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

Diff of /code/trunk/ChangeLog

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

revision 691 by ph10, Sun Sep 11 14:31:21 2011 UTC revision 989 by zherczeg, Sat Jul 7 11:11:02 2012 UTC
# Line 1  Line 1 
1  ChangeLog for PCRE  ChangeLog for PCRE
2  ------------------  ------------------
3    
4  Version 8.20 12-Sep-2011  Version 8.32
5    ------------
6    
7    1.  Improved JIT compiler optimizations for first character search and single
8        character iterators.
9    
10    2.  Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
11        Patch by Daniel Richard G.
12    
13    3.  Single character iterator optimizations in the JIT compiler.
14    
15    
16    Version 8.31 06-July-2012
17    -------------------------
18    
19    1.  Fixing a wrong JIT test case and some compiler warnings.
20    
21    2.  Removed a bashism from the RunTest script.
22    
23    3.  Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
24        to unsigned type, result still unsigned" that was given by an MS compiler
25        on encountering the code "-sizeof(xxx)".
26    
27    4.  Partial matching support is added to the JIT compiler.
28    
29    5.  Fixed several bugs concerned with partial matching of items that consist
30        of more than one character:
31    
32        (a) /^(..)\1/ did not partially match "aba" because checking references was
33            done on an "all or nothing" basis. This also applied to repeated
34            references.
35    
36        (b) \R did not give a hard partial match if \r was found at the end of the
37            subject.
38    
39        (c) \X did not give a hard partial match after matching one or more
40            characters at the end of the subject.
41    
42        (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
43            a partial match for the string "\r".
44    
45        (e) When newline was set to CRLF, the metacharacter "." did not recognize
46            a partial match for a CR character at the end of the subject string.
47    
48    6.  If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
49        running pcretest, the text "(JIT)" added to the output whenever JIT is
50        actually used to run the match.
51    
52    7.  Individual JIT compile options can be set in pcretest by following -s+[+]
53        or /S+[+] with a digit between 1 and 7.
54    
55    8.  OP_NOT now supports any UTF character not just single-byte ones.
56    
57    9.  (*MARK) control verb is now supported by the JIT compiler.
58    
59    10. The command "./RunTest list" lists the available tests without actually
60        running any of them. (Because I keep forgetting what they all are.)
61    
62    11. Add PCRE_INFO_MAXLOOKBEHIND.
63    
64    12. Applied a (slightly modified) user-supplied patch that improves performance
65        when the heap is used for recursion (compiled with --disable-stack-for-
66        recursion). Instead of malloc and free for each heap frame each time a
67        logical recursion happens, frames are retained on a chain and re-used where
68        possible. This sometimes gives as much as 30% improvement.
69    
70    13. As documented, (*COMMIT) is now confined to within a recursive subpattern
71        call.
72    
73    14. As documented, (*COMMIT) is now confined to within a positive assertion.
74    
75    15. It is now possible to link pcretest with libedit as an alternative to
76        libreadline.
77    
78    16. (*COMMIT) control verb is now supported by the JIT compiler.
79    
80    17. The Unicode data tables have been updated to Unicode 6.1.0.
81    
82    18. Added --file-list option to pcregrep.
83    
84    19. Added binary file support to pcregrep, including the -a, --binary-files,
85        -I, and --text options.
86    
87    20. The madvise function is renamed for posix_madvise for QNX compatibility
88        reasons. Fixed by Giuseppe D'Angelo.
89    
90    21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
91    
92    22. Changed the option for creating symbolic links for 16-bit man pages from
93        -s to -sf so that re-installing does not cause issues.
94    
95    23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
96    
97    24. Fixed a very old bug in pcretest that caused errors with restarted DFA
98        matches in certain environments (the workspace was not being correctly
99        retained). Also added to pcre_dfa_exec() a simple plausibility check on
100        some of the workspace data at the beginning of a restart.
101    
102    25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
103        was not doing so when it should - probably a typo introduced by SVN 528
104        (change 8.10/14).
105    
106    26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
107        \w+ when the character tables indicated that \x{c4} was a word character.
