| Commit message (Collapse) | Author | Age | Files | Lines |
... | |
| |
| |
| |
| | |
more streamlining.
|
| | |
|
| |
| |
| |
| |
| |
| | |
that ended up not really optimising (limited benchmarks). Now
calling UnsetVarStruct (streamlined old code is #ifdef'ed out, in
case better benchmarks do show a difference).
|
| |
| |
| |
| | |
last commit.
|
| |
| |
| |
| |
| |
| |
| |
| |
| | |
slightly clearer, eliminated some duplicated code.
*** POTENTIAL INCOMPATIBILITY *** (tclInt.h and Var struct users)
The core never builds VAR_LINK variable to have traces. Such a
"monster", should one exist, will now have its unset traces called
*before* it is unlinked.
|
| |
| |
| |
| | |
compute the variable's full name if the variable is traced.
|
| |
| |
| |
| |
| |
| |
| |
| |
| | |
* generic/tclNamesp.c:
* generic/tclObj.c:
* generic/tclvar.c: new macros TclGetCurrentNamespace() and
TclGetGlobalNamespace(); Tcl_GetCommandFromObj and
TclGetNamespaceFromObj rewritten to make the logic clearer;
slightly faster too.
|
| | |
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
* generic/tclLink.c (LinkTraceProc): Update Tcl_VarTraceProcs so that
* generic/tclNamesp.c (Error*Read): they call Tcl_InterpDeleted()
* generic/tclTrace.c (Trace*Proc): for themselves, and do not rely
* generic/tclUtil.c (TclPrecTraceProc): on (frequently buggy) setting
of the TCL_INTERP_DESTROYED flag by the trace core.
* generic/tclVar.c: Update callers of TclCallVarTraces to not
pass in the TCL_INTERP_DESTROYED flag. Also apply filters so that
public routines only pass documented flag values down to lower level
routines.
* generic/tclTrace.c (TclCallVarTraces): The setting of the
TCL_INTERP_DESTROYED flag is now done entirely within the
TclCallVarTraces routine, the only place it can be done right.
|
| |
| |
| |
| |
| | |
had refCount 0 and was used for appending (but not
lappending). [Bug 1710710], thanks to mistachkin and kbk.
|
| |
| |
| |
| |
| | |
TCL_INTERP_DESTROYED flags gets passed to unset trace routines
so they can respond appropriately. [Bug 1705778, leak #9]
|
| |
| |
| |
| | |
memory detected by valgrind.
|
| |
| |
| |
| | |
1515234]
|
| |
| |
| |
| |
| |
| |
| |
| | |
* generic/tclCompCmds.c: variable linking commands: 'global',
* generic/tclCompile.h: 'variable', 'upvar', 'namespace upvar'
* generic/tclExecute.c: [Patch 1688593]
* generic/tclInt.h:
* generic/tclVar.c:
|
| |
| |
| |
| |
| | |
* tests/var.test (var-17.1): the "list" argument to [array set] to
avoid crash due to shimmering invalidating pointers. [Bug 1669489].
|
| | |
|
| | |
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
* generic/tclBasic.c: there is always a valid CallFrame, even
* generic/tclCmdIL.c: at level 0 [Patch 1577278]. Most of the
* generic/tclInt.h: changes involve removing tests for
* generic/tclNamesp.c: iPtr->(var)framePtr==NULL. There is now a
* generic/tclObj.c: CallFrame pushed at interp creation
* generic/tclProc.c: with a pointer to it stored in
* generic/tclTrace.c: iPtr->rootFramePtr. A second unused
* generic/tclVar.c: field in Interp is hijacked to enable
further functionality, currently unused (but with several FRQs
depending on it).
|
| | |
|
| |
| |
| |
| | |
other developers
|
| |
| |
| |
| |
| | |
* tests/append.test(4.21-22): fix for longstanding [Bug 1570718],
lappending nothing to non-list. Reported by lvirden
|
| | |
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
* doc/incr.n: Revised [incr] to auto-initialize when varName
* generic/tclExecute.c: argument is unset. [Patch 1413115].
* generic/tclVar.c:
* tests/compile.test:
* tests/incr-old.test:
* tests/incr.test:
* tests/set.test:
|
| | |
|
| |
| |
| |
| |
| |
| |
| |
| |
| | |
* doc/namespace.n: New command [namespace upvar]. [Patch 1275435]
* generic/tclInt.h:
* generic/tclNamesp.c:
* generic/tclVar.c:
* tests/namespace.test:
* tests/upvar.test:
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
add CFLAGS to SHLIB_LD to support passing -isysroot in env(CFLAGS) to
configure (flag can't be present twice, so can't be in both CFLAGS and
LDFLAGS during configure), don't use -prebind when deploying on 10.4,
define TCL_IO_TRACK_OS_FOR_DRIVER_WITH_BAD_BLOCKING (rdar://3171542).
(SC_ENABLE_LANGINFO, SC_TIME_HANDLER): add/fix caching, fix obsolete
autoconf macros. Sync with tk/unix/tcl.m4.
