diff options
author | dkf <donal.k.fellows@manchester.ac.uk> | 2010-02-01 11:12:00 (GMT) |
---|---|---|
committer | dkf <donal.k.fellows@manchester.ac.uk> | 2010-02-01 11:12:00 (GMT) |
commit | f231dc527e35148e1fb38301a10ccb8614fd1c72 (patch) | |
tree | 39bc26fb3acf944bf541192fef351f5c0a327c73 /ChangeLog | |
parent | 31c2cd054d791e99d70c6ff41a6c160142229d1c (diff) | |
download | tcl-f231dc527e35148e1fb38301a10ccb8614fd1c72.zip tcl-f231dc527e35148e1fb38301a10ccb8614fd1c72.tar.gz tcl-f231dc527e35148e1fb38301a10ccb8614fd1c72.tar.bz2 |
[Bug 2942697]: Rework the RE engine so that certain pathological patterns are
matched much more rapidly. Many thanks to Tom Lane for dianosing this issue
and providing an initial patch.
Diffstat (limited to 'ChangeLog')
-rw-r--r-- | ChangeLog | 169 |
1 files changed, 87 insertions, 82 deletions
@@ -1,30 +1,38 @@ +2010-02-01 Donal K. Fellows <dkf@users.sf.net> + + * generic/regexec.c (ccondissect, crevdissect): [Bug 2942697]: Rework + these functions so that certain pathological patterns are matched much + more rapidly. Many thanks to Tom Lane for dianosing this issue and + providing an initial patch. + 2010-02-01 Jan Nijtmans <nijtmans@users.sf.net> - * generic/tclInt.decls Various CYGWIN-related fixes - * generic/tclInt.h backported from HEAD. Still - * generic/tclIntPlatDecls.h configure script not modified, - * generic/tclPort.h so CYGWIN build is still - * generic/tclTest/c disabled. Reason: although the - * win/cat.c build succeeds with those changes, - * win/tclWinDde.c many tests still fail. - * win/tclWinError.c - * win/tclWinFile.c - * win/tclWinPipe.c - * win/tclWinPort.h - * win/tclWinReg.c - * win/tclWinSerial.c - * win/tclWinSock.c - * win/tclWinTest.c - * win/tclWinThrd.c + * generic/tclInt.decls: Various CYGWIN-related fixes + * generic/tclInt.h: backported from HEAD. Still + * generic/tclIntPlatDecls.h: configure script not modified, + * generic/tclPort.h: so CYGWIN build is still + * generic/tclTest.c: disabled. Reason: although the + * win/cat.c: build succeeds with those changes, + * win/tclWinDde.c: many tests still fail. + * win/tclWinError.c: + * win/tclWinFile.c: + * win/tclWinPipe.c: + * win/tclWinPort.h: + * win/tclWinReg.c: + * win/tclWinSerial.c: + * win/tclWinSock.c: + * win/tclWinTest.c: + * win/tclWinThrd.c: 2010-01-29 Jan Nijtmans <nijtmans@users.sf.net> - * generic/tcl.h Use correct TCL_LL_MODIFIER for CYGWIN. - Formatting (all backported from HEAD) - * generic/rege_dfa.c: Fix macro conflict on CYGWIN: don't use "small". - * generic/tclTest.c: Fix gcc 4.4 warning: ignoring return value of - * unix/tclUnixPipe.c ‘write’ - * unix/tclUnixNotify.c + * generic/tcl.h: Use correct TCL_LL_MODIFIER for CYGWIN. + Formatting (all backported from HEAD) + * generic/rege_dfa.c: Fix macro conflict on CYGWIN: don't use + "small". + * generic/tclTest.c: Fix gcc 4.4 warning: ignoring return value of + * unix/tclUnixPipe.c: ‘write’ + * unix/tclUnixNotify.c: 2010-01-19 Donal K. Fellows <dkf@users.sf.net> @@ -47,7 +55,6 @@ internal representation when not needed. Thanks to Alexandre Ferrieux for this fix. ->>>>>>> 1.3975.2.291 2010-01-06 Jan Nijtmans <nijtmans@users.sf.net> * generic/tclCompExpr.c: Warning: array subscript has type 'char' @@ -55,22 +62,22 @@ * libtommath/bn_mp_read_radix.c: * unix/tclUnixCompat.c: Fix gcc warning: signed and unsigned type in conditional expression. - * unix/tcl.m4 Add support for Haiku and CYGWIN dynamical loading - * unix/configure (regenerated) - * unix/Makefile.in - * unix/.cvsignore - * tests/stack.test Reduced minimum required C-stack size to 2034: - CYGWIN has this stack size and the test runs fine! - * generic/tclEnv.c Fix environment tests under CYGWIN - * generic/tclPort.h - * tests/env.test + * unix/tcl.m4: Add support for Haiku and CYGWIN dynamical loading + * unix/configure: (regenerated) + * unix/Makefile.in: + * unix/.cvsignore: + * tests/stack.test: Reduced minimum required C-stack size to 2034: + CYGWIN has this stack size and the test runs fine! + * generic/tclEnv.c: Fix environment tests under CYGWIN + * generic/tclPort.h: + * tests/env.test: 2010-01-05 Don Porter <dgp@users.sourceforge.net> - * generic/tclPathObj.c (TclPathPart): Correct inconsistency between - * tests/fileName.test (filename-14.31): the string rep and the intrep - of a path value created by [file rootname]. Thanks to Vitaly Magerya - for reporting. [Bug 2918610] + * generic/tclPathObj.c (TclPathPart): [Bug 2918610]: Correct + * tests/fileName.test (filename-14.31): inconsistency between the + string rep and the intrep of a path value created by [file rootname]. + Thanks to Vitaly Magerya for reporting. 