summaryrefslogtreecommitdiffstats
path: root/generic
Commit message (Collapse)AuthorAgeFilesLines
* * generic/tclExecute.c (TclExecuteByteCode INST_STR_CMP):hobbs2002-05-302-15/+13
| | | | | | | * generic/tclCmdMZ.c (Tcl_StringObjCmd): changed the case for choosing the Tcl_UniCharNcmp compare to when both objs are of StringType, as benchmarks show that is the optimal check (both bigendian and littleendian systems).
* * Removed "dummy" reference to Tcl_LinkVar.dgp2002-05-291-11/+1
| | | | | It is no longer needed since Tcl_Main() now actually calls Tcl_LinkVar(). Thanks to Joe English for pointing that out.
* Made Tcl_UniCharNcmp faster on big-endian machines; the system memcmp()isdkf2002-05-294-9/+33
| | | | | probably optimized far in excess of anything we could do! Little-endian just use the old code...
* * generic/tclInt.decls:hobbs2002-05-296-153/+193
| | | | | | | | | | | | | | | | | | | | | | | * generic/tclIntDecls.h: * generic/tclStubInit.c: * generic/tclUtf.c: added TclpUtfNcmp2 private command that mirrors Tcl_UtfNcmp, but takes n in bytes, not utf-8 chars. This provides a faster alternative for comparing utf strings internally. (Tcl_UniCharNcmp, Tcl_UniCharNcasecmp): removed the explicit end of string check as it wasn't correct for the function (by doc and logic). * generic/tclCmdMZ.c (Tcl_StringObjCmd): reworked the string equal comparison code to use TclpUtfNcmp2 as well as short-circuit for equal objects or unequal length strings in the equal case. Removed the use of goto and streamlined the other parts. * generic/tclExecute.c (TclExecuteByteCode): added check for object equality in the comparison instructions. Added short-circuit for != length strings in INST_EQ, INST_NEQ and INST_STR_CMP. Reworked INST_STR_CMP to use TclpUtfNcmp2 where appropriate, and only use Tcl_UniCharNcmp when at least one of the objects is a Unicode obj with no utf bytes.
* * generic/tclCompCmds.c (TclCompileStringCmd): removed errorhobbs2002-05-291-7/+1
| | | | creation in code that no longer throws an error.
* * tests/clock.test: added clock-9.1hobbs2002-05-293-9/+10
| | | | | | | | | | | | * compat/strftime.c: * generic/tclClock.c: * generic/tclInt.decls: * generic/tclIntDecls.h: * unix/tclUnixTime.c: fix for Windows msvcrt mem leak caused by using an env(TZ) setting trick for in clock format -gmt 1. This also makes %s seem to work correctly with -gmt 1 as well as making it a lot faster by avoid the env(TZ) hack. TclpStrftime now takes useGMT as an arg. [Bug #559376]
* * generic/tclThreadAlloc.c (TclpRealloc, TclpFree): protecthobbs2002-05-291-26/+32
| | | | against the case when NULL is based.
* load cleanup in vfsvincentdarley2002-05-281-7/+56
|
* Better handling of EOVERFLOW on Win32 systems (as EFBIG or EINVAL.)dkf2002-05-271-3/+3
|
* TIP#91 implementation; makes old style channels binary compatible withdkf2002-05-246-90/+298
| | | | new TIP#72-enabled Tcl. See http://purl.org/tcl/tip/91 for details.
* * generic/tclInt.h: added prototype for TclpFilesystemPathType().das2002-05-201-1/+3
| | | | | * mac/tclMacChan.c: use MSL provided creator type if available instead of the default 'MPW '.
* memory cleanupvincentdarley2002-05-142-12/+20
|
* memory cleanupvincentdarley2002-05-133-4/+63
|
* commentvincentdarley2002-05-131-3/+5
|
* fix to bug 553320vincentdarley2002-05-071-2/+9
|
* fix to 551306vincentdarley2002-05-022-37/+51
|
* Hex values on 64-bit machines can be booleans too! [Bug 548686]dkf2002-04-261-3/+25
|
* * generic/tclInt.h: corrected TclRememberJoinableThread decl tohobbs2002-04-241-2/+2
| | | | use VOID instead of void.
* noted that this code isn't needed on unixhobbs2002-04-241-2/+6
|
* * generic/tclAlloc.c:hobbs2002-04-233-2/+954
| | | | | | | | | | | | | | | * generic/tclInt.h: * generic/tclThreadAlloc.c (new): * unix/Makefile.in: * unix/tclUnixThrd.c: * win/Makefile.in: * win/tclWinInt.h: * win/tclWinThrd.c: added new threaded allocator contributed by AOL that significantly reduces lock contention when multiple threads are in use. Only Windows and Unix implementations are ready, and the Windows one may need work. It is only used by default on Unix for now, and requires that USE_THREAD_ALLOC be defined (--enable-threads on Unix will define this).
* cleanup of testfilesystem commandhobbs2002-04-231-19/+11
|
* * generic/tclIOUtil.c (Tcl_FSRegister, Tcl_FSUnregister):hobbs2002-04-231-10/+13
| | | | | corrected calling of Tcl_ConditionWait to ensure that there would be a condition to wait upon.
