diff options
Diffstat (limited to 'ChangeLog.2003')
| -rw-r--r-- | ChangeLog.2003 | 2539 |
1 files changed, 1200 insertions, 1339 deletions
diff --git a/ChangeLog.2003 b/ChangeLog.2003 index 5c32878..b12cd2c 100644 --- a/ChangeLog.2003 +++ b/ChangeLog.2003 @@ -1,17 +1,16 @@ 2003-12-25 Mo DeJong <mdejong@users.sourceforge.net> - * win/tclWin32Dll.c (DllMain): Add HAVE_NO_SEH - blocks in place of __try and __except statements - to support gcc builds. This is needed after - David's changes on 2003-12-21. [Tcl patch 858493] + * win/tclWin32Dll.c (DllMain): Add HAVE_NO_SEH blocks in place of + __try and __except statements to support gcc builds. This is needed + after David's changes on 2003-12-21. [Patch 858493] 2003-12-23 David Gravereaux <davygrvy@pobox.com> - * generic/tclAlloc.c: All uses of 'panic' (the macro) changed - * generic/tclBasic.c: to 'Tcl_Panic' (the function). The - * generic/tclBinary.c: #define of panic in tcl.h clearly states - * generic/tclCkalloc.c: it is deprecated in the comments. - * generic/tclCmdAH.c: [Patch 865264] + * generic/tclAlloc.c: All uses of 'panic' (the macro) changed to + * generic/tclBasic.c: 'Tcl_Panic' (the function). The #define of + * generic/tclBinary.c: panic in tcl.h clearly states it is deprecated + * generic/tclCkalloc.c: in the comments. [Patch 865264] + * generic/tclCmdAH.c: * generic/tclCmdIL.c: * generic/tclCmdMZ.c: * generic/tclCompCmds.c: @@ -65,64 +64,61 @@ 2003-12-22 David Gravereaux <davygrvy@pobox.com> * win/nmakehlp.c: - * win/rules.vc: New feature for extensions that use rules.vc. - Now reads header files for version strings. No more hard coding + * win/rules.vc: New feature for extensions that use rules.vc. Now + reads header files for version strings. No more hard coding TCL_VERSION = 8.5 and having to edit it when you swap cores. * win/makefile.vc: VERSION macro now set by reading tcl.h for it. - * generic/tcl.h: Removed note that makefile.vc needs to have a - version number changed. + * generic/tcl.h: Removed note that makefile.vc needs to have a version + number changed. 2003-12-21 David Gravereaux <davygrvy@pobox.com> * win/tclWin32Dll.c: Structured Exception Handling added around - Tcl_Finalize called from DllMain's DLL_PROCESS_DETACH. We can't - be 100% assured that Tcl is being unloaded by the OS in a stable - condition and we need to protect the exit handlers should the - stack be in a hosed state. AT&T style assembly for SEH under - MinGW has not been added yet. This is a first part change for - [Patch 858493] + Tcl_Finalize called from DllMain's DLL_PROCESS_DETACH. We can't be + 100% assured that Tcl is being unloaded by the OS in a stable + condition and we need to protect the exit handlers should the stack be + in a hosed state. AT&T style assembly for SEH under MinGW has not been + added yet. This is a first part change for [Patch 858493] 2003-12-17 Daniel Steffen <das@users.sourceforge.net> - * generic/tclBinary.c (DeleteScanNumberCache): fixed crashing bug - when numeric scan-value cache contains NULL value. + * generic/tclBinary.c (DeleteScanNumberCache): fixed crashing bug when + numeric scan-value cache contains NULL value. 2003-12-17 Vince Darley <vincentdarley@users.sourceforge.net> - * generic/tclCmdAH.c: - * unix/tclUnixFile.c: - * win/tclWinFCmd.c: - * tests/fCmd.test: - * tests/fileSystem.test: - * doc/file.n: final fix to support for relative links and - its implications on normalization and other parts of the - filesystem code. Fixes [Bug 859251] and some Windows - problems with recursive file delete/copy and symbolic links. + * generic/tclCmdAH.c: + * unix/tclUnixFile.c: + * win/tclWinFCmd.c: + * tests/fCmd.test: + * tests/fileSystem.test: + * doc/file.n: final fix to support for relative links and its + implications on normalization and other parts of the filesystem code. + Fixes [Bug 859251] and some Windows problems with recursive file + delete/copy and symbolic links. 2003-12-17 Vince Darley <vincentdarley@users.sourceforge.net> - * generic/tclPathObj.c: - * tests/fileSystem.test: fix and tests for [Bug 860402] in new - file normalization code. + * generic/tclPathObj.c: + * tests/fileSystem.test: fix and tests for [Bug 860402] in new file + normalization code. 2003-12-17 Zoran Vasiljevic <zv@archiware.com> - * generic/tclIOUtil.c: fixed 2 memory (object) leaks. - This fixes Tcl Bug #839519. + * generic/tclIOUtil.c: fixed 2 memory (object) leaks. [Bug 839519] - * generic/tclPathObj.c: fixed Tcl_FSGetTranslatedPath - to always return properly refcounted path object. - This fixes Tcl Bug #861515. + * generic/tclPathObj.c: fixed Tcl_FSGetTranslatedPath to always return + properly refcounted path object. [Bug 861515] 2003-12-16 Vince Darley <vincentdarley@users.sourceforge.net> - * tests/fCmd.test: marking fCmd-9.14.2, as nonPortable, since - on Solaris one can change the name of the current directory - with 'file rename'. - * doc/FileSystem.3: clarified documentation on ownership - of return objects/strings of some Tcl_FS* calls. + * tests/fCmd.test: marking fCmd-9.14.2, as nonPortable, since on + Solaris one can change the name of the current directory with 'file + rename'. + * doc/FileSystem.3: clarified documentation on ownership of return + objects/strings of some Tcl_FS* calls. 2003-12-16 Donal K. Fellows <donal.k.fellows@man.ac.uk> @@ -131,70 +127,67 @@ 2003-12-15 David Gravereaux <davygrvy@pobox.com> * win/tcl.rc: - * win/tclsh.rc: Slight modification to the STRINGIFY macro to - support Borland's rc tool. + * win/tclsh.rc: Slight modification to the STRINGIFY macro to support + Borland's rc tool. - * win/tclWinFile.c (TclpUtime) : utimbuf struct not a problem - with Borland. + * win/tclWinFile.c (TclpUtime) : utimbuf struct not a problem with + Borland. - * win/tclWinTime.c (TclpGetDate) : Borland's localtime() has - a slight behavioral difference. + * win/tclWinTime.c (TclpGetDate) : Borland's localtime() has a slight + behavioral difference. From Helmut Giese <hgiese@ratiosoft.com> [Patch 758097]. 2003-12-14 David Gravereaux <davygrvy@pobox.com> - * generic/tclInt.decls: commented-out entry for - TclpCheckStackSpace, removing it from the Stubs table. It's - already declared in tclInt.h and labeled as a function that is - not to be exported. Regened tables. + * generic/tclInt.decls: commented-out entry for TclpCheckStackSpace, + removing it from the Stubs table. It's already declared in tclInt.h + and labeled as a function that is not to be exported. Regened tables. 2003-12-14 Donal K. Fellows <donal.k.fellows@man.ac.uk> * generic/tclCmdMZ.c (Tcl_SwitchObjCmd): TIP#75 Implementation - * tests/switch.test: Can now get submatch information when - * doc/switch.n: using -regexp matching in [switch]. + * tests/switch.test: Can now get submatch information when using + * doc/switch.n: -regexp matching in [switch]. 2003-12-14 Vince Darley <vincentdarley@users.sourceforge.net> - * generic/tclPathObj.c: complete rewrite of generic file - normalization code to cope with links followed by '..'. - [Bug 849514], and parts of [859251] + * generic/tclPathObj.c: complete rewrite of generic file normalization + code to cope with links followed by '..'. [Bug 849514], and parts of + [Bug 859251] 2003-12-12 David Gravereaux <davygrvy@pobox.com> - * win/tclWinChan.c: Win32's SetFilePointer() takes LONGs not - DWORDs (a signed/unsigned mismatch). Redid local vars to - avoid all casting except where truly required. + * win/tclWinChan.c: Win32's SetFilePointer() takes LONGs not DWORDs (a + signed/unsigned mismatch). Redid local vars to avoid all casting + except where truly required. 2003-12-12 Vince Darley <vincentdarley@users.sourceforge.net> - * generic/tclCmdAH.c: fix to normalization of non-existent user - name ('file normalize ~nobody') [Bug 858937] - * doc/file.n: clarify behaviour of 'file link' when the target - is not an absolute path. - * doc/filename.n: correct documentation to say that Windows Tcl - does handle '~user', for recent Windows releases, and clarified - distinction between MacOS 'classic' and MacOS X. + * generic/tclCmdAH.c: fix to normalization of non-existent user name + ('file normalize ~nobody') [Bug 858937] + * doc/file.n: clarify behaviour of 'file link' when the target is not + an absolute path. + * doc/filename.n: correct documentation to say that Windows Tcl does + handle '~user', for recent Windows releases, and clarified distinction + between MacOS 'classic' and MacOS X. * doc/glob.n: clarification of glob's behaviour when returning filenames starting with a '~'. * tests/fileSystem.test: - * tests/fileName.test: new tests added for the normalization - problem above and other recentlt reported issues. + * tests/fileName.test: new tests added for the normalization problem + above and other recentlt reported issues. * win/tclWinFile.c: corrected unclear comments - * unix/tclUnixFile.c: allow creation of relative links - [Bug 833713] + * unix/tclUnixFile.c: allow creation of relative links. [Bug 833713] 2003-12-11 David Gravereaux <davygrvy@pobox.com> - * win/tclWinSock.c (SocketThreadExitHandler) : added a - TerminateThread fallback just in case the socket handler thread - is really in a paused state. This can happen when Tcl is being - unloaded by the OS from an exception handler. See MSDN docs on - DllMain, it states this behavior. + * win/tclWinSock.c (SocketThreadExitHandler) : added a TerminateThread + fallback just in case the socket handler thread is really in a paused + state. This can happen when Tcl is being unloaded by the OS from an + exception handler. See MSDN docs on DllMain, it states this behavior. 2003-12-09 Jeff Hobbs <jeffh@ActiveState.com> @@ -204,24 +197,24 @@ 2003-12-09 Donal K. Fellows <donal.k.fellows@man.ac.uk> - * unix/tclUnixPort.h: #ifdef'd out declarations of errno which - * tools/man2tcl.c: are known to cause problems with recent - glibc. [Bug 852369] + * unix/tclUnixPort.h: #ifdef'd out declarations of errno which are + * tools/man2tcl.c: known to cause problems with recent glibc. + [Bug 852369] 2003-12-09 Vince Darley <vincentdarley@users.sourceforge.net> * win/tclWinFile.c: fix to NT file permissions code [Bug 855923] - * tests/winFile.test: added tests for NT file permissions - patch - and test scripts supplied by Benny. + * tests/winFile.test: added tests for NT file permissions - patch and + test scripts supplied by Benny. * tests/winFCmd.test: fixed one test for when not running in C:/ 2003-12-02 Donal K. Fellows <donal.k.fellows@man.ac.uk> - * generic/tclBinary.c (DeleteScanNumberCache, ScanNumber): Made - the numeric scan-value cache have proper references to the objects - within it so strange patterns of writes won't cause references to - freed objects. Thanks to Paul Obermeir for the report. [Bug 851747] + * generic/tclBinary.c (DeleteScanNumberCache, ScanNumber): Made the + numeric scan-value cache have proper references to the objects within + it so strange patterns of writes won't cause references to freed + objects. Thanks to Paul Obermeir for the report. [Bug 851747] 2003-12-01 Miguel Sofer <msofer@users.sf.net> @@ -230,14 +223,13 @@ 2003-11-24 Don Porter <dgp@users.sourceforge.net> * generic/tclParse.c: Corrected faulty check for trailing white - space in {expand} parsing. Thanks Andreas Leitgeb. [Bug 848262]. + space in {expand} parsing. Thanks Andreas Leitgeb. [Bug 848262] * tests/parse.test: New tests for the bug. 2003-11-24 Vince Darley <vincentdarley@users.sourceforge.net> - * generic/tclPathObj.c: fix to [Bug 845778] - Infinite recursion - on [cd] (Windows only bug), for which new tests have just been - added. + * generic/tclPathObj.c: fix to [Bug 845778] - Infinite recursion on + [cd] (Windows only bug), for which new tests have just been added. 2003-11-21 Don Porter <dgp@users.sourceforge.net> @@ -246,20 +238,20 @@ 2003-11-20 Miguel Sofer <msofer@users.sf.net> - * generic/tclVar.c: fix flag bit collision between - LOOKUP_FOR_UPVAR and TCL_PARSE_PART1 (deprecated) [Bug 835020] + * generic/tclVar.c: fix flag bit collision between LOOKUP_FOR_UPVAR + and TCL_PARSE_PART1 (deprecated) [Bug 835020] 2003-11-19 Don Porter <dgp@users.sourceforge.net> - * tests/compile.test (compile-16.22.0): Improved test for the - recent fix for Bug 845412. + * tests/compile.test (compile-16.22.0): Improved test for the recent + fix for Bug 845412. 2003-11-19 Donal K. Fellows <donal.k.fellows@man.ac.uk> * generic/tclCompile.c (TclCompileScript): Added a guard for the - expansion code so that long non-expanding commands don't get - expansion infrastructure inserted in them, especially when that - infrastructure isn't initialised. [Bug 845412] + expansion code so that long non-expanding commands don't get expansion + infrastructure inserted in them, especially when that infrastructure + isn't initialised. [Bug 845412] 2003-11-18 David Gravereaux <davygrvy@pobox.com> @@ -273,44 +265,43 @@ 2003-11-17 Don Porter <dgp@users.sourceforge.net> * tests/reg.test: Added tests for [Bugs 230589, 504785, 505048, 840258] - recently fixed by 2003-11-15 commit to regcomp.c by Pavel Goran. - His notes on the fix: This bug results from an error in code that - splits states into "progress" and "no-progress" ones. This error - causes an interesting situation with the pre-collected single-linked - list of states to be splitted: many items were added to the list, but - only several of them are accessible from the list beginning, - since the "tmp" member of struct state (which is used here to - hold a pointer to the next list item) gets overwritten, which - results in a "looped" chain. As a result, not all of states are - splitted, and one state is splitted two times, causing incorrect - "no-progress" flag values. + recently fixed by 2003-11-15 commit to regcomp.c by Pavel Goran. His + notes on the fix: This bug results from an error in code that splits + states into "progress" and "no-progress" ones. This error causes an + interesting situation with the pre-collected single-linked list of + states to be splitted: many items were added to the list, but only + several of them are accessible from the list beginning, since the + "tmp" member of struct state (which is used here to hold a pointer to + the next list item) gets overwritten, which results in a "looped" + chain. As a result, not all of states are splitted, and one state is + splitted two times, causing incorrect "no-progress" flag values. 2003-11-16 Donal K. Fellows <donal.k.fellows@man.ac.uk> * generic/tclExecute.c (TclExecuteByteCode): Make sure that Tcl_AsyncInvoke is called regularly when processing bytecodes. - * generic/tclTest.c (AsyncThreadProc, TestasyncCmd): Extended - testing harness to send an asynchronous marking without relying on - UNIX signals. - * tests/async.test (async-4.*): Tests to check that async events - are handled by the bytecode core. [Bug 746722] + * generic/tclTest.c (AsyncThreadProc, TestasyncCmd): Extended testing + harness to send an asynchronous marking without relying on UNIX + signals. + * tests/async.test (async-4.*): Tests to check that async events are + handled by the bytecode core. [Bug 746722] 2003-11-15 Donal K. Fellows <donal.k.fellows@man.ac.uk> - * generic/tclTest.c (TestHashSystemHashCmd): Removed 'const' - modifier from hash type structure; it should be const and the hash - code assumes it behaves like const, but that's not how the API is - defined. Like this, we are following in the same footsteps as - Tcl_RegisterObjType() which has the same conditions on its - argument. Stops VC++5.2 warning. [Bug 842511] + * generic/tclTest.c (TestHashSystemHashCmd): Removed 'const' modifier + from hash type structure; it should be const and the hash code assumes + it behaves like const, but that's not how the API is defined. Like + this, we are following in the same footsteps as Tcl_RegisterObjType() + which has the same conditions on its argument. Stops VC++5.2 warning. + [Bug 842511] 2003-11-14 Donal K. Fellows <donal.k.fellows@man.ac.uk> - * generic/tclHash.c (Tcl_DeleteHashTable,Tcl_HashStats,RebuildTable): - * generic/tclTest.c (TestHashSystemHashCmd): TIP#138 implementation, + * generic/tclHash.c (Tcl_DeleteHashTable,Tcl_HashStats,RebuildTable): + * generic/tclTest.c (TestHashSystemHashCmd): TIP#138 implementation, * tests/misc.test: plus a new chunk of stuff to test the hash - functions more thoroughly in the test - suite. [Patch 731356, modified] + functions more thoroughly in the test suite. + [Patch 731356, modified] * doc/Tcl.n: Updated Tcl version number and changebars. @@ -319,23 +310,23 @@ * doc/ParseCmd.3: Implementation of TIP 157. Adds recognition * doc/Tcl.n: of the new leading {expand} syntax on words. * generic/tcl.h: Parses such words as the new Tcl_Token type - * generic/tclBasic.c: TCL_TOKEN_EXPAND_WORD. Updated Tcl_EvalEx - * generic/tclCompile.c: and the bytecode compiler/execution engine - * generic/tclCompile.h: to recognize the new token type. New opcodes + * generic/tclBasic.c: TCL_TOKEN_EXPAND_WORD. Updated Tcl_EvalEx and + * generic/tclCompile.c: the bytecode compiler/execution engine to + * generic/tclCompile.h: recognize the new token type. New opcodes * generic/tclExecute.c: INST_LIST_VERIFY and INST_INVOKE_EXP and a new - * generic/tclParse.c: operand type OPERAND_ULIST1 are defined. Docs + * generic/tclParse.c: operand type OPERAND_ULIST1 are defined. Docs * generic/tclTest.c: and tests are included. * tests/basic.test: * tests/compile.test: * tests/parse.test: * library/auto.tcl: Replaced several [eval]s used to perform - * library/package.tcl: argument expansion with the new syntax. - * library/safe.tcl: In the test files lindex.test and lset.test, - * tests/cmdInfo.test: replaced use of [eval] to force direct - * tests/encoding.test: string evaluation with use of [testevalex] - * tests/execute.test: which more directly and robustly serves the - * tests/fCmd.test: same purpose. + * library/package.tcl: argument expansion with the new syntax. In the + * library/safe.tcl: test files lindex.test and lset.test, replaced + * tests/cmdInfo.test: use of [eval] to force direct string + * tests/encoding.test: evaluation with use of [testevalex] which more + * tests/execute.test: directly and robustly serves the same purpose. + * tests/fCmd.test: * tests/http.test: * tests/init.test: * tests/interp.test: @@ -357,16 +348,16 @@ 2003-11-12 Jeff Hobbs <jeffh@ActiveState.com> - * tests/cmdMZ.test (cmdMZ-1.4): change to nonPortable as more - systems are using permissions caching, and this isn't really a Tcl - controlled issue. + * tests/cmdMZ.test (cmdMZ-1.4): change to nonPortable as more systems + are using permissions caching, and this isn't really a Tcl controlled + issue. 2003-11-11 Jeff Hobbs <jeffh@ActiveState.com> * unix/configure: * unix/tcl.m4: improve AIX --enable-64bit handling - remove -D__NO_STRING_INLINES -D__NO_MATH_INLINES from - CFLAGS_OPTIMIZE on Linux. Make default opt -O2 (was -O). + remove -D__NO_STRING_INLINES -D__NO_MATH_INLINES from CFLAGS_OPTIMIZE + on Linux. Make default opt -O2 (was -O). 2003-11-11 David Gravereaux <davygrvy@pobox.com> @@ -381,10 +372,10 @@ * win/tclWinInit.c (TclpInitLibraryPath): Fix for [Bug 832657] that should not run afoul of startup constraints. - * library/dde/pkgIndex.tcl: Added safeguards so that registry - * library/reg/pkgIndex.tcl: and dde packages are not offered - * win/tclWinDde.c: on non-Windows platforms. Bumped to - * win/tclWinReg.c: registry 1.1.3 and dde 1.3. + * library/dde/pkgIndex.tcl: Added safeguards so that registry and + * library/reg/pkgIndex.tcl: dde packages are not offered on + * win/tclWinDde.c: non-Windows platforms. Bumped to + * win/tclWinReg.c: registry 1.1.3 and dde 1.3. * win/Makefile.in: * win/configure.in: * win/makefile.bc: @@ -401,13 +392,13 @@ 2003-11-10 Vince Darley <vincentdarley@users.sourceforge.net> * generic/tclCmdAH.c: - * tests/fCmd.test: fix to misleading error message in 'file link' + * tests/fCmd.test: fix to misleading error message in 'file link'. [Bug 836208] 2003-11-07 Vince Darley <vincentdarley@users.sourceforge.net> - * generic/tclIOUtil.c: fix to compiler warning/error with - some compilers [Bug 835918] + * generic/tclIOUtil.c: fix to compiler warning/error with some + compilers. [Bug 835918] 2003-11-07 Daniel Steffen <das@users.sourceforge.net> @@ -427,10 +418,9 @@ 2003-11-03 Vince Darley <vincentdarley@users.sourceforge.net> * generic/tclIOUtil.c - * generic/tclInt.h: added comments and re-arranged code to - clarify distinction between Tcl_LoadHandle, ClientData for - 'load'ed code, and point out limitations of the design - introduced with Tcl 8.4. + * generic/tclInt.h: added comments and re-arranged code to clarify + distinction between Tcl_LoadHandle, ClientData for 'load'ed code, and + point out limitations of the design introduced with Tcl 8.4. * unix/tclUnixFile.c: fix to memory leak @@ -440,14 +430,13 @@ * generic/tclCmdIL.c (Tcl_LrepeatObjCmd): Check for sensible list lengths and allow for soft failure of the memory subsystem in the - [lconcat] command [Bug 829027]. Uses direct list creation to - avoid extra copies when working near the limit of available - memory. Also reorganized to encourage optimizing compilers to - optimize heavily. + [lconcat] command [Bug 829027]. Uses direct list creation to avoid + extra copies when working near the limit of available memory. Also + reorganized to encourage optimizing compilers to optimize heavily. * generic/tclListObj.c (TclNewListObjDirect): New list constructor that does not copy the array of objects. Useful for creating - potentially very large lists or where you are about to throw away - the array argument which is being used in its entirety. + potentially very large lists or where you are about to throw away the + array argument which is being used in its entirety. 2003-10-28 Miguel Sofer <msofer@users.sf.net> @@ -457,14 +446,14 @@ 2003-10-23 Andreas Kupries <andreask@activestate.com> * unix/tclUnixChan.c (Tcl_MakeFileChannel): Applied [Patch 813606] - fixing [Bug 813087]. Detection of sockets was off for Mac OS X - which implements pipes as local sockets. The new code ensures - that only IP sockets are detected as such. + fixing [Bug 813087]. Detection of sockets was off for Mac OS X which + implements pipes as local sockets. The new code ensures that only IP + sockets are detected as such. - * win/tclWinSock.c (TcpWatchProc): Watch for FD_CLOSE too when - asked for writable events by the generic layer. - (SocketEventProc): Generate a writable event too when a close is - detected. + * win/tclWinSock.c (TcpWatchProc): Watch for FD_CLOSE too when asked + for writable events by the generic layer. + (SocketEventProc): Generate a writable event too when a close is + detected. Together the changes fix [Bug 599468]. @@ -487,42 +476,42 @@ * tools/tcltk-man2html.tcl: fixed incorrect html generated for .IP/.TP lists, now use <DL><DT>...