ii) The code for \cx, which used the "case flipping" table to upper case
2045     lower case letters, now just substracts 32. This is ASCII-specific,
2046     but the whole concept of \cx is ASCII-specific, so it seems
2047     reasonable.
2048    
2049     (iii) PCRE was using its character types table to recognize decimal and
2050     hexadecimal digits in the pattern. This is silly, because it handles
2051     only 0-9, a-f, and A-F, but the character types table is locale-
2052     specific, which means strange things might happen. A private
2053     table is now used for this - though it costs 256 bytes, a table is
2054     much faster than multiple explicit tests. Of course, the standard
2055     character types table is still used for matching digits in subject
2056     strings against \d.
2057    
2058     (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
2059     ending in _t are). So I've renamed it as ESC_tee.
2060    
2061     3. The first argument for regexec() in the POSIX wrapper should have been
2062     defined as "const".
2063    
2064     4. Changed pcretest to use malloc() for its buffers so that they can be
2065     Electric Fenced for debugging.
2066    
2067     5. There were several places in the code where, in UTF-8 mode, PCRE would try
2068     to read one or more bytes before the start of the subject string. Often this
2069     had no effect on PCRE's behaviour, but in some circumstances it could
2070     provoke a segmentation fault.
2071    
2072     6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
2073     to try to read one or more bytes before the start of the subject string.
2074    
2075     7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
2076     UTF-8 support could misbehave in various ways if the subject string
2077     contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
2078     area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
2079     back over UTF-8 characters.)
2080    
2081    
2082 nigel 67 Version 4.2 14-Apr-03
2083     ---------------------
2084    
2085     1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
2086    
2087     2. Changes to the building process, supplied by Ronald Landheer-Cieslak
2088     [ON_WINDOWS]: new variable, "#" on non-Windows platforms
2089     [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
2090     [WIN_PREFIX]: new variable, "cyg" for Cygwin
2091     * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
2092     and BUILD_EXEEXT
2093     Note: automatic setting of the BUILD variables is not yet working
2094     set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
2095     compile-time but not at link-time
2096     [LINK]: use for linking executables only
2097     make different versions for Windows and non-Windows
2098     [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
2099     libraries
2100     [LINK_FOR_BUILD]: new variable
2101     [OBJEXT]: use throughout
2102     [EXEEXT]: use throughout
2103     <winshared>: new target
2104     <wininstall>: new target
2105     <dftables.o>: use native compiler
2106     <dftables>: use native linker
2107     <install>: handle Windows platform correctly
2108     <clean>: ditto
2109     <check>: ditto
2110     copy DLL to top builddir before testing
2111    
2112     As part of these changes, -no-undefined was removed again. This was reported
2113     to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
2114     in any case.
2115    
2116     3. Some tidies to get rid of compiler warnings:
2117    
2118     . In the match_data structure, match_limit was an unsigned long int, whereas
2119     match_call_count was an int. I've made them both unsigned long ints.
2120    
2121     . In pcretest the fact that a const uschar * doesn't automatically cast to
2122     a void * provoked a warning.
2123    
2124     . Turning on some more compiler warnings threw up some "shadow" variables
2125     and a few more missing casts.
2126    
2127     4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
2128     option, a class that contained a single character with a value between 128
2129     and 255 (e.g. /[\xFF]/) caused PCRE to crash.
2130    
2131     5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
2132     option, a class that contained several characters, but with at least one
2133     whose value was between 128 and 255 caused PCRE to crash.
2134    
2135    
2136 nigel 65 Version 4.1 12-Mar-03
2137     ---------------------
2138 nigel 63
2139 nigel 65 1. Compiling with gcc -pedantic found a couple of places where casts were
2140     needed, and a string in dftables.c that was longer than standard compilers are
2141     required to support.
2142    
2143     2. Compiling with Sun's compiler found a few more places where the code could
2144     be tidied up in order to avoid warnings.
2145    
2146     3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
2147     first of these names is deprecated in the latest Autoconf in favour of the name
2148     CC_FOR_BUILD, because "host" is typically used to mean the system on which the
2149     compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
2150     analogy I have changed it to CFLAGS_FOR_BUILD.
2151    
2152     4. Added -no-undefined to the linking command in the Makefile, because this is
2153     apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
2154     linking step for the pcreposix library.
2155    
2156     5. PCRE was failing to diagnose the case of two named groups with the same
2157     name.
2158    
2159     6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
2160     literal character that is needed in the subject for a match, and scans along to
2161     ensure that it is present before embarking on the full matching process. This
2162     saves time in cases of nested unlimited repeats that are never going to match.
2163     Problem: the scan can take a lot of time if the subject is very long (e.g.
2164     megabytes), thus penalizing straightforward matches. It is now done only if the
2165     amount of subject to be scanned is less than 1000 bytes.
2166    
2167     7. A lesser problem with the same optimization is that it was recording the
2168     first character of an anchored pattern as "needed", thus provoking a search
2169     right along the subject, even when the first match of the pattern was going to
2170     fail. The "needed" character is now not set for anchored patterns, unless it
2171     follows something in the pattern that is of non-fixed length. Thus, it still
2172     fulfils its original purpose of finding quick non-matches in cases of nested
2173     unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
2174    
2175    
2176     Version 4.0 17-Feb-03
2177     ---------------------
2178    
2179 nigel 63 1. If a comment in an extended regex that started immediately after a meta-item
2180     extended to the end of string, PCRE compiled incorrect data. This could lead to
2181     all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
2182    
2183     2. Moved to autoconf 2.53 and libtool 1.4.2.
2184    
2185     3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
2186     the special perltest8 script is no longer needed - all the tests can be run
2187     from a single perltest script.
2188    
2189     4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
2190     by \s. It has now been removed in PCRE. This means it isn't recognized as
2191     whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
2192     class [:space:] *does* include VT, thereby creating a mess.
2193    
2194     5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
2195     space and tab.
2196    
2197     6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
2198     its new features into the main test script, reducing the number of scripts.
2199    
2200     7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
2201     were backward compatible, and made the (?i) apply to the whole pattern, as if
2202     /i were given. Now it behaves more logically, and applies the option setting
2203     only to what follows. PCRE has been changed to follow suit. However, if it
2204     finds options settings right at the start of the pattern, it extracts them into
2205     the global options, as before. Thus, they show up in the info data.