* generic/tclCmdAH.c (Tcl_FileObjCmd): added cast in FILE_SIZEhobbs2002-04-231-3/+4
|
* * Free the memory allocated for thedgp2002-04-201-2/+3
| | | | startup script path. [Bug 543549]
* 2002-04-20 Daniel Steffen <das@users.sourceforge.net>das2002-04-193-3/+16
| | | | | | | | | | | | | | * generic/tclInt.decls: * generic/tclIntPlatDecls.h: * generic/tclStubInit.c: * mac/tclMacFCmd.c: * mac/tclMacFile.c: * mac/tclMacUtil.c: Modified TclpObjNormalizePath to be alias file aware, and replaced various calls to FSpLocationFrom*Path by calls to new alias file aware versions FSpLLocationFrom*Path. The alias file aware routines don't resolve the last component of a path if it is an alias. This allows [file copy/delete] etc. to act correctly on alias files. (c.f. discussion in Bug #511666)
* list/string indexes don't destroy wide int rep [Bug #526717]dkf2002-04-191-48/+80
|
* [Bug 545325 ] info level didn't report namespace eval.Miguel Sofer2002-04-181-4/+6
|
* Fixed Bug #545644; [info body] always gives a proper string now!dkf2002-04-181-1/+8
|
* Partial resolution of Bug #536831; the comment is a bit clearer!dkf2002-04-181-2/+10
|
* fix for [Bug #542588], where "too large integers" were reported asMiguel Sofer2002-04-181-12/+20
| | | | "floating-point value" in [expr] error messages.
* * generic/tclEncoding.c (EscapeFromUtfProc):hobbs2002-04-182-34/+81
| | | | | | * generic/tclIO.c (WriteChars, Tcl_Close): corrected the handling of outputting end escapes for escape-based encodings. [Bug #526524] (yamamoto)
* Improved stack trace for TCL_BREAK and TCL_CONTINUE returns from procs. [Bug ↵Miguel Sofer2002-04-151-24/+23
| | | | 536955].
* made bytecodes check for a catch before returning; the compiled [return] is ↵Miguel Sofer2002-04-151-10/+12
| | | | otherwise non-catchable. [Bug 542588]
* * generic/tclNotify.c (TclFinalizeNotifier): remove remaininghobbs2002-04-121-1/+12
| | | | unserviced events on finalization.
* * generic/tclParse.c (Tcl_ParseVar): conditionally incr objhobbs2002-04-121-2/+10
| | | | refcount to prevent possible mem leak.
* 2002-04-08 Daniel Steffen <das@users.sourceforge.net>das2002-04-081-1/+5
| | | | | | | | | | | | | | | | | | | * generic/tcl.h: no <sys/types.h> on mac. * mac/tclMacFile.c: minor fixes to Vince's changes from 03-24. * mac/tclMacOSA.c: * mac/tclMacResource.c: added missing Tcl_UtfToExternalDString conversions of resource file names. * mac/tclMacSock.c (TcpGetOptionProc): fixed bug introduced by Andreas on 02-25; changed strcmp's to strncmp's so that option comparison behaves like on other platforms. * mac/tcltkMacBuildSupport.sea.hqx (CW Pro6 changes): added support to allow Tk to hookup C library stderr/stdout to TkConsole. * tests/basic.test: * tests/cmdAH.test: * tests/encoding.test: * tests/fileSystem.test: * tests/ioCmd.test: fixed tests failing on mac: check for existence of [exec], changed some result strings.
* * Corrected [subst] so that return codesdgp2002-04-051-3/+9
| | | | | | TCL_BREAK and TCL_CONTINUE returned by variable substitution have the same effect as when those codes are returned by command substitution. [Bug 536879]
* tilde expansion cache-clearingvincentdarley2002-04-032-21/+37
|
* * generic/tclCmdMZ.c (Tcl_TraceObjCmd, TraceVarProc)hobbs2002-03-291-13/+23
| | | | | | | | | (TraceCommandProc, TclTraceCommandObjCmd): corrected potential double-free of traces on variables by flagging in Trace*Proc that it will free the var in case the eval wants to delete the var trace as well. [Bug #536937] Also converted Tcl_UntraceVar -> Tcl_UntraceVar2 and Tcl_Eval to Tcl_EvalEx in Trace*Proc for slight efficiency improvement.
* * Corrected problems with Tcl_AllowExceptionsdgp2002-03-293-36/+42
| | | | | having influence over the wrong scope of Tcl_*Eval* calls. Patch from Miguel Sofer. Report from Jean-Claude Wippler. [Bug 219181]
* * Refactored CallTraces to collect repeateddgp2002-03-291-257/+124
| | | | handling of its returned value into CallTraces itself.
* passing the correct commandSize to TclEvalObjvInternal. [Bug 219362],Miguel Sofer2002-03-271-2/+3
| | | | fix by David Knoll.
* avoid exceptional returns at level 0 [Bug 219181]Miguel Sofer2002-03-271-1/+11
|
* Allow NULL callback on trace deletions [Bug 534728]Miguel Sofer2002-03-251-2/+4
|
* fix for [Bug 533907] in tclBasic.c (Tcl_EvalObjv).Miguel Sofer2002-03-241-27/+34
|
* fixed typo and compiler warning in last commit; all tests now pass on unixvincentdarley2002-03-241-2/+2
|
* 4 fs fixesvincentdarley2002-03-245-52/+83
|
* fixed the errorInfo for return codes other than (TCL_OK, TCL_ERROR) toMiguel Sofer2002-03-222-66/+19
| | | | | runLevel 0 [Bug 533758]. Removed the static RecordTracebackInfo(), as its functionality is easily replicated by Tcl_LogCommandInfo.
* * Updated interfaces of generic/tclVar.c accordingdgp2002-03-2011-74/+79
| | | | to TIP 27. In particular, the "part2" arguments were CONSTified.
* Fixed buffer overrun reported in 530320; luckily it is not likely todkf2002-03-151-3/+4
| | | | | be exploitable in any meaningful way, but crashing Tcl instead of triggering an error still isn't good.
generated by cgit v0.12 at 2024-11-06 02:01:47 (GMT)