<DD>...<P><DT>...<DD>...</DL> instead of illegal <DL><P><DT>...<DD>...<P><DT>...<DD>...</DL>. - Added skipping of directives directly after .TP to avoid them - being used as item descriptions, e.g. .TP\n.VS in clock.n. + Added skipping of directives directly after .TP to avoid them being + used as item descriptions, e.g. .TP\n.VS in clock.n. 2003-10-21 Andreas Kupries <andreask@pliers.activestate.com> - * win/tclWinPipe.c (BuildCommandLine): Applied the patch coming - with [Bug 805605] to the code, fixing the incorrect use of - ispace noted by Ronald Dauster <ronaldd@users.sourceforge.net>. + * win/tclWinPipe.c (BuildCommandLine): Applied the patch coming with + [Bug 805605] to the code, fixing the incorrect use of ispace noted by + Ronald Dauster <ronaldd@users.sourceforge.net>. 2003-10-20 Kevin B. Kenny <kennykb@users.sourceforge.net> * doc/msgcat.n: * library/msgcat/msgcat.tcl (mclocale,mcload): * tools/tcl.wse.in: - * unix/Makefile.in: Implementation of TIP#156 - * win/makefile.bc: adding a "root locale" to - * win/Makefile.in: the 'msgcat' package. Advanced - * win/Makefile.vc: msgcat version number to 1.4. + * unix/Makefile.in: Implementation of TIP#156, add a "root locale" + * win/makefile.bc: to the 'msgcat' package. Advanced msgcat + * win/Makefile.in: version number to 1.4 + * win/Makefile.vc: 2003-10-15 Donal K. Fellows <fellowsd@cs.man.ac.uk> * generic/tclCmdIL.c (SortInfo,etc): Reorganized so that SortInfo - carries an array of integer indices instead of a Tcl list. This - nips shimmering problems in the bud and simplifies SelectObjFromSublist - at the cost of making setup slightly more complex. [Bug 823768] + carries an array of integer indices instead of a Tcl list. This nips + shimmering problems in the bud and simplifies SelectObjFromSublist at + the cost of making setup slightly more complex. [Bug 823768] 2003-10-14 David Gravereaux <davygrvy@pobox.com> - * win/tclAppInit.c (sigHandler): Punt gracefully if exitToken - has already been destroyed. + * win/tclAppInit.c (sigHandler): Punt gracefully if exitToken has + already been destroyed. 2003-10-14 Vince Darley <vincentdarley@users.sourceforge.net> * generic/tclCmdMZ.c: - * tests/regexp.test: fix to [Bug 823524] in regsub; added three - new tests. + * tests/regexp.test: fix to [Bug 823524] in regsub; added three new + tests. 2003-10-14 Don Porter <dgp@users.sourceforge.net> @@ -534,7 +523,7 @@ routine that supports truncated appends with optional ellipsis marking. This single routine supports UTF-8-safe truncated appends needed in several places throughout the Tcl source code, mostly for error and - stack messages. Clean fix for [Bug 760872]. + stack messages. Clean fix for [Bug 760872]. * generic/tclInt.h: Declarations for new internal routines. @@ -551,23 +540,23 @@ * mac/tclMacResource.c: * library/init.tcl: Updated ::errorInfo cleanup in [unknown] to - reflect slight modifications to Tcl_LogCommandInfo(). Corrects - failing init-4.* tests. + reflect slight modifications to Tcl_LogCommandInfo(). Corrects failing + init-4.* tests. 2003-10-14 Donal K. Fellows <fellowsd@cs.man.ac.uk> TIP#127 IMPLEMENTATION FROM JOE MICHAEL SCHLENKER * generic/tclCmdIL.c (SelectObjFromSublist): Element selection engine. - * generic/tclCmdIL.c (Tcl_LsearchObjCmd, Tcl_LsortObjCmd): + * generic/tclCmdIL.c (Tcl_LsearchObjCmd, Tcl_LsortObjCmd): * tests/lsearch.test: Set up and use of element selection engine, * tests/cmdIL.test: plus tests and documentation. * doc/lsearch.n: Based on [Patch 693836] - * doc/lsort.n: + * doc/lsort.n: 2003-10-13 Vince Darley <vincentdarley@users.sourceforge.net> - * generic/tcl.h: + * generic/tcl.h: * generic/tclFileName.c: * generic/tclIOUtil.c: * generic/tclPathObj.c: @@ -579,28 +568,27 @@ * win/tclWin32Dll.c: * win/tclWinFCmd.c: * win/tclWinFile.c: - * win/tclFileInt.h: + * win/tclFileInt.h: Fixed [Bug 800106] in which 'glob' was incapable of merging the results of a directory listing (real or virtual) and any virtual - filesystem mountpoints in that directory (the latter were - ignored). This meant boundaries between different filesystems - were not seamless (e.g. 'glob */*' across a filesystem boundary - was wrong). Added new entry to Tcl_GlobTypeData in a totally - backwards compatible way. To allow listing of mounts, registered - filesystems must support the 'TCL_GLOB_TYPE_MOUNT' flag. If this - is not supported (e.g. in tclvfs 1.2) then mounts will simply not - be listed for that filesystem. - - Fixed [Bug 749876] 'file writable/readable/etc' (NativeAccess) - using correct permission checking code for Windows NT/2000/XP - where more complex user-based security/access priveleges are - available, particularly on shared volumes. The performance - impact of this extra checking will need further investigation. - Note: Win 95,98,ME have no support for this. - - Also made better use of normalized rather than translated paths - in the platform specific code. + filesystem mountpoints in that directory (the latter were ignored). + This meant boundaries between different filesystems were not seamless + (e.g. 'glob */*' across a filesystem boundary was wrong). Added new + entry to Tcl_GlobTypeData in a totally backwards compatible way. To + allow listing of mounts, registered filesystems must support the + 'TCL_GLOB_TYPE_MOUNT' flag. If this is not supported (e.g. in tclvfs + 1.2) then mounts will simply not be listed for that filesystem. + + Fixed [Bug 749876] 'file writable/readable/etc' (NativeAccess) using + correct permission checking code for Windows NT/2000/XP where more + complex user-based security/access priveleges are available, + particularly on shared volumes. The performance impact of this extra + checking will need further investigation. Note: Win 95,98,ME have no + support for this. + + Also made better use of normalized rather than translated paths in the + platform specific code. 2003-10-12 Jeff Hobbs <jeffh@ActiveState.com> @@ -618,7 +606,7 @@ * generic/tclBasic.c: Save and restore the iPtr->flag bits that control the state of errorCode and errorInfo management when calling "leave" execution traces, so that all error information of the traced - command is still available whether traced or not. [Bug 760947] + command is still available whether traced or not. [Bug 760947] Thanks to Yahalom Emet. 2003-10-08 Donal K. Fellows <fellowsd@cs.man.ac.uk> @@ -626,27 +614,27 @@ * generic/tclTest.c (TestNumUtfCharsCmd): Command to allow finer access to Tcl_NumUtfChars for testing. * generic/tclUtf.c (Tcl_NumUtfChars): Corrected string length - determining when the length parameter is negative; the terminator - is a zero byte, not (necessarily) a \u0000 character. [Bug 769812] + determining when the length parameter is negative; the terminator is a + zero byte, not (necessarily) a \u0000 character. [Bug 769812] 2003-10-07 Don Porter <dgp@users.sourceforge.net> * tests/cmdAH.test: * tests/exec.test: Corrected temporary file management * tests/fileSystem.test: issues uncovered by -debug 1 test - * tests/io.test: operations. Also backported some + * tests/io.test: operations. Also backported some * tests/ioCmd.test: other fixes from the HEAD. * tests/main.test: * tests/pid.test: [Bugs 675605, 675655, 675659] * tests/socket.test: * tests/source.test: - * tests/fCmd.test: Run tests with the [temporaryDirectory] as - the current directory, so that tests can depend on ability to write - files. [Bug 575837] + * tests/fCmd.test: Run tests with the [temporaryDirectory] as the + current directory, so that tests can depend on ability to write files. + [Bug 575837] - * doc/OpenFileChnl.3: Updated Tcl_Tell and Tcl_Seek documentation - to reflect that they now return Tcl_WideInt (TIP 72) [Bug 787537] + * doc/OpenFileChnl.3: Updated Tcl_Tell and Tcl_Seek documentation to + reflect that they now return Tcl_WideInt (TIP 72). [Bug 787537] * tests/io.test: Corrected several tests that failed when paths * tests/ioCmd.test: included regexp-special chars. [Bug 775394] @@ -654,9 +642,9 @@ 2003-10-06 Jeff Hobbs <jeffh@ActiveState.com> * win/configure: - * win/tcl.m4: removed incorrect checks for existence of - optimization. TCL_CFG_OPTIMIZED is now defined whenever the user - does not build with --enable-symbols. + * win/tcl.m4: removed incorrect checks for existence of optimization. + TCL_CFG_OPTIMIZED is now defined whenever the user does not build with + --enable-symbols. 2003-10-06 Don Porter <dgp@users.sourceforge.net> @@ -666,8 +654,8 @@ * tests/fCmd.test (fCmd-8.2): Test only that tilde-substitution happens, not for any particular result. [Bug 685991] - * unix/tcl.m4 (SC_PATH_TCLCONFIG): Corrected search path so - that alpha and beta releases of Tcl are not favored. [Bug 608698] + * unix/tcl.m4 (SC_PATH_TCLCONFIG): Corrected search path so that + alpha and beta releases of Tcl are not favored. [Bug 608698] * tests/reg.test: Corrected duplicate test names. * tests/resource.test: [Bugs 710370, 710358] @@ -691,12 +679,12 @@ * generic/tclBasic.c: Fixed error in ref count management of command * generic/tclCmdMZ.c: and execution traces that caused access to - freed memory in trace-32.1. [Bug 811483]. + freed memory in trace-32.1. [Bug 811483] 2003-10-02 Don Porter <dgp@users.sourceforge.net> * generic/tclTrace.c: Corrected comingling of introspection results of - [trace info command] and [trace info execution]. [Bug 807243] + [trace info command] and [trace info execution]. [Bug 807243] Thanks to Mark Saye. 2003-10-01 Daniel Steffen <das@users.sourceforge.net> @@ -712,10 +700,10 @@ 2003-09-29 Donal K. Fellows <fellowsd@cs.man.ac.uk> - * generic/tclPathObj.c (TclNativePathInFilesystem,TclFSGetPathType): + * generic/tclPathObj.c (TclNativePathInFilesystem,TclFSGetPathType): * generic/tclIOUtil.c (TclNativeDupInternalRep,TclGetPathType): Rename - to make sure function names won't interfere with other non-Tcl - code (reported by George Staplin) + to make sure function names won't interfere with other non-Tcl code + (reported by George Staplin) TIP#121 IMPLEMENTATION FROM JOE MISTACHKIN @@ -727,14 +715,13 @@ TIP#112 IMPLEMENTATION * generic/tclNamesp.c: Core of implementation. - * generic/tclInt.h (Namespace,TclInvalidateNsCmdLookup): Add - command list epoch counter and list of ensembles to namespace - structure, and define a macro to ease update of the epoch - counter. - * generic/tclBasic.c (Tcl_CreateObjCommand,etc.): Update epoch - counter when list of commands in a namespace changes. - * generic/tclObj.c (TclInitObjSubsystem): Register ensemble - subcommand type. + * generic/tclInt.h (Namespace,TclInvalidateNsCmdLookup): Add command + list epoch counter and list of ensembles to namespace structure, and + define a macro to ease update of the epoch counter. + * generic/tclBasic.c (Tcl_CreateObjCommand,etc.): Update epoch counter + when list of commands in a namespace changes. + * generic/tclObj.c (TclInitObjSubsystem): Register ensemble subcommand + type. * tests/namespace.test (42.1-47.6): Tests. * doc/namespace.n: Documentation. @@ -747,38 +734,37 @@ 2003-09-28 David Gravereaux <davygrvy@pobox.com> * win/tclWinPipe.c: The windows port of expect can call - TclWinAddProcess before any of the other pipe functions. - Added a missing PipeInit() call to make sure the - initialization happens. + TclWinAddProcess before any of the other pipe functions. Added a + missing PipeInit() call to make sure the initialization happens. 2003-09-25 Daniel Steffen <das@users.sourceforge.net> - * macosx/Makefile: ensure SYMROOT exists if OBJROOT is overridden - on command line. Replaced explict use of /usr/bin by ${BINDIR}. + * macosx/Makefile: ensure SYMROOT exists if OBJROOT is overridden on + command line. Replaced explict use of /usr/bin by ${BINDIR}. 2003-09-24 Vince Darley <vincentdarley@users.sourceforge.net> - * library/package.tcl (tcl::MacPkgUnknown, tcl::MacOSXPkgUnknown): - Minor performance tweaks to reduce the number of [file] invocations. - Meant to improve startup times, at least a little bit. - (The generic equivalent patch was applied on 2003-02-21). + * library/package.tcl (tcl::MacPkgUnknown, tcl::MacOSXPkgUnknown): + Minor performance tweaks to reduce the number of [file] invocations. + Meant to improve startup times, at least a little bit. (The generic + equivalent patch was applied on 2003-02-21). 2003-09-24 Vince Darley <vincentdarley@users.sourceforge.net> - * trace.test: removed 'knownBug' from a test which doesn't - illustrate a bug, just a bad test. + * trace.test: removed 'knownBug' from a test which doesn't illustrate + a bug, just a bad test. 2003-09-23 Miguel Sofer <msofer@users.sf.net> * generic/tclExecute.c: - * generic/tclInt.h: changed the evaluation-stack addressing mode, - from array-style to pointer-style; the catch stack and evaluation - stack are now contiguous in memory. [Patch 457449] + * generic/tclInt.h: changed the evaluation-stack addressing mode, from + array-style to pointer-style; the catch stack and evaluation stack are + now contiguous in memory. [Patch 457449] 2003-09-23 Don Porter <dgp@users.sourceforge.net> - * tests/trace.test (trace-31,32-*): Added tests for [Bug 807243] - and [Bug 811483]. + * tests/trace.test (trace-31,32-*): Added tests for [Bug 807243] and + [Bug 811483]. * library/init.tcl (auto_load, auto_import): Expanded Eric Melski's 2000-01-28 fix for [Bug 218871] to all potentially troubled uses of @@ -787,29 +773,29 @@ 2003-09-20 Donal K. Fellows <fellowsd@cs.man.ac.uk> - * tests/expr.test (expr-23.4): Prevented accidental wrapping round - of exponential operation; it isn't portable, and not what I - intended to test either. [Bug 808244] + * tests/expr.test (expr-23.4): Prevented accidental wrapping round of + exponential operation; it isn't portable, and not what I intended to + test either. [Bug 808244] 2003-09-19 Miguel Sofer <msofer@users.sf.net> - * generic/tclExecute.c: adding (DE)CACHE_STACK_INFO() pairs to - protect all calls that may cause traces on ::errorInfo or - ::errorCode to corrupt the stack [Bug 804681] + * generic/tclExecute.c: adding (DE)CACHE_STACK_INFO() pairs to protect + all calls that may cause traces on ::errorInfo or ::errorCode to + corrupt the stack. [Bug 804681] 2003-09-17 Vince Darley <vincentdarley@users.sourceforge.net> - * tclPathObj.c: fix to test-suite problem introduced by the bug - fix below. No problem in ordinary code, just test suite code - which manually adjusts tclPlatform. [Bug 808247] + * tclPathObj.c: fix to test-suite problem introduced by the bug fix + below. No problem in ordinary code, just test suite code which + manually adjusts tclPlatform. [Bug 808247] 2003-09-16 Vince Darley <vincentdarley@users.sourceforge.net> * doc/filename.n: documentation of Windows-specific feature as discussed in [Bug 541989] - * generic/tclPathObj.c: fix for normalization of volume-relative - paths [Bug 767834] - * tests/winFCmd.test: new tests for both of the above. + * generic/tclPathObj.c: fix for normalization of volume-relative paths + [Bug 767834] + * tests/winFCmd.test: new tests for both of the above. * tests/cmdAH.test: fix for AFS problem in test suite [Bug 748960] 2003-09-13 Donal K. Fellows <fellowsd@cs.man.ac.uk> @@ -818,35 +804,35 @@ * generic/tclCompile.h (INST_EXPON): Implementation of * generic/tclCompile.c (tclInstructionTable): exponential operator. - * generic/tclCompExpr.c (operatorTable): - * generic/tclParseExpr.c (ParseExponentialExpr, GetLexeme): - * generic/tclExecute.c (TclExecuteByteCode, ExponWide, ExponLong): - (IllegalExprOperandType): - * tests/expr.test: - * tests/compExpr-old.test: - * doc/expr.n: + * generic/tclCompExpr.c (operatorTable): + * generic/tclParseExpr.c (ParseExponentialExpr, GetLexeme): + * generic/tclExecute.c (TclExecuteByteCode, ExponWide, ExponLong): + (IllegalExprOperandType): + * tests/expr.test: + * tests/compExpr-old.test: + * doc/expr.n: 2003-09-10 Don Porter <dgp@users.sourceforge.net> * library/opt/optparse.tcl: Latest revisions caused [OptGuessType] - to guess "int" instead of "string" for empty strings. Missed the - required "-strict" option to [string is]. Thanks to Revar Desmera. + to guess "int" instead of "string" for empty strings. Missed the + required "-strict" option to [string is]. Thanks to Revar Desmera. [Bug 803968] 2003-09-08 David Gravereaux <davygrvy@pobox.com> * win/tclWinLoad.c (TclpDlopen): Changed the error message for ERROR_PROC_NOT_FOUND to be a bit more helpful in giving us clues. - "can't find specified procedure" means a function in the import - table, for implicit loading, couldn't be resolved and that's why - the load failed. + "can't find specified procedure" means a function in the import table, + for implicit loading, couldn't be resolved and that's why the load + failed. 2003-09-04 Don Porter <dgp@users.sourceforge.net> * doc/Tcl_Main.3: * doc/FileSystem.3: Implementation of - * doc/source.n: TIPs 137/151. Adds - * doc/tclsh.1: a -encoding option to + * doc/source.n: TIPs 137/151. Adds a + * doc/tclsh.1: -encoding option to * generic/tcl.decls: the [source] command * generic/tclCmdMZ.c (Tcl_SourceObjCmd): and a new C routine, * generic/tclIOUtil.c (Tcl_FSEvalFileEx): Tcl_FSEvalFileEx(), @@ -854,20 +840,20 @@ * mac/tclMacResource.c (Tcl_MacSourceObjCmd): to the same function. * tests/cmdMZ.test: Also adds command line * tests/main.test: option handling in Tcl_Main() so that tclsh - * tests/source.test: and other apps built on Tcl_Main() respect - a -encoding command line option before a script filename. Docs and - tests updated as well. [Patch 742683] + * tests/source.test: and other apps built on Tcl_Main() respect a + -encoding command line option before a script filename. Docs and tests + updated as well. [Patch 742683] This is a ***POTENTIAL INCOMPATIBILITY*** only for those C programs that embed Tcl, build on Tcl_Main(), and make use of Tcl_Main's former ability to pass a leading "-encoding" option to interactive shell operations. * generic/tclInt.decls: Added internal stub - * generic/tclMain.c (Tcl*StartupScript*): table entries for - two new functions Tcl_SetStartupScript() and Tcl_GetStartupScript() - that set/get the path and encoding for the startup script to be - evaluated by either Tcl_Main() or Tk_Main(). Given public names in - anticipation of their exposure by a followup TIP. + * generic/tclMain.c (Tcl*StartupScript*): table entries for two + new functions Tcl_SetStartupScript() and Tcl_GetStartupScript() that + set/get the path and encoding for the startup script to be evaluated + by either Tcl_Main() or Tk_Main(). Given public names in anticipation + of their exposure by a followup TIP. * generic/tclDecls.h: make genstubs * generic/tclIntDecls.h: @@ -893,8 +879,8 @@ 2003-09-01 Donal K. Fellows <fellowsd@cs.man.ac.uk> - * doc/Namespace.3: Basic documentation for the TIP#139 functions. - This will need improving, but the basic bits are there at least. + * doc/Namespace.3: Basic documentation for the TIP#139 functions. This + will need improving, but the basic bits are there at least. 2003-08-31 Don Porter <dgp@users.sourceforge.net> @@ -902,102 +888,100 @@ 2003-08-29 Don Porter <dgp@users.sourceforge.net> - * generic/tclCmdAH.c: Corrected bug in TIP 90 implementation - * tests/cmdMZ.test: where the default -errorcode NONE value - was not copied into the return options dictionary. This correction - modified one test result. + * generic/tclCmdAH.c: Corrected bug in TIP 90 implementation where + * tests/cmdMZ.test: the default -errorcode NONE value was not + copied into the return options dictionary. This correction modified + one test result. 2003-08-27 David Gravereaux <davygrvy@pobox.com> * compat/strftime.c (_fmt): Removed syst array intializer that couldn't take variables within it under the watcom compiler: - 'Initializers must be constant'. I believe Borland has this - strictness as well. VC++ must be non-standard about this. + 'Initializers must be constant'. I believe Borland has this strictness + as well. VC++ must be non-standard about this. - Changed Win32 platform #ifdef from 'WIN32' to '__WIN32__' as - this is the correct one to use across the Tcl sources. Even - though we do force it in tcl.h, the true parent one is __WIN32__. + Changed Win32 platform #ifdef from 'WIN32' to '__WIN32__' as this is + the correct one to use across the Tcl sources. Even though we do force + it in tcl.h, the true parent one is __WIN32__. - Added missing CONST'ification usage to match prototype listed - in tclInt.decls. + Added missing CONST'ification usage to match prototype listed in + tclInt.decls. - * win/tclWinPort.h: Added a block for OpenWatcom adjustments - that fixes 1) the same issue Mo did for MinGW lack of missing LPFN_* + * win/tclWinPort.h: Added a block for OpenWatcom adjustments that + fixes 1) the same issue Mo did for MinGW lack of missing LPFN_* typedefs in their WINE derived <winsock2.h> and 2) The need to be strict about how the char type needs to be signed by default. * win/tclWinSock.c: Added OpenWatcom to the commentary about the #ifdef HAVE_NO_LPFN_DECLS block. - * win/tclWinTime.c: Changed use of '_timezone' to 'timezone' as - this difference is already adjusted for in tclWinPort.h. Removed + * win/tclWinTime.c: Changed use of '_timezone' to 'timezone' as this + difference is already adjusted for in tclWinPort.h. Removed unreferenced posixEpoch file-scope global. - * win/tclWinFile.c (WinReadLinkDirectory): Fix for 'Initializers - must be constant' with the driveSpec array using OpenWatcom. + * win/tclWinFile.c (WinReadLinkDirectory): Fix for 'Initializers must + be constant' with the driveSpec array using OpenWatcom. 2003-08-27 Don Porter <dgp@users.sourceforge.net> * generic/tclUtil.c: Corrected [Bug 411825] and other bugs in TclNeedSpace() where non-breaking space (\u00A0) and backslash-escaped - spaces were handled incorrectly. + spaces were handled incorrectly. * tests/util.test: Added new tests util-8.[2-6]. 