* unix/configure.in: fix obsolete autoconf macros, sync gratuitous
formatting/ordering differences with tk/unix/configure.in.
* unix/Makefile.in: add CFLAGS to tclsh/tcltest link to make executable
linking the same as during configure (needed to avoid loosing any linker
relevant flags in CFLAGS, in particular flags that can't be in LDFLAGS).
Avoid concurrent linking of tclsh and compiling of tclTestInit.o or
xtTestInit.o during parallel make.
(checkstubs, checkdoc, checkexports): dependency and Darwin fixes
(dist): add new macosx files.
* unix/tclLoadDyld.c (TclpDlopen): use NSADDIMAGE_OPTION_WITH_SEARCHING
on second NSAddImage only. [Bug 1204237]
(TclGuessPackageName): should not be MODULE_SCOPE.
(TclpLoadMemory): ppc64 and endian (i386) fixes, add support for loading
universal (fat) bundles from memory.
* unix/tclUnixFCmd.c:
* macosx/tclMacOSXFCmd.c: ppc64 and endian (i386) fixes.
(TclMacOSXCopyFileAttributes): add support for new Tiger copyfile() API
to enable copying of xattrs & ACLs by [file copy].
* generic/tcl.h: add Darwin specifc configure overrides for TCL_WIDE
defines to support fat compiles of ppc and ppc64 at the same time,
(replaces Darwin CVS fix by emoy, rdar://3693001).
add/correct location of version numbers in macosx files.
* generic/tclInt.h: clarify fat compile comment.
* unix/tclUnixPort.h: add Darwin specifc configure overrides to support
fat compiles, where configure runs only once for multiple architectures
(replaces Darwin CVS fix by emoy, rdar://3693001).
* macosx/tclMacOSXBundle.c:
* macosx/tclMacOSXNotify.c:
* unix/tclUnixNotfy.c:
* unix/tclUnixPort.h: fix #include order to support compile time
override of HAVE_COREFOUNDATION in tclUnixPort.h when building for ppc64
* macosx/Tcl.pbproj/default.pbxuser (new file):
* macosx/Tcl.pbproj/jingham.pbxuser:
* macosx/Tcl.pbproj/project.pbxproj:
* macosx/Tcl.xcode/default.pbxuser (new file):
* macosx/Tcl.xcode/project.pbxproj (new file):
* macosx/Tcl.xcodeproj/default.pbxuser (new file):
* macosx/Tcl.xcodeproj/project.pbxproj (new file): new/updated projects
for Xcode 2.2 on 10.4, Xcode 1.5 on 10.3 & ProjectBuilder on 10.2, with
native tcltest targets and support for universal (fat) compiles.
* macosx/README: clarification/cleanup, document new Xcode projects and
universal (fat) builds via CFLAGS (i.e. all of ppc ppc64 i386 at once).
* unix/Makefile.in:
* unix/aclocal.m4:
* unix/configure.in:
* unix/dltest/Makefile.in:
* macosx/configure.ac (new file): add support for inclusion of
unix/configure.in by macosx/configure.ac, allows generation of a config
headers enabled configure script in macosx (required by Xcode projects).
* macosx/GNUmakefile: rename from Makefile to avoid overwriting by
configure run in tcl/macosx, add support for reusing configure cache,
build target fixes, remove GENERIC_FLAGS override now handled by tcl.m4.
* generic/tcl.decls: add Tcl_Main declaration as comment to avoid
'checkstubs' target complaining about it missing from stubs.
* generic/regex.h:
* generic/tclDate.c:
* generic/tclEnv.c:
* generic/tclGetDate.y:
* generic/tclIOUtil.c:
* generic/tclObj.c:
* generic/tclStubInit.c:
* generic/tclStubLib.c:
* generic/tclPathObj.c:
* generic/tclThreadAlloc.c:
* generic/tclThreadStorage.c:
* generic/tclTrace.c:
* generic/tclVar.c:
* generic/tommath.h:
* tools/fix_tommath_h.tcl:
* unix/tclUnixFCmd.c: ensure externally visible symbols not contained in
stubs table are declared as MODULE_SCOPE (or as static if not used
outside of own source file). These changes allow 'make checkstubs' to
complete without error on Darwin with gcc 4.
* generic/rege_dfa.c (getvacant):
* generic/regexec.c (cfind):
* generic/tclCompExpr.c (CompileSubExpr):
* generic/tclNamesp.c (NamespaceEnsembleCmd):
* unix/tclUnixChan.c (TclUnixWaitForFile): initialise variables to
silence gcc 4 warnings.
* generic/tclExecute.c (TclExecuteByteCode): fix unused variable warning
when NO_WIDE_TYPE is defined.
* generic/regguts.h: only #define NDEBUG if not already #defined.
* unix/tclUnixNotfy.c:
* macosx/tclMacOSXNotify.c: sync whitespace & comments.
