| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
|
| |
* generic/tclIOGT.c: Channel version fixed, must be 3, to have
wideseekProc. Thanks to David Graveraux <davygrvy@pobox.com>.
|
| |
|
|
|
|
|
|
|
|
|
| |
longstanding inconsistency in the treatment of the TCL_NAMESPACE_ONLY
flag revealed by testing the 2004-09-09 commits against Itcl.
TCL_NAMESPACE_ONLY now acts as specified in the pre-function
comment, forcing resolution in the passed in context namespace.
It has been incorrectly forcing resolution in the interp's current
namespace.
|
|
|
|
|
|
| |
optimisation for concatting an empty string. This enables
replacing the idiom 'K $x [set x {}]' by '$x[set x {}]' for
fastest execution.
|
|
|
|
| |
[Bug 1017299]
|
|
|
|
|
| |
* tests/namespace.test: logic that relied exclusively on string
matching and failed in the presence of [rename]s. [Bug 560297]
|
|
|
|
|
| |
field for storing the string-based command procedure of built-in
commands. We no longer have any string-based built-in commands!
|
|
|
|
|
|
| |
the month was scanned incorrectly in -timezone :localtime.
* tests/clock.test (clock-40.1): Added regression test case for
the bug where month was scanned incorrectly in -timezone :localtime.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
attempted to store the assoc data in the client data; the
optimisation caused a bug that [after] would overwrite
its imports. [Bug 1016167]
* library/clock.tcl (InitTZData, ClearCaches): Changed so that the
in-memory time zone :UTC (and its aliases) always gets
reinitialised, in case tzdata is absent. [Bug 1019537, 1023779]
* library/tzdata/*: Regenerated.
* tests/clock.test (clock-31.*, clock-39.1): Corrected a problem
where the 'system' locale tests fail on a non-English Windows
machine. [Bug 1023761]. Added a test to make sure that alias
time zones load correctly. [Bug 1023779].
* tools/tclZIC.tcl (writeLinks): Corrected a problem where
alias time zone names were written incorrectly, causing them
to fail to load at run time. [Bug 1023779].
* win/tclWinTime.c (Tcl_GetTime): Eliminated CPUID tests on
Win64 - assuming that HAL vendors now do a better job of
keeping the performance counters synchronized among CPU's.
[Bug 1020445]
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
|
|
|
|
|
| |
* generic/tclTimer.c: Workaround for situation when a
[namespace import] causes the objv[0] value to be something
other than what Tcl_AfterObjCmd expects. [Bug 1016167].
|
|
|
|
|
| |
disrupting things horribly.
Problem found by Don Porter when investigating [Bug 1016167].
|
|
|
|
|
|
|
|
|
|
|
|
| |
specific
* generic/tclTestProcBodyObj.c: instance of the general service already provided
by TclObjInvokeProc. Removed TclProcInterpProc and TclGetInterpProc from the
code...
* generic/tclInt.decls ...and from the internal stubs table.
* generic/tclIntDecls.h
* generic/tclStubInit.c
|
| |
|
|
|
|
| |
1011860]
|
| |
|
| |
|
| |
|
| |
|
| |
|
|
|
|
|
|
|
|
|
|
| |
* doc/interp.n: Added support for a -namespace option to the
* generic/tclBasic.c: [interp invokehidden] command. Also added an
* generic/tclInt.h: internal routine TclObjInvokeNamespace() and
* generic/tclInterp.c: corrected the flag names TCL_FIND_ONLY_NS and
* generic/tclNamesp.c: TCL_CREATE_NS_IF_UNKNOWN that are passed to the
* generic/tclTrace.c: internal routine TclGetNamespaceForQualName().
* tests/interp.test: [Patch 981841]
|
|
|
|
|
|
| |
* tests/compile.test (compile-12.4): flawed deletion of literal
internal reps that could lead to accessing of freed memory.