2003-08-26 David Gravereaux <davygrvy@pobox.com> * generic/tcl.h: Added some support for the LCC-Win32 compiler. - Unfortunetly, this compiler has a bug in its preprocessor and - can't build Tcl even with this minor patch. Also added some - support for the OpenWatcom compiler. A new win/makefile.wc to - follow soon. + Unfortunetly, this compiler has a bug in its preprocessor and can't + build Tcl even with this minor patch. Also added some support for the + OpenWatcom compiler. A new win/makefile.wc to follow soon. 2003-08-25 Donal K. Fellows <fellowsd@cs.man.ac.uk> * tools/genStubs.tcl (genStubs::makeDecl): A more subtle way of - generating stubbed declarations allows us to have declarations of - a function in multiple interfaces simultaneously. + generating stubbed declarations allows us to have declarations of a + function in multiple interfaces simultaneously. * generic/tcl.decls: Duplicated some namespace declarations from - tclInt.decls here, as mandated by TIP #139. This is OK since the + tclInt.decls here, as mandated by TIP #139. This is OK since the declarations match and will end up using the declarations in the - public code from now on because of #include ordering. Keeping the - old declarations in tclInt.decls; there's no need to gratuitously - break compatability for those extensions which are already clients - of the namespace code. + public code from now on because of #include ordering. Keeping the old + declarations in tclInt.decls; there's no need to gratuitously break + compatability for those extensions which are already clients of the + namespace code. 2003-08-23 Zoran Vasiljevic <zoran@archiwrae.com> - * generic/tclIOUtil.c: merged fixes for thread-unsafe - handling of filesystem records [Bug #753315]. - This also fixed the Bug #788780 - * generic/tclPathObj.c: merged fixes for thread-unsafe - handling of filesystem records [Bug #753315]. + * generic/tclIOUtil.c: merged fixes for thread-unsafe handling of + filesystem records [Bug 753315]. This also fixed the [Bug 788780] + * generic/tclPathObj.c: merged fixes for thread-unsafe handling of + filesystem records. [Bug 753315] - * generic/tclFileSystem.h: merged fixes for thread-unsafe - handling of filesystem records [Bug #753315]. + * generic/tclFileSystem.h: merged fixes for thread-unsafe handling of + filesystem records. [Bug 753315] 2003-08-19 Pat Thoyts <patthoyts@users.sourceforge.net> - * win/tclWinSerial.c (SerialErrorStr): Fixed a syntax error - created in the previous code cleanup. + * win/tclWinSerial.c (SerialErrorStr): Fixed a syntax error created in + the previous code cleanup. 2003-08-19 Donal K. Fellows <fellowsd@cs.man.ac.uk> - * win/tclWinSerial.c: Adjusted commenting and spacing usage to - follow the principles of the Style Guide better. + * win/tclWinSerial.c: Adjusted commenting and spacing usage to follow + the principles of the Style Guide better. 2003-08-18 Mo DeJong <mdejong@users.sourceforge.net> * win/configure: Regen. - * win/tcl.m4 (SC_ENABLE_SYMBOLS): Use test instead - of -eq, which does not work. [Bug 781109] + * win/tcl.m4 (SC_ENABLE_SYMBOLS): Use test instead of -eq, which does + not work. [Bug 781109] 2003-08-13 Chengye Mao <chengye.geo@yahoo.com> - * win/tclWinPipe.c: fixed a bug in BuildCommandLine. - This bug built a command line with a missing space between - tclpipe.dll and the following arguments. It caused error - in Windows 98 when exec command.com (e.g. dir) [Bug 789040] + * win/tclWinPipe.c: fixed a bug in BuildCommandLine. This bug built a + command line with a missing space between tclpipe.dll and the + following arguments. It caused error in Windows 98 when exec + command.com (e.g. dir). [Bug 789040] 2003-08-11 Donal K. Fellows <fellowsd@cs.man.ac.uk> @@ -1006,7 +990,7 @@ * doc/lrepeat.n: patch, updated to the HEAD * tests/lrepeat.test: and matching the core style. * generic/tclBasic.c (buildIntCmds): Splice into core. - * generic/tclInt.h: + * generic/tclInt.h: * doc/list.n: Cross-reference. 2003-08-06 Jeff Hobbs <jeffh@ActiveState.com> @@ -1017,50 +1001,45 @@ * library/msgcat/msgcat.tcl: Added escape so that non-Windows * library/msgcat/pkgIndex.tcl: platforms do not try to use the - registry package. This can save a costly and pointless package - search. Bumped to 1.3.1. Thanks to Dave Bodenstab. [Bug 781609]. + registry package. This can save a costly and pointless package search. + Bumped to 1.3.1. Thanks to Dave Bodenstab. [Bug 781609] 2003-08-05 Miguel Sofer <msofer@users.sf.net> * generic/tclExecute.c (INST_INVOKE, INST_EVAL, INST_PUSH_RESULT): added a Tcl_ResetResult(interp) at each point where the interp's - result is pushed onto the stack, to avoid keeping an extra - reference that may cause costly Tcl_Obj duplication [Bug 781585] - Detected by Franco Violi, analyzed by Peter Spjuth and Donal - Fellows. + result is pushed onto the stack, to avoid keeping an extra reference + that may cause costly Tcl_Obj duplication. Detected by Franco Violi, + analyzed by Peter Spjuth and Donal Fellows. [Bug 781585] 2003-07-28 Vince Darley <vincentdarley@users.sourceforge.net> * doc/FileSystem.3: - * doc/Translate.3: better documentation of Tcl_TranslateFileName - and related functions [Bug 775220] + * doc/Translate.3: better documentation of Tcl_TranslateFileName and + related functions. [Bug 775220] 2003-07-24 Mo DeJong <mdejong@users.sourceforge.net> - * generic/tcl.h: Revert change made on 2003-07-21 - since it made the sizeof(Tcl_Obj) different for - regular vs mem debug builds. - * generic/tclInt.h: Define TclDecrRefCount in terms - of Tcl_DbDecrRefCount which removes one layer of - inderection. + * generic/tcl.h: Revert change made on 2003-07-21 since it made the + sizeof(Tcl_Obj) different for regular vs mem debug builds. + * generic/tclInt.h: Define TclDecrRefCount in terms of + Tcl_DbDecrRefCount which removes one layer of inderection. * generic/tclObj.c (TclDbInitNewObj, Tcl_DbIncrRefCount, - Tcl_DbDecrRefCount, Tcl_DbIsShared): - Define ThreadSpecificData that contains a hashtable. - The table is used to ensure that a Tcl_Obj is only - acted upon in the thread that allocated it. This - checking code is enabled only when mem debug and - threads are enabled. + (Tcl_DbDecrRefCount, Tcl_DbIsShared): Define ThreadSpecificData that + contains a hashtable. The table is used to ensure that a Tcl_Obj is + only acted upon in the thread that allocated it. This checking code is + enabled only when mem debug and threads are enabled. 2003-07-24 Don Porter <dgp@users.sourceforge.net> - * tests/async.test: Added several tests that demonstrate Tcl - * tests/basic.test: Bug 489537, Tcl's longstanding failure to - * tests/dict.test: properly quote any leading '#' character - * tests/dstring.test: when generating the string rep of a list - * tests/list.test: so that the comment-power of that character - * tests/parse.test: is hidden from any [eval], in order to - * tests/util.test: satisfy the documentation that [list] does - [eval]-safe quoting. + * tests/async.test: Added several tests that demonstrate [Bug + * tests/basic.test: 489537], Tcl's longstanding failure to + * tests/dict.test: properly quote any leading '#' character when + * tests/dstring.test: generating the string rep of a list so that + * tests/list.test: the comment-power of that character is hidden + * tests/parse.test: from any [eval], in order to satisfy the + * tests/util.test: documentation that [list] does [eval]-safe + quoting. 2003-07-24 Reinhard Max <max@suse.de> @@ -1069,12 +1048,12 @@ * ChangeLog.2002 (new file): * ChangeLog: broke changes from 2002 into ChangeLog.2002 to reduce - size of the main ChangeLog. + size of the main ChangeLog. 2003-07-23 Daniel Steffen <das@users.sourceforge.net> - * unix/Makefile.in: changes to html-tcl & html-tk - targets for compatibility with non-gnu makes. + * unix/Makefile.in: changes to html-tcl & html-tk targets for + compatibility with non-gnu makes. * unix/Makefile.in: added macosx/README to dist target. @@ -1086,48 +1065,41 @@ 2003-07-21 Mo DeJong <mdejong@users.sourceforge.net> - Check that the thread incrementing or decrementing - the ref count of a Tcl_Obj is the thread that - originally allocated the thread. This fail fast - behavior will catch programming errors that - allow a single Tcl_Obj to be accessed from multiple - threads. - - * generic/tcl.h (Tcl_Obj): Add allocThread member - to Tcl_Obj. This member records the thread id the - Tcl_Obj was allocated. It is used to check that - any future ref count incr or decr is done from - the same thread that allocated the Tcl_Obj. - This member is defined only when threads and - mem debug are enabled. - * generic/tclInt.h (TclNewObj, TclDbNewObj, - TclDecrRefCount): - Define TclNewObj and TclDbNewObj using TclDbInitNewObj - when mem debug is enabled. This fixes a problem where - TclNewObj calls did not work the same as TclDbNewObj - when mem debug was enabled. + Check that the thread incrementing or decrementing the ref count of a + Tcl_Obj is the thread that originally allocated the thread. This fail + fast behavior will catch programming errors that allow a single + Tcl_Obj to be accessed from multiple threads. + + * generic/tcl.h (Tcl_Obj): Add allocThread member to Tcl_Obj. This + member records the thread id the Tcl_Obj was allocated. It is used to + check that any future ref count incr or decr is done from the same + thread that allocated the Tcl_Obj. This member is defined only when + threads and mem debug are enabled. + * generic/tclInt.h (TclNewObj, TclDbNewObj, TclDecrRefCount): + Define TclNewObj and TclDbNewObj using TclDbInitNewObj when mem debug + is enabled. This fixes a problem where TclNewObj calls did not work + the same as TclDbNewObj when mem debug was enabled. * generic/tclObj.c (TclDbInitNewObj, Tcl_DbIncrRefCount, - Tcl_DbDecrRefCount): Add new helper to init Tcl_Obj - members when mem debug is enabled. Init the allocThread - member in TclDbInitNewObj and check it in - Tcl_DbIncrRefCount and Tcl_DbDecrRefCount to make sure - a Tcl_Obj allocated in one thread is not being acted - upon in another thread. + (Tcl_DbDecrRefCount): Add new helper to init Tcl_Obj members when mem + debug is enabled. Init the allocThread member in TclDbInitNewObj and + check it in Tcl_DbIncrRefCount and Tcl_DbDecrRefCount to make sure a + Tcl_Obj allocated in one thread is not being acted upon in another + thread. 2003-07-21 Vince Darley <vincentdarley@users.sourceforge.net> - * test/cmdAH.test: ensure certain tests run in local filesystem - [Bug 748960] + * test/cmdAH.test: ensure certain tests run in local filesystem. [Bug + 748960] 2003-07-18 Daniel Steffen <das@users.sourceforge.net> - * macosx/Makefile: added option to allow installing manpages - in addition to default html help. + * macosx/Makefile: added option to allow installing manpages in + addition to default html help. 2003-07-18 Donal K. Fellows <fellowsd@cs.man.ac.uk> - * doc/Utf.3: Tightened up documentation of Tcl_UtfNext and - Tcl_UtfPrev to better match the behaviour. [Bug 769895] + * doc/Utf.3: Tightened up documentation of Tcl_UtfNext and Tcl_UtfPrev + to better match the behaviour. [Bug 769895] 2003-07-18 Jeff Hobbs <jeffh@ActiveState.com> @@ -1146,18 +1118,18 @@ only in the caller's namespace. Documentation made more precise on the subject. [Bug 706359] - * doc/AddErrInfo.3: Improved consistency of documentation - * doc/CrtTrace.3: by using "null" everywhere to refer to - * doc/Encoding.3: the character '\0', and using "NULL" - * doc/Eval.3: everywhere to refer to the value of a - * doc/GetIndex.3: pointer that points to nowhere. - * doc/Hash.3: Also dropped references to ASCII that - * doc/LinkVar.3: are no longer true, and standardized on - * doc/Macintosh.3: the hyphenated spelling of "null-terminated". - * doc/OpenFileChnl.3: - * doc/SetVar.3: - * doc/StringObj.3: - * doc/Utf.3: + * doc/AddErrInfo.3: Improved consistency of documentation by + * doc/CrtTrace.3: using "null" everywhere to refer to the + * doc/Encoding.3: character '\0', and using "NULL" everywhere + * doc/Eval.3: to refer to the value of a pointer that points + * doc/GetIndex.3: to nowhere. Also dropped references to ASCII + * doc/Hash.3: that are no longer true, and standardized on + * doc/LinkVar.3: the hyphenated spelling of "null-terminated". + * doc/Macintosh.3: + * doc/OpenFileChnl.3: + * doc/SetVar.3: + * doc/StringObj.3: + * doc/Utf.3: * doc/CrtSlave.3 (Tcl_MakeSafe): Removed warning about possible deprecation (no TIP on that). @@ -1170,17 +1142,16 @@ * macosx/Makefile: Rewrote buildsystem for Mac OS X framework build to be purely make driven; in order to become independent of Apple's closed-source IDE and build tool. The changes are intended to be - transparent to the Makefile user, all existing make targets and - cmd line variable overrides should continue to work. - Changed build to only include tcl specific html help in Tcl.framework, - the tk specific html help is now included in Tk.framework. - Added var to allow overriding of tclsh used during html help - building (Landon Fuller). + transparent to the Makefile user, all existing make targets and cmd + line variable overrides should continue to work. Changed build to only + include tcl specific html help in Tcl.framework, the tk specific html + help is now included in Tk.framework. Added var to allow overriding of + tclsh used during html help building (Landon Fuller). * macosx/Tcl.pbproj/project.pbxproj: - * macosx/Tcl.pbproj/jingham.pbxuser: Changed to purely call through - to the make driven buildsystem; Tcl.framework is no longer assembled - by ProjectBuilder. + * macosx/Tcl.pbproj/jingham.pbxuser: Changed to purely call through to + the make driven buildsystem; Tcl.framework is no longer assembled by + ProjectBuilder. Set default SYMROOT in target options to simplify setting up PB (manually setting common build folder for tcl & tk no longer needed). @@ -1188,17 +1159,17 @@ tcl or tk html help files; the default behaviour with none of the new options is to build both, as before. - * unix/Makefile.in: Added targets for building only the tcl or tk help. + * unix/Makefile.in: Added targets for building only the tcl or tk help - * macosx/README (new): Tcl specific excerpts of tk/macosx/README. + * macosx/README (new): Tcl specific excerpts of tk/macosx/README. * generic/tcl.h: Updated reminder comment about editing macosx/Tcl.pbproj/project.pbxproj when version number changes. 2003-07-16 Mumit Khan <khan@nanotech.wisc.edu> - * generic/tclPathObj.c (SetFsPathFromAny): Add Cygwin specific - code to convert POSIX filename to native format. + * generic/tclPathObj.c (SetFsPathFromAny): Add Cygwin specific code to + convert POSIX filename to native format. * generic/tclFileName.c (Tcl_TranslateFileName): And remove from here. (TclDoGlob): Adjust for cygwin and append / for dirs instead of \ * win/tclWinFile.c (TclpObjChdir): Use chdir on Cygwin. @@ -1209,12 +1180,12 @@ * library/safe.tcl (FileInAccessPath): normalize paths before comparison. [Bug 759607] (myers) - * unix/tclUnixNotfy.c (NotifierThreadProc): correct size of found - and word vars from int to long. [Bug 767578] (hgo) + * unix/tclUnixNotfy.c (NotifierThreadProc): correct size of found and + word vars from int to long. [Bug 767578] (hgo) - * generic/tcl.h: add recognition of -DTCL_UTF_MAX=6 on the - * generic/regcustom.h: make line to support UCS-4 mode. No config - arg at this time, as it is not the recommended build mode. + * generic/tcl.h: Add recognition of -DTCL_UTF_MAX=6 on the make + * generic/regcustom.h: line to support UCS-4 mode. No config arg at + this time, as it is not the recommended build mode. * generic/tclPreserve.c: In Result and Preserve'd routines, do not * generic/tclUtil.c: assume that ckfree == free, as that is not @@ -1223,8 +1194,8 @@ 2003-07-16 Donal K. Fellows <fellowsd@cs.man.ac.uk> * doc/CrtSlave.3 (Tcl_MakeSafe): Updated documentation to strongly - discourage use. IMHO code outside the core that uses this - function is a bug... [Bug 655300] + discourage use. IMHO code outside the core that uses this function is + a bug... [Bug 655300] 2003-07-16 Don Porter <dgp@users.sourceforge.net> @@ -1234,13 +1205,13 @@ Silence compiler warnings about unreached lines. * library/tcltest/tcltest.tcl (ProcessFlags): Corrected broken call - * library/tcltest/pkgIndex.tcl: to [lrange]. Bumped - to version 2.2.4. [Bug 772333] + * library/tcltest/pkgIndex.tcl: to [lrange]. Bumped to + version 2.2.4. [Bug 772333] 2003-07-15 Mo DeJong <mdejong@users.sourceforge.net> - * unix/dltest/pkga.c (Pkga_EqObjCmd): Fix typo - that was causing a crash in load.test. + * unix/dltest/pkga.c (Pkga_EqObjCmd): Fix typo that was causing a + crash in load.test. 2003-07-15 Donal K. Fellows <fellowsd@cs.man.ac.uk> @@ -1248,20 +1219,20 @@ 2003-07-15 Don Porter <dgp@users.sourceforge.net> - * doc/http.n: Updated SYNOPSIS to match actual syntax of - commands. [Bug 756112] + * doc/http.n: Updated SYNOPSIS to match actual syntax of commands. + [Bug 756112] * unix/dltest/pkga.c: Updated to not use Tcl_UtfNcmp and counted - strings instead of strcmp (not defined in any #include'd header) - and presumed NULL-terminated strings. + strings instead of strcmp (not defined in any #include'd header) and + presumed NULL-terminated strings. - * generic/tclCompCmds.c (TclCompileIfCmd): Prior fix of Bug 711371 - on 2003-04-07 introduced a buffer overflow. Corrected. [Bug 771613] + * generic/tclCompCmds.c (TclCompileIfCmd): Prior fix of Bug 711371 on + 2003-04-07 introduced a buffer overflow. Corrected. [Bug 771613] 2003-07-15 Kevin B. Kenny <kennykb@acm.org> - * win/rules.vc: Added a missing $(OPTDEFINES) which broke the - build if STATS=memdbg was specified. + * win/rules.vc: Added a missing $(OPTDEFINES) which broke the build if + STATS=memdbg was specified. 2003-07-15 Donal K. Fellows <fellowsd@cs.man.ac.uk> @@ -1270,16 +1241,16 @@ 2003-07-11 Donal K. Fellows <fellowsd@cs.man.ac.uk> - * tests/binary.test (binary-46.*): Tests to help enforce the - current behaviour. + * tests/binary.test (binary-46.*): Tests to help enforce the current + behaviour. * doc/binary.n: Documented that [binary format a] and [binary scan a] - do encoding conversion by dropping high bytes, unlike the rest of - the core. [Bug 735364] + do encoding conversion by dropping high bytes, unlike the rest of the + core. [Bug 735364] 2003-07-11 Don Porter <dgp@users.sourceforge.net> * library/package.tcl: Corrected [pkg_mkIndex] bug reported on - comp.lang.tcl. The indexer was searching for newly indexed packages + comp.lang.tcl. The indexer was searching for newly indexed packages instead of newly provided packages. 2003-07-08 Vince Darley <vincentdarley@users.sourceforge.net> @@ -1302,15 +1273,14 @@ filesystems. [Patch 760768] Also a little general cleanup. * generic/tclCmdMZ.c (Tcl_StringObjCmd): Made [string map] accept - dictionaries for maps. This is much trickier than it looks, since - map entry ordering is significant. [Bug 759936] + dictionaries for maps. This is much trickier than it looks, since map + entry ordering is significant. [Bug 759936] - * generic/tclVar.c (Tcl_ArrayObjCmd, TclArraySet): Made [array - get] and [array set] work with dictionaries, producing them and - consuming them. Note that for compatability reasons, you will - never get a dict from feeding a string literal to [array set] - since that alters the trace behaviour of "multi-key" sets. - [Bug 759935] + * generic/tclVar.c (Tcl_ArrayObjCmd, TclArraySet): Made [array get] + and [array set] work with dictionaries, producing them and consuming + them. Note that for compatability reasons, you will never get a dict + from feeding a string literal to [array set] since that alters the + trace behaviour of "multi-key" sets. [Bug 759935] 2003-06-23 Vince Darley <vincentdarley@users.sourceforge.net> @@ -1330,16 +1300,15 @@ 2003-06-25 Donal K. Fellows <fellowsd@cs.man.ac.uk> * generic/tclTrace.c: New file, factoring out of virtually all the - various trace-related things from tclBasic.c and tclCmdMZ.c with - the goal of making this a separate maintenance area. + various trace-related things from tclBasic.c and tclCmdMZ.c with the + goal of making this a separate maintenance area. 2003-06-25 Mo DeJong <mdejong@users.sourceforge.net> * unix/configure: Regen. - * unix/tcl.m4 (SC_CONFIG_CFLAGS): Add -ieee when - compiling with cc and add -mieee when compiling - with gcc under OSF1-V5 "Tru64" systems. - [Bug 748957] + * unix/tcl.m4 (SC_CONFIG_CFLAGS): Add -ieee when compiling with cc and + add -mieee when compiling with gcc under OSF1-V5 "Tru64" systems. [Bug + 748957] 2003-06-24 Donal K. Fellows <fellowsd@cs.man.ac.uk> @@ -1353,15 +1322,14 @@ 2003-06-24 Mo DeJong <mdejong@users.sourceforge.net> - * unix/tclUnixPort.h: #undef inet_ntoa before - #define to avoid compiler warning under freebsd. - [Bug 745844] + * unix/tclUnixPort.h: #undef inet_ntoa before #define to avoid + compiler warning under freebsd. [Bug 745844] 2003-06-23 Pat Thoyts <patthoyts@users.sourceforge.net> * doc/dde.n: Committed TIP #135 which changes the - * win/tclWinDde.c: -exact option to -force. Also cleaned - * tests/winDde.test: a bug in the tests. + * win/tclWinDde.c: -exact option to -force. Also cleaned a + * tests/winDde.test: bug in the tests. * library/dde/pkgIndex.tcl: Incremented version to 1.2.5 * doc/dde.n: Committed TIP #120 which provides the @@ -1371,70 +1339,70 @@ 2003-06-23 Vince Darley <vincentdarley@users.sourceforge.net> - * generic/tclFCmd.c: fix to bad error message when trying to - do 'file copy foo ""'. [Bug 756951] + * generic/tclFCmd.c: fix to bad error message when trying to do 'file + copy foo ""'. [Bug 756951] * tests/fCmd.test: added two new tests for the bug. * win/tclWinFile.c: - * win/tclWin32Dll.c: recommitted some filesystem globbing - speed-ups, but disabled some on the older Win 95/98/ME where - they don't seem to work. + * win/tclWin32Dll.c: recommitted some filesystem globbing speed-ups, + but disabled some on the older Win 95/98/ME where they don't seem to + work. * doc/FileSystem.3: documentation fix [Bug 720634] 2003-06-18 Miguel Sofer <msofer@users.sf.net> - * generic/tclNamesp.c (Tcl_Export): removed erroneous comments - [Bug 756744] + * generic/tclNamesp.c (Tcl_Export): removed erroneous comments. [Bug + 756744] 2003-06-17 Vince Darley <vincentdarley@users.sourceforge.net> - * win/makefile.vc: fixes to check-in below so compilation now - works again on Windows. + * win/makefile.vc: fixes to check-in below so compilation now works + again on Windows. * generic/tclCmdMZ.c: * tests/regexp.test: fixing of bugs related to regexp and regsub - matching of empty strings. Addition of a number of new tests. - [Bug 755335] + matching of empty strings. Addition of a number of new tests. [Bug + 755335] 2003-06-16 Andreas Kupries <andreask@activestate.com> - * win/Makefile.in: Haven't heard back from David for a week. - * win/configure: Now committing the remaining changes. - * win/configure.in: Note: In active contact with Helmut Giese - * win/makefile.vc: about the borland relatedchanges. This part - * win/rules.vc: will see future updates. - * win/tcl.m4: + * win/Makefile.in: Haven't heard back from David for a week. Now + * win/configure: committing the remaining changes. + * win/configure.in: Note: In active contact with Helmut Giese about + * win/makefile.vc: the borland relatedchanges. This part will see + * win/rules.vc: future updates. + * win/tcl.m4: * win/makefile.bc: 2003-06-10 Andreas Kupries <andreask@activestate.com> * generic/tclConfig.c (ASSOC_KEY): Changed the key to - "tclPackageAboutDict" (tcl prefix) to make collisions with the - keys of other packages more unlikely. + "tclPackageAboutDict" (tcl prefix) to make collisions with the keys of + other packages more unlikely. 