| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
|
|
|
|
|
| |
generated ifneeded scripts of Tcl Modules, for early detection of
conflicts between the version specified through the file name and
a 'provide' command in the module implementation, if any. Note
that this change also now allows Tcl Modules to not provide a
'provide' command at all, and declaring their version only through
their filename.
|
|
|
|
|
|
| |
* tests/proc.test: by procbody::test::proc. See [Bug 2043636].
Added a test case demonstrating the leak before the fix. Fixed a
few spelling errors in test descriptions as well.
|
|
|
|
|
|
|
|
| |
* library/http/pkgIndex.tcl: for [Bug 2046486] bug fix. This
* unix/Makefile.in: release of http now requires a
* win/Makefile.in: dependency on Tcl 8.5 to be able to
* win/makefile.bc: use the unsigned formats in the
* win/makefile.vc: [binary scan] command.
|
|
|
|
| |
[Bug 2046846]
|
|
|
|
|
|
| |
a fix for [Bug 1482718]. This is not needed at least since
varReform, where the local variable data at runtime is read from
the CallFrame and/or the LocalCache.
|
|
|
|
|
| |
* generic/tclCompile.h:
* generic/tclExecute.c:
|
|
|
|
|
|
|
| |
* generic/tclOOMethod.c (PushMethodCallFrame): fix uninitialized efi
name field.
* tests/lrange.test (lrange-1.17): add test cleanup; whitespace.
|
| |
|
| |
|
| |
|
|
|
|
|
| |
* generic/tclExecute.c: Tcl_EvalEx [Bug 2017946]
* generic/tclInt.h:
|
|
|
|
|
|
|
|
|
|
| |
being deleted during handling of object command deletion. The
old code was relying on documented features of command delete
traces that do not in fact work. [Bug 2039178].
* tests/oo.test (oo-26.*): Added tests that demonstrate
failure of TclOO to check for various kinds of invalid bytecode
during method dispatch. [Bug 2037727].
|
| |
|
|
|
|
|
|
| |
could not trigger before TclOO: the number of locals was being
read from the Proc, which can under some circumstance be out of
sync with the localCache's.
|
|
|
|
|
|
|
|
|
| |
hack that was maintained for Itcl
*** POTENTIAL INCOMPATIBILITY *** for Itcl
Itcl users will need a new release with Itcl's [Patch 2040295], or
else load the tiny script in that patch by themselves (rewrite
::unknown). Note that it is a script-only patch.
|
|
|
|
|
|
|
| |
Consolidate error reporting paths in CreateSocket();
Restore blocking mode immediately after connect()
FD blocking mode is now consistent with [fconfigure ... -blocking]
setting while async connect is in progress.
|
| |
|
|
|
|
|
| |
for [foreach] has been added and marked as knownbug, awaiting for
it to be NR-enabled.
|
|
|
|
|
|
|
| |
* generic/tclCompile.h: inconditionally, streamlined
* generic/tclExecute.c: atProcExit/tailcall processing
* generic/tclProc.c: in TEBC.
* tests/unsupported.test:
|
|
|
|
|
|
| |
* tests/execute.test: stack trace when a compile epoch bump triggers
fallback to direct evaluation of commands in a compiled script.
[Bug 2037338]
|
| |
|
| |
|
|
|
|
|
|
|
|
| |
his thing in macosx/Tcl.xcodeproj/project.pbxproj
* tests/nre.test: migrated tests to standard locations,
* tests/unsupported.test: separating core functionality from the
experimental commands. These are new files.
|
|
|
|
|
|
|
| |
* generic/tclEvent.c: on DLL_PROCESS_DETACH as it may lead
* win/tclWin32Dll.c (DllMain): to issues and the user should be
explicitly calling Tcl_Finalize before unloading regardless.
Clarify the docs to note the explicit need in embedded use.
|
|
|
|
|
|
|
|
| |
* tests/info.test: in TclEvalEx so that the CmdFrame will still
be on the stack at the time Tcl_LogCommandInfo is called to append
another level of -errorinfo information. Sets the stage to add
file and line data to the stack trace. Added test to check that
[info frame] functioning remains unchanged by the revision.
|
|
|
|
|
| |
tests that actually measure the C-stack depth. This makes them
bearable again (even under memdebug) and avoid crashing on failure.
|
| |
|
| |
|
| |
|
| |
|
|
|
|
|
|
| |
* generic/tclCmdAH.c:
* generic/tclInt.h:
* tests/NRE.test:
|
| |
|
|
|
|
|
|
| |
unrefernced everywhere but in macosx/Tcl.xcodeproj/project.pbxproj: some
knowledgeable maintainer please remove tclNRE.h after making sure it
doesn't break the build on macosx
|
|
|
|
|
|
|
| |
* generic/tclCompile.h:
* generic/tclExecute.c:
* generic/tclTest.c:
* tests/NRE.test:
|
|
|
|
| |
to clarify what is happening.
|
|
|
|
|
| |
changing between the times where TEOV and TEOV_exception
run. Thanks dgp for catching this.
|
| |
|
| |
|
|
|
|
|
|
|
|
|
| |
* generic/tclInt.h: stored in the ekeko at interp creation
* generic/tclNRE.h: to avoid hitting the TSD each time an
* generic/tclThreadAlloc.c: NRE callback is pushed or pulled; the
approach is suitably general to extend to evry other obj
allocation where an interp is know; this is left for some other
time, requires a lot of grunt work.
|
|
|
|
|
| |
TclStackRealloc to panic on rare corner cases. Thx ajpasadyn for
diagnose and patch.
|
| |
|
| |
|
|
|
|
|
|
|
|
|
|
|
| |
* generic/tcl.decls: which all are supposed to be a constant, but
* generic/tclDecls.h: this was not reflected in the API:
* generic/tclFileSystem.h: Tcl_FSGetInternalRep
* generic