2206    
2207     8. Added support for the \Q...\E escape sequence. Characters in between are
2208     treated as literals. This is slightly different from Perl in that $ and @ are
2209     also handled as literals inside the quotes. In Perl, they will cause variable
2210     interpolation. Note the following examples:
2211    
2212     Pattern PCRE matches Perl matches
2213    
2214     \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
2215     \Qabc\$xyz\E abc\$xyz abc\$xyz
2216     \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
2217    
2218     For compatibility with Perl, \Q...\E sequences are recognized inside character
2219     classes as well as outside them.
2220    
2221     9. Re-organized 3 code statements in pcretest to avoid "overflow in
2222     floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
2223     (size_t) cast to one statement in pcretest and one in pcreposix to avoid
2224     signed/unsigned warnings.
2225    
2226     10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
2227     option for pcretest, so I've replaced it by a simple function that does just
2228     that job.
2229    
2230     11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
2231     "pcregrep -".
2232    
2233     12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
2234     Java package. This provides some syntactic sugar for simple cases of what my
2235     documentation calls "once-only subpatterns". A pattern such as x*+ is the same
2236     as (?>x*). In other words, if what is inside (?>...) is just a single repeated
2237     item, you can use this simplified notation. Note that only makes sense with
2238     greedy quantifiers. Consequently, the use of the possessive quantifier forces
2239     greediness, whatever the setting of the PCRE_UNGREEDY option.
2240    
2241     13. A change of greediness default within a pattern was not taking effect at
2242     the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
2243     subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
2244     was abstracted outside.
2245    
2246     14. PCRE now supports the \G assertion. It is true when the current matching
2247     position is at the start point of the match. This differs from \A when the
2248     starting offset is non-zero. Used with the /g option of pcretest (or similar
2249     code), it works in the same way as it does for Perl's /g option. If all
2250     alternatives of a regex begin with \G, the expression is anchored to the start
2251     match position, and the "anchored" flag is set in the compiled expression.
2252    
2253     15. Some bugs concerning the handling of certain option changes within patterns
2254     have been fixed. These applied to options other than (?ims). For example,
2255     "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
2256     been the other way round. Some of this was related to change 7 above.
2257    
2258     16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
2259     features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
2260     and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
2261     POSIX classes only within a class (e.g. /[[:alpha:]]/).
2262    
2263     17. Added support for Perl's \C escape. This matches one byte, even in UTF8
2264     mode. Unlike ".", it always matches newline, whatever the setting of
2265     PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
2266     assertions. Perl allows it, but it doesn't (in general) work because it can't
2267     calculate the length of the lookbehind. At least, that's the case for Perl
2268     5.8.0 - I've been told they are going to document that it doesn't work in
2269     future.
2270    
2271     18. Added an error diagnosis for escapes that PCRE does not support: these are
2272     \L, \l, \N, \P, \p, \U, \u, and \X.
2273    
2274     19. Although correctly diagnosing a missing ']' in a character class, PCRE was
2275     reading past the end of the pattern in cases such as /[abcd/.
2276    
2277     20. PCRE was getting more memory than necessary for patterns with classes that
2278     contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
2279    
2280     21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
2281     compiling PCRE for use with Virtual Pascal.
2282    
2283     22. Small fix to the Makefile to make it work properly if the build is done
2284     outside the source tree.
2285    
2286     23. Added a new extension: a condition to go with recursion. If a conditional
2287     subpattern starts with (?(R) the "true" branch is used if recursion has
2288     happened, whereas the "false" branch is used only at the top level.
2289    
2290     24. When there was a very long string of literal characters (over 255 bytes
2291     without UTF support, over 250 bytes with UTF support), the computation of how
2292     much memory was required could be incorrect, leading to segfaults or other
2293     strange effects.
2294    
2295     25. PCRE was incorrectly assuming anchoring (either to start of subject or to
2296     start of line for a non-DOTALL pattern) when a pattern started with (.*) and
2297     there was a subsequent back reference to those brackets. This meant that, for
2298     example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
2299     possible to check for precisely this case. All we can do is abandon the
2300     optimization if .* occurs inside capturing brackets when there are any back
2301     references whatsoever. (See below for a better fix that came later.)
2302    
2303     26. The handling of the optimization for finding the first character of a
2304     non-anchored pattern, and for finding a character that is required later in the
2305     match were failing in some cases. This didn't break the matching; it just
2306     failed to optimize when it could. The way this is done has been re-implemented.
2307    
2308     27. Fixed typo in error message for invalid (?R item (it said "(?p").
2309    
2310     28. Added a new feature that provides some of the functionality that Perl
2311     provides with (?{...}). The facility is termed a "callout". The way it is done
2312     in PCRE is for the caller to provide an optional function, by setting
2313     pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
2314     global variable. By default it is unset, which disables all calling out. To get
2315     the function called, the regex must include (?C) at appropriate points. This
2316     is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
2317     This provides a means of identifying different callout points. When PCRE
2318     reaches such a point in the regex, if pcre_callout has been set, the external
2319     function is called. It is provided with data in a structure called
2320     pcre_callout_block, which is defined in pcre.h. If the function returns 0,
2321     matching continues; if it returns a non-zero value, the match at the current
2322     point fails. However, backtracking will occur if possible. [This was changed
2323     later and other features added - see item 49 below.]
2324    
2325     29. pcretest is upgraded to test the callout functionality. It provides a
2326     callout function that displays information. By default, it shows the start of
2327     the match and the current position in the text. There are some new data escapes
2328     to vary what happens:
2329    
2330     \C+ in addition, show current contents of captured substrings
2331     \C- do not supply a callout function
2332     \C!n return 1 when callout number n is reached
2333     \C!n!m return 1 when callout number n is reached for the mth time
2334    
2335     30. If pcregrep was called with the -l option and just a single file name, it
2336     output "<stdin>" if a match was found, instead of the file name.
2337    
2338     31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
2339     slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
2340     pcre_exec(). This saves a malloc/free per call. The default value of
2341     POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
2342     when configuring.