2003-06-10 Miguel Sofer <msofer@users.sf.net> * generic/tclBasic.c: * generic/tclExecute.c: let TclExecuteObjvInternal call - TclInterpReady instead of relying on its callers to do so; fix for - the part of [Bug 495830] that is new in 8.4. + TclInterpReady instead of relying on its callers to do so; fix for the + part of [Bug 495830] that is new in 8.4. * tests/interp.test: Added tests 18.9 (knownbug) and 18.10 2003-06-09 Andreas Kupries <andreask@activestate.com> * generic/tcl.decls: Ported the changes from the * generic/tcl.h: 'tip-59-implementation' branch into the CVS - * generic/tclBasic.c: head. Regenerated stub table. Regenerated - * generic/tclInt.h: the configure's scripts, with help from Joe - * generic/tclDecls.h English. + * generic/tclBasic.c: head. Regenerated stub table. Regenerated the + * generic/tclInt.h: configure's scripts, with help from Joe English. + * generic/tclDecls.h: * generic/tclStubInit.c: * generic/tclConfig.c: * generic/tclPkgConfig.c: - * unix/Makefile.in: - * unix/configure.in: The changes in the windows section are not - * unix/tcl.m4: yet committed, they await feedback from - * unix/mkLinks: David Gravereaux. + * unix/Makefile.in: + * unix/configure.in: The changes in the windows section are not yet + * unix/tcl.m4: committed, they await feedback from David + * unix/mkLinks: Gravereaux. * doc/RegConfig.3: * mac/tclMacPkgConfig.c: * tests/config.test: @@ -1447,17 +1415,16 @@ 2003-06-04 Joe Mistachkin <joe@mistachkin.com> - * tools/man2help.tcl: Added duplicate help section checking - * tools/index.tcl: and corrected a comment typo for the - getTopics proc in index.tcl [Bug #748700]. + * tools/man2help.tcl: Added duplicate help section checking and + * tools/index.tcl: corrected a comment typo for the getTopics proc + in index.tcl. [Bug 748700] 2003-06-02 Vince Darley <vincentdarley@users.sourceforge.net> * win/tclWinFCmd.c: - * tests/fCmd.test: fix to [Bug #747575] in which a bad error - message is given when trying to rename a busy directory to - one with the same prefix, but not the same name. Added three - new tests. + * tests/fCmd.test: fix to [Bug #747575] in which a bad error message + is given when trying to rename a busy directory to one with the same + prefix, but not the same name. Added three new tests. 2003-05-23 D. Richard Hipp <drh@hwaci.com> @@ -1467,30 +1434,28 @@ 2003-05-23 Don Porter <dgp@users.sourceforge.net> * generic/tclObj.c (tclCmdNameType): Converted internal rep - management of the cmdName Tcl_ObjType the opposite way, to always - use the twoPtrValue instead of always using the otherValuePtr. - Previous fix on 2003-05-12 broke several extensions that wanted - to poke around with the twoPtrValue.ptr2 value of a cmdName - Tcl_Obj, like TclBlend and e4graph. [Bug 726018] - Thanks to George Petasis for the bug report and Jacob Levy for - testing assistance. + management of the cmdName Tcl_ObjType the opposite way, to always use + the twoPtrValue instead of always using the otherValuePtr. Previous + fix on 2003-05-12 broke several extensions that wanted to poke around + with the twoPtrValue.ptr2 value of a cmdName Tcl_Obj, like TclBlend + and e4graph. [Bug 726018] + Thanks to George Petasis for the bug report and Jacob Levy for testing + assistance. 2003-05-23 Mo DeJong <mdejong@users.sourceforge.net> - * unix/mkLinks: Set the var S to "" at the top - of the file to avoid error when user has set S - to something. - [Tk Bug #739833] + * unix/mkLinks: Set the var S to "" at the top of the file to avoid + error when user has set S to something. [Tk Bug 739833] 2003-05-22 Daniel Steffen <das@users.sourceforge.net> - * macosx/Tcl.pbproj/project.pbxproj: added missing references to - new source files tclPathObj.c and tclMacOSXFCmd.c. + * macosx/Tcl.pbproj/project.pbxproj: added missing references to new + source files tclPathObj.c and tclMacOSXFCmd.c. - * macosx/tclMacOSXBundle.c: fixed a problem that caused only the - first call to Tcl_MacOSXOpenVersionedBundleResources() for a given - bundle identifier to succeed. This caused the tcl runtime library - not to be found in all interps created after the inital one. + * macosx/tclMacOSXBundle.c: fixed a problem that caused only the first + call to Tcl_MacOSXOpenVersionedBundleResources() for a given bundle + identifier to succeed. This caused the tcl runtime library not to be + found in all interps created after the inital one. 2003-05-19 Kevin B. Kenny <kennykb@hippolyta> @@ -1499,31 +1464,30 @@ 2003-05-19 Daniel Steffen <das@users.sourceforge.net> - * macosx/Tcl.pbproj/project.pbxproj: changed tclConfig.sh location - in versioned framework subdirectories to be identical to location - in framework toplevel; fixed stub library symbolic links to be - tcl version specific. + * macosx/Tcl.pbproj/project.pbxproj: changed tclConfig.sh location in + versioned framework subdirectories to be identical to location in + framework toplevel; fixed stub library symbolic links to be tcl + version specific. * unix/tclUnixTime.c: fixed typo. 2003-05-18 Kevin Kenny <kennykb@acm.org> - * compat/strftime.c: Modified TclpStrftime to return its - * generic/tclClock.c: result in UTF-8 encoding, and removed - * mac/tclMacTime.c: the conversion from system encoding to - * unix/tclUnixTime.c: UTF-8 from [clock format]. Needed to - * win/tclWinTime.c: avoid double conversion of the timezone - name on Windows systems. [Bug 624408] + * compat/strftime.c: Modified TclpStrftime to return its result in + * generic/tclClock.c: UTF-8 encoding, and removed the conversion from + * mac/tclMacTime.c: system encoding to UTF-8 from [clock format]. + * unix/tclUnixTime.c: Needed to avoid double conversion of the + * win/tclWinTime.c: timezone name on Windows systems. [Bug 624408] 2003-05-16 Pat Thoyts <patthoyts@users.sourceforge.net> - * library/dde/pkgIndex.tcl: Applied TIP #130 which provides - * tests/winDde.test: for unique dde server names. Added - * win/tclWinDde.c: some more tests. Fixes [Bug 219293] + * library/dde/pkgIndex.tcl: Applied TIP #130 which provides for + * tests/winDde.test: unique dde server names. Added some more + * win/tclWinDde.c: tests. Fixes [Bug 219293] * doc/dde.n: Updated documentation re TIP #130. - * tests/winDde.test: Applied patch for [Bug 738929] by KKB and - changed to new-style tests. + * tests/winDde.test: Applied patch for [Bug 738929] by KKB and changed + to new-style tests. 2003-05-16 Kevin B. Kenny <kennykb@acm.org> @@ -1536,33 +1500,33 @@ 2003-05-15 Kevin B. Kenny <kennykb@acm.org> * generic/tclGetDate.y: added further hackery to the yacc - * generic/tclDate.c: post-processing to arrange for the - * unix/Makefile.in: code to set up exit handlers to free - the stacks [Bug 736425]. + * generic/tclDate.c: post-processing to arrange for the code to set + * unix/Makefile.in: up exit handlers to free the stacks. [Bug + 736425] 2003-05-15 Jeff Hobbs <jeffh@ActiveState.com> - * win/tclWinFile.c (TclpMatchInDirectory): revert glob code to - r1.44 as 2003-04-11 optimizations broke Windows98 glob'ing. + * win/tclWinFile.c (TclpMatchInDirectory): revert glob code to r1.44 + as 2003-04-11 optimizations broke Windows98 glob'ing. * doc/socket.n: nroff font handling correction * library/encoding/gb2312-raw.enc (new): This is the original - gb2312.enc renamed to allow for it to still be used. This is - needed by Tk (unix) because X fonts with gb2312* charsets really - do want the original gb2312 encoding. [Bug 557030] + gb2312.enc renamed to allow for it to still be used. This is needed by + Tk (unix) because X fonts with gb2312* charsets really do want the + original gb2312 encoding. [Bug 557030] 2003-05-14 Donal K. Fellows <fellowsd@cs.man.ac.uk> - * generic/tclCmdAH.c (Tcl_FormatObjCmd): Stop unwarranted demotion - of wide values to longs by formatting of int values. [Bug 699060] + * generic/tclCmdAH.c (Tcl_FormatObjCmd): Stop unwarranted demotion of + wide values to longs by formatting of int values. [Bug 699060] 2003-05-14 Jeff Hobbs <jeffh@ActiveState.com> * library/encoding/gb2312.enc: copy euc-cn.enc over original - gb2312.enc. gb2312.enc appeared to not work as expected, and most - uses of gb2312 really mean euc-cn (which may be the cause of the - problem). [Bug 557030] + gb2312.enc. gb2312.enc appeared to not work as expected, and most uses + of gb2312 really mean euc-cn (which may be the cause of the problem). + [Bug 557030] 2003-05-14 Daniel Steffen <das@users.sourceforge.net> @@ -1602,50 +1566,48 @@ * tests/unixFCmd.test: added tests of -readonly attribute. - * tests/macOSXFCmd.test (new): tests of macosx file attributes and - of preservation of attributes & resource fork during [file copy]. + * tests/macOSXFCmd.test (new): tests of macosx file attributes and of + preservation of attributes & resource fork during [file copy]. * tests/macFCmd.test: restore -readonly attribute of test dir, as otherwise its removal can fail on unices supporting -readonly. 2003-05-13 David Gravereaux <davygrvy@pobox.com> - * generic/tclEnv.c: Another putenv() copy behavior problem - repaired when compiling on windows and using microsoft's runtime. - [Bug 736421] + * generic/tclEnv.c: Another putenv() copy behavior problem repaired + when compiling on windows and using microsoft's runtime. [Bug 736421] 2003-05-13 Jeff Hobbs <jeffh@ActiveState.com> * generic/tclIOUtil.c: ensure cd is thread-safe. - [Bug #710642] (vasiljevic) + [Bug 710642] (vasiljevic) 2003-05-13 Donal K. Fellows <fellowsd@cs.man.ac.uk> - * generic/tclEvent.c (Tcl_Finalize): Removed unused variable to - reduce compiler warnings. [Bug 664745] + * generic/tclEvent.c (Tcl_Finalize): Removed unused variable to reduce + compiler warnings. [Bug 664745] 2003-05-13 Joe Mistachkin <joe@mistachkin.com> * generic/tcl.decls: Changed Tcl_JoinThread parameter name from * generic/tclDecls.h: "id" to "threadId". [Bug 732477] * unix/tclUnixThrd.c: - * win/tclWinThrd.c: + * win/tclWinThrd.c: * mac/tclMacThrd.c: 2003-05-13 Daniel Steffen <das@users.sourceforge.net> * generic/tcl.decls: - * macosx/tclMacOSXBundle.c: added extended version of the + * macosx/tclMacOSXBundle.c: added extended version of the Tcl_MacOSXOpenBundleResources() API taking an extra version number - argument: Tcl_MacOSXOpenVersionedBundleResources(). - This is needed to be able to access bundle resources in versioned - frameworks such as Tcl and Tk, otherwise if multiple versions were - installed, only the latest version's resources could be accessed. - [Bug 736774] + argument: Tcl_MacOSXOpenVersionedBundleResources(). This is needed to + be able to access bundle resources in versioned frameworks such as Tcl + and Tk, otherwise if multiple versions were installed, only the latest + version's resources could be accessed. [Bug 736774] * unix/tclUnixInit.c (Tcl_MacOSXGetLibraryPath): use new versioned - bundle resource API to get tcl runtime library for TCL_VERSION. - [Bug 736774] + bundle resource API to get tcl runtime library for TCL_VERSION. [Bug + 736774] * generic/tclPlatDecls.h: * generic/tclStubInit.c: regen. @@ -1658,13 +1620,13 @@ * tests/cmdAH.test: General clean-up of tests so that all tcltest-specific commands are protected by constraints and all - platforms see the same number of tests. [Bug 736431] + platforms see the same number of tests. [Bug 736431] 2003-05-12 Don Porter <dgp@users.sourceforge.net> * generic/tclInterp.c: (AliasObjCmd): Added refCounting of the words * tests/interp.test (interp-33.1): of the target of an interp - alias during its execution. Also added test. [Bug 730244]. + alias during its execution. Also added test. [Bug 730244] * generic/tclBasic.c (TclInvokeObjectCommand): objv[argc] is no longer set to NULL (Tcl_CreateObjCommand docs already say that it @@ -1674,7 +1636,7 @@ * generic/tclObj.c (tclCmdNameType): Corrected variable use of the otherValuePtr or the twoPtrValue.ptr1 fields to store a - (ResolvedCmdName *) as the internal rep. [Bug 726018]. + (ResolvedCmdName *) as the internal rep. [Bug 726018] * doc/Eval.3: Corrected prototype for Tcl_GlobalEvalObj [Bug 727622]. @@ -1686,26 +1648,26 @@ 2003-05-10 Jeff Hobbs <jeffh@ActiveState.com> - * win/tclWinSerial.c (SerialCloseProc): correct mem leak on - closing a Windows serial port [Bug #718002] (schroedter) + * win/tclWinSerial.c (SerialCloseProc): correct mem leak on closing a + Windows serial port [Bug 718002] (schroedter) - * generic/tclCmdMZ.c (Tcl_StringObjCmd): prevent string repeat - crash when overflow sizes were given (throws error). [Bug #714106] + * generic/tclCmdMZ.c (Tcl_StringObjCmd): prevent string repeat crash + when overflow sizes were given (throws error). [Bug 714106] 2003-05-09 Joe Mistachkin <joe@mistachkin.com> * generic/tclThreadAlloc.c (TclFreeAllocCache): Fixed memory leak - caused by treating cachePtr as a TLS index [Bug 731754]. + caused by treating cachePtr as a TLS index. [Bug 731754] * win/tclAppInit.c (Tcl_AppInit): Fixed memory leaks caused by not freeing the memory allocated by setargv and the async handler created by Tcl_AppInit. An exit handler has been created that takes care of both leaks. In addition, Tcl_AppInit now uses ckalloc instead of Tcl_Alloc to allow for easier leak tracking and to be more consistent - with the rest of the Tcl core [Bugs 733156, 733221]. + with the rest of the Tcl core. [Bugs 733156, 733221] * tools/encoding/txt2enc.c (main): Fixed memory leak caused by failing - to free the memory used by the toUnicode array of strings [Bug 733221]. + to free the memory used by the toUnicode array of strings [Bug 733221] 2003-05-09 Miguel Sofer <msofer@users.sf.net> @@ -1716,17 +1678,17 @@ 2003-05-09 Donal K. Fellows <fellowsd@cs.man.ac.uk> - * generic/tclCmdMZ.c (Tcl_ReturnObjCmd): The array of strings - passed to Tcl_GetIndexFromObj must be NULL terminated. [Bug 735186] + * generic/tclCmdMZ.c (Tcl_ReturnObjCmd): The array of strings passed + to Tcl_GetIndexFromObj must be NULL terminated. [Bug 735186] Thanks to Joe Mistachkin for spotting this. 2003-05-07 Donal K. Fellows <fellowsd@cs.man.ac.uk> - * doc/trace.n: Fixed very strange language in the documentation - for 'trace add execution'. [Bug 729821] + * doc/trace.n: Fixed very strange language in the documentation for + 'trace add execution'. [Bug 729821] - * generic/tclCmdMZ.c (Tcl_TraceObjCmd): Made error message for - 'trace info' more consistent with documentation. [Bug 706961] + * generic/tclCmdMZ.c (Tcl_TraceObjCmd): Made error message for 'trace + info' more consistent with documentation. [Bug 706961] * generic/tclDictObj.c (DictInfoCmd): Fixed memory leak caused by confusion about string ownership. [Bug 731706] @@ -1755,89 +1717,85 @@ 2003-04-28 Donal K. Fellows <fellowsd@cs.man.ac.uk> - * generic/tclDictObj.c (DictIncrCmd): Updated to reflect the - behaviour with wide increments of the normal [incr] command. - * generic/tclInt.decls: Added TclIncrWideVar2 to internal stub - table and cleaned up. + * generic/tclDictObj.c (DictIncrCmd): Updated to reflect the behaviour + with wide increments of the normal [incr] command. + * generic/tclInt.decls: Added TclIncrWideVar2 to internal stub table + and cleaned up. * tests/incr.test (incr-3.*): - * generic/tclVar.c (TclIncrWideVar2, TclPtrIncrWideVar): - * generic/tclExecute.c (TclExecuteByteCode): - * generic/tclCmdIL.c (Tcl_IncrObjCmd): Make [incr] work when - trying to increment by wide values. [Bug 728838] + * generic/tclVar.c (TclIncrWideVar2, TclPtrIncrWideVar): + * generic/tclExecute.c (TclExecuteByteCode): + * generic/tclCmdIL.c (Tcl_IncrObjCmd): Make [incr] work when trying to + increment by wide values. [Bug 728838] * generic/tclCompCmds.c (TclCompileSwitchCmd): Default mode of - operation of [switch] is exact matching. [Bug 727563] + operation of [switch] is exact matching. [Bug 727563] 2003-04-25 Don Porter <dgp@users.sourceforge.net> * generic/tclBasic.c: Tcl_EvalObjv() failed to honor the - TCL_EVAL_GLOBAL flag when resolving command names. Tcl_EvalEx - passed a string rep including leading whitespace and comments - to TclEvalObjvInternal(). + TCL_EVAL_GLOBAL flag when resolving command names. Tcl_EvalEx passed a + string rep including leading whitespace and comments to + TclEvalObjvInternal(). 2003-04-25 Andreas Kupries <andreask@activestate.com> - * win/tclWinThrd.c: Applied SF patch #727271. This patch changes - the code to catch any errors returned by the windows functions - handling TLS ASAP instead of waiting to get some mysterious - crash later on due to bogus pointers. Patch provided by Joe - Mistachkin. + * win/tclWinThrd.c: Applied SF patch #727271. This patch changes the + code to catch any errors returned by the windows functions handling + TLS ASAP instead of waiting to get some mysterious crash later on due + to bogus pointers. Patch provided by Joe Mistachkin. - This is a stop-gap measure to deal with the low number of ?TLS - slots provided by some of the variants of Windows (60-80). + This is a stop-gap measure to deal with the low number of ?TLS slots + provided by some of the variants of Windows (60-80). 2003-04-24 Vince Darley <vincentdarley@users.sourceforge.net> - * generic/tclFileName.c: fix to bug reported privately by - Jeff where, for example, 'glob -path {[tcl]} *' gets confused - by the leading special character (which is escaped internally), - and instead lists files in '/'. Bug only occurs on Windows - where '\' is also a directory separator. + * generic/tclFileName.c: fix to bug reported privately by Jeff where, + for example, 'glob -path {[tcl]} *' gets confused by the leading + special character (which is escaped internally), and instead lists + files in '/'. Bug only occurs on Windows where '\' is also a directory + separator. * tests/fileName.test: added test for the above bug. 2003-04-22 Andreas Kupries <andreask@activestate.com> * The changes below fix SF bugs [593810], and [718045]. - * generic/tclIO.c (Tcl_CutChannel, Tcl_SpliceChannel): - Invoke TclpCutSockChannel and TclpSpliceSockChannel. + * generic/tclIO.c (Tcl_CutChannel, Tcl_SpliceChannel): Invoke + TclpCutSockChannel and TclpSpliceSockChannel. * generic/tclInt.h: Declare TclpCutSockChannel and - TclpSpliceSockChannel. + TclpSpliceSockChannel. * unix/tclUnixSock.c (TclpCutSockChannel, TclpSpliceSockChannel): - Dummy functions, on unix the sockets are _not_ handled - specially. + Dummy functions, on unix the sockets are _not_ handled specially. * mac/tclMacSock.c (TclpCutSockChannel, TclpSpliceSockChannel): - * win/tclWinSock.c (TclpCutSockChannel, TclpSpliceSockChannel): - New functions to handle socket specific cut/splice operations: - auto-initi of socket system for thread on splice, management of - the module internal per-thread list of sockets, management of - association of sockets with HWNDs for event notification. + * win/tclWinSock.c (TclpCutSockChannel, TclpSpliceSockChannel): New + functions to handle socket specific cut/splice operations: auto-init + of socket system for thread on splice, management of the module + internal per-thread list of sockets, management of association of + sockets with HWNDs for event notification. * win/tclWinSock.c (NewSocketInfo): Extended initialization - assignments to cover all items of the structure. During - debugging of the new code mentioned above I found that two - fileds could contain bogus data. + assignments to cover all items of the structure. During debugging of + the new code mentioned above I found that two fileds could contain + bogus data. * win/tclWinFile.c: Added #undef HAVE_NO_FINDEX_ENUMS before - definition because when compiling in debug mode the compiler - complains about a redefinition, and this warning is also treated - as an error. + definition because when compiling in debug mode the compiler complains + about a redefinition, and this warning is also treated as an error. 