2010-01-03 Donal K. Fellows <dkf@users.sf.net> @@ -152,8 +159,8 @@ 2009-12-02 Jan Nijtmans <nijtmans@users.sf.net> - * tools/genStubs.tcl: Add support for win32 CALLBACK functions - (needed for Tk bugfix). + * tools/genStubs.tcl: Add support for win32 CALLBACK functions (needed + for Tk bugfix). 2009-11-30 Donal K. Fellows <dkf@users.sf.net> @@ -194,22 +201,21 @@ 2009-11-12 Andreas Kupries <andreask@activestate.com> - * generic/tclIO.c (CopyData): [Bug 2895565]. Dropped bogosity - which used the number of _written_ bytes or character to update - the counters for the read bytes/characters. See last entry for the - test case. + * generic/tclIO.c (CopyData): [Bug 2895565]: Dropped bogosity which + used the number of _written_ bytes or character to update the counters + for the read bytes/characters. See last entry for the test case. 2009-11-11 Pat Thoyts <patthoyts@users.sourceforge.net> - * tests/fCmd.test: Fixed a number of issues for Vista - * tests/registry.test: and Win7 that are due to restricted - * tests/winFCmd.test: permissions. + * tests/fCmd.test: Fixed a number of issues for Vista and Win7 + * tests/registry.test: that are due to restricted permissions. + * tests/winFCmd.test: 2009-11-11 Don Porter <dgp@users.sourceforge.net> - * library/http/http.tcl: Update the URL syntax check to - RFC 3986 compliance on the subject of non-encoded question mark - characters. [Bug 2891171]. + * library/http/http.tcl: [Bug 2891171]: Update the URL syntax + check to RFC 3986 compliance on the subject of non-encoded question + mark characters. * library/http/pkgIndex.tcl: Bump to http 2.7.5 to avoid any * unix/Makefile.in: confusion with snapshot "releases" @@ -258,17 +264,16 @@ 2009-11-03 Andreas Kupries <andreask@activestate.com> - * library/safe.tcl (::safe::InterpSetConfig): [Bug 2854929]. Added + * library/safe.tcl (::safe::InterpSetConfig): [Bug 2854929]: Added code to recursively find deeper paths which may contain modules. - Required to handle modules with names like 'platform::shell', - which translate into 'platform/shell-X.tm', i.e arbitrarily deep + Required to handle modules with names like 'platform::shell', which + translate into 'platform/shell-X.tm', i.e arbitrarily deep subdirectories. 2009-11-03 Kevin B. Kenny <kennykb@acm.org> - * library/tzdata/Asia/Novokuznetsk: New tzdata locale for - Kemerovo oblast', which now keeps Novosibirsk time and not - Kranoyarsk time. + * library/tzdata/Asia/Novokuznetsk: New tzdata locale for Kemerovo + oblast', which now keeps Novosibirsk time and not Kranoyarsk time. * library/tzdata/Asia/Damascus: Syrian DST changes. * library/tzdata/Asia/Hong_Kong: Hong Kong historic DST corrections. Olson tzdata2009q. @@ -290,12 +295,12 @@ to: typedef unsigned int mp_digit; For 32-bit builds where "long" and "int" are two names for the same - thing, this is no change at all. For 64-bit builds, though, this + thing, this is no change at all. For 64-bit builds, though, this causes the dp[] array of an mp_int to be made up of 32-bit elements - instead of 64-bit elements. This is a huge improvement because - details elsewhere in the mp_int implementation cause only 28 bits of - each element to be actually used storing number data. Without this - change bignums are over 50% wasted space on 64-bit systems. + instead of 64-bit elements. This is a huge improvement because details + elsewhere in the mp_int implementation cause only 28 bits of each + element to be actually used storing number data. Without this change + bignums are over 50% wasted space on 64-bit systems. ***POTENTIAL INCOMPATIBILITY*** For 64-bit builds, callers of routines with (mp_digit) or (mp_digit *) @@ -321,12 +326,12 @@ * tests/fileName.test (fileName-20.