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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 206 - (hide annotations) (download)
Fri Aug 3 14:53:04 2007 UTC (6 years, 8 months ago) by ph10
File size: 134642 byte(s)
Fix loop for null-matching condition nested in an outer unlimited repeat.

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

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