2003-04-21 Don Porter <dgp@users.sourceforge.net> * library/tcltest/tcltest.tcl: When the return code of a test does - not meet expectations, report that as the reason for test failure, - and do not attempt to check the test result for correctness. - [Bug 725253] + not meet expectations, report that as the reason for test failure, and + do not attempt to check the test result for correctness. [Bug 725253] 2003-04-18 Jeff Hobbs <jeffh@ActiveState.com> * win/tclWinInt.h (VER_PLATFORM_WIN32_CE): conditionally define. - * win/tclWinInit.c: recognize Windows CE as a Win platform. - This just recognizes CE - full support will come later. + * win/tclWinInit.c: recognize Windows CE as a Win platform. This just + recognizes CE - full support will come later. * win/configure: regen * win/configure.in (SHELL): force it to /bin/sh as autoconf 2.5x @@ -1848,26 +1806,26 @@ 2003-04-18 Donal K. Fellows <fellowsd@cs.man.ac.uk> - * doc/open.n: Moved serial port options from [fconfigure] - * doc/fconfigure.n: to [open] as it is up to the creator of a - channel to describe the channel's special - config options. [Bug 679010] + * doc/open.n: Moved serial port options from [fconfigure] to + * doc/fconfigure.n: [open] as it is up to the creator of a channel + to describe the channel's special config + options. [Bug 679010] 2003-04-16 Don Porter <dgp@users.sourceforge.net> * generic/tcl.h: Made changes so that the "wideInt" Tcl_ObjType * generic/tclObj.c: is defined on all platforms, even those where - * generic/tclPort.h: TCL_WIDE_INT_IS_LONG is defined. Also made - the Tcl_Value struct have a wideValue field on all platforms. This is - a ***POTENTIAL INCOMPATIBILITY*** for TCL_WIDE_INT_IS_LONG platforms - because that struct changes size. This is the same TIP 72 + * generic/tclPort.h: TCL_WIDE_INT_IS_LONG is defined. Also made the + Tcl_Value struct have a wideValue field on all platforms. This is a + ***POTENTIAL INCOMPATIBILITY*** for TCL_WIDE_INT_IS_LONG platforms + because that struct changes size. This is the same TIP 72 incompatibility that was seen on other platforms at the 8.4.0 release, - when this change should have happened as well. [Bug 713562] + when this change should have happened as well. [Bug 713562] * generic/tclInt.h: New internal macros TclGetWide() and TclGetLongFromWide() to deal with both forms of the "wideInt" - Tcl_ObjType, so that conditional TCL_WIDE_INT_IS_LONG code - is confined to the header file. + Tcl_ObjType, so that conditional TCL_WIDE_INT_IS_LONG code is confined + to the header file. * generic/tclCmdAH.c: Replaced most coding that was conditional * generic/tclCmdIL.c: on TCL_WIDE_INT_IS_LONG with code that @@ -1878,35 +1836,36 @@ 2003-04-17 Donal K. Fellows <fellowsd@cs.man.ac.uk> - * doc/socket.n: Added a paragraph to remind people to specify - their encodings when using sockets. [Bug 630621] + * doc/socket.n: Added a paragraph to remind people to specify their + encodings when using sockets. [Bug 630621] 2003-04-16 Donal K. Fellows <fellowsd@cs.man.ac.uk> - * doc/CrtMathFnc.3: Functions also have to deal with wide ints, - but this was not documented. [Bug 709720] + * doc/CrtMathFnc.3: Functions also have to deal with wide ints, but + this was not documented. [Bug 709720] 2003-04-16 Vince Darley <vincentdarley@users.sourceforge.net> - * generic/tclPathObj.c: removed undesired 'static' for function - which is now shared (previously it was duplicated). + * generic/tclPathObj.c: removed undesired 'static' for function which + is now shared (previously it was duplicated). 2003-04-15 Joe English <jenglish@users.sourceforge.net> - * doc/namespace.n: added example section "SCOPED SCRIPTS", - supplied by Kevin Kenny. (Fixes [Bug 219183]) + + * doc/namespace.n: added example section "SCOPED SCRIPTS", supplied by + Kevin Kenny. [Bug 219183] 2003-04-15 Kevin Kenny <kennykb@acm.org> - * makefile.vc: Updated makefile.vc to conform with Mo DeJong's - changes to Makefile.in and tclWinPipe.c on 2003-04-14. Now passes - TCL_PIPE_DLL in place of TCL_DBGX. + * makefile.vc: Updated makefile.vc to conform with Mo DeJong's changes + to Makefile.in and tclWinPipe.c on 2003-04-14. Now passes TCL_PIPE_DLL + in place of TCL_DBGX. * win/tclWinTime.c: Corrected use of types to make compilation compatible with VC++5. 2003-04-15 Vince Darley <vincentdarley@users.sourceforge.net> - * generic/tclIOUtil.c: finished check-in from yesterday, - removing duplicate function definition. + * generic/tclIOUtil.c: finished check-in from yesterday, removing + duplicate function definition. 2003-04-14 Don Porter <dgp@users.sourceforge.net> @@ -1915,19 +1874,14 @@ 2003-04-14 Mo DeJong <mdejong@users.sourceforge.net> - * win/Makefile.in: Don't define TCL_DBGX - symbol for every compile. Instead, define - TCL_PIPE_DLL only when compiling tclWinPipe.c. - This will break other build systems, so - they will need to remove the TCL_DBGX define - and replace it with a define for TCL_PIPE_DLL. - * win/tclWinPipe.c (TclpCreateProcess): - Remove PREFIX_IDENT and DEBUG_IDENT from - top of file. Use TCL_PIPE_DLL passed in - from build env instead of trying to construct - the dll name from already defined symbols. - This approach is more flexible and better - in the long run. + * win/Makefile.in: Don't define TCL_DBGX symbol for every compile. + Instead, define TCL_PIPE_DLL only when compiling tclWinPipe.c. This + will break other build systems, so they will need to remove the + TCL_DBGX define and replace it with a define for TCL_PIPE_DLL. + * win/tclWinPipe.c (TclpCreateProcess): Remove PREFIX_IDENT and + DEBUG_IDENT from top of file. Use TCL_PIPE_DLL passed in from build + env instead of trying to construct the dll name from already defined + symbols. This approach is more flexible and better in the long run. 2003-04-14 Kevin Kenny <kennykb@acm.org> @@ -1936,29 +1890,26 @@ 2003-04-14 Vince Darley <vincentdarley@users.sourceforge.net> - * generic/tclIOUtil.c: - * generic/tclPathObj.c: - * generic/tclFileSystem.h: overlooked one function which - was duplicated, so this is now shared between modules. + * generic/tclIOUtil.c: + * generic/tclPathObj.c: + * generic/tclFileSystem.h: overlooked one function which was + duplicated, so this is now shared between modules. * win/tclWinFile.c: allow this file to compile with VC++ 5.2 again since Mingw build fixes broke that. 2003-04-13 Mo DeJong <mdejong@users.sourceforge.net> * win/configure: Regen. - * win/configure.in: Add check for FINDEX_INFO_LEVELS - from winbase.h, known to be a problem in VC++ 5.2. - Define HAVE_NO_FINDEX_ENUMS if the define does not - exist. - * win/tclWinFile.c: Put declarations for - FINDEX_INFO_LEVELS and FINDEX_SEARCH_OPS inside - a check for HAVE_NO_FINDEX_ENUMS so that these are - not declared twice. This fixes the Mingw build. - * win/tclWinTime.c: Rework the init of timeInfo - so that the number or initializers matches the - declaration. This was broken under Mingw. Add - cast to avoid compile warning when calling the - AccumulateSample function. + * win/configure.in: Add check for FINDEX_INFO_LEVELS from winbase.h, + known to be a problem in VC++ 5.2. Define HAVE_NO_FINDEX_ENUMS if the + define does not exist. + * win/tclWinFile.c: Put declarations for FINDEX_INFO_LEVELS and + FINDEX_SEARCH_OPS inside a check for HAVE_NO_FINDEX_ENUMS so that + these are not declared twice. This fixes the Mingw build. + * win/tclWinTime.c: Rework the init of timeInfo so that the number or + initializers matches the declaration. This was broken under Mingw. Add + cast to avoid compile warning when calling the AccumulateSample + function. 2003-04-12 Jeff Hobbs <jeffh@ActiveState.com> @@ -1967,76 +1918,73 @@ 2003-04-12 Kevin Kenny <kennykb@acm.org> * doc/clock.n: - * generic/tclClock.c (Tcl_ClockObjCmd): - * tests/clock.test: Implementation of TIP #124. Also renumbered - test cases to avoid duplicates [Bug 710310]. + * generic/tclClock.c (Tcl_ClockObjCmd): + * tests/clock.test: Implementation of TIP #124. Also renumbered test + cases to avoid duplicates. [Bug 710310] * tests/winTime.test: * win/tclWinTest.c (TestwinclockCmd, TestwinsleepCmd): * win/tclWinTime.c (Tcl_WinTime, UpdateTimeEachSecond, - ResetCounterSamples, AccumulateSample, - SAMPLES, TimeInfo): Made substantial changes - to the phase-locked loop (replaced an IIR filter with an FIR one) - in a quest for improved loop stability (Bug not logged at SF, but - cited in private communication from Jeff Hobbs). + (ResetCounterSamples, AccumulateSample, SAMPLES, TimeInfo): Made + substantial changes to the phase-locked loop (replaced an IIR filter + with an FIR one) in a quest for improved loop stability (Bug not + logged at SF, but cited in private communication from Jeff Hobbs). 2003-04-11 Don Porter <dgp@users.sourceforge.net> * generic/tclCmdMZ.c (Tcl_StringObjCmd,STR_IS_INT): Corrected - inconsistent results of [string is integer] observed on systems - where sizeof(long) != sizeof(int). [Bug 718878] + inconsistent results of [string is integer] observed on systems where + sizeof(long) != sizeof(int). [Bug 718878] * tests/string.test: Added tests for Bug 718878. - * doc/string.n: Clarified that [string is integer] accepts - 32-bit integers. + * doc/string.n: Clarified that [string is integer] accepts 32-bit + integers. 2003-04-11 Andreas Kupries <andreask@activestate.com> * generic/tclIO.c (UpdateInterest): When dropping interest in - TCL_READABLE now dropping interest in TCL_EXCEPTION too. This - fixes a bug where Expect detects eof on a file prematurely on - solaris 2.6 and higher. A much more complete explanation is in - the code itself (40 lines of comments for a one-line change :) + TCL_READABLE now dropping interest in TCL_EXCEPTION too. This fixes a + bug where Expect detects eof on a file prematurely on solaris 2.6 and + higher. A much more complete explanation is in the code itself (40 + lines of comments for a one-line change :) 2003-04-11 Vince Darley <vincentdarley@users.sourceforge.net> - * tests/cmdAH.test: fix test suite problem if /home is a symlink - [Bug #703264] - * generic/tclIOUtil.c: fix bad error message with 'cd ""' - [Bug #704917] - * win/tclWinFile.c: - * win/tclWin32Dll.c: - * win/tclWinInt.h: allow Tcl to differentiate between reparse - points which are symlinks and mounted volumes, and correctly - handle the latter. This involves some elaborate code to find - the actual drive letter (if possible) corresponding to a mounted - volume. [Bug #697862] - * tests/fileSystem.test: add constraints to stop tests running - in ordinary tcl interpreter. [Bug #705675] + * tests/cmdAH.test: fix test suite problem if /home is a symlink. [Bug + 703264] + * generic/tclIOUtil.c: fix bad error message with 'cd ""'. [Bug + 704917] + * win/tclWinFile.c, win/tclWin32Dll.c: + * win/tclWinInt.h: allow Tcl to differentiate between reparse points + which are symlinks and mounted volumes, and correctly handle the + latter. This involves some elaborate code to find the actual drive + letter (if possible) corresponding to a mounted volume. [Bug 697862] + * tests/fileSystem.test: add constraints to stop tests running in + ordinary tcl interpreter. [Bug 705675] * generic/tclIOUtil.c: * generic/tclPathObj.c: (new file) * generic/tclFileSystem.h: (new file) * win/makefile.vc: - Split path object handling out of the virtual filesystem layer, - into tclPathObj.c. This refactoring cleans up the internal - filesystem code, and will make any future optimisations and - forthcoming better thread-safety much easier. + Split path object handling out of the virtual filesystem layer, into + tclPathObj.c. This refactoring cleans up the internal filesystem code, + and will make any future optimisations and forthcoming better + thread-safety much easier. * generic/tclTest.c: - * tests/reg.test: added some 'knownBug' tests for problems in - Tcl's regexp code with the TCL_REG_CAN_MATCH flag (see Bug #703709). - Code too impenetrable to fix right now, but a fix is needed - for tip113 to work correctly. + * tests/reg.test: added some 'knownBug' tests for problems in Tcl's + regexp code with the TCL_REG_CAN_MATCH flag (see Bug 703709). Code too + impenetrable to fix right now, but a fix is needed for tip113 to work + correctly. * tests/fCmd.test - * win/tclWinFile.c: added some filesystem optimisation to the - 'glob' implementation, and some new tests. + * win/tclWinFile.c: added some filesystem optimisation to the 'glob' + implementation, and some new tests. * generic/tclCmdMZ.c: fix typo in comment * tests/winFile.test: * tests/ioUtil.test: - * tests/unixFCmd.test: renumbered tests with duplicate numbers. - (Bug #710361) + * tests/unixFCmd.test: renumbered tests with duplicate numbers. [Bug + 710361] 2003-04-10 Donal K. Fellows <fellowsd@cs.man.ac.uk> @@ -2045,8 +1993,8 @@ 2003-04-08 Donal K. Fellows <fellowsd@cs.man.ac.uk> * generic/tclCmdAH.c (Tcl_ErrorObjCmd): Strings are only empty if - they have zero length, not if their first byte is zero, so fix - test guarding Tcl_AddObjErrorInfo to take this into account. [Bug + they have zero length, not if their first byte is zero, so fix test + guarding Tcl_AddObjErrorInfo to take this into account. [Bug reported by Don Porter; no bug-id.] 2003-04-07 Don Porter <dgp@users.sourceforge.net> @@ -2066,58 +2014,56 @@ 2003-04-07 Donal K. Fellows <fellowsd@cs.man.ac.uk> - * tests/dict.test (dict-2.6): + * tests/dict.test (dict-2.6): * generic/tclDictObj.c (Tcl_NewDictObj, Tcl_DbNewDictObj): Oops! Failed to fully initialise the Dict structure. - (DictIncrCmd): Moved valueAlreadyInDictionary label to stop - compiler complaints. [Bug 715751] + (DictIncrCmd): Moved valueAlreadyInDictionary label to stop compiler + complaints. [Bug 715751] * generic/tclDictObj.c (DictIncrCmd): Followed style in the rest of - the core by commenting out wide-specific operations on platforms - where wides are longs, and used longs more thoroughly than ints - through [dict incr] anyway to forestall further bugs. + the core by commenting out wide-specific operations on platforms where + wides are longs, and used longs more thoroughly than ints through + [dict incr] anyway to forestall further bugs. * generic/tclObj.c: Made sure there's always a tclWideIntType - implementation available, not that it is always useful. [Bug 713562] + implementation available, not that it is always useful. [Bug 713562] 2003-04-05 Donal K. Fellows <fellowsd@cs.man.ac.uk> - * generic/tclDictObj.c: Removed commented out notes on - declarations to be moved to elsewhere in the Tcl core. + * generic/tclDictObj.c: Removed commented out notes on declarations to + be moved to elsewhere in the Tcl core. * generic/tclInt.h: Final stages of plumbing in. - * generic/tclBasic.c: - * generic/tclObj.c (TclInitObjSubsystem): + * generic/tclBasic.c: + * generic/tclObj.c (TclInitObjSubsystem): * unix/Makefile.in, win/Makefile.in, win/makefile.[bv]c: Build support. * generic/tcl.decls: Added dict public API to stubs table. - * generic/tcl.h (Tcl_DictSearch): Added declaration of structure - to allow user code to iterate over dictionaries. + * generic/tcl.h (Tcl_DictSearch): Added declaration of structure to + allow user code to iterate over dictionaries. - * doc/DictObj.3: New files containing dictionary - * doc/dict.n: implementation, documentation and tests - * generic/tclDictObj.c: as mandated by TIP #111. + * doc/DictObj.3: New files containing dictionary implementation + * doc/dict.n: documentation and tests as as mandated by TIP + * generic/tclDictObj.c: #111. * tests/dict.test: 2003-04-03 Mo DeJong <mdejong@users.sourceforge.net> * unix/configure: - * unix/tcl.m4 (SC_CONFIG_CFLAGS): Don't set - TCL_LIBS if it is already set to support - use of TCL_LIBS var from tclConfig.sh in - the Tk configure script. + * unix/tcl.m4 (SC_CONFIG_CFLAGS): Don't set TCL_LIBS if it is already + set to support use of TCL_LIBS var from tclConfig.sh in the Tk + configure script. 2003-04-03 Mo DeJong <mdejong@users.sourceforge.net> - * unix/Makefile.in: Don't subst MATH_LIBS, - LIBS, and DL_LIBS separately. Instead, just - subst TCL_LIBS since it includes the others. + * unix/Makefile.in: Don't subst MATH_LIBS, LIBS, and DL_LIBS + separately. Instead, just subst TCL_LIBS since it includes the + others. * unix/configure: Regen. - * unix/tcl.m4 (SC_CONFIG_CFLAGS, SC_TCL_LINK_LIBS): - Set and subst TCL_LIBS in SC_CONFIG_CFLAGS instead - of SC_TCL_LINK_LIBS. Don't subst MATH_LIBS - since it is now covered by TCL_LIBS. - * unix/tclConfig.sh.in: Use TCL_LIBS instead - of DL_LIBS, LIBS, and MATH_LIBS. + * unix/tcl.m4 (SC_CONFIG_CFLAGS, SC_TCL_LINK_LIBS): Set and subst + TCL_LIBS in SC_CONFIG_CFLAGS instead of SC_TCL_LINK_LIBS. Don't subst + MATH_LIBS since it is now covered by TCL_LIBS. + * unix/tclConfig.sh.in: Use TCL_LIBS instead of DL_LIBS, LIBS, and + MATH_LIBS. * unix/dltest/Makefile.in: Ditto. 2003-04-03 Don Porter <dgp@users.sourceforge.net> @@ -2128,56 +2074,51 @@ 2003-04-02 Mo DeJong <mdejong@users.sourceforge.net> * win/configure: Regen. - * win/configure.in: Set stub lib flag based - on new LIBFLAGSUFFIX variable. - * win/tcl.m4 (SC_CONFIG_CFLAGS): Set new - LIBFLAGSUFFIX that works like LIBSUFFIX, - it is used when creating library names. - The previous implementation would generate - -ltclstub85 instead of -ltclstub85s when + * win/configure.in: Set stub lib flag based on new LIBFLAGSUFFIX + variable. + * win/tcl.m4 (SC_CONFIG_CFLAGS): Set new LIBFLAGSUFFIX that works like + LIBSUFFIX, it is used when creating library names. The previous + implementation would generate -ltclstub85 instead of -ltclstub85s when configured with --disable-shared. 2003-04-02 Don Porter <dgp@users.sourceforge.net> * generic/tclParse.c (TclSubstTokens): Moved declaration of - utfCharBytes to beginning of procedure so that it does not go - out of scope (get free()d) while append is still pointing to it. - [Bugs 703167, 713754] + utfCharBytes to beginning of procedure so that it does not go out of + scope (get free()d) while append is still pointing to it. [Bugs + 703167, 713754] 2003-04-01 Mo DeJong <mdejong@users.sourceforge.net> * unix/configure: Regen. - * unix/tcl.m4 (SC_CONFIG_CFLAGS): Check for - inet_ntoa in -lbind inside the BeOS block since - doing it later broke the build under SuSE 7.3. + * unix/tcl.m4 (SC_CONFIG_CFLAGS): Check for inet_ntoa in -lbind inside + the BeOS block since doing it later broke the build under SuSE 7.3. [Bug 713128] 2003-04-01 Don Porter <dgp@users.sourceforge.net> * tests/README: Direct [source] of *.test files is no longer - recommended. The tests/*.test files should only be evaluated under - the control of the [runAllTests] command in tests/all.tcl. + recommended. The tests/*.test files should only be evaluated under the + control of the [runAllTests] command in tests/all.tcl. - * generic/tclExecute.c (INST_RETURN): Bytecompiled [return] failed - to reset iPtr->returnCode, causing tests parse-18.17 and parse-18.21 - to fail strangely. + * generic/tclExecute.c (INST_RETURN): Bytecompiled [return] failed to + reset iPtr->returnCode, causing tests parse-18.17 and parse-18.21 to + fail strangely. * tests/parse.test (parse-18.21): Corrected now functioning test. Added further coverage tests. 2003-03-31 Don Porter <dgp@users.sourceforge.net> * tests/parse.test (parse-18.*): Coverage tests for the new - implementation of Tcl_SubstObj(). Note that tests parse-18.17 and + implementation of Tcl_SubstObj(). Note that tests parse-18.17 and parse-18.21 demonstrate some bugs left to fix in the current code. 2003-03-27 Mo DeJong <mdejong@users.sourceforge.net> * unix/configure: Regen. - * unix/tcl.m4 (SC_CONFIG_CFLAGS): Use -Wl,--export-dynamic - instead of -rdynamic for LDFLAGS. The -rdynamic is - not documented so it seems better to pass the - --export-dynamic flag to the linker. - [Patch 573395] + * unix/tcl.m4 (SC_CONFIG_CFLAGS): Use -Wl,--export-dynamic instead of + -rdynamic for LDFLAGS. The -rdynamic is not documented so it seems + better to pass the --export-dynamic flag to the linker. [Patch 573395] 2003-03-27 Miguel Sofer <msofer@users.sf.net> @@ -2206,19 +2147,18 @@ 2003-03-26 Mo DeJong <mdejong@users.sourceforge.net> * unix/configure: Regen. - * unix/tcl.m4 (SC_CONFIG_CFLAGS, SC_TCL_LINK_LIBS): - Add BeOS system to SC_CONFIG_CFLAGS. Check for - inet_ntoa in -lbind, needed for BeOS. + * unix/tcl.m4 (SC_CONFIG_CFLAGS, SC_TCL_LINK_LIBS): Add BeOS system to + SC_CONFIG_CFLAGS. Check for inet_ntoa in -lbind, needed for BeOS. 2003-03-26 Don Porter <dgp@users.sourceforge.net> * doc/tcltest.n: - * library/tcltest/tcltest.tcl: Added reporting during - [configure -debug 1] operations to warn about multiple uses of - the same test name. [FR 576693] + * library/tcltest/tcltest.tcl: Added reporting during [configure + -debug 1] operations to warn about multiple uses of the same test + name. [FRQ 576693] * tests/msgcat.test (msgcat-2.2.1): changed test name to avoid - duplication. [Bug 710356] + duplication. [Bug 710356] * unix/dltest/pkg?.c: Changed all Tcl_InitStubs calls to pass argument exact = 0, so that rebuilds are not required when Tcl @@ -2228,13 +2168,13 @@ * generic/tclVar.c: * tests/var.test: fixing ObjMakeUpvar's lookup algorithm for the - created local variable, bugs #631741 (Chris Darroch) and #696893 - (David Hilker). + created local variable. [Bug 631741] (Chris Darroch) and [Bug 696893] + (David Hilker) 2003-03-24 Pat Thoyts <patthoyts@users.sourceforge.net> * library/dde/pkgIndex.tcl: bumped version to 1.2.2 in tclWinDde.c, - now adding here too. + now adding here too. 2003-03-22 Kevin Kenny <kennykb@acm.org> @@ -2243,20 +2183,19 @@ or [package require registry] attempted to load the release version of the DLL into a debug build. [Bug 708218] Thanks to Joe Mistachkin for the patch. - * win/makefile.vc: Added quoting around the script name in the - 'test' target; Joe Mistachkin insists that he has a configuration - that fails to launch tcltest without it, and it appears harmless - otherwise. + * win/makefile.vc: Added quoting around the script name in the 'test' + target; Joe Mistachkin insists that he has a configuration that fails + to launch tcltest without it, and it appears harmless otherwise. 2003-03-22 Pat Thoyts <patthoyts@users.sourceforge.net> * win/tclWinDde.c: Make dde services conform the the documentation - such that giving only a topic name really returns all services - with that topic. [Bug 219155] + such that giving only a topic name really returns all services with + that topic. [Bug 219155] Prevent hangup caused by dde server applications failing to process - messages [Bug 707822] - * tests/winDde.test: Corrected labels and added a test for search - by topic name. + messages. [Bug 707822] + * tests/winDde.test: Corrected labels and added a test for search by + topic name. 2003-03-20 Don Porter <dgp@users.sourceforge.net> @@ -2264,30 +2203,30 @@ * generic/tclStubInit.c (tclOriginalNotifier): * mac/tclMacNotify.c (Tcl_SetTimer,Tcl_WaitForEvent): * unix/tclUnixNotfy.c (Tcl_SetTimer,Tcl_WaitForEvent, - Tcl_CreateFileHandler,Tcl_DeleteFileHandler): + (Tcl_CreateFileHandler,Tcl_DeleteFileHandler): * win/tclWinNotify.c (Tcl_SetTimer,Tcl_WaitForEvent): Some linkers apparently use a different representation for a pointer to a function within the same compilation unit and a pointer to a function in a - different compilation unit. This causes checks like those in the - original notifier procedures to fall into infinite loops. The fix - is to store pointers to the original notifier procedures in a struct + different compilation unit. This causes checks like those in the + original notifier procedures to fall into infinite loops. The fix is + to store pointers to the original notifier procedures in a struct defined in the same compilation unit as the stubs tables, and compare - against those values. [Bug 707174] + against those values. [Bug 707174] - * generic/tclInt.h: Removed definition of ParseValue struct that - is no longer used. + * generic/tclInt.h: Removed definition of ParseValue struct that is + no longer used. 2003-03-19 Miguel Sofer <msofer@users.sf.net> * generic/tclCompile.c: - * tests/compile.test: bad command count on TCL_OUT_LINE_COMPILE - [Bug 705406] (Don Porter). + * tests/compile.test: bad command count on TCL_OUT_LINE_COMPILE. + [Bug 705406] (Don Porter) 2003-03-19 Don Porter <dgp@users.sourceforge.net> * library/auto.tcl: Replaced [regexp] and [regsub] with * library/history.tcl: [string map] where possible. Thanks - * library/ldAout.tcl: to David Welton. [Bugs 667456,667558] + * library/ldAout.tcl: to David Welton. [Bugs 667456,667558] * library/safe.tcl: Bumped to http 2.4.3, opt 0.4.5, and * library/http/http.tcl: tcltest 2.2.3. * library/http/pkgIndex.tcl: @@ -2300,67 +2239,64 @@ * unix/mkLinks.tcl: * doc/Eval.3 (Tcl_EvalObjEx): Corrected CONST and - * doc/ParseCmd.3 (Tcl_EvalTokensStandard): return type errors - in documentation. [Bug 683994] + * doc/ParseCmd.3 (Tcl_EvalTokensStandard): return type errors in + documentation. [Bug 683994] * generic/tclCompCmds.c (TclCompileReturnCmd): Alternative fix for * generic/tclCompile.c (INST_RETURN): [Bug 633204] that uses a new * generic/tclCompile.h (INST_RETURN): bytecode INST_RETURN to - * generic/tclExecute.c (INST_RETURN): properly bytecode the - [return] command to something that returns TCL_RETURN. + * generic/tclExecute.c (INST_RETURN): properly bytecode the [return] + command to something that returns TCL_RETURN. 