2343    
2344     32. The default maximum size of a compiled pattern is 64K. There have been a
2345     few cases of people hitting this limit. The code now uses macros to handle the
2346     storing of links as offsets within the compiled pattern. It defaults to 2-byte
2347     links, but this can be changed to 3 or 4 bytes by --with-link-size when
2348     configuring. Tests 2 and 5 work only with 2-byte links because they output
2349     debugging information about compiled patterns.
2350    
2351     33. Internal code re-arrangements:
2352    
2353     (a) Moved the debugging function for printing out a compiled regex into
2354     its own source file (printint.c) and used #include to pull it into
2355     pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
2356     separate copies.
2357    
2358     (b) Defined the list of op-code names for debugging as a macro in
2359     internal.h so that it is next to the definition of the opcodes.
2360    
2361     (c) Defined a table of op-code lengths for simpler skipping along compiled
2362     code. This is again a macro in internal.h so that it is next to the
2363     definition of the opcodes.
2364    
2365     34. Added support for recursive calls to individual subpatterns, along the
2366     lines of Robin Houston's patch (but implemented somewhat differently).
2367    
2368     35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
2369     allow it to read and process whole directories in Win32. This code was
2370     contributed by Lionel Fourquaux; it has not been tested by me.
2371    
2372     36. Added support for named subpatterns. The Python syntax (?P<name>...) is
2373     used to name a group. Names consist of alphanumerics and underscores, and must
2374     be unique. Back references use the syntax (?P=name) and recursive calls use
2375     (?P>name) which is a PCRE extension to the Python extension. Groups still have
2376     numbers. The function pcre_fullinfo() can be used after compilation to extract
2377     a name/number map. There are three relevant calls:
2378    
2379     PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
2380     PCRE_INFO_NAMECOUNT yields the number of entries
2381     PCRE_INFO_NAMETABLE yields a pointer to the map.
2382    
2383     The map is a vector of fixed-size entries. The size of each entry depends on
2384     the length of the longest name used. The first two bytes of each entry are the
2385     group number, most significant byte first. There follows the corresponding
2386     name, zero terminated. The names are in alphabetical order.
2387    
2388     37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
2389     case instead of 255. Making it the same both with and without UTF-8 support
2390     means that the same test output works with both.
2391    
2392     38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
2393     calling malloc() with a zero argument.
2394    
2395     39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
2396     optimization. I've improved things by keeping a bitmap of backreferences with
2397     numbers 1-31 so that if .* occurs inside capturing brackets that are not in
2398     fact referenced, the optimization can be applied. It is unlikely that a
2399     relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
2400     the match to follow \n) will appear inside brackets with a number greater than
2401     31, but if it does, any back reference > 31 suppresses the optimization.
2402    
2403     40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
2404     of disabling numbered capturing parentheses. Any opening parenthesis that is
2405     not followed by ? behaves as if it were followed by ?: but named parentheses
2406     can still be used for capturing (and they will acquire numbers in the usual
2407     way).
2408    
2409     41. Redesigned the return codes from the match() function into yes/no/error so
2410     that errors can be passed back from deep inside the nested calls. A malloc
2411     failure while inside a recursive subpattern call now causes the
2412     PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
2413    
2414     42. It is now possible to set a limit on the number of times the match()
2415     function is called in a call to pcre_exec(). This facility makes it possible to
2416     limit the amount of recursion and backtracking, though not in a directly
2417     obvious way, because the match() function is used in a number of different
2418     circumstances. The count starts from zero for each position in the subject
2419     string (for non-anchored patterns). The default limit is, for compatibility, a
2420     large number, namely 10 000 000. You can change this in two ways:
2421    
2422     (a) When configuring PCRE before making, you can use --with-match-limit=n
2423     to set a default value for the compiled library.
2424    
2425     (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
2426     a different value is set. See 45 below.
2427    
2428     If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
2429    
2430     43. Added a new function pcre_config(int, void *) to enable run-time extraction
2431     of things that can be changed at compile time. The first argument specifies
2432     what is wanted and the second points to where the information is to be placed.
2433     The current list of available information is:
2434    
2435     PCRE_CONFIG_UTF8
2436    
2437     The output is an integer that is set to one if UTF-8 support is available;
2438     otherwise it is set to zero.
2439    
2440     PCRE_CONFIG_NEWLINE
2441    
2442     The output is an integer that it set to the value of the code that is used for
2443     newline. It is either LF (10) or CR (13).
2444    
2445     PCRE_CONFIG_LINK_SIZE
2446    
2447     The output is an integer that contains the number of bytes used for internal
2448     linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
2449    
2450     PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
2451    
2452     The output is an integer that contains the threshold above which the POSIX
2453     interface uses malloc() for output vectors. See item 31 above.
2454    
2455     PCRE_CONFIG_MATCH_LIMIT
2456    
2457     The output is an unsigned integer that contains the default limit of the number
2458     of match() calls in a pcre_exec() execution. See 42 above.
2459    
2460     44. pcretest has been upgraded by the addition of the -C option. This causes it
2461     to extract all the available output from the new pcre_config() function, and to
2462     output it. The program then exits immediately.
2463    
2464     45. A need has arisen to pass over additional data with calls to pcre_exec() in
2465     order to support additional features. One way would have been to define
2466     pcre_exec2() (for example) with extra arguments, but this would not have been
2467     extensible, and would also have required all calls to the original function to
2468     be mapped to the new one. Instead, I have chosen to extend the mechanism that
2469     is used for passing in "extra" data from pcre_study().
2470    
2471     The pcre_extra structure is now exposed and defined in pcre.h. It currently
2472     contains the following fields:
2473    
2474     flags a bitmap indicating which of the following fields are set
2475     study_data opaque data from pcre_study()
2476     match_limit a way of specifying a limit on match() calls for a specific
2477     call to pcre_exec()
2478     callout_data data for callouts (see 49 below)
2479    
2480     The flag bits are also defined in pcre.h, and are
2481    
2482     PCRE_EXTRA_STUDY_DATA
2483     PCRE_EXTRA_MATCH_LIMIT
2484     PCRE_EXTRA_CALLOUT_DATA
2485    
2486     The pcre_study() function now returns one of these new pcre_extra blocks, with
2487     the actual study data pointed to by the study_data field, and the
2488     PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
2489     before. That is, this change is entirely upwards-compatible and requires no
2490     change to existing code.