108        There were several related cases, all because the tests for doing a table
109        lookup were testing for characters less than 127 instead of 255.
110    
111    27. If a pattern contains capturing parentheses that are not used in a match,
112        their slots in the ovector are set to -1. For those that are higher than
113        any matched groups, this happens at the end of processing. In the case when
114        there were back references that the ovector was too small to contain
115        (causing temporary malloc'd memory to be used during matching), and the
116        highest capturing number was not used, memory off the end of the ovector
117        was incorrectly being set to -1. (It was using the size of the temporary
118        memory instead of the true size.)
119    
120    28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
121        ovector size, it uses memory at the end of the block that it has got.
122    
123    29. Check for an overlong MARK name and give an error at compile time. The
124        limit is 255 for the 8-bit library and 65535 for the 16-bit library.
125    
126    30. JIT compiler update.
127    
128    31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
129        Rill for the patch.
130    
131    32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
132    
133    33. Variable renamings in the PCRE-JIT compiler. No functionality change.
134    
135    34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
136        SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
137        was enabled.
138    
139    35. Improve JIT code generation for greedy plus quantifier.
140    
141    36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
142        1 to "aa" instead of to an empty string. The bug affected repeated groups
143        that could potentially match an empty string.
144    
145    37. Optimizing single character iterators in JIT.
146    
147    38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
148        the same checks as \x{...} characters in non-JavaScript mode. Specifically,
149        codepoints that are too big for the mode are faulted, and in a UTF mode,
150        disallowed codepoints are also faulted.
151    
152    39. If PCRE was compiled with UTF support, in three places in the DFA
153        matcher there was code that should only have been obeyed in UTF mode, but
154        was being obeyed unconditionally. In 8-bit mode this could cause incorrect
155        processing when bytes with values greater than 127 were present. In 16-bit
156        mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
157        both cases the values are those that cannot be the first data item in a UTF
158        character. The three items that might have provoked this were recursions,
159        possessively repeated groups, and atomic groups.
160    
161    40. Ensure that libpcre is explicitly listed in the link commands for pcretest
162        and pcregrep, because some OS require shared objects to be explicitly
163        passed to ld, causing the link step to fail if they are not.
164    
165    41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
166        mode, patterns that started with \h* or \R* might be incorrectly matched.
167    
168    
169    Version 8.30 04-February-2012
170    -----------------------------
171    
172    1.  Renamed "isnumber" as "is_a_number" because in some Mac environments this
173        name is defined in ctype.h.
174    
175    2.  Fixed a bug in fixed-length calculation for lookbehinds that would show up
176        only in quite long subpatterns.
177    
178    3.  Removed the function pcre_info(), which has been obsolete and deprecated
179        since it was replaced by pcre_fullinfo() in February 2000.
180    
181    4.  For a non-anchored pattern, if (*SKIP) was given with a name that did not
182        match a (*MARK), and the match failed at the start of the subject, a
183        reference to memory before the start of the subject could occur. This bug
184        was introduced by fix 17 of release 8.21.
185    
186    5.  A reference to an unset group with zero minimum repetition was giving
187        totally wrong answers (in non-JavaScript-compatibility mode). For example,
188        /(another)?(\1?)test/ matched against "hello world test". This bug was
189        introduced in release 8.13.
190    
191    6.  Add support for 16-bit character strings (a large amount of work involving
192        many changes and refactorings).
193    
194    7.  RunGrepTest failed on msys because \r\n was replaced by whitespace when the
195        command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
196        from a file.
197    
198    8.  Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
199        rounding is not applied in this particular case).
200    
201    9.  The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
202        if they appear, or are escaped, in patterns.
203    
204    10. Get rid of a number of -Wunused-but-set-variable warnings.
205    
206    11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
207        "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
208        Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
209        also returns the mark "x". This bug applied to capturing parentheses,
210        non-capturing parentheses, and atomic parentheses. It also applied to some
211        assertions.
212    
213    12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
214        information out of configure.ac instead of relying on pcre.h.generic, which
215        is not stored in the repository.
216    
217    13. Applied Dmitry V. Levin's patch for a more portable method for linking with
218        -lreadline.
