summaryrefslogtreecommitdiffstats
path: root/generic
Commit message (Collapse)AuthorAgeFilesLines
* reverted earlier rename from tcl*Stubs tonijtmans2010-02-156-27/+27
| | | | | | | | tcl*ConstStubs, it's not necessary at all. tclEnsemble.c: Fix signed-unsigned mismatch make tclWinProcs "const" Add first part of mslu support, See [Feature Request #2819611]
* Fix [Bug 2950259] so that deleting an object by killing its namespace willdkf2010-02-153-4/+69
| | | | reliably call the object's destructor.
* Reduce code nesting in [namespace ensemble] implementation.dkf2010-02-141-139/+116
|
* Corrected a comment.dkf2010-02-141-3/+2
|
* Hive off the ensemble code into its own file.dkf2010-02-135-3496/+3596
| | | | Split the [switch] compiler for sanity's sake.
* [Bug 2949740]: Do not try to put a NULL pipeline channel into binary mode.dkf2010-02-111-2/+2
|
* Tcl Bug 2826551 regexp bugs related to -all -line and -start and newlinesdkf2010-02-111-20/+29
|
* [Bug 2949397]: Prevent destructors from running on the two core class objectsdkf2010-02-111-2/+9
| | | | when the whole interpreter is being destroyed.
* Forgot the magic bias values. FNV is wildly magical...dkf2010-02-102-4/+4
|
* Minor corrections as recommended by Joe English.dkf2010-02-091-4/+4
|
* Small cleanups of formatting issuesdkf2010-02-091-72/+88
|
* Compilation of [try] now enabled!dkf2010-02-093-4/+646
|
* [Bug 2947783]: Ensure that result is an unshared object before appending to it.dkf2010-02-081-1/+19
|
* Upgrade Tcl's hash function to use the FNV-32 algorithm. This is marginallydkf2010-02-073-76/+30
| | | | | faster and gives a bit better distribution of keys (especially in large hash tables) but does change hash iteration order.
* Added basic compilation of [error] (the most common case only).dkf2010-02-053-4/+52
|
* Follow-up to earlier commit today:nijtmans2010-02-056-41/+30
| | | | | | Eliminate the need for an extra Stubs Pointer for adressing a static stub table: Just change the exported table from static to MODULE_SCOPE.
* More consistency in errorcode generation.dkf2010-02-051-14/+25
|
* NRE-enabled destructors! Also more generation of errorcodes.dkf2010-02-051-17/+49
|
* Use 'const' more often for pointers to read-only structures.dkf2010-02-051-177/+203
|
* Follow-up to [2010-01-29] commit:nijtmans2010-02-0514-851/+851
| | | | | | | | | | prevent space within stub table function parameters if the parameter type is a pointer. Minor formatting, and VOID -> void. Change signature of TclNRInterpProcCore, and TclOONewProc(Instance|)MethodEx, indicating that errorProc is a function pointer tclVar.c: fixed two gcc warnings
* Use the object RE interface for faster matching in [array names -regexp].dkf2010-02-041-121/+136
|
* Make [array get] work again with a trivial pattern.dkf2010-02-041-4/+4
|
* Minor updates: more errorcodes, less C stack levels for old APIsdkf2010-02-041-98/+191
|
* More corrections to the [array unset] command.dkf2010-02-031-26/+30
|
* Turned the [array] command into a true ensemble. Test changes indicate somedkf2010-02-023-560/+1184
| | | | alteration to error messages, otherwise no change.
* Chisel away at reducing the cost of recent changes.dkf2010-02-021-59/+55
|
* [Bug 2944404] Be careful in case an object deletes itself in its destructor.dkf2010-02-021-3/+6
|
* Switch to using the new faster var-ref internal API inside [array unset] whendkf2010-02-021-4/+5
| | | | it makes sense.
* Fix [Bug 2939073]: dangling ref when an unset trace triggered by [array unset]dkf2010-02-021-20/+59
| | | | hits the next element to be deleted.
* Reduce obscurity of function names. Do a little recursion unrolling.dkf2010-02-012-352/+338
|
* Use attemptckalloc instead of ckalloc; RE engine knows how to deal with failuredkf2010-02-011-2/+2
|
* [Bug 2942697]: Rework the RE engine so that certain pathological patterns aredkf2010-02-011-57/+67
| | | | | matched much more rapidly. Many thanks to Tom Lane for dianosing this issue and providing an initial patch.