2491    
2492     If you want to pass in additional data to pcre_exec(), you can either place it
2493     in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
2494     block.
2495    
2496     46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
2497     data string contains the escape sequence \M, pcretest calls pcre_exec() several
2498     times with different match limits, until it finds the minimum value needed for
2499     pcre_exec() to complete. The value is then output. This can be instructive; for
2500     most simple matches the number is quite small, but for pathological cases it
2501     gets very large very quickly.
2502    
2503     47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
2504     returns the size of the data block pointed to by the study_data field in a
2505     pcre_extra block, that is, the value that was passed as the argument to
2506     pcre_malloc() when PCRE was getting memory in which to place the information
2507     created by pcre_study(). The fourth argument should point to a size_t variable.
2508     pcretest has been extended so that this information is shown after a successful
2509     pcre_study() call when information about the compiled regex is being displayed.
2510    
2511     48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
2512     because what follows is always an absolute path. (Later: it turns out that this
2513     is more than cosmetic for MinGW, because it doesn't like empty path
2514     components.)
2515    
2516     49. Some changes have been made to the callout feature (see 28 above):
2517    
2518     (i) A callout function now has three choices for what it returns:
2519    
2520     0 => success, carry on matching
2521     > 0 => failure at this point, but backtrack if possible
2522     < 0 => serious error, return this value from pcre_exec()
2523    
2524     Negative values should normally be chosen from the set of PCRE_ERROR_xxx
2525     values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
2526     "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
2527     use by callout functions. It will never be used by PCRE itself.
2528    
2529     (ii) The pcre_extra structure (see 45 above) has a void * field called
2530     callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
2531     pcre_callout_block structure has a field of the same name. The contents of
2532     the field passed in the pcre_extra structure are passed to the callout
2533     function in the corresponding field in the callout block. This makes it
2534     easier to use the same callout-containing regex from multiple threads. For
2535     testing, the pcretest program has a new data escape
2536    
2537     \C*n pass the number n (may be negative) as callout_data
2538    
2539     If the callout function in pcretest receives a non-zero value as
2540     callout_data, it returns that value.
2541    
2542     50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
2543     there were some redundant $(CFLAGS) in commands that are now specified as
2544     $(LINK), which already includes $(CFLAGS).
2545    
2546     51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
2547     has been compiled with UTF-8 support *and* pcre_compile() has been compiled
2548     with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
2549     one-byte characters throughout. Note that case-insensitive matching applies
2550     only to characters whose values are less than 256. PCRE doesn't support the
2551     notion of cases for higher-valued characters.
2552    
2553     (i) A character class whose characters are all within 0-255 is handled as
2554     a bit map, and the map is inverted for negative classes. Previously, a
2555     character > 255 always failed to match such a class; however it should
2556     match if the class was a negative one (e.g. [^ab]). This has been fixed.
2557    
2558     (ii) A negated character class with a single character < 255 is coded as
2559     "not this character" (OP_NOT). This wasn't working properly when the test
2560     character was multibyte, either singly or repeated.
2561    
2562     (iii) Repeats of multibyte characters are now handled correctly in UTF-8
2563     mode, for example: \x{100}{2,3}.
2564    
2565     (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
2566     singly or repeated) now correctly test multibyte characters. However,
2567     PCRE doesn't recognize any characters with values greater than 255 as
2568     digits, spaces, or word characters. Such characters always match \D, \S,
2569     and \W, and never match \d, \s, or \w.
2570    
2571     (v) Classes may now contain characters and character ranges with values
2572     greater than 255. For example: [ab\x{100}-\x{400}].
2573    
2574     (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
2575     PCRE in UTF-8 mode.
2576    
2577     52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
2578     PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
2579     retained for backwards compatibility. (Note that LASTLITERAL is also a byte
2580     value.)
2581    
2582     53. The single man page has become too large. I have therefore split it up into
2583     a number of separate man pages. These also give rise to individual HTML pages;
2584     these are now put in a separate directory, and there is an index.html page that
2585     lists them all. Some hyperlinking between the pages has been installed.
2586    
2587     54. Added convenience functions for handling named capturing parentheses.
2588    
2589     55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
2590     aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
2591     true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
2592     are faulted.
2593    
2594     56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
2595     calling configure. These values are used when compiling the dftables.c program
2596     which is run to generate the source of the default character tables. They
2597     default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
2598     you will need to set these values.
2599    
2600     57. Updated the building process for Windows DLL, as provided by Fred Cox.
2601    
2602    
2603     Version 3.9 02-Jan-02
2604 nigel 61 ---------------------
2605    
2606     1. A bit of extraneous text had somehow crept into the pcregrep documentation.
2607    
2608     2. If --disable-static was given, the building process failed when trying to
2609     build pcretest and pcregrep. (For some reason it was using libtool to compile
2610     them, which is not right, as they aren't part of the library.)
2611    
2612    
2613 nigel 59 Version 3.8 18-Dec-01
2614     ---------------------
2615    
2616     1. The experimental UTF-8 code was completely screwed up. It was packing the
2617     bytes in the wrong order. How dumb can you get?
2618    
2619    
2620 nigel 57 Version 3.7 29-Oct-01
2621     ---------------------
2622    
2623     1. In updating pcretest to check change 1 of version 3.6, I screwed up.
2624     This caused pcretest, when used on the test data, to segfault. Unfortunately,
2625     this didn't happen under Solaris 8, where I normally test things.
2626    
2627 nigel 59 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
2628     doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
2629     isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
2630     this fix an hour or so after the initial 3.7 release.)
2631 nigel 57
2632 nigel 59
2633 nigel 55 Version 3.6 23-Oct-01
2634     ---------------------
2635    
2636     1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
2637     offsets passed as NULL with zero offset count.
2638    
2639     2. The config.guess and config.sub files had not been updated when I moved to
2640     the latest autoconf.
2641    
2642    
2643 nigel 53 Version 3.5 15-Aug-01
2644     ---------------------
2645 nigel 5
2646 nigel 53 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
2647     had been forgotten.
2648    
2649     2. By using declared but undefined structures, we can avoid using "void"
2650     definitions in pcre.h while keeping the internal definitions of the structures
2651     private.