* unix/tclUnixPort.h:
* win/tclWinPort.h: remove declaration of obsolete&unused TclpMutex API.
* unix/configure:
* unix/tclConfig.h.in: regen.
|
| | |
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
* generic/tclCmdAH.c:
* generic/tclCmdIL.c:
* generic/tclCmdMZ.c:
* generic/tclDictObj.c:
* generic/tclExecute.c:
* generic/tclIOCmd.c:
* generic/tclLink.c:
* generic/tclTest.c:
* generic/tclVar.c: fix for [Bug 1334947]. The functions
TclPtrSetVar, Tcl_ObjSetVar2 and Tcl_SetVar2Ex now always consume
the newValuePtr argument - i.e., they will free a 0-refCount
object if they failed to set the variable. Fixed all callers in
the core.
|
| |
| |
| |
| |
| |
| |
| |
| | |
* generic/tclNamesp.c:
* generic/tclVar.c:
* tests/trace.test: fix for [Bugs 1338280/1337229]; changed to use
the same approach as the 8.4 patch in the ticket (i.e., removed
the patch committed on 2005-31-10).
|
| | |
|
| | |
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
* generic/tclCmdIL.c:
* generic/tclCompCmds.c:
* generic/tclDictObj.c:
* generic/tclExecute.c:
* generic/tclLiteral.c:
* generic/tclParseExpr.c:
* generic/tclScan.c:
* generic/tclUtil.c:
* generic/tclVar.c:
|
| |
| |
| |
| |
| |
| | |
[kennykb-numerics-branch] Resynchronized with the HEAD; at this
checkpoint [-rkennykb-numerics-branch-20051008], the HEAD and
kennykb-numerics-branch contain identical code.
|
| | |
|
| | |
|
| |
| |
| |
| |
| |
| |
| | |
* generic/tclIndexObj.c: "enembleCmd", "localVarName", and
* generic/tclNamesp.c: "levelReference" to file static scope.
* generic/tclProc.c:
* generic/tclVar.c:
|
| | |
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
* generic/tclInt.h:
* generic/tclUtil.c:
* generic/tclIntDecls.h: `make genstubs`
* generic/tclStubInit.c:
* generic/tclBasic.c: Added callers of TclMatchIsTrivial where
* generic/tclCmdIL.c: a search can be done more efficiently
* generic/tclCompCmds.c:when it is recognized that a pattern match
* generic/tclDictObj.c: is really an exact match. [Patch 1076088]
* generic/tclIO.c:
* generic/tclNamesp.c:
* generic/tclVar.c:
|
| |
| |
| |
| | |
[Bug 1084111]
|
| | |
|
| |
| |
| |
| |
| |
| |
| |
| | |
calls at list creation (from 2 to 1), (b) reduce the cost of handling empty
lists (we now never create a list internal rep for them), (c) allow
refcounting of the list internal rep. The latter permits insuring that the
pointers returned by Tcl_ListObjGetElements remain valid even if the object
shimmers away from its original list type. This is [Patch 1158008]
|
| |
| |
| |
| | |
to flags. Should be perfectly backwards compatible.
|
| |
| |
| |
| | |
1084111] - thanks to Rolf Ade.
|
| |
| |
| |
| |
| | |
* generic/tclExecute.c: multiple .c files into one common header where
* generic/tclVar.c: it is verifiably after tclConfig.h inclusion.
|
| |
| |
| |
| |
| |
| | |
relevant to users. [Patch 1056864]
Also patches to core to take advantage of this
Also other general cleaning up of Tcl_WrongNumArgs usage
|
| |
| |
| |
| | |
macros.
|
| |
| |
| |
| |
| |
| | |
causing a leak of Proc structures and failure of compile-12.1. Two
lines were 'zombies' from the previous way localVarNames
worked. Credit dgp for finding this.
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
* generic/tclBinary.c:
* generic/tclCmdAH.c:
* generic/tclCmdIL.c:
* generic/tclCmdMZ.c:
* generic/tclCompExpr.c:
* generic/tclDictObj.c:
* generic/tclEncoding.c:
* generic/tclExecute.c:
* generic/tclFCmd.c:
* generic/tclHistory.c:
* generic/tclIndexObj.c:
* generic/tclInterp.c:
* generic/tclIO.c:
* generic/tclIOCmd.c:
* generic/tclNamesp.c:
* generic/tclObj.c:
* generic/tclPkg.c:
* generic/tclResult.c:
* generic/tclScan.c:
* generic/tclTimer.c:
* generic/tclTrace.c:
* generic/tclUtil.c:
* generic/tclVar.c:
It is a poor practice to directly set or append to the value
of the objResult of an interp, because that value might be
shared, and in that circumstance a Tcl_Panic() will be the
result. Searched for example of this practice and replaced
with safer alternatives, often using the Tcl_AppendResult()
routine that dkf just rehabilitated.
|
| | |
|
| |
| |
| |
| | |
as strings every time through. [Patch 1037357]
|