2003-03-18 Mo DeJong <mdejong@users.sourceforge.net> * win/configure: Regen. - * win/configure.in: Don't run the AC_CYGWIN - macro since it uses AC_CANONICAL_HOST under - autoconf 2.5X. Just check to see if __CYGWIN__ - is defined by the compiler and set the - ac_cv_cygwin variable based on that. - [Bug 705912] + * win/configure.in: Don't run the AC_CYGWIN macro since it uses + AC_CANONICAL_HOST under autoconf 2.5X. Just check to see if __CYGWIN__ + is defined by the compiler and set the ac_cv_cygwin variable based on + that. [Bug 705912] 2003-03-18 Kevin Kenny <kennykb@users.sourceforge.net> - * tests/registry.test: Changed the conditionals to avoid an - abort if [testlocale] is missing, as when running the test in - tclsh rather than tcltest. [Bug #705677] + * tests/registry.test: Changed the conditionals to avoid an abort if + [testlocale] is missing, as when running the test in tclsh rather than + tcltest. [Bug 705677] 2003-03-18 Daniel Steffen <das@users.sourceforge.net> * tools/tcltk-man2html.tcl: added support for building 'make html' - from inside distribution directories named with 8.x.x version - numbers. tcltk-man2html now uses the latest tcl8.x.x resp. tk8.x.x - directories found inside its --srcdir argument. + from inside distribution directories named with 8.x.x version numbers. + tcltk-man2html now uses the latest tcl8.x.x resp. tk8.x.x directories + found inside its --srcdir argument. 2003-03-17 Mo DeJong <mdejong@users.sourceforge.net> - * tests/format.test: Renumber tests, a bunch of - tests all had the same id. + * tests/format.test: Renumber tests, a bunch of tests all had the same + id. 2003-03-17 Donal K. Fellows <fellowsd@cs.man.ac.uk> * doc/lsearch.n: Altered documentation of -ascii options so - * doc/lsort.n: they don't specify that they operate on - ASCII strings, which they never did - anyway. [Bug #703807] + * doc/lsort.n: they don't specify that they operate on ASCII + strings, which they never did anyway. [Bug + 703807] 2003-03-14 Donal K. Fellows <fellowsd@cs.man.ac.uk> - * generic/tclCmdAH.c (Tcl_FormatObjCmd): Only add the modifier - that indicates we've got a wide int when we're formatting in an - integer style. Stops some libc's from going mad. [Bug #702622] - Also tidied whitespace. + * generic/tclCmdAH.c (Tcl_FormatObjCmd): Only add the modifier that + indicates we've got a wide int when we're formatting in an integer + style. Stops some libc's from going mad. [Bug 702622] Also tidied + whitespace. 2003-03-13 Mo DeJong <mdejong@users.sourceforge.net> - * win/tcl.m4 (SC_WITH_TCL): Port version number - fix that was made in tk instead of tcl sources. + * win/tcl.m4 (SC_WITH_TCL): Port version number fix that was made in + tk instead of tcl sources. 2003-03-13 Mo DeJong <mdejong@users.sourceforge.net> - Require autoconf 2.57 or newer, see TIP 34 - for a detailed explanation of why this is good. - This will no doubt break the build on some + Require autoconf 2.57 or newer, see TIP 34 for a detailed explanation + of why this is good. This will no doubt break the build on some platforms, let the flaming begin. * tools/configure: Regen with autoconf 2.57. @@ -2371,8 +2307,8 @@ * unix/tcl.m4: Ditto. * win/configure: Regen with autoconf 2.57. * win/configure.in: Require autoconf 2.57. - Don't subst LIBOBJS since this happens by - default, this avoids an autoconf error. + Don't subst LIBOBJS since this happens by default, this avoids an + autoconf error. 2003-03-12 Don Porter <dgp@users.sourceforge.net> @@ -2381,7 +2317,7 @@ * generic/tclCompCmds.c (TclCompileSwitchCmd): * generic/tclCompExpr.c (CompileSubExpr): * generic/tclCompile.c (TclSetByteCodeFromAny,TclCompileScript, - TclCompileTokens,TclCompileCmdWord): + (TclCompileTokens,TclCompileCmdWord): * generic/tclCompile.h (TclCompileScript): * generic/tclExecute.c (TclCompEvalObj): * generic/tclInt.h (Interp,TCL_BRACKET_TERM,TclSubstTokens): @@ -2389,16 +2325,16 @@ * tests/subst.test (2.4, 8.7, 8.8, 11.4, 11.5): Substantial refactoring of Tcl_SubstObj to make use of the same parsing and substitution procedures as normal script evaluation. - Tcl_SubstObj() moved to tclParse.c. New routine TclSubstTokens() - created in tclParse.c which implements all substantial functioning - of Tcl_EvalTokensStandard(). TclCompileScript() loses its - "nested" argument, the Tcl_Interp struct loses its termOffset - field and the TCL_BRACKET_TERM flag in the evalFlags field, all - of which were only used (indirectly) by Tcl_SubstObj(). Tests - subst-8.7,8.8,11.4,11.5 modified to accomodate the only behavior - change: reporting of parse errors now takes precedence over - [return] and [continue] exceptions. All other behavior should - remain compatible. [RFE 536831,684982] [Bug 685106] + Tcl_SubstObj() moved to tclParse.c. New routine TclSubstTokens() + created in tclParse.c which implements all substantial functioning of + Tcl_EvalTokensStandard(). TclCompileScript() loses its "nested" + argument, the Tcl_Interp struct loses its termOffset field and the + TCL_BRACKET_TERM flag in the evalFlags field, all of which were only + used (indirectly) by Tcl_SubstObj(). Tests subst-8.7,8.8,11.4,11.5 + modified to accomodate the only behavior change: reporting of parse + errors now takes precedence over [return] and [continue] exceptions. + All other behavior should remain compatible. [RFE 536831,684982] [Bug + 685106] * generic/tcl.h: Removed TCL_PREFIX_IDENT and TCL_DEBUG_IDENT * win/tclWinPipe.c: from tcl.h -- they are not part of Tcl's @@ -2406,7 +2342,7 @@ * generic/tclInterp.c (Tcl_InterpObjCmd): Corrected and added * tests/interp.test (interp-2.13): test for option - parsing beyond objc for [interp create --]. Thanks to Marco Maggi. + parsing beyond objc for [interp create --]. Thanks to Marco Maggi. [Bug 702383] 2003-03-11 Kevin Kenny <kennykb@users.sourceforge.net> @@ -2416,52 +2352,48 @@ 2003-03-09 Kevin Kenny <kennykb@users.sourceforge.net> - * generic/tclTest.c (TestChannelCmd): Removed an unused local - variable that caused compilation problems on some platforms. + * generic/tclTest.c (TestChannelCmd): Removed an unused local variable + that caused compilation problems on some platforms. 2003-03-08 Don Porter <dgp@users.sourceforge.net> - * doc/tcltest.n: Added missing "-body" to example. Thanks to - Helmut Giese. [Bug 700011] + * doc/tcltest.n: Added missing "-body" to example. Thanks to Helmut + Giese. [Bug 700011] 2003-03-07 Mo DeJong <mdejong@users.sourceforge.net> * tests/io.test: - * tests/ioCmd.test: Define a fcopy constraint and add - it to the constraint list of any test that depends - on the fcopy command. This is only useful to - Jacl which does not support fcopy. + * tests/ioCmd.test: Define a fcopy constraint and add it to the + constraint list of any test that depends on the fcopy command. This is + only useful to Jacl which does not support fcopy. 2003-03-07 Mo DeJong <mdejong@users.sourceforge.net> - * tests/encoding.test: Name temp files *.tcltestout - instead of *.out so that when they are removed later, - we don't accidently toast any files named *.out that - the user has created in the build directory. + * tests/encoding.test: Name temp files *.tcltestout instead of *.out + so that when they are removed later, we don't accidently toast any + files named *.out that the user has created in the build directory. 2003-03-07 Donal K. Fellows <fellowsd@cs.man.ac.uk> * generic/tclCmdAH.c (Tcl_FileObjCmd): Fix the setting of a file's - mtime and atime on 64-bit platforms. [Bug #698146] + mtime and atime on 64-bit platforms. [Bug 698146] 2003-03-06 Mo DeJong <mdejong@users.sourceforge.net> - * tests/io.test: Doh! Undo accidental commenting - out of a couple of tests. + * tests/io.test: Doh! Undo accidental commenting out of a couple of + tests. 2003-03-06 Mo DeJong <mdejong@users.sourceforge.net> - * tests/io.test: Define a fileevent constraint and add - it to the constraint list of any test that depends - on the fileevent command. This is only useful to - Jacl which does not support fileevent. + * tests/io.test: Define a fileevent constraint and add it to the + constraint list of any test that depends on the fileevent command. + This is only useful to Jacl which does not support fileevent. 2003-03-06 Mo DeJong <mdejong@users.sourceforge.net> - * tests/io.test: Define an openpipe constraint and add - it to the constraint list of any test that creates - a pipe using the open command. This is only useful to - Jacl which does not support pipes. + * tests/io.test: Define an openpipe constraint and add it to the + constraint list of any test that creates a pipe using the open + command. This is only useful to Jacl which does not support pipes. 2003-03-06 Don Porter <dgp@users.sourceforge.net> @@ -2472,72 +2404,57 @@ 2003-03-06 Kevin Kenny <kennykb@users.sourceforge.net> * generic/tclCompCmds.c (TclCompileSwitchCmd): - Replaced a non-portable 'bzero' with a portable 'memset'. - [Bug 698442]. + Replaced a non-portable 'bzero' with a portable 'memset'. [Bug 698442] 2003-03-06 Mo DeJong <mdejong@users.sourceforge.net> - * generic/tclIO.c (Tcl_Seek, Tcl_OutputBuffered): - If there is data buffered in the statePtr->curOutPtr - member then set the BUFFER_READY flag in Tcl_Seek. - This is needed so that the next call to FlushChannel - will write any buffered bytes before doing the seek. - The existing code would set the BUFFER_READY flag - inside the Tcl_OutputBuffered function. This was a - programming error made when Tcl_OutputBuffered - was originally created in CVS revision 1.35. The - setting of the BUFFER_READY flag should not have - been included in the Tcl_OutputBuffered function. - * generic/tclTest.c (TestChannelCmd): Use the - Tcl_InputBuffered and Tcl_OutputBuffered - util methods to query the amount of buffered - input and output. + * generic/tclIO.c (Tcl_Seek, Tcl_OutputBuffered): If there is data + buffered in the statePtr->curOutPtr member then set the BUFFER_READY + flag in Tcl_Seek. This is needed so that the next call to FlushChannel + will write any buffered bytes before doing the seek. The existing code + would set the BUFFER_READY flag inside the Tcl_OutputBuffered + function. This was a programming error made when Tcl_OutputBuffered + was originally created in CVS revision 1.35. The setting of the + BUFFER_READY flag should not have been included in the + Tcl_OutputBuffered function. + * generic/tclTest.c (TestChannelCmd): Use the Tcl_InputBuffered and + Tcl_OutputBuffered util methods to query the amount of buffered input + and output. 2003-03-06 Mo DeJong <mdejong@users.sourceforge.net> - * generic/tclIO.c (Tcl_Flush): Compare the - nextAdded member of the ChannelBuffer to the - nextRemoved member to determine if any output - has been buffered. The previous check against - the value 0 seems to have just been a coding - error. See other methods like Tcl_OutputBuffered - for examples where nextAdded is compared to - nextRemoved to find the number of bytes buffered. + * generic/tclIO.c (Tcl_Flush): Compare the nextAdded member of the + ChannelBuffer to the nextRemoved member to determine if any output has + been buffered. The previous check against the value 0 seems to have + just been a coding error. See other methods like Tcl_OutputBuffered + for examples where nextAdded is compared to nextRemoved to find the + number of bytes buffered. 2003-03-06 Mo DeJong <mdejong@users.sourceforge.net> - * generic/tclIO.c (Tcl_GetsObj): Check that - the eol pointer has not gone past the end - of the string when in auto translation - mode and the INPUT_SAW_CR flag is set. - The previous code worked because the - end of string value \0 was being compared - to \n, this patch just skips that pointless - check. + * generic/tclIO.c (Tcl_GetsObj): Check that the eol pointer has not + gone past the end of the string when in auto translation mode and the + INPUT_SAW_CR flag is set. The previous code worked because the end of + string value \0 was being compared to \n, this patch just skips that + pointless check. 2003-03-06 Mo DeJong <mdejong@users.sourceforge.net> - * generic/tclIO.c (WriteBytes, WriteChars, - Tcl_GetsObj, ReadBytes): Rework calls to - TranslateOutputEOL to make it clear that - a boolean value is being returned. - Add some comments in an effort to make - the code more clear. This patch makes - no functional changes. + * generic/tclIO.c (WriteBytes, WriteChars, Tcl_GetsObj, ReadBytes): + Rework calls to TranslateOutputEOL to make it clear that a boolean + value is being returned. Add some comments in an effort to make the + code more clear. This patch makes no functional changes. 2003-03-06 Mo DeJong <mdejong@users.sourceforge.net> - * generic/tclIO.c (Tcl_SetChannelOption): - Invoke the Tcl_SetChannelBufferSize method - as a result of changing the -buffersize - option to fconfigure. The previous - implementation used some inlined code that - reset the buffer size to the default size - instead of ignoring the request as - implemented in Tcl_SetChannelBufferSize. - * tests/io.test: Update test case so that - it actually checks the implementation of + * generic/tclIO.c (Tcl_SetChannelOption): Invoke the + Tcl_SetChannelBufferSize method as a result of changing the + -buffersize option to fconfigure. The previous implementation used + some inlined code that reset the buffer size to the default size + instead of ignoring the request as implemented in Tcl_SetChannelBufferSize. + * tests/io.test: Update test case so that it actually checks the + implementation of Tcl_SetChannelBufferSize. 2003-03-05 David Gravereaux <davygrvy@pobox.com> @@ -2546,17 +2463,17 @@ 2003-03-05 Donal K. Fellows <fellowsd@cs.man.ac.uk> * generic/tclCompCmds.c (TclCompileSwitchCmd): First attempt at a - bytecode-compiled switch command. It only handles the most common - case of switching, but that should be enough for this to speed up - a lot of people's code. It is expected that the speed gains come - from two things: better handling of the switch itself, and - integrated compilation of the arms instead of embedding separate - bytecode sequences (i.e. better local variable handling.) + bytecode-compiled switch command. It only handles the most common case + of switching, but that should be enough for this to speed up a lot of + people's code. It is expected that the speed gains come from two + things: better handling of the switch itself, and integrated + compilation of the arms instead of embedding separate bytecode + sequences (i.e. better local variable handling.) * tests/switch.test (switch-10.*): Tests of both uncompiled and compiled switch behaviour. [Patch #644819] - * generic/tclCompile.h (TclFixupForwardJumpToHere): Additional - macro to make the most common kind of jump fixup a bit easier. + * generic/tclCompile.h (TclFixupForwardJumpToHere): Additional macro + to make the most common kind of jump fixup a bit easier. 2003-03-04 Don Porter <dgp@users.sourceforge.net> @@ -2589,16 +2506,16 @@ 2003-03-03 Daniel Steffen <das@users.sourceforge.net> Mac OS Classic specific fixes: - * generic/tclIOUtil.c (TclNewFSPathObj): on TCL_PLATFORM_MAC, - skip potential directory separator at the beginning of addStrRep. - * mac/tclMacChan.c (OpenFileChannel, CommonWatch): followup - fixes to cut and splice implementation for file channels. + * generic/tclIOUtil.c (TclNewFSPathObj): on TCL_PLATFORM_MAC, skip + potential directory separator at the beginning of addStrRep. + * mac/tclMacChan.c (OpenFileChannel, CommonWatch): followup fixes to + cut and splice implementation for file channels. * mac/tclMacFile.c (TclpUtime): pass native path to utime(). - * mac/tclMacFile.c (TclpObjLink): correctly implemented creation - of alias files via new static proc CreateAliasFile(). + * mac/tclMacFile.c (TclpObjLink): correctly implemented creation of + alias files via new static proc CreateAliasFile(). * mac/tclMacPort.h: define S_ISLNK macro to fix stat'ing of links. - * mac/tclMacUtil.c (FSpLocationFromPathAlias): fix to enable - stat'ing of broken links. + * mac/tclMacUtil.c (FSpLocationFromPathAlias): fix to enable stat'ing + of broken links. 2003-03-03 Kevin Kenny <kennykb@users.sourceforge.net> @@ -2618,8 +2535,8 @@ * win/configure: * win/configure.in: check for 'g' for debug build type, not 'd'. - * win/rules.vc (DBGX): correct to use 'g' for nmake win makefile - to match the cygwin makefile for debug builds. [Bug #635107] + * win/rules.vc (DBGX): correct to use 'g' for nmake win makefile to + match the cygwin makefile for debug builds. [Bug 635107] 2003-02-28 Vince Darley <vincentdarley@users.sourceforge.net> @@ -2630,14 +2547,14 @@ * generic/tclIOUtil.c (MakeFsPathFromRelative): removed dead code check of typePtr (darley). - * tests/winTime.test: added note about PCI hardware dependency - issues with high performance clock. + * tests/winTime.test: added note about PCI hardware dependency issues + with high performance clock. 2003-02-27 Donal K. Fellows <fellowsd@cs.man.ac.uk> - * tests/lsearch.test (lsearch-10.7): - * generic/tclCmdIL.c (Tcl_LsearchObjCmd): Stopped -start option - from causing an option when used with an empty list. [Bug #694232] + * tests/lsearch.test (lsearch-10.7): + * generic/tclCmdIL.c (Tcl_LsearchObjCmd): Stopped -start option from + causing an option when used with an empty list. [Bug 694232] 2003-02-26 Chengye Mao <chengye.geo@yahoo.com> @@ -2648,21 +2565,20 @@ 2003-02-26 Jeff Hobbs <jeffh@ActiveState.com> - * generic/tclCmdMZ.c (TraceCommandProc): Fix mem leak when - deleting a command that had trace on it. [Bug #693564] (sofer) + * generic/tclCmdMZ.c (TraceCommandProc): Fix mem leak when deleting a + command that had trace on it. [Bug 693564] (sofer) 2003-02-25 Don Porter <dgp@users.sourceforge.net> * doc/pkgMkIndex.n: Modified [pkg_mkIndex] to use -nocase matching - * library/package.tcl: of -load patterns, to better accomodate - common user errors due to confusion between [package names] names - and [info loaded] names. + * library/package.tcl: of -load patterns, to better accomodate common + user errors due to confusion between [package names] names and [info + loaded] names. 2003-02-25 Andreas Kupries <andreask@pliers.activestate.com> - * tests/pid.test: See below [Bug #678412]. - * tests/io.test: Made more robust against spaces in paths - [Bug #678400]. + * tests/pid.test: See below [Bug 678412]. + * tests/io.test: Made more robust against spaces in paths [Bug 678400] 2003-02-25 Miguel Sofer <msofer@users.sf.net> @@ -2671,13 +2587,13 @@ 2003-02-22 Zoran Vasiljevic <zoran@archiwrae.com> - * generic/tclEvent.c (Tcl_FinalizeThread): Fix [Bug #571002] + * generic/tclEvent.c (Tcl_FinalizeThread): Fix [Bug 571002] 2003-02-21 Donal K. Fellows <fellowsd@cs.man.ac.uk> - * tests/binary.test (binary-44.[34]): + * tests/binary.test (binary-44.[34]): * generic/tclBinary.c (ScanNumber): Fixed problem with unwanted - sign-bit propagation when scanning wide ints. [Bug #690774] + sign-bit propagation when scanning wide ints. [Bug 690774] 2003-02-21 Daniel Steffen <das@users.sourceforge.net> @@ -2686,16 +2602,16 @@ 2003-02-21 Don Porter <dgp@users.sourceforge.net> - * library/package.tcl (tclPkgUnknown): Minor performance tweaks - to reduce the number of [file] invocations. Meant to improve - startup times, at least a little bit. [Patch 687906] + * library/package.tcl (tclPkgUnknown): Minor performance tweaks to + reduce the number of [file] invocations. Meant to improve startup + times, at least a little bit. [Patch 687906] 2003-02-20 Daniel Steffen <das@users.sourceforge.net> * unix/tcl.m4: - * unix/tclUnixPipe.c: (macosx) use vfork() instead of fork() to - create new processes, as recommended by Apple (vfork can be up to - 100 times faster thank fork on macosx). + * unix/tclUnixPipe.c: (macosx) use vfork() instead of fork() to create + new processes, as recommended by Apple (vfork can be up to 100 times + faster thank fork on macosx). * unix/configure: regen. 2003-02-20 Jeff Hobbs <jeffh@ActiveState.com> @@ -2703,53 +2619,52 @@ * generic/tclEncoding.c (LoadTableEncoding): * library/encoding/cp932.enc: Correct jis round-trip encoding * library/encoding/euc-jp.enc: by adding 'R' type to .enc files. - * library/encoding/iso2022-jp.enc: [Patch #689341] (koboyasi, taguchi) + * library/encoding/iso2022-jp.enc: [Patch 689341] (koboyasi, taguchi) * library/encoding/jis0208.enc: * library/encoding/shiftjis.enc: * tests/encoding.test: * unix/tclUnixChan.c (Tcl_MakeTcpClientChannel): add - MakeTcpClientChannelMode that takes actual mode flags to avoid - hang on OS X (may be OS X bug, but patch works x-plat). - [Bug #689835] (steffen) + MakeTcpClientChannelMode that takes actual mode flags to avoid hang on + OS X (may be OS X bug, but patch works x-plat). [Bug 689835] (steffen) 2003-02-20 Donal K. Fellows <fellowsd@cs.man.ac.uk> - * doc/regsub.n: Typo fix [Bug #688943] + * doc/regsub.n: Typo fix [Bug 688943] 2003-02-19 Jeff Hobbs <jeffh@ActiveState.com> * unix/tclUnixThrd.c (TclpReaddir): - * unix/tclUnixPort.h: update to Bug 689100 patch to ensure that - there is a defined value of MAXNAMLEN (aka NAME_MAX in POSIX) and - that we have some buffer allocated. + * unix/tclUnixPort.h: update to Bug 689100 patch to ensure that there + is a defined value of MAXNAMLEN (aka NAME_MAX in POSIX) and that we + have some buffer allocated. 