2652    
2653     3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
2654     user point of view, this means that both static and shared libraries are built
2655     by default, but this can be individually controlled. More of the work of
2656     handling this static/shared cases is now inside libtool instead of PCRE's make
2657     file.
2658    
2659     4. The pcretest utility is now installed along with pcregrep because it is
2660     useful for users (to test regexs) and by doing this, it automatically gets
2661     relinked by libtool. The documentation has been turned into a man page, so
2662     there are now .1, .txt, and .html versions in /doc.
2663    
2664     5. Upgrades to pcregrep:
2665     (i) Added long-form option names like gnu grep.
2666     (ii) Added --help to list all options with an explanatory phrase.
2667     (iii) Added -r, --recursive to recurse into sub-directories.
2668     (iv) Added -f, --file to read patterns from a file.
2669    
2670     6. pcre_exec() was referring to its "code" argument before testing that
2671     argument for NULL (and giving an error if it was NULL).
2672    
2673     7. Upgraded Makefile.in to allow for compiling in a different directory from
2674     the source directory.
2675    
2676     8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
2677     options bits, the pointer it was passed was to an int instead of to an unsigned
2678     long int. This mattered only on 64-bit systems.
2679    
2680     9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
2681     generated) instead of pcre.in, which it its source. Also made the same change
2682     in several of the .c files.
2683    
2684     10. A new release of gcc defines printf() as a macro, which broke pcretest
2685     because it had an ifdef in the middle of a string argument for printf(). Fixed
2686     by using separate calls to printf().
2687    
2688     11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
2689     script, to force use of CR or LF instead of \n in the source. On non-Unix
2690     systems, the value can be set in config.h.
2691    
2692     12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
2693     absolute limit. Changed the text of the error message to make this clear, and
2694     likewise updated the man page.
2695    
2696     13. The limit of 99 on the number of capturing subpatterns has been removed.
2697     The new limit is 65535, which I hope will not be a "real" limit.
2698    
2699    
2700 nigel 51 Version 3.4 22-Aug-00
2701     ---------------------
2702    
2703     1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
2704    
2705     2. Diagnose condition (?(0) as an error instead of crashing on matching.
2706    
2707    
2708 nigel 49 Version 3.3 01-Aug-00
2709     ---------------------
2710    
2711     1. If an octal character was given, but the value was greater than \377, it
2712     was not getting masked to the least significant bits, as documented. This could
2713     lead to crashes in some systems.
2714    
2715     2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
2716     the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
2717    
2718     3. Added the functions pcre_free_substring() and pcre_free_substring_list().
2719     These just pass their arguments on to (pcre_free)(), but they are provided
2720     because some uses of PCRE bind it to non-C systems that can call its functions,
2721     but cannot call free() or pcre_free() directly.
2722    
2723     4. Add "make test" as a synonym for "make check". Corrected some comments in
2724     the Makefile.
2725    
2726     5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
2727     Makefile.
2728    
2729     6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
2730     command called pgrep for grepping around the active processes.
2731    
2732     7. Added the beginnings of support for UTF-8 character strings.
2733    
2734     8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
2735     RANLIB to ./ltconfig so that they are used by libtool. I think these are all
2736     the relevant ones. (AR is not passed because ./ltconfig does its own figuring
2737     out for the ar command.)
2738    
2739    
2740 nigel 47 Version 3.2 12-May-00
2741     ---------------------
2742    
2743     This is purely a bug fixing release.
2744    
2745     1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
2746     of ZA. This was just one example of several cases that could provoke this bug,
2747     which was introduced by change 9 of version 2.00. The code for breaking
2748     infinite loops after an iteration that matches an empty string was't working
2749     correctly.
2750    
2751     2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
2752     when matched against abbab (for example). After matching an empty string, it
2753     wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
2754     caused it to match further down the string than it should.
2755    
2756     3. The code contained an inclusion of sys/types.h. It isn't clear why this
2757     was there because it doesn't seem to be needed, and it causes trouble on some
2758     systems, as it is not a Standard C header. It has been removed.
2759    
2760     4. Made 4 silly changes to the source to avoid stupid compiler warnings that
2761     were reported on the Macintosh. The changes were from
2762    
2763     while ((c = *(++ptr)) != 0 && c != '\n');
2764     to
2765     while ((c = *(++ptr)) != 0 && c != '\n') ;
2766    
2767     Totally extraordinary, but if that's what it takes...
2768    
2769     5. PCRE is being used in one environment where neither memmove() nor bcopy() is
2770     available. Added HAVE_BCOPY and an autoconf test for it; if neither
2771     HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
2772     assumes the way PCRE uses memmove() (always moving upwards).
2773    
2774     6. PCRE is being used in one environment where strchr() is not available. There
2775     was only one use in pcre.c, and writing it out to avoid strchr() probably gives
2776     faster code anyway.
2777    
2778    
2779 nigel 45 Version 3.1 09-Feb-00
2780     ---------------------
2781    
2782     The only change in this release is the fixing of some bugs in Makefile.in for
2783     the "install" target:
2784    
2785     (1) It was failing to install pcreposix.h.
2786    
2787     (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
2788    
2789    
2790 nigel 43 Version 3.0 01-Feb-00
2791     ---------------------
2792 nigel 41
2793     1. Add support for the /+ modifier to perltest (to output $` like it does in
2794     pcretest).
2795    
2796     2. Add support for the /g modifier to perltest.
2797    
2798     3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
2799     matches null strings.
2800    
2801     4. Fix perltest so that it doesn't do unwanted things when fed an empty
2802     pattern. Perl treats empty patterns specially - it reuses the most recent
2803     pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
2804     effect.
2805    
2806     5. The POSIX interface was broken in that it was just handing over the POSIX
2807     captured string vector to pcre_exec(), but (since release 2.00) PCRE has
2808     required a bigger vector, with some working space on the end. This means that
2809     the POSIX wrapper now has to get and free some memory, and copy the results.
2810    
2811 nigel 43 6. Added some simple autoconf support, placing the test data and the
2812     documentation in separate directories, re-organizing some of the
2813     information files, and making it build pcre-config (a GNU standard). Also added
2814     libtool support for building PCRE as a shared library, which is now the
2815     default.
2816 nigel 41
2817 nigel 43 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
2818     09 are not valid octal constants. Single digits will be used for minor values
2819     less than 10.