* One more function that needs the LVT index passing into it now.dkf2010-01-311-5/+7
|
* Unbreak the builddkf2010-01-311-31/+35
|
* Make the [unset] command be bytecode compiled.dkf2010-01-307-376/+689
|
* Reverted Tcl_ThreadDataKey type change,nijtmans2010-01-291-12/+11
| | | | | | | see Bug #2942081 Changed some Tcl_CallFrame fields from "char *" to "void *". This saves unnecessary space on Cray's (and it's simply more correct).
* - genStubs.tcl: No longer generate a space after "*" andnijtmans2010-01-2916-1766/+1690
| | | | | | | | | | immediately after a function name, so the format of function definitions in tcl*Decls.h matches all other tcl*.h header files. - Change Tcl_ArgvFuncProc, Tcl_ArgvGenFuncProc and GetFrameInfoValueProc to be function definitions, not pointers, for consistency with all other Tcl function definitions.
* Make things compile... D'oh!dkf2010-01-282-5/+7
|
* Improvements to destructor handling.dkf2010-01-283-18/+91
| | | | Stop crashes from odd destruction routes.
* Remove double includes (which causes anijtmans2010-01-252-5/+2
| | | | | warning in CYGWIN compiles) Add confdefs.h to unix/.cvsignore
* Revert [2009-12-21] change in tcl.h, in steadnijtmans2010-01-221-8/+2
| | | | | | | | resolve the CYGWIN inclusion problems by re-arranging the inclusions at other places. Make cygwin configuration error into a warning: CYGWIN compilation works although there still are test failures.
* Revert [2009-12-21] change in tcl.h, in steadnijtmans2010-01-224-28/+28
| | | | | | | | resolve the CYGWIN inclusion problems by re-arranging the inclusions at other places. Make cygwin configuration error into a warning: CYGWIN compilation works although there still are test failures.
* Improve error code generation from some of the tailcall-related bits of TEBC.dkf2010-01-221-29/+32
|
* * generic/tclCompile.h: NRE-enable direct eval on BC spoilageMiguel Sofer2010-01-212-25/+25
| | | | | * generic/tclExecute.c: [Bug 2910748] * tests/nre.test:
* * generic/tclIO.c (CreateScriptRecord): [Bug 2918110]: Initializeandreas_kupries2010-01-181-6/+23
| | | | | | | | the EventScriptRecord (esPtr) fully before handing it to Tcl_CreateChannelHandler for registration. Otherwise a reflected channel calling 'chan postevent' (== Tcl_NotifyChannel) in its 'watchProc' will cause the function 'TclChannelEventScriptInvoker' to be run on an uninitialized structure.
* Ensure that objects are released in all cases (hard to test this particulardkf2010-01-181-1/+4
| | | | one without a very large system, but the leak path existed)
* [Bug 2932421]: Make [format] less likely to smash intreps.dkf2010-01-181-40/+48
|
* Fix TCL_LL_MODIFIER for Cygwinnijtmans2010-01-134-12/+17
| | | | | and various other minor CYGWIN compilation problems
* * win/tclWinDde.c: VC++ 6.0 doesn't havenijtmans2010-01-103-9/+9
| | | | | | | | | | | | | | | | | | | | * win/tclWinReg.c PDWORD_PTR * win/tclWinThrd.c: Fix various minor gcc warnings. * win/tclWinTime.c * win/tclWinConsole.c Put channel type definitions * win/tclWinChan.c in static const memory * win/tclWinPipe.c * win/tclWinSerial.c * win/tclWinSock.c * generic/tclIOGT.c * generic/tclIORChan.c * generic/tclIORTrans.c * unix/tclUnixChan.c * unix/tclUnixPipe.c * unix/tclUnixSock.c * unix/configure (regenerated with autoconf 2.59) * tests/info.test: Make test independant from tcltest implementation.
* * generic/tclPathObj.c (TclPathPart): Correct inconsistency betweendgp2010-01-051-26/+10
| | | | | | * tests/fileName.test (filename-14.31): the string rep and the intrep of a path value created by [file rootname]. Thanks to Vitaly Magerya for reporting. [Bug 2918610]