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

Diff of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log | View Patch Patch

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

Legend:
Removed from v.31  
changed lines
  Added in v.265

webmaster@exim.org
ViewVC Help
Powered by ViewVC 1.1.12