2820    
2821     8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
2822     existing programs that set these in the POSIX interface can use PCRE without
2823     modification.
2824    
2825     9. Added a new function, pcre_fullinfo() with an extensible interface. It can
2826     return all that pcre_info() returns, plus additional data. The pcre_info()
2827     function is retained for compatibility, but is considered to be obsolete.
2828    
2829     10. Added experimental recursion feature (?R) to handle one common case that
2830     Perl 5.6 will be able to do with (?p{...}).
2831    
2832     11. Added support for POSIX character classes like [:alpha:], which Perl is
2833     adopting.
2834    
2835    
2836 nigel 39 Version 2.08 31-Aug-99
2837     ----------------------
2838    
2839     1. When startoffset was not zero and the pattern began with ".*", PCRE was not
2840     trying to match at the startoffset position, but instead was moving forward to
2841     the next newline as if a previous match had failed.
2842    
2843     2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
2844     and could get into a loop if a null string was matched other than at the start
2845     of the subject.
2846    
2847     3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
2848     be distinguished at compile time, and for completeness also added PCRE_DATE.
2849    
2850     5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
2851     in GnuWin32 environments.
2852    
2853    
2854 nigel 37 Version 2.07 29-Jul-99
2855     ----------------------
2856    
2857     1. The documentation is now supplied in plain text form and HTML as well as in
2858     the form of man page sources.
2859    
2860     2. C++ compilers don't like assigning (void *) values to other pointer types.
2861     In particular this affects malloc(). Although there is no problem in Standard
2862     C, I've put in casts to keep C++ compilers happy.
2863    
2864     3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
2865     should be (const char *).
2866    
2867     4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
2868     be useful for non-Unix systems who don't want to bother with the POSIX stuff.
2869     However, I haven't made this a standard facility. The documentation doesn't
2870     mention it, and the Makefile doesn't support it.
2871    
2872     5. The Makefile now contains an "install" target, with editable destinations at
2873     the top of the file. The pcretest program is not installed.
2874    
2875     6. pgrep -V now gives the PCRE version number and date.
2876    
2877     7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
2878     causing the entire string to be ignored, instead of just the last character.
2879    
2880     8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
2881     non-matching string, it can take a very, very long time, even for strings of
2882     quite modest length, because of the nested recursion. PCRE now does better in
2883     some of these cases. It does this by remembering the last required literal
2884     character in the pattern, and pre-searching the subject to ensure it is present
2885     before running the real match. In other words, it applies a heuristic to detect
2886     some types of certain failure quickly, and in the above example, if presented
2887     with a string that has no trailing " it gives "no match" very quickly.
2888    
2889     9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
2890     other alternatives are tried instead.
2891    
2892    
2893 nigel 35 Version 2.06 09-Jun-99
2894     ----------------------
2895    
2896     1. Change pcretest's output for amount of store used to show just the code
2897     space, because the remainder (the data block) varies in size between 32-bit and
2898     64-bit systems.
2899    
2900     2. Added an extra argument to pcre_exec() to supply an offset in the subject to
2901     start matching at. This allows lookbehinds to work when searching for multiple
2902     occurrences in a string.
2903    
2904     3. Added additional options to pcretest for testing multiple occurrences:
2905    
2906     /+ outputs the rest of the string that follows a match
2907     /g loops for multiple occurrences, using the new startoffset argument
2908     /G loops for multiple occurrences by passing an incremented pointer
2909    
2910     4. PCRE wasn't doing the "first character" optimization for patterns starting
2911     with \b or \B, though it was doing it for other lookbehind assertions. That is,
2912     it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
2913     the letter 'x'. On long subject strings, this gives a significant speed-up.
2914    
2915    
2916 nigel 33 Version 2.05 21-Apr-99
2917     ----------------------
2918    
2919     1. Changed the type of magic_number from int to long int so that it works
2920     properly on 16-bit systems.
2921    
2922     2. Fixed a bug which caused patterns starting with .* not to work correctly
2923     when the subject string contained newline characters. PCRE was assuming
2924     anchoring for such patterns in all cases, which is not correct because .* will
2925     not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
2926     DOTALL is set at top level; otherwise it knows that patterns starting with .*
2927     must be retried after every newline in the subject.
2928    
2929    
2930 nigel 31 Version 2.04 18-Feb-99
2931     ----------------------
2932    
2933     1. For parenthesized subpatterns with repeats whose minimum was zero, the
2934     computation of the store needed to hold the pattern was incorrect (too large).
2935     If such patterns were nested a few deep, this could multiply and become a real
2936     problem.
2937    
2938     2. Added /M option to pcretest to show the memory requirement of a specific
2939     pattern. Made -m a synonym of -s (which does this globally) for compatibility.
2940    
2941     3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
2942     compiled in such a way that the backtracking after subsequent failure was
2943     pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
2944     ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
2945    
2946    
2947 nigel 29 Version 2.03 02-Feb-99
2948     ----------------------
2949    
2950     1. Fixed typo and small mistake in man page.
2951    
2952 nigel 31 2. Added 4th condition (GPL supersedes if conflict) and created separate
2953     LICENCE file containing the conditions.
2954 nigel 29
2955     3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
2956     Perl, that is the internal \ allows the delimiter to be included in the
2957     pattern. Locked out the use of \ as a delimiter. If \ immediately follows
2958     the final delimiter, add \ to the end of the pattern (to test the error).
2959    
2960     4. Added the convenience functions for extracting substrings after a successful
2961     match. Updated pcretest to make it able to test these functions.
2962    
2963    
2964 nigel 27 Version 2.02 14-Jan-99
2965     ----------------------
2966    
2967     1. Initialized the working variables associated with each extraction so that
2968     their saving and restoring doesn't refer to uninitialized store.
2969    
2970     2. Put dummy code into study.c in order to trick the optimizer of the IBM C
2971     compiler for OS/2 into generating correct code. Apparently IBM isn't going to
2972     fix the problem.
2973    
2974     3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
2975     calls, and wasn't printing the correct value for compiling calls. Increased the
2976     default value of LOOPREPEAT, and the number of significant figures in the
2977     times.
2978    
2979     4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
2980    
2981     5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
2982     a building problem on Windows NT with a FAT file system.