[78]): Corrected poor test hygiene (failure to save and restore the working directory) that - caused these two tests to fail on Windows (and [Bug 2806250] - to be reopened). + caused these two tests to fail on Windows (and [Bug 2806250] to be + reopened). 2009-10-27 Don Porter <dgp@users.sourceforge.net> - * generic/tclPathObj.c: [Bug 2884203]: Missing refcount on cached + * generic/tclPathObj.c: [Bug 2884203]: Missing refcount on cached normalized path caused crashes. 2009-10-27 Kevin B. Kenny <kennykb@acm.org> @@ -345,9 +350,9 @@ 2009-10-24 Kevin B. Kenny <kennykb@acm.org> * library/clock.tcl (ProcessPosixTimeZone): - Corrected a regression in the fix to [Bug 2207436] that caused - [clock] to apply EU daylight saving time rules in the US. Thanks - to Karl Lehenbauer for reporting this regression. + Corrected a regression in the fix to [Bug 2207436] that caused [clock] + to apply EU daylight saving time rules in the US. Thanks to Karl + Lehenbauer for reporting this regression. * tests/clock.test (clock-52.4): Added a regression test for the above regression. * library/tzdata/Asia/Dhaka: @@ -357,9 +362,9 @@ 2009-10-23 Andreas Kupries <andreask@activestate.com> * generic/tclIO.c (FlushChannel): Skip OutputProc for low-level - 0-length writes. When closing pipes which have already been closed - not skipping leads to spurious SIG_PIPE signals. Reported by - Mikhail Teterin <mi+thun@aldan.algebra.com>. + 0-length writes. When closing pipes which have already been closed not + skipping leads to spurious SIG_PIPE signals. Reported by Mikhail + Teterin <mi+thun@aldan.algebra.com>. 2009-10-21 Donal K. Fellows <dkf@users.sf.net> @@ -368,27 +373,27 @@ 2009-10-19 Don Porter <dgp@users.sourceforge.net> - * generic/tclIO.c: Revised ReadChars and FilterInputBytes routines - to permit reads to continue up to the string limits of Tcl values. - Before revisions, large read attempts could panic when as little as - half the limiting value length was reached. [Patch 2107634] + * generic/tclIO.c: Revised ReadChars and FilterInputBytes + routines to permit reads to continue up to the string limits of Tcl + values. Before revisions, large read attempts could panic when as + little as half the limiting value length was reached. [Patch 2107634] Thanks to Sean Morrison and Bob Parker for their roles in the fix. 2009-10-18 Joe Mistachkin <joe@mistachkin.com> * tests/thread.test (thread-4.[345]): [Bug 1565466]: Correct tests to - save their error state before the final call to threadReap just in case - it triggers an "invalid thread id" error. This error can occur if one - or more of the target threads has exited prior to the attempt to send - it an asynchronous exit command. + save their error state before the final call to threadReap just in + case it triggers an "invalid thread id" error. This error can occur + if one or more of the target threads has exited prior to the attempt + to send it an asynchronous exit command. * doc/memory.n: [Bug 988703]: Add mechanism for finding what Tcl_Objs - * generic/tclCkalloc.c (MemoryCmd): are allocated when built for memory - * generic/tclInt.decls: debugging. This was previously backported from - * generic/tclInt.h: Tcl 8.6 with the corrections to fix [Bug 2871908]. - * generic/tclObj.c (ObjData, TclFinalizeThreadObjects): However, there - were key elements missing. These changes make things consistent between - branches. + * generic/tclCkalloc.c (MemoryCmd): are allocated when built for + * generic/tclInt.decls: memory debugging. This was previously + * generic/tclInt.h: backported from Tcl 8.6 with the corrections to + * generic/tclObj.c (ObjData, TclFinalizeThreadObjects): fix [Bug + 2871908]. However, there were key elements missing. These changes make + things consistent between branches. 2009-10-17 Donal K. Fellows <dkf@users.sf.net> |