219    
220    14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
221    
222    15. Applied Graycode's patch to put the top-level frame on the stack rather
223        than the heap when not using the stack for recursion. This gives a
224        performance improvement in many cases when recursion is not deep.
225    
226    16. Experimental code added to "pcretest -C" to output the stack frame size.
227    
228    
229    Version 8.21 12-Dec-2011
230    ------------------------
231    
232    1.  Updating the JIT compiler.
233    
234    2.  JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
235        are added as well.
236    
237    3.  Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
238        PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
239        calling _pcre_jit_exec. Some extra comments are added.
240    
241    4.  (*MARK) settings inside atomic groups that do not contain any capturing
242        parentheses, for example, (?>a(*:m)), were not being passed out. This bug
243        was introduced by change 18 for 8.20.
244    
245    5.  Supporting of \x, \U and \u in JavaScript compatibility mode based on the
246        ECMA-262 standard.
247    
248    6.  Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
249        erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
250        This bug was probably introduced by change 9 of 8.13.
251    
252    7.  While fixing 6 above, I noticed that a number of other items were being
253        incorrectly rejected as "not fixed length". This arose partly because newer
254        opcodes had not been added to the fixed-length checking code. I have (a)
255        corrected the bug and added tests for these items, and (b) arranged for an
256        error to occur if an unknown opcode is encountered while checking for fixed
257        length instead of just assuming "not fixed length". The items that were
258        rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
259        (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
260        repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
261    
262    8.  A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
263        being incorrectly compiled and would have given unpredicatble results.
264    
265    9.  A possessively repeated subpattern with minimum repeat count greater than
266        one behaved incorrectly. For example, (A){2,}+ behaved as if it was
267        (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
268        the first (A) could occur when it should not.
269    
270    10. Add a cast and remove a redundant test from the code.
271    
272    11. JIT should use pcre_malloc/pcre_free for allocation.
273    
274    12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
275        best practice nowadays, and helps with cross-compiling. (If the exec_prefix
276        is anything other than /usr, -L is still shown).
277    
278    13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
279    
280    14. Perl does not support \N without a following name in a [] class; PCRE now
281        also gives an error.
282    
283    15. If a forward reference was repeated with an upper limit of around 2000,
284        it caused the error "internal error: overran compiling workspace". The
285        maximum number of forward references (including repeats) was limited by the
286        internal workspace, and dependent on the LINK_SIZE. The code has been
287        rewritten so that the workspace expands (via pcre_malloc) if necessary, and
288        the default depends on LINK_SIZE. There is a new upper limit (for safety)
289        of around 200,000 forward references. While doing this, I also speeded up
290        the filling in of repeated forward references.
291    
292    16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
293        incorrectly expecting the subject to contain another "a" after the start.
294    
295    17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
296        in the match, the SKIP should be ignored. This was not happening; instead
297        the SKIP was being treated as NOMATCH. For patterns such as
298        /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
299        tested.
300    
301    18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
302        now much more compatible with Perl, in particular in cases where the result
303        is a non-match for a non-anchored pattern. For example, if
304        /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
305        "m", where previously it did not return a name. A side effect of this
306        change is that for partial matches, the last encountered mark name is
307        returned, as for non matches. A number of tests that were previously not
308        Perl-compatible have been moved into the Perl-compatible test files. The
309        refactoring has had the pleasing side effect of removing one argument from
310        the match() function, thus reducing its stack requirements.
311    
312    19. If the /S+ option was used in pcretest to study a pattern using JIT,
313        subsequent uses of /S (without +) incorrectly behaved like /S+.
314    
315    21. Retrieve executable code size support for the JIT compiler and fixing
316        some warnings.
317    
318    22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
319        not work when the shorter character appeared right at the end of the
320        subject string.
321    
322    23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
323        systems.
324    
325    24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
326        output it when the /M option is used in pcretest.
327    
328    25. The CheckMan script was not being included in the distribution. Also, added
329        an explicit "perl" to run Perl scripts from the PrepareRelease script
330        because this is reportedly needed in Windows.