2983    
2984    
2985 nigel 25 Version 2.01 21-Oct-98
2986     ----------------------
2987    
2988     1. Changed the API for pcre_compile() to allow for the provision of a pointer
2989     to character tables built by pcre_maketables() in the current locale. If NULL
2990     is passed, the default tables are used.
2991    
2992    
2993 nigel 23 Version 2.00 24-Sep-98
2994 nigel 21 ----------------------
2995    
2996 nigel 23 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
2997     it any more.
2998    
2999     2. Allow quantification of (?>) groups, and make it work correctly.
3000    
3001     3. The first character computation wasn't working for (?>) groups.
3002    
3003     4. Correct the implementation of \Z (it is permitted to match on the \n at the
3004     end of the subject) and add 5.005's \z, which really does match only at the
3005     very end of the subject.
3006    
3007     5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
3008    
3009     6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
3010     DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
3011     localized options. All options to pcre_study() were also removed.
3012    
3013     7. Add other new features from 5.005:
3014    
3015     $(?<= positive lookbehind
3016     $(?<! negative lookbehind
3017     (?imsx-imsx) added the unsetting capability
3018     such a setting is global if at outer level; local otherwise
3019     (?imsx-imsx:) non-capturing groups with option setting
3020     (?(cond)re|re) conditional pattern matching
3021    
3022     A backreference to itself in a repeated group matches the previous
3023     captured string.
3024    
3025     8. General tidying up of studying (both automatic and via "study")
3026     consequential on the addition of new assertions.
3027    
3028     9. As in 5.005, unlimited repeated groups that could match an empty substring
3029     are no longer faulted at compile time. Instead, the loop is forcibly broken at
3030     runtime if any iteration does actually match an empty substring.
3031    
3032     10. Include the RunTest script in the distribution.
3033    
3034     11. Added tests from the Perl 5.005_02 distribution. This showed up a few
3035     discrepancies, some of which were old and were also with respect to 5.004. They
3036     have now been fixed.
3037    
3038    
3039     Version 1.09 28-Apr-98
3040     ----------------------
3041    
3042 nigel 21 1. A negated single character class followed by a quantifier with a minimum
3043     value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
3044     program crashes, or just wrong answers. This did not apply to negated classes
3045     containing more than one character, or to minima other than one.
3046    
3047    
3048 nigel 19 Version 1.08 27-Mar-98
3049     ----------------------
3050    
3051     1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
3052    
3053     2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
3054     latter must appear before anything that relies on it in the pattern.
3055    
3056    
3057 nigel 17 Version 1.07 16-Feb-98
3058     ----------------------
3059    
3060     1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
3061     repeat of a potentially empty string).
3062    
3063    
3064 nigel 15 Version 1.06 23-Jan-98
3065     ----------------------
3066    
3067     1. Added Markus Oberhumer's little patches for C++.
3068    
3069     2. Literal strings longer than 255 characters were broken.
3070    
3071    
3072 nigel 13 Version 1.05 23-Dec-97
3073     ----------------------
3074    
3075     1. Negated character classes containing more than one character were failing if
3076     PCRE_CASELESS was set at run time.
3077    
3078    
3079 nigel 11 Version 1.04 19-Dec-97
3080     ----------------------
3081    
3082     1. Corrected the man page, where some "const" qualifiers had been omitted.
3083    
3084     2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
3085     input syntax.
3086    
3087     3. Fixed memory leak which occurred when a regex with back references was
3088     matched with an offsets vector that wasn't big enough. The temporary memory
3089     that is used in this case wasn't being freed if the match failed.
3090    
3091     4. Tidied pcretest to ensure it frees memory that it gets.
3092    
3093     5. Temporary memory was being obtained in the case where the passed offsets
3094     vector was exactly big enough.
3095    
3096     6. Corrected definition of offsetof() from change 5 below.
3097    
3098     7. I had screwed up change 6 below and broken the rules for the use of
3099     setjmp(). Now fixed.
3100    
3101    
3102 nigel 9 Version 1.03 18-Dec-97
3103     ----------------------
3104    
3105     1. A erroneous regex with a missing opening parenthesis was correctly
3106     diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
3107     on some systems.
3108    
3109     2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
3110     it was reported that one broken compiler failed on the former because "code" is
3111     also an independent variable.
3112    
3113     3. The erroneous regex a[]b caused an array overrun reference.
3114    
3115     4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
3116     fail on data ending with that character. (It was going on too far, and checking
3117     the next character, typically a binary zero.) This was specific to the
3118     optimized code for single-character negative classes.
3119    
3120     5. Added a contributed patch from the TIN world which does the following:
3121    
3122     + Add an undef for memmove, in case the the system defines a macro for it.
3123    
3124     + Add a definition of offsetof(), in case there isn't one. (I don't know
3125     the reason behind this - offsetof() is part of the ANSI standard - but
3126     it does no harm).
3127    
3128     + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
3129     most of the places where whitespace preceded '#'. I have given up and
3130     allowed the remaining 2 cases to be at the margin.
3131    
3132     + Rename some variables in pcre to eliminate shadowing. This seems very
3133     pedantic, but does no harm, of course.
3134    
3135     6. Moved the call to setjmp() into its own function, to get rid of warnings
3136     from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
3137    
3138     7. Constructs such as \d{8,} were compiling into the equivalent of
3139 nigel 11 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
3140 nigel 9 outcome, but in this particular case used more store than had been allocated,
3141     which caused the bug to be discovered because it threw up an internal error.
3142    
3143     8. The debugging code in both pcre and pcretest for outputting the compiled
3144     form of a regex was going wrong in the case of back references followed by
3145     curly-bracketed repeats.
3146    
3147    
3148 nigel 7 Version 1.02 12-Dec-97
3149     ----------------------
3150    
3151     1. Typos in pcre.3 and comments in the source fixed.
3152    
3153     2. Applied a contributed patch to get rid of places where it used to remove
3154     'const' from variables, and fixed some signed/unsigned and uninitialized
3155     variable warnings.
3156    
3157     3. Added the "runtest" target to Makefile.
3158    
3159     4. Set default compiler flag to -O2 rather than just -O.
3160    
3161    
3162 nigel 5 Version 1.01 19-Nov-97
3163     ----------------------
3164    
3165     1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
3166     like /([ab]*)*/, that is, for classes with more than one character in them.