Thanks to Kevin Kenny for test case and fix [Bug 1001997].
|
|
|
|
|
| |
so that Tcl_ExitProc's that call Tcl_Finalize recursively do not
cause deadlock. [Patch 999084 fixes Tk Bug 714956]
|
|
|
|
|
|
| |
reference to the corresponding proc. The reference is now seen as
unnecessary, and it may cause leaking circular references under
some circumstances (see for example [Bug 994838]).
|
|
|
|
|
|
|
| |
* generic/tclInt.h: fully cleanup on exit and allow for
* generic/tclThreadAlloc.c: reinitialization. [Bug #736426]
* unix/tclUnixThrd.c: (mistachkin, kenny)
* win/tclWinThrd.c:
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
* generic/tclBasic.c (DeleteInterpProc):
* generic/tclLiteral.c (TclCleanupLiteralTable):
* generic/tclInt.h: added a TclCleanupLiteralTable function,
called from DeleteInterpProc, that frees internal representations
of shared literals early when an interpreter is being deleted.
This change corrects a number of memory mismanagement issues in
the cases where the internal representation of one literal
contains a reference to another, and avoids conditions such as
resolved variable names referring to procedure and namespace
contexts that no longer exist. [Bug 994838]
|
| |
|
|
|
|
|
| |
generation of error messages and simplified by reusing data in a
variable instead of retrieving the string again. Fixes [Tcl SF Bug 835289].
|
|
|
|
|
| |
the TclFinalizeNotifier in order to correct broken ref-counting
of the notifier thread.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
* generic/tclIO.c (Tcl_UnregisterChannel): 'Tcl_Close' while the
* generic/tclIO.c (Tcl_Close): close callbacks are
run. Checked in 'Tcl_Close' and 'Tcl_Unregister' to prevent
recursive call of 'close' in the close-callbacks. This is a
possible error made by implementors of virtual filesystems based
on 'tclvfs', thinking that they have to close the channel in the
close handler for the filesystem.
* generic/tclIO.c:
* generic/tclIO.h:
* Not reverting, but #ifdef'ing the changes from May 19, 2004 out
of the core. This removes the ***POTENTIAL INCOMPATIBILITY***
for channel drivers it introduced. This has become possible due
to Expect gaining a BlockModeProc and now handling blockingg and
non-blocking modes correctly. Thus [SF Tcl Bug 943274] is still
fixed if a recent enough version of Expect is used.
* doc/CrtChannel.3: Added warning about usage of a channel without
a BlockModeProc.
|
|
|
|
|
|
| |
old depreceated newline syntax, to ensure that only "nonewline"
is accepted. [Tcl SF Bug 985869], reported by Joe Mistachkin
<mistachkin@users.sourceforge.net>.
|
|
|
|
|
|
|
|
|
|
|
|
| |
after the last call to TclFinalizeThreadData (done from within
Tcl_FinalizeThread()). We basically just repeat the
TclFinalizeThreadData() once more before tearing down TSD keys
in TclFinalizeSynchronization().
There should be more elaborate mechanism in place for handling
such issues, based on thread cleanup handlers registered on the
OS level. Such change requires much more work and would also
require TIP because some visible parts of Tcl API would have to
be modified. In the meantime, this will do.
|
|
|
|
| |
variable 'codePtr' to silence a message from VC++.
|
|
|
|
|
|
|
|
|
| |
* generic/tclLiteral.c (TclReleaseLiteral): fix for [Bug 467523],
which resurfaced with the latest changes. The previous strategy
was to have special code in TclReleaseLiteral to handle the
self-references generated by empty scripts. The new approach
avoids the self-reference altogether, by having empty scripts
return an unshared literal.
|
|
|
|
| |
this will result in TclInitNotifier never being called.
|
| |
|
|
|
|
| |
See SF bugreport for more info.
|
|
|
|
| |
objv[objectsUsed] on error.
|
|
|
|
| |
result on error.
|
| |
|
| |
|
|
|
|
|
| |
when there is an error while an expansion is in progress (code
added at checkForCatch).
|