2003-02-19 Daniel Steffen <das@users.sourceforge.net> - * generic/tclStringObj.c: restored Tcl_SetObjLength() side-effect - of always invalidating unicode rep (if the obj has a string rep). - Added hasUnicode flag to String struct, allows decoupling of - validity of unicode rep from buffer size allocated to it (improves - memory allocation efficiency). [Bugs #686782, #671138, #635200] + * generic/tclStringObj.c: restored Tcl_SetObjLength() side-effect of + always invalidating unicode rep (if the obj has a string rep). Added + hasUnicode flag to String struct, allows decoupling of validity of + unicode rep from buffer size allocated to it (improves memory + allocation efficiency). [Bugs 686782, 671138, 635200] * macosx/Tcl.pbproj/project.pbxproj: * macosx/Makefile: reworked embedded build to no longer require relinking but to use install_name_tool instead to change the - install_names for embedded frameworks. [Bug #644510] + install_names for embedded frameworks. [Bug 644510] - * macosx/Tcl.pbproj/project.pbxproj: preserve mod dates when - running 'make install' to build framework (avoids bogus rebuilds - of dependent frameworks because tcl headers appear changed). + * macosx/Tcl.pbproj/project.pbxproj: preserve mod dates when running + 'make install' to build framework (avoids bogus rebuilds of dependent + frameworks because tcl headers appear changed). - * tests/ioCmd.test (iocmd-1.8): fix failure when system encoding - is utf-8: use iso8859-1 encoding explicitly. + * tests/ioCmd.test (iocmd-1.8): fix failure when system encoding is + utf-8: use iso8859-1 encoding explicitly. 2003-02-18 Miguel Sofer <msofer@users.sf.net> - * generic/tclCompile.c (TclCompileExprWords): remove unused - variable "range" [Bug 664743] - * generic/tclExecute.c (ExprSrandFunc): remove unused - variable "result" [Bug 664743] + * generic/tclCompile.c (TclCompileExprWords): remove unused variable + "range" [Bug 664743] + * generic/tclExecute.c (ExprSrandFunc): remove unused variable + "result" [Bug 664743] * generic/tclStringObj.c (UpdateStringOfString): remove unused variable "length" [Bug 664751] * tests/execute.test (execute-7.30): fix for [Bug 664775] @@ -2757,9 +2672,9 @@ 2003-02-18 Andreas Kupries <andreask@activestate.com> * unix/tcl.m4: [Bug #651811] Added definition of _XOPEN_SOURCE and - linkage of 'xnet' library to HP 11 branch. This kills a lot of - socket-related failures in the testsuite when Tcl was compiled - in 64 bit mode (both PA-RISC 2.0W, and IA 64). + linkage of 'xnet' library to HP 11 branch. This kills a lot of + socket-related failures in the testsuite when Tcl was compiled in 64 + bit mode (both PA-RISC 2.0W, and IA 64). * unix/configure: Regenerated. @@ -2767,25 +2682,24 @@ * generic/tclIO.c (HaveVersion): correctly decl static - * unix/tclUnixThrd.c (TclpReaddir): reduce size of name string in - tsd to NAME_MAX instead of PATH_MAX. [Bug #689100] (waters) + * unix/tclUnixThrd.c (TclpReaddir): reduce size of name string in tsd + to NAME_MAX instead of PATH_MAX. [Bug 689100] (waters) 2003-02-18 Mo DeJong <mdejong@users.sourceforge.net> * unix/configure: Regen. - * unix/tcl.m4 (SC_ENABLE_THREADS): Make sure - -lpthread gets passed on the link line when - checking for the pthread_attr_setstacksize symbol. + * unix/tcl.m4 (SC_ENABLE_THREADS): Make sure -lpthread gets passed on + the link line when checking for the pthread_attr_setstacksize symbol. 2003-02-18 Vince Darley <vincentdarley@users.sourceforge.net> - * generic/tclTest.c: cleanup of new 'simplefs' test code, and - better documentation. + * generic/tclTest.c: cleanup of new 'simplefs' test code, and better + documentation. 2003-02-17 Miguel Sofer <msofer@users.sf.net> * generic/tclBasic.c (TclRenameCommand): fixing error in previous - commit. + commit. 2003-02-17 Jeff Hobbs <jeffh@ActiveState.com> @@ -2794,24 +2708,22 @@ * generic/tclUtf.c (TclUniCharMatch): * generic/tclInt.decls: add private TclUniCharMatch function that * generic/tclIntDecls.h: does string match on counted unicode - * generic/tclStubInit.c: strings. Tcl_UniCharCaseMatch has the - * tests/string.test: failing that it can't handle strings or - * tests/stringComp.test: patterns with embedded NULLs. Added - tests that actually try strings/pats with NULLs. TclUniCharMatch - should be TIPed and made public in the next minor version rev. + * generic/tclStubInit.c: strings. Tcl_UniCharCaseMatch has the failing + * tests/string.test: that it can't handle strings or patterns with + * tests/stringComp.test: embedded NULLs. Added tests that actually try + strings/pats with NULLs. TclUniCharMatch should be TIPed and made + public in the next minor version rev. 2003-02-17 Miguel Sofer <msofer@users.sf.net> - * generic/tclBasic.c (TclRenameCommand): 'oldFullName' object was - not being freed on all function exits, causing a memory leak - [Bug 684756] + * generic/tclBasic.c (TclRenameCommand): 'oldFullName' object was not + being freed on all function exits, causing a memory leak. [Bug 684756] 2003-02-17 Mo DeJong <mdejong@users.sourceforge.net> - * generic/tclIO.c (Tcl_GetsObj): Minor change - so that eol is only assigned at the top of the - TCL_TRANSLATE_AUTO case block. The other cases - assign eol so this does not change any functionality. + * generic/tclIO.c (Tcl_GetsObj): Minor change so that eol is only + assigned at the top of the TCL_TRANSLATE_AUTO case block. The other + cases assign eol so this does not change any functionality. 2003-02-17 Kevin Kenny <kennykb@users.sourceforge.net> @@ -2829,23 +2741,23 @@ * tests/misc.test (1.2): * tests/parse.test (6.18): * tests/parseExpr.test (15.35): - * tests/subst.test (8.6): Don Porter's fix for bad parsing of - nested scripts [Bug 681841]. + * tests/subst.test (8.6): Don Porter's fix for bad parsing of nested + scripts. [Bug 681841] 2003-02-15 Kevin Kenny <kennykb@users.sourceforge.net> - * tests/notify.test (new-file): + * tests/notify.test (new-file): * generic/tclTest.c (TclTest_Init, EventtestObjCmd, EventtestProc, - EventTestDeleteProc): - * generic/tclNotify.c (Tcl_DeleteEvents): Fixed Tcl_DeleteEvents - not to get a pointer smash when deleting the last event in the - queue. Added test code in 'tcltest' and a new file of test cases - 'notify.test' to exercise this functionality; several of the new - test cases fail for the original code and pass for the corrected - code. [Bug 673714] + (EventTestDeleteProc): + * generic/tclNotify.c (Tcl_DeleteEvents): Fixed Tcl_DeleteEvents not + to get a pointer smash when deleting the last event in the queue. + Added test code in 'tcltest' and a new file of test cases + 'notify.test' to exercise this functionality; several of the new test + cases fail for the original code and pass for the corrected code. [Bug + 673714] - * unix/tclUnixTest.c (TestfilehandlerCmd): Corrected a couple - of typos in error messages. [Bug 596027] + * unix/tclUnixTest.c (TestfilehandlerCmd): Corrected a couple of typos + in error messages. [Bug 596027] 2003-02-14 Jeff Hobbs <jeffh@ActiveState.com> @@ -2867,40 +2779,38 @@ 2003-02-14 Kevin Kenny <kennykb@users.sourceforge.net> - * win/tclWinTime.c: Added code to test and compensate for - forward leaps of the performance counter. See the MSDN Knowledge - Base article Q274323 for the hardware problem that makes this - necessary on certain machines. - * tests/winTime.test: Revised winTime-2.1 - it had a tolerance - of thousands of seconds, rather than milliseconds. (What's six - orders of magnitude among friends? - Both the above changes are triggered by a problem reported at + * win/tclWinTime.c: Added code to test and compensate for forward + leaps of the performance counter. See the MSDN Knowledge Base article + Q274323 for the hardware problem that makes this necessary on certain + machines. + * tests/winTime.test: Revised winTime-2.1 - it had a tolerance of + thousands of seconds, rather than milliseconds. (What's six orders of + magnitude among friends?) Both the above changes are triggered by a + problem reported at: http://aspn.activestate.com/ASPN/Mail/Message/ActiveTcl/1536811 - although the developers find it difficult to believe that it - accounts for the observed behavior and suspect a fault in the - RTC chip. + although the developers find it difficult to believe that it accounts + for the observed behavior and suspect a fault in the RTC chip. 2003-02-13 Kevin Kenny <kennykb@users.sourceforge.net> - * win/tclWinInit.c: Added conversion from the system encoding - to tcl_platform(user), so that it works with non-ASCII7 user names. - [Bug 685926] + * win/tclWinInit.c: Added conversion from the system encoding to + tcl_platform(user), so that it works with non-ASCII7 user names. [Bug + 685926] * doc/tclsh.1: Added language to describe the handling of the - end-of-file character \u001a embedded in a script file. - [Bug 685485] + end-of-file character \u001a embedded in a script file. [Bug 685485] 2003-02-11 Vince Darley <vincentdarley@users.sourceforge.net> * tests/fileName.test: - * unix/tclUnixFile.c: fix for [Bug 685445] when using 'glob -l' - on broken symbolic links. Added two new tests for this bug. + * unix/tclUnixFile.c: fix for [Bug 685445] when using 'glob -l' on + broken symbolic links. Added two new tests for this bug. 2003-02-11 Kevin Kenny <kennykb@users.sourceforge.net> - * tests/http.test: Corrected a problem where http-4.14 would fail - when run in an environment with a proxy server. Replaced references - to scriptics.com by tcl.tk. + * tests/http.test: Corrected a problem where http-4.14 would fail when + run in an environment with a proxy server. Replaced references to + scriptics.com by tcl.tk. 2003-02-11 Jeff Hobbs <jeffh@ActiveState.com> @@ -2909,47 +2819,44 @@ that lsearch -regepx list and pattern objects are equal. * tests/stringObj.test: - * generic/tclStringObj.c (Tcl_GetCharLength): correct ascii char - opt of 2002-11-11 to not stop early on \x00. [Bug #684699] + * generic/tclStringObj.c (Tcl_GetCharLength): correct ascii char opt + of 2002-11-11 to not stop early on \x00. [Bug 684699] * tests.parse.test: remove excess EOF whitespace - * generic/tclParse.c (CommandComplete): more paranoid check to - break on (p >= end) instead of just (p == end). + * generic/tclParse.c (CommandComplete): more paranoid check to break + on (p >= end) instead of just (p == end). 2003-02-11 Miguel Sofer <msofer@users.sf.net> - * generic/tclParse.c (CommandComplete): + * generic/tclParse.c (CommandComplete): * tests/parse.test: fix for [Bug 684744], by Don Porter. 2003-02-11 Jeff Hobbs <jeffh@ActiveState.com> - * generic/tclIOUtil.c (Tcl_FSJoinPath, Tcl_FSGetNormalizedPath): + * generic/tclIOUtil.c (Tcl_FSJoinPath, Tcl_FSGetNormalizedPath): (UpdateStringOfFsPath): revert the cwdLen == 0 check and instead follow a different code path in Tcl_FSJoinPath. (Tcl_FSConvertToPathType, Tcl_FSGetNormalizedPath): - (Tcl_FSGetFileSystemForPath): Update string rep of path objects - before freeing the internal object. (darley) + (Tcl_FSGetFileSystemForPath): Update string rep of path objects before + freeing the internal object. (darley) * tests/fileSystem.test: added test 8.3 - * generic/tclIOUtil.c (Tcl_FSGetNormalizedPath): - (UpdateStringOfFsPath): handle the cwdLen == 0 case + * generic/tclIOUtil.c (Tcl_FSGetNormalizedPath): + (UpdateStringOfFsPath): handle the cwdLen == 0 case - * unix/tclUnixFile.c (TclpMatchInDirectory): simplify the hidden - file match check. + * unix/tclUnixFile.c (TclpMatchInDirectory): simplify the hidden file + match check. 2003-02-10 Mo DeJong <mdejong@users.sourceforge.net> * win/configure: - * win/configure.in: Generate error when attempting - to build under Cygwin. The Cygwin port of Tcl/Tk - does not build and people are filing bug reports - under the mistaken impression that someone is - actually maintaining the Cygwin port. A post to - comp.lang.tcl asking someone to volunteer as an - area maintainer has generated no results. - Closing bugs 680840, 630199, and 634772 and - marking as "Won't fix". + * win/configure.in: Generate error when attempting to build under + Cygwin. The Cygwin port of Tcl/Tk does not build and people are filing + bug reports under the mistaken impression that someone is actually + maintaining the Cygwin port. A post to comp.lang.tcl asking someone to + volunteer as an area maintainer has generated no results. Closing bugs + 680840, 630199, and 634772 and marking as "Won't fix". 2003-02-10 Donal K. Fellows <fellowsd@cs.man.ac.uk> @@ -2963,17 +2870,16 @@ * tests/fileSystem.test: * unix/tclUnixFCmd.c: * unix/tclUnixFile.c: - * win/tclWinFile.c: further filesystem optimization, applying - [Patch 682500]. In particular, these code examples are - faster now: - foreach f $flist { if {[file exists $f]} {file stat $f arr;...}} - foreach f [glob -dir $dir *] { # action and/or recursion on $f } - cd $dir - foreach f [glob *] { # action and/or recursion on $f } - cd .. - - * generic/tclTest.c: Fix for [Bug 683181] where test suite - left files in 'tmp'. + * win/tclWinFile.c: further filesystem optimization, applying [Patch + 682500]. In particular, these code examples are faster now: + foreach f $flist { if {[file exists $f]} {file stat $f arr;...}} + foreach f [glob -dir $dir *] { # action and/or recursion on $f } + cd $dir + foreach f [glob *] { # action and/or recursion on $f } + cd .. + + * generic/tclTest.c: Fix for [Bug 683181] where test suite left files + in 'tmp'. 2003-02-08 Jeff Hobbs <jeffh@ActiveState.com> @@ -2985,58 +2891,52 @@ * win/tclWinFile.c: sped up pure 'glob' by a factor of 2.5 ('foreach f [glob *] { file exists $f }' is still slow) * tests/fileSystem.text: - * tests/fileName.test: added new tests to ensure correct - behaviour in optimized filesystem code. + * tests/fileName.test: added new tests to ensure correct behaviour in + optimized filesystem code. 2003-02-07 Vince Darley <vincentdarley@users.sourceforge.net> * generic/tclTest.c: - * tests/fileSystem.text: fixed test 7.2 to avoid a possible - crash, and not change the pwd. + * tests/fileSystem.text: fixed test 7.2 to avoid a possible crash, and + not change the pwd. - * tests/http.text: added comment to test 4.15, that it may - fail if you use a proxy server. + * tests/http.text: added comment to test 4.15, that it may fail if you + use a proxy server. 2003-02-06 Mo DeJong <mdejong@users.sourceforge.net> * generic/tclCompCmds.c (TclCompileIncrCmd): - * tests/incr.test: Don't include the text - "(increment expression)" in the errorInfo - generated by the compiled version of the - incr command since it does not match the - message generated by the non-compiled version - of incr. It is also not possible to match - this error output under Jacl, which does - not support a compiler. + * tests/incr.test: Don't include the text "(increment expression)" in + the errorInfo generated by the compiled version of the incr command + since it does not match the message generated by the non-compiled + version of incr. It is also not possible to match this error output + under Jacl, which does not support a compiler. 2003-02-06 Mo DeJong <mdejong@users.sourceforge.net> - * generic/tclExecute.c (TclExecuteByteCode): When an - error is encountered reading the increment value during - a compiled call to incr, add a "(reading increment)" - error string to the errorInfo variable. This makes - the errorInfo variable set by the compiled incr command - match the value set by the non-compiled version. - * tests/incr-old.test: Change errorInfo result for - the compiled incr command case to match the modified - implementation. - * tests/incr.test: Add tests to make sure the compiled - and non-compiled errorInfo messages are the same. + * generic/tclExecute.c (TclExecuteByteCode): When an error is + encountered reading the increment value during a compiled call to + incr, add a "(reading increment)" error string to the errorInfo + variable. This makes the errorInfo variable set by the compiled incr + command match the value set by the non-compiled version. + * tests/incr-old.test: Change errorInfo result for the compiled incr + command case to match the modified implementation. + * tests/incr.test: Add tests to make sure the compiled and + non-compiled errorInfo messages are the same. 2003-02-06 Don Porter <dgp@users.sourceforge.net> * library/tcltest/tcltest.tcl: Filename arguments to [outputChannel] and [errorChannel] (also -outfile and -errfile) were [open]ed but - never [closed]. Also, [cleanupTests] could remove output or error - files. [Bug 676978]. + never [closed]. Also, [cleanupTests] could remove output or error + files. [Bug 676978]. * library/tcltest/pkgIndex.tcl: Bumped to version 2.2.2. 2003-02-05 Mo DeJong <mdejong@users.sourceforge.net> * tests/interp.test: - * tests/set-old.test: Run test cases that depend - on hash order through lsort so that the tests - also pass under Jacl. Does not change test + * tests/set-old.test: Run test cases that depend on hash order through + lsort so that the tests also pass under Jacl. Does not change test results under Tcl. 2003-02-04 Vince Darley <vincentdarley@users.sourceforge.net> @@ -3050,172 +2950,158 @@ * win/tclWinFCmd.c: * win/tclWinInit.c: * win/tclWinInt.h: - * tests/fileSystem.test: fix to finalization/unloading/encoding - issues to make filesystem much less dependent on encodings for - its cleanup, and therefore allow it to be finalized later in the - exit process. This fixes fileSystem.test-7.1. Also fixed one - more bug in setting of modification dates of files which have - undergone cross-platform copies. [Patch 676271] + * tests/fileSystem.test: fix to finalization/unloading/encoding issues + to make filesystem much less dependent on encodings for its cleanup, + and therefore allow it to be finalized later in the exit process. This + fixes fileSystem.test-7.1. Also fixed one more bug in setting of + modification dates of files which have undergone cross-platform + copies. [Patch 676271] * tests/basic.test: * tests/exec.test: * tests/fileName.test: - * tests/io.test: fixed some test failures when tests are run - from a directory containing spaces. + * tests/io.test: fixed some test failures when tests are run from a + directory containing spaces. * tests/fileSystem.test: - * generic/tclTest.c: added regression test for the modification - date setting of cross-platform file copies. + * generic/tclTest.c: added regression test for the modification date + setting of cross-platform file copies. 2003-02-03 Kevin Kenny <kennykb@users.sourceforge.net> * generic/tclBasic.c: Changed [trace add command] so that 'rename' - callbacks get fully qualified names of the command. [Bug - 651271]. ***POTENTIAL INCOMPATIBILITY*** - * tests/trace.test: Modified the test cases for [trace add - command] to expect fully qualified names on the 'rename' - callbacks. Added a case for renaming a proc within a namespace. - * doc/trace.n: Added language about use of fully qualified names - in trace callbacks. + callbacks get fully qualified names of the command. [Bug 651271]. + ***POTENTIAL INCOMPATIBILITY*** + * tests/trace.test: Modified the test cases for [trace add command] to + expect fully qualified names on the 'rename' callbacks. Added a case + for renaming a proc within a namespace. + * doc/trace.n: Added language about use of fully qualified names in + trace callbacks. 2003-02-01 Kevin Kenny <kennykb@users.sourceforge.net> * generic/tclCompCmds.c: Removed an unused variable that caused compiler warnings on SGI. [Bug 664379] - * generic/tclLoad.c: Changed the code so that if Tcl_StaticPackage - is called to report the same package as being loaded in two interps, - it shows up in [info loaded {}] in both of them (previously, - it didn't appear in the static package list in the second. + * generic/tclLoad.c: Changed the code so that if Tcl_StaticPackage is + called to report the same package as being loaded in two interps, it + shows up in [info loaded {}] in both of them (previously, it didn't + appear in the static package list in the second). - * tests/load.test Added regression test for the above bug. - [Bug 670042] + * tests/load.test Added regression test for the above bug. [Bug + 670042] - * generic/tclClock.c: Fixed a bug that incorrectly allowed - [clock clicks {}] and [clock clicks -] to be accepted as if - they were [clock clicks -milliseconds]. + * generic/tclClock.c: Fixed a bug that incorrectly allowed [clock + clicks {}] and [clock clicks -] to be accepted as if they were [clock + clicks -milliseconds]. - * tests/clock.test: Added regression tests for the above bug. - [Bug 675356] + * tests/clock.test: Added regression tests for the above bug. [Bug + 675356] - * tests/unixNotfy.test: Added cleanup of working files - [Bug 675609] + * tests/unixNotfy.test: Added cleanup of working files. [Bug 675609] * doc/Tcl.n: Added headings to the eleven paragraphs, to improve formatting in the tools that attempt to extract tables of contents from the manual pages. [Bug 627455] - * generic/tclClock.c: Expanded mutex protection around the setting - of env(TZ) and the thread-unsafe call to tzset(). [Bug 656660] + * generic/tclClock.c: Expanded mutex protection around the setting of + env(TZ) and the thread-unsafe call to tzset(). [Bug 656660] 2003-01-31 Don Porter <dgp@users.sourceforge.net> * tests/tcltest.test: Cleaned up management of file/directory - creation/deletion to improve "-debug 1" output. [Bug 675614] + creation/deletion to improve "-debug 1" output. [Bug 675614] The utility [slave] command failed to properly [list]-quote a constructed [open] command, causing failure when the pathname - contained whitespace. [Bug 678415] + contained whitespace. [Bug 678415] - * tests/main.test: Stopped main.test from deleting existing file. - Test suite should not delete files that already exist. [Bug 675660] + * tests/main.test: Stopped main.test from deleting existing file. Test + suite should not delete files that already exist. [Bug 675660] 2003-01-28 Don Porter <dgp@users.sourceforge.net> - * tests/main.test: Constrain tests that do not work on Windows. - [Bug 674387] + * tests/main.test: Constrain tests that do not work on Windows. [Bug + 674387] 2003-01-28 Vince Darley <vincentdarley@users.sourceforge.net> - * generic/tclIOUtil.c: fix to setting modification date - in TclCrossFilesystemCopy. Also added 'panic' in - Tcl_FSGetFileSystemForPath under illegal calling circumstances - which lead to hard-to-track-down bugs. + * generic/tclIOUtil.c: fix to setting modification date in + TclCrossFilesystemCopy. Also added 'panic' in + Tcl_FSGetFileSystemForPath under illegal calling circumstances which + lead to hard-to-track-down bugs. - * generic/tclTest.c: added test suite code to allow - exercising a vfs-crash-on-exit bug in Tcl's finalization caused - by the encodings being cleaned up before unloading occurs. - * tests/fileSystem.test: added new 'knownBug' test 7.1 - to demonstrate the crash on exit. + * generic/tclTest.c: added test suite code to allow exercising a + vfs-crash-on-exit bug in Tcl's finalization caused by the encodings + being cleaned up before unloading occurs. + * tests/fileSystem.test: added new 'knownBug' test 7.1 to demonstrate + the crash on exit. 2003-01-28 Mo DeJong <mdejong@users.sourceforge.net> - * generic/tcl.h: Add TCL_PREFIX_IDENT and - TCL_DEBUG_IDENT, used only by TclpCreateProcess. + * generic/tcl.h: Add TCL_PREFIX_IDENT and TCL_DEBUG_IDENT, used only + by TclpCreateProcess. * unix/Makefile.in: Define TCL_DBGX. * win/Makefile.in: Define TCL_DBGX. - * win/tclWinPipe.c (TclpCreateProcess): - Check that the Tcl pipe dll actually exists - in the Tcl bin directory and panic if it - is not found. Incorporate TCL_DBGX into - the Tcl pipe dll name. This fixes a really - mysterious error that would show up when - exec'ing a 16 bit application under Win95 - or Win98 when Tcl was compiled with symbols. - The error seemed to indicate that the executable - could not be found, but it was actually the - Tcl pipe dll that could not be found. + * win/tclWinPipe.c (TclpCreateProcess): Check that the Tcl pipe dll + actually exists in the Tcl bin directory and panic if it is not found. + Incorporate TCL_DBGX into the Tcl pipe dll name. This fixes a really + mysterious error that would show up when exec'ing a 16 bit application + under Win95 or Win98 when Tcl was compiled with symbols. The error + seemed to indicate that the executable could not be found, but it was + actually the Tcl pipe dll that could not be found. 