3167    
3168     2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
3169     as /((?>a*))*/ (a PCRE_EXTRA facility).
3170    
3171    
3172     Version 1.00 18-Nov-97
3173     ----------------------
3174    
3175     1. Added compile-time macros to support systems such as SunOS4 which don't have
3176     memmove() or strerror() but have other things that can be used instead.
3177    
3178     2. Arranged that "make clean" removes the executables.
3179    
3180    
3181 nigel 3 Version 0.99 27-Oct-97
3182     ----------------------
3183    
3184     1. Fixed bug in code for optimizing classes with only one character. It was
3185     initializing a 32-byte map regardless, which could cause it to run off the end
3186     of the memory it had got.
3187    
3188     2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
3189    
3190    
3191     Version 0.98 22-Oct-97
3192     ----------------------
3193    
3194     1. Fixed bug in code for handling temporary memory usage when there are more
3195     back references than supplied space in the ovector. This could cause segfaults.
3196    
3197    
3198     Version 0.97 21-Oct-97
3199     ----------------------
3200    
3201     1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
3202    
3203     2. Optimized negated single characters not to use a bit map.
3204    
3205     3. Brought error texts together as macro definitions; clarified some of them;
3206     fixed one that was wrong - it said "range out of order" when it meant "invalid
3207     escape sequence".
3208    
3209     4. Changed some char * arguments to const char *.
3210    
3211     5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
3212    
3213     6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
3214     pcretest.
3215    
3216    
3217     Version 0.96 16-Oct-97
3218     ----------------------
3219    
3220     1. Added a simple "pgrep" utility to the distribution.
3221    
3222     2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
3223     unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
3224     where "ddd" means "one or more decimal digits".
3225    
3226     3. Fixed serious bug. If a pattern had a back reference, but the call to
3227     pcre_exec() didn't supply a large enough ovector to record the related
3228     identifying subpattern, the match always failed. PCRE now remembers the number
3229     of the largest back reference, and gets some temporary memory in which to save
3230     the offsets during matching if necessary, in order to ensure that
3231     backreferences always work.
3232    
3233     4. Increased the compatibility with Perl in a number of ways:
3234    
3235     (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
3236     to request this handling. The option can be set at compile or exec time.
3237    
3238     (b) $ matches before a terminating newline by default; an option
3239     PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
3240     mode). The option can be set at compile or exec time.
3241    
3242     (c) The handling of \ followed by a digit other than 0 is now supposed to be
3243     the same as Perl's. If the decimal number it represents is less than 10
3244     or there aren't that many previous left capturing parentheses, an octal
3245     escape is read. Inside a character class, it's always an octal escape,
3246     even if it is a single digit.
3247    
3248     (d) An escaped but undefined alphabetic character is taken as a literal,
3249     unless PCRE_EXTRA is set. Currently this just reserves the remaining
3250     escapes.
3251    
3252     (e) {0} is now permitted. (The previous item is removed from the compiled
3253     pattern).
3254    
3255     5. Changed all the names of code files so that the basic parts are no longer
3256     than 10 characters, and abolished the teeny "globals.c" file.
3257    
3258     6. Changed the handling of character classes; they are now done with a 32-byte
3259     bit map always.
3260    
3261     7. Added the -d and /D options to pcretest to make it possible to look at the
3262     internals of compilation without having to recompile pcre.
3263    
3264    
3265     Version 0.95 23-Sep-97
3266     ----------------------
3267    
3268     1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
3269     \x20 at the start of a run of normal characters. These were being treated as
3270     real characters, instead of the source characters being re-checked.
3271    
3272    
3273     Version 0.94 18-Sep-97
3274     ----------------------
3275    
3276     1. The functions are now thread-safe, with the caveat that the global variables
3277     containing pointers to malloc() and free() or alternative functions are the
3278     same for all threads.
3279    
3280     2. Get pcre_study() to generate a bitmap of initial characters for non-
3281     anchored patterns when this is possible, and use it if passed to pcre_exec().
3282    
3283    
3284     Version 0.93 15-Sep-97
3285     ----------------------
3286    
3287     1. /(b)|(:+)/ was computing an incorrect first character.
3288    
3289     2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
3290     but not actually doing anything yet.
3291    
3292     3. Treat "-" characters in classes that cannot be part of ranges as literals,
3293     as Perl does (e.g. [-az] or [az-]).
3294    
3295     4. Set the anchored flag if a branch starts with .* or .*? because that tests
3296     all possible positions.
3297    
3298     5. Split up into different modules to avoid including unneeded functions in a
3299     compiled binary. However, compile and exec are still in one module. The "study"
3300     function is split off.
3301    
3302     6. The character tables are now in a separate module whose source is generated
3303     by an auxiliary program - but can then be edited by hand if required. There are
3304     now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
3305     toupper() in the code.
3306    
3307     7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
3308     make them global. Abolish the function for setting them, as the caller can now
3309     set them directly.
3310    
3311    
3312     Version 0.92 11-Sep-97
3313     ----------------------
3314    
3315     1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
3316     (e.g. /a{1,3}/) was broken (I mis-optimized it).
3317    
3318     2. Caseless matching was not working in character classes if the characters in
3319     the pattern were in upper case.
3320    
3321     3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
3322    
3323     4. Make PCRE_ANCHORED public and accept as a compile option.
3324    
3325     5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
3326     PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
3327     pass them.
3328    
3329     6. Give an error if bad option bits passed at compile or run time.
3330    
3331     7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
3332     pcretest to cause it to pass that flag.
3333    
3334     8. Add pcre_info(), to get the number of identifying subpatterns, the stored
3335     options, and the first character, if set.
3336    
3337     9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
3338    
3339    
3340     Version 0.91 10-Sep-97
3341     ----------------------
3342    
3343     1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
3344     match the empty string as in /(a*)*/. It was looping and ultimately crashing.
3345    
3346     2. PCRE was looping on encountering an indefinitely repeated back reference to
3347     a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
3348     Perl does - treats the match as successful.
3349    
3350     ****

Properties

Name Value
svn:eol-style native
svn:keywords "Author Date Id Revision Url"

webmaster@exim.org
ViewVC Help
Powered by ViewVC 1.1.12