331    
332    26. If study data was being save in a file and studying had not found a set of
333        "starts with" bytes for the pattern, the data written to the file (though
334        never used) was taken from uninitialized memory and so caused valgrind to
335        complain.
336    
337    27. Updated RunTest.bat as provided by Sheri Pierce.
338    
339    28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
340    
341    29. Computation of memory usage for the table of capturing group names was
342        giving an unnecessarily large value.
343    
344    
345    Version 8.20 21-Oct-2011
346  ------------------------  ------------------------
347    
348  1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had  1.  Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
349     a POSIX class. After further experiments with Perl, which convinced me that      a POSIX class. After further experiments with Perl, which convinced me that
350     Perl has bugs and confusions, a closing square bracket is no longer allowed      Perl has bugs and confusions, a closing square bracket is no longer allowed
351     in a POSIX name. This bug also affected patterns with classes that started      in a POSIX name. This bug also affected patterns with classes that started
352     with full stops.      with full stops.
353    
354  2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no captured  2.  If a pattern such as /(a)b|ac/ is matched against "ac", there is no
355     substring, but while checking the failing first alternative, substring 1 is      captured substring, but while checking the failing first alternative,
356     temporarily captured. If the output vector supplied to pcre_exec() was not      substring 1 is temporarily captured. If the output vector supplied to
357     big enough for this capture, the yield of the function was still zero      pcre_exec() was not big enough for this capture, the yield of the function
358     ("insufficient space for captured substrings"). This cannot be totally fixed      was still zero ("insufficient space for captured substrings"). This cannot
359     without adding another stack variable, which seems a lot of expense for a      be totally fixed without adding another stack variable, which seems a lot
360     edge case. However, I have improved the situation in cases such as      of expense for a edge case. However, I have improved the situation in cases
361     /(a)(b)x|abc/ matched against "abc", where the return code indicates that      such as /(a)(b)x|abc/ matched against "abc", where the return code
362     fewer than the maximum number of slots in the ovector have been set.      indicates that fewer than the maximum number of slots in the ovector have
363        been set.
364  3. Related to (2) above: when there are more back references in a pattern than  
365     slots in the output vector, pcre_exec() uses temporary memory during  3.  Related to (2) above: when there are more back references in a pattern than
366     matching, and copies in the captures as far as possible afterwards. It was      slots in the output vector, pcre_exec() uses temporary memory during
367     using the entire output vector, but this conflicts with the specification      matching, and copies in the captures as far as possible afterwards. It was
368     that only 2/3 is used for passing back captured substrings. Now it uses only      using the entire output vector, but this conflicts with the specification
369     the first 2/3, for compatibility. This is, of course, another edge case.      that only 2/3 is used for passing back captured substrings. Now it uses
370        only the first 2/3, for compatibility. This is, of course, another edge
371  4. Zoltan Herczeg's just-in-time compiler support has been integrated into the      case.
372     main code base, and can be used by building with --enable-jit. When this is  
373     done, pcregrep automatically uses it unless --disable-pcregrep-jit or the  4.  Zoltan Herczeg's just-in-time compiler support has been integrated into the
374     runtime --no-jit option is given.      main code base, and can be used by building with --enable-jit. When this is
375        done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
376  5. When the number of matches in a pcre_dfa_exec() run exactly filled the      runtime --no-jit option is given.
377     ovector, the return from the function was zero, implying that there were  
378     other matches that did not fit. The correct "exactly full" value is now  5.  When the number of matches in a pcre_dfa_exec() run exactly filled the
379     returned.      ovector, the return from the function was zero, implying that there were
380        other matches that did not fit. The correct "exactly full" value is now
381  6. If a subpattern that was called recursively or as a subroutine contained      returned.
382     (*PRUNE) or any other control that caused it to give a non-standard return,  
383     invalid errors such as "Error -26 (nested recursion at the same subject  6.  If a subpattern that was called recursively or as a subroutine contained
384     position)" or even infinite loops could occur.      (*PRUNE) or any other control that caused it to give a non-standard return,
385        invalid errors such as "Error -26 (nested recursion at the same subject
386        position)" or even infinite loops could occur.
