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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 261 - (hide annotations) (download)
Fri Sep 21 08:37:48 2007 UTC (7 years, 3 months ago) by ph10
File size: 141931 byte(s)
Final doc updates and file tidies for 7.4.

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

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