2003-01-26 Mo DeJong <mdejong@users.sourceforge.net> - * win/README: Update msys+mingw URL to release 6. - This version bundles gcc 3. + * win/README: Update msys+mingw URL to release 6. This version bundles + gcc 3. 2003-01-26 Mo DeJong <mdejong@users.sourceforge.net> * win/configure: Regen. - * win/configure.in: Add test that checks to - see if the compiler can cast to a union type. - * win/tclWinTime.c: Squelch compiler warning - about union initializer by casting to union - type when compiling with gcc. + * win/configure.in: Add test that checks to see if the compiler can + cast to a union type. + * win/tclWinTime.c: Squelch compiler warning about union initializer + by casting to union type when compiling with gcc. 2003-01-25 Mo DeJong <mdejong@users.sourceforge.net> - * generic/tclIO.c (Tcl_CutChannel, Tcl_SpliceChannel): - Invoke TclpCutFileChannel and TclpSpliceFileChannel. - * generic/tclInt.h: Declare TclpCutFileChannel - and TclpSpliceFileChannel. + * generic/tclIO.c (Tcl_CutChannel, Tcl_SpliceChannel): Invoke + TclpCutFileChannel and TclpSpliceFileChannel. + * generic/tclInt.h: Declare TclpCutFileChannel and + TclpSpliceFileChannel. * unix/tclUnixChan.c (FileCloseProc, TclpOpenFileChannel, - Tcl_MakeFileChannel, TclpCutFileChannel, - TclpSpliceFileChannel): Implement thread load data - cut and splice for file channels. This avoids - an invalid memory ref when compiled with -DDEPRECATED. + (Tcl_MakeFileChannel, TclpCutFileChannel, TclpSpliceFileChannel): + Implement thread load data cut and splice for file channels. This + avoids an invalid memory ref when compiled with -DDEPRECATED. * win/tclWinChan.c (FileCloseProc, TclpCutFileChannel, - TclpSpliceFileChannel): Implement thread load data - cut and splice for file channels. This avoids - an invalid memory ref that was showing up in the - thread extension. + (TclpSpliceFileChannel): Implement thread load data cut and splice for + file channels. This avoids an invalid memory ref that was showing up + in the thread extension. 2003-01-25 Mo DeJong <mdejong@users.sourceforge.net> * win/tclWin32Dll.c (TclpCheckStackSpace, squelch_warnings): * win/tclWinChan.c (Tcl_MakeFileChannel, squelch_warnings): * win/tclWinFCmd.c (DoRenameFile, DoCopyFile, squelch_warnings): - Re-implement inline ASM SEH handlers for gcc. - The esp and ebp registers are now saved on the - stack instead of in global variables so that - the code is thread safe. Add additional checks - when TCL_MEM_DEBUG is defined to be sure the - values were recovered from the stack properly. - Remove squelch_warnings functions and add - a dummy call in the handler methods to squelch - compiler warnings. + Re-implement inline ASM SEH handlers for gcc. The esp and ebp + registers are now saved on the stack instead of in global variables so + that the code is thread safe. Add additional checks when TCL_MEM_DEBUG + is defined to be sure the values were recovered from the stack + properly. Remove squelch_warnings functions and add a dummy call in + the handler methods to squelch compiler warnings. 2003-01-25 Mo DeJong <mdejong@users.sourceforge.net> * win/configure: - * win/configure.in: Define HAVE_ALLOCA_GCC_INLINE - when we detect that no alloca function is found - in malloc.h and we are compiling with GCC. + * win/configure.in: Define HAVE_ALLOCA_GCC_INLINE when we detect that + no alloca function is found in malloc.h and we are compiling with GCC. Remove HAVE_NO_ALLOC_DECL define. - * win/tclWin32Dll.c (TclpCheckStackSpace): - Don't define alloca as a cdecl function. - Doing this caused a tricky runtime bug because - the _alloca function expects the size argument - to be passed in a register and not on the stack. - To fix this problem, we use inline ASM when - compiling with gcc to invoke _alloca with the - size argument loaded into a register. + * win/tclWin32Dll.c (TclpCheckStackSpace): Don't define alloca as a + cdecl function. Doing this caused a tricky runtime bug because the + _alloca function expects the size argument to be passed in a register + and not on the stack. To fix this problem, we use inline ASM when + compiling with gcc to invoke _alloca with the size argument loaded + into a register. 2003-01-24 Jeff Hobbs <jeffh@ActiveState.com> @@ -3223,57 +3109,53 @@ (DdeServerProc): better refcount handling of returnPackagePtr. * generic/tclEvent.c (Tcl_Finalize): revert finalize change on - 2002-12-04 to correct the issue with extensions that have TSD - needing to finalize that before they are unloaded. This issue - needs further clarification. + 2002-12-04 to correct the issue with extensions that have TSD needing + to finalize that before they are unloaded. This issue needs further + clarification. * tests/unixFCmd.test: only do groups check on unix 2003-01-24 Vince Darley <vincentdarley@users.sourceforge.net> - * generic/tclStringObj.c: proper fixes for Tcl_SetObjLength and - Tcl_AttemptSetObjectLength dealing with string objects with - both pure-unicode and normal internal representations. - Previous fix didn't handle all cases correctly. - * generic/tclIO.c: Add 'Tcl_GetString()' to ensure the object has - a valid 'objPtr->bytes' field before manipulating it directly. + * generic/tclStringObj.c: proper fixes for Tcl_SetObjLength and + Tcl_AttemptSetObjectLength dealing with string objects with both + pure-unicode and normal internal representations. Previous fix didn't + handle all cases correctly. + * generic/tclIO.c: Add 'Tcl_GetString()' to ensure the object has a + valid 'objPtr->bytes' field before manipulating it directly. This fixes [Bug 635200] and [Bug 671138], but may reduce performance - of Unicode string handling in some cases. A further patch will - be applied to address this, once the code is known to be correct. + of Unicode string handling in some cases. A further patch will be + applied to address this, once the code is known to be correct. 2003-01-24 Mo DeJong <mdejong@users.sourceforge.net> * win/configure: Regen. - * win/configure.in: Add test to see if alloca - is undefined in malloc.h. - * win/tclWin32Dll.c (TclpCheckStackSpace): Rework - the SEH exception handler logic to avoid using - the stack since alloca will modify the stack. - This was causing a nasty bug that would set the - exception handler to 0 because it tried to pop - the previous exception handler off the top of - the stack. + * win/configure.in: Add test to see if alloca is undefined in + malloc.h. + * win/tclWin32Dll.c (TclpCheckStackSpace): Rework the SEH exception + handler logic to avoid using the stack since alloca will modify the + stack. This was causing a nasty bug that would set the exception + handler to 0 because it tried to pop the previous exception handler + off the top of the stack. 2003-01-23 Donal K. Fellows <fellowsd@cs.man.ac.uk> - * doc/lset.n: Fixed fault in return values from lset in - documentation examples [SF Bug #658463] and tidied up a bit at the - same time. + * doc/lset.n: Fixed fault in return values from lset in documentation + examples [SF Bug #658463] and tidied up a bit at the same time. 2003-01-21 Joe English <jenglish@users.sourceforge.net> + * doc/namespace.n (namespace inscope): Clarified documentation - [SF Patch #670110] + [Patch 670110] 2003-01-21 Mo DeJong <mdejong@users.sourceforge.net> * win/configure: Regen. - * win/tcl.m4 (SC_CONFIG_CFLAGS): Set SHLIB_SUFFIX - so that TCL_SHLIB_SUFFIX will be set to a useful - value in the generated tclConfig.sh. - Set SHLIB_LD_LIBS to "" or '${LIBS}' based on - the --enable-shared flag. This matches the - UNIX implementation. + * win/tcl.m4 (SC_CONFIG_CFLAGS): Set SHLIB_SUFFIX so that + TCL_SHLIB_SUFFIX will be set to a useful value in the generated + tclConfig.sh. Set SHLIB_LD_LIBS to "" or '${LIBS}' based on the + --enable-shared flag. This matches the UNIX implementation. 2003-01-18 Jeff Hobbs <jeffh@ActiveState.com> @@ -3281,158 +3163,139 @@ 2003-01-17 Mo DeJong <mdejong@users.sourceforge.net> - * win/tclWinDde.c (DdeServerProc): Deallocate - the Tcl_Obj returned by ExecuteRemoteObject - if it was not saved in a connection object. + * win/tclWinDde.c (DdeServerProc): Deallocate the Tcl_Obj returned by + ExecuteRemoteObject if it was not saved in a connection object. 2003-01-17 Mo DeJong <mdejong@users.sourceforge.net> - * generic/tcl.h: Revert earlier change that - defined TCL_WIDE_INT_TYPE as long long and - TCL_LL_MODIFIER as L when compiling with - mingw. This change ended up causing some - test case failures when compiling with mingw. - * generic/tclObj.c (UpdateStringOfWideInt): - Describe the warning generated by mingw and - why it needs to be ignored so that someone - is not tempted to "fix" this problem again - in the future. + * generic/tcl.h: Revert earlier change that defined TCL_WIDE_INT_TYPE + as long long and TCL_LL_MODIFIER as L when compiling with mingw. This + change ended up causing some test case failures when compiling with + mingw. + * generic/tclObj.c (UpdateStringOfWideInt): Describe the warning + generated by mingw and why it needs to be ignored so that someone is + not tempted to "fix" this problem again in the future. 2003-01-16 Vince Darley <vincentdarley@users.sourceforge.net> - * generic/tclStringObj.c: Tcl_SetObjLength fix for when - the object has a unicode string rep. Fixes [Bug 635200] - * tests/stringObj.test: removed 'knownBug' constraint from - test 14.1 now that this bug is fixed. + * generic/tclStringObj.c: Tcl_SetObjLength fix for when the object has + a unicode string rep. [Bug 635200] + * tests/stringObj.test: removed 'knownBug' constraint from test 14.1 + now that this bug is fixed. * generic/tclInt.h: * generic/tclBasic.c: * generic/tclCmdMZ.z: * tests/trace.test: execution and command tracing bug fixes and - cleanup. In particular fixed [Bug 655645], [Bug 615043], - [Bug 571385] - - fixed some subtle cleanup problems with tracing. This - required replacing Tcl_Preserve/Tcl_Release with a more - robust refCount approach. Solves at least one known crash - caused by memory corruption. - - fixed some confusion in the code between new style traces - (Tcl 8.4) and the very limited 'Tcl_CreateTrace' which existed - before. - - made behaviour consistent with documentation (several - tests even contradicted the documentation before). + cleanup. In particular fixed [Bug 655645], [Bug 615043], [Bug 571385] + - fixed some subtle cleanup problems with tracing. This required + replacing Tcl_Preserve/Tcl_Release with a more robust refCount + approach. Solves at least one known crash caused by memory + corruption. + - fixed some confusion in the code between new style traces (Tcl + 8.4) and the very limited 'Tcl_CreateTrace' which existed before. + - made behaviour consistent with documentation (several tests even + contradicted the documentation before). - fixed some minor error message details - added a number of new tests 2003-01-16 Jeff Hobbs <jeffh@ActiveState.com> - * win/tclWinSerial.c (SerialOutputProc): add casts for - bytesWritten to allow strict compilation (no warnings). + * win/tclWinSerial.c (SerialOutputProc): add casts for bytesWritten to + allow strict compilation (no warnings). * tests/winDde.test: - * win/tclWinDde.c (Tcl_DdeObjCmd): Prevent crash when empty - service name is passed to 'dde eval' and goto errorNoResult in - request and poke error cases to free up any allocated data. + * win/tclWinDde.c (Tcl_DdeObjCmd): Prevent crash when empty service + name is passed to 'dde eval' and goto errorNoResult in request and + poke error cases to free up any allocated data. 2003-01-16 Mo DeJong <mdejong@users.sourceforge.net> - * win/tclWin32Dll.c (squelch_warnings): Squelch - compiler warnings from SEH ASM code. - * win/tclWinChan.c (squelch_warnings): Squelch - compiler warnings from SEH ASM code. - * win/tclWinDde.c: Add casts to avoid compiler - warnings. Pass pointer to DWORD instead of int - to avoid compiler warnings. - * win/tclWinFCmd.c (squelch_warnings): Add casts - and fixup decls to avoid compiler warnings. - Squelch compiler warnings from SEH ASM code. - * win/tclWinFile.c: Add casts and fixup decls - to avoid compiler warnings. Remove unused variable. - * win/tclWinNotify.c: Declare as DWORD instead - of int to avoid compiler warning. - * win/tclWinReg.c: Add casts to avoid compiler - warning. Fix assignment in if expression bug. - * win/tclWinSerial.c: Add casts to avoid compiler + * win/tclWin32Dll.c (squelch_warnings): Squelch compiler warnings from + SEH ASM code. + * win/tclWinChan.c (squelch_warnings): Squelch compiler warnings from + SEH ASM code. + * win/tclWinDde.c: Add casts to avoid compiler warnings. Pass pointer + to DWORD instead of int to avoid compiler warnings. + * win/tclWinFCmd.c (squelch_warnings): Add casts and fixup decls to + avoid compiler warnings. Squelch compiler warnings from SEH ASM code. + * win/tclWinFile.c: Add casts and fixup decls to avoid compiler warnings. Remove unused variable. - * win/tclWinSock.c: Add casts and fixup decls - to avoid compiler warnings. + * win/tclWinNotify.c: Declare as DWORD instead of int to avoid + compiler warning. + * win/tclWinReg.c: Add casts to avoid compiler warning. Fix assignment + in if expression bug. + * win/tclWinSerial.c: Add casts to avoid compiler warnings. Remove + unused variable. + * win/tclWinSock.c: Add casts and fixup decls to avoid compiler + warnings. 2003-01-14 Jeff Hobbs <jeffh@ActiveState.com> - * generic/tclClock.c (FormatClock): corrected typo that - incorrectly conditionally defined savedTZEnv and savedTimeZone. + * generic/tclClock.c (FormatClock): corrected typo that incorrectly + conditionally defined savedTZEnv and savedTimeZone. 2003-01-13 Mo DeJong <mdejong@users.sourceforge.net> Fix mingw build problems and compiler warnings. - * generic/tcl.h: Add if defined(__MINGW32__) - check to code that sets the TCL_WIDE_INT_TYPE - and TCL_LL_MODIFIER. - * generic/tclClock.c (FormatClock): Don't - define savedTimeZone and savedTZEnv if - we are not going to use them. + * generic/tcl.h: Add if defined(__MINGW32__) check to code that sets + the TCL_WIDE_INT_TYPE and TCL_LL_MODIFIER. + * generic/tclClock.c (FormatClock): Don't define savedTimeZone and + savedTZEnv if we are not going to use them. * generic/tclEnv.c: Add cast to avoid warning. - * win/tclWinChan.c: Use DWORD instead of int - to avoid compiler warning. - * win/tclWinThrd.c: Only define allocLock, - allocLockPtr, and dataKey when TCL_THREADS - is defined. This avoid a compiler warning - about unused variables. + * win/tclWinChan.c: Use DWORD instead of int to avoid compiler warning + * win/tclWinThrd.c: Only define allocLock, allocLockPtr, and dataKey + when TCL_THREADS is defined. This avoid a compiler warning about + unused variables. 2003-01-12 Mo DeJong <mdejong@users.sourceforge.net> - * win/README: Update msys + mingw URL, the - new release includes the released 1.0.8 - version of msys which includes a number - of bug fixes. + * win/README: Update msys + mingw URL, the new release includes the + released 1.0.8 version of msys which includes a number of bug fixes. 2003-01-12 Mo DeJong <mdejong@users.sourceforge.net> * win/configure: Regen. - * win/tcl.m4 (SC_CONFIG_CFLAGS): Pull in - addition of shell32.lib to LIBS_GUI that - was added to the Tk tcl.m4 but never made - it back into the Tcl version. + * win/tcl.m4 (SC_CONFIG_CFLAGS): Pull in addition of shell32.lib to + LIBS_GUI that was added to the Tk tcl.m4 but never made it back into + the Tcl version. 2003-01-12 Mo DeJong <mdejong@users.sourceforge.net> - * generic/tcl.h: Skip Tcl's define of CHAR, - SHORT, and LONG when HAVE_WINNT_IGNORE_VOID - is defined. This avoids a bunch of compiler + * generic/tcl.h: Skip Tcl's define of CHAR, SHORT, and LONG when + HAVE_WINNT_IGNORE_VOID is defined. This avoids a bunch of compiler warnings when building with Cygwin or Mingw. * win/configure: Regen. - * win/configure.in: Define HAVE_WINNT_IGNORE_VOID - when we detect a winnt.h that still defines - CHAR, SHORT, and LONG when VOID has already + * win/configure.in: Define HAVE_WINNT_IGNORE_VOID when we detect a + winnt.h that still defines CHAR, SHORT, and LONG when VOID has already been defined. - * win/tcl.m4 (SC_LOAD_TCLCONFIG): Subst the - TCL_DEFS loaded from tclConfig.sh so that - Tcl defines can make it into the Tk Makefile. + * win/tcl.m4 (SC_LOAD_TCLCONFIG): Subst the TCL_DEFS loaded from + tclConfig.sh so that Tcl defines can make it into the Tk Makefile. 2003-01-12 Mo DeJong <mdejong@users.sourceforge.net> * win/configure: Regen. - * win/configure.in: Check for typedefs like LPFN_ACCEPT - in winsock2.h and define HAVE_NO_LPFN_DECLS if not found. - * win/tclWinSock.c: Define LPFN_* typedefs if - HAVE_NO_LPFN_DECLS is defined. This fixes the build - under Mingw and Cygwin, it was broken by the changes - made on 2002-11-26. + * win/configure.in: Check for typedefs like LPFN_ACCEPT in winsock2.h + and define HAVE_NO_LPFN_DECLS if not found. + * win/tclWinSock.c: Define LPFN_* typedefs if HAVE_NO_LPFN_DECLS is + defined. This fixes the build under Mingw and Cygwin, it was broken by + the changes made on 2002-11-26. 2003-01-10 Vince Darley <vincentdarley@users.sourceforge.net> * generic/tclIOUtil.c: * win/tclWinInt.h: * win/tclWinInit.c: fix to new WinTcl crash on exit with vfs, - introduced on 2002-12-06. Encodings must be cleaned up after - the filesystem. + introduced on 2002-12-06. Encodings must be cleaned up after the + filesystem. * win/makefile.vc: fix to minor VC++ 5.2 syntax problem 2003-01-09 Don Porter <dgp@users.sourceforge.net> * generic/tclCompCmds.c (TclCompileReturnCmd): Corrected off-by-one - problem with recent commit. [Bug 633204] + problem with recent commit. [Bug 633204] 2003-01-09 Vince Darley <vincentdarley@users.sourceforge.net> @@ -3445,40 +3308,38 @@ * mac/tclMacFile.c: * win/tclWinFile.c: * win/tclWinInt.h: - * win/tclWin32Dll.c: - * tests/cmdAH.test: fix to non-ascii chars in paths when - setting mtime and atime through 'file (a|m)time $path $time' - [Bug 634151] + * win/tclWin32Dll.c: + * tests/cmdAH.test: fix to non-ascii chars in paths when setting mtime + and atime through 'file (a|m)time $path $time'. [Bug 634151] 2003-01-08 Don Porter <dgp@users.sourceforge.net> - * generic/tclExecute.c (TclExprFloatError): Use the IS_NAN macro - for greater clarity of code. + * generic/tclExecute.c (TclExprFloatError): Use the IS_NAN macro for + greater clarity of code. 2003-01-07 Don Porter <dgp@users.sourceforge.net> - * generic/tclCompCmds.c (TclCompileReturnCmd): - * tests/compile.test: Corrects failure of bytecompiled - [catch {return}] to have result TCL_RETURN (not TCL_OK) [Bug 633204]. - This patch is a workaround for 8.4.X. A new opcode INST_RETURN is a - better long term solution for 8.5 and later. + * generic/tclCompCmds.c (TclCompileReturnCmd): + * tests/compile.test: Corrects failure of bytecompiled [catch + {return}] to have result TCL_RETURN (not TCL_OK) [Bug 633204]. This + patch is a workaround for 8.4.X. A new opcode INST_RETURN is a better + long term solution for 8.5 and later. 2003-01-04 David Gravereaux <davygrvy@pobox.com> * win/makefile.vc: - * win/rules.vc: Fixed INSTALLDIR macro problem that blanked itself - by accident causing the install target to put the tree at the root - of the drive built on. Whoops.. - - Renamed the 'linkexten' option to be 'staticpkg'. Added 'thrdalloc' - to allow the switching _on_ of the thread allocator. Under testing, - I found it not to be benificial under windows for the purpose of the - application I was using it for. It was more important for this app + * win/rules.vc: Fixed INSTALLDIR macro problem that blanked itself by + accident causing the install target to put the tree at the root of the + drive built on. Whoops.. + + Renamed the 'linkexten' option to be 'staticpkg'. Added 'thrdalloc' to + allow the switching _on_ of the thread allocator. Under testing, I + found it not to be benificial under windows for the purpose of the + application I was using it for. It was more important for this app that resources for tcl threads be returned to the system rather than - saved/moved to the global recycler. Be extra clean or extra fast - for the default threaded build? Let's move to clean and allow it to - be switched on for users who find it benificial for their use of - threads. + saved/moved to the global recycler. Be extra clean or extra fast for + the default threaded build? Let's move to clean and allow it to be + switched on for users who find it benificial for their use of threads. ****************************************************************** *** CHANGELOG ENTRIES FOR 2002 IN "ChangeLog.2002" *** |