387    
388    7.  If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
389        computing the minimum length on reaching *ACCEPT, and so ended up with the
390        wrong value of 1 rather than 0. Further investigation indicates that
391        computing a minimum subject length in the presence of *ACCEPT is difficult
392        (think back references, subroutine calls), and so I have changed the code
393        so that no minimum is registered for a pattern that contains *ACCEPT.
394    
395    8.  If (*THEN) was present in the first (true) branch of a conditional group,
396        it was not handled as intended. [But see 16 below.]
397    
398    9.  Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
399        Sheri Pierce.
400    
401    10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
402        the first byte in a match must be "a".
403    
404    11. Change 17 for 8.13 increased the recursion depth for patterns like
405        /a(?:.)*?a/ drastically. I've improved things by remembering whether a
406        pattern contains any instances of (*THEN). If it does not, the old
407        optimizations are restored. It would be nice to do this on a per-group
408        basis, but at the moment that is not feasible.
409    
410    12. In some environments, the output of pcretest -C is CRLF terminated. This
411        broke RunTest's code that checks for the link size. A single white space
412        character after the value is now allowed for.
413    
414    13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
415        For "fr", it uses the Windows-specific input and output files.
416    
417    14. If (*THEN) appeared in a group that was called recursively or as a
418        subroutine, it did not work as intended. [But see next item.]
419    
420    15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
421        pattern fragments (but not containing any | characters). If A and B are
422        matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
423        was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
424        D. In other words, Perl considers parentheses that do not contain any |
425        characters to be part of a surrounding alternative, whereas PCRE was
426        treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
427        differently. PCRE now behaves in the same way as Perl, except in the case
428        of subroutine/recursion calls such as (?1) which have in any case always
429        been different (but PCRE had them first :-).
430    
431    16. Related to 15 above: Perl does not treat the | in a conditional group as
432        creating alternatives. Such a group is treated in the same way as an
433        ordinary group without any | characters when processing (*THEN). PCRE has
434        been changed to match Perl's behaviour.
435    
436    17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
437        RunGrepTest script failed.
438    
439    18. Change 22 for version 13 caused atomic groups to use more stack. This is
440        inevitable for groups that contain captures, but it can lead to a lot of
441        stack use in large patterns. The old behaviour has been restored for atomic
442        groups that do not contain any capturing parentheses.
443    
444    19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
445        suppress the check for a minimum subject length at run time. (If it was
446        given to pcre_exec() or pcre_dfa_exec() it did work.)
447    
448    20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
449        fail to work when decoding hex characters in data strings in EBCDIC
450        environments.
451    
452    21. It appears that in at least one Mac OS environment, the isxdigit() function
453        is implemented as a macro that evaluates to its argument more than once,
454        contravening the C 90 Standard (I haven't checked a later standard). There
455        was an instance in pcretest which caused it to go wrong when processing
456        \x{...} escapes in subject strings. The has been rewritten to avoid using
457        things like p++ in the argument of isxdigit().
458    
459    
460  Version 8.13 16-Aug-2011  Version 8.13 16-Aug-2011
# Line 140  Version 8.13 16-Aug-2011 Line 554  Version 8.13 16-Aug-2011
554      tail recursion to cut down on stack usage. Unfortunately, now that there is      tail recursion to cut down on stack usage. Unfortunately, now that there is
555      the possibility of (*THEN) occurring in these branches, tail recursion is      the possibility of (*THEN) occurring in these branches, tail recursion is
556      no longer possible because the return has to be checked for (*THEN). These      no longer possible because the return has to be checked for (*THEN). These
557      two optimizations have therefore been removed.      two optimizations have therefore been removed. [But see 8.20/11 above.]
558    
559  18. If a pattern containing \R was studied, it was assumed that \R always  18. If a pattern containing \R was studied, it was assumed that \R always
560      matched two bytes, thus causing the minimum subject length to be      matched two bytes, thus causing the minimum subject length to be

Legend:
Removed from v.691  
changed lines
  Added in v.989

  ViewVC Help
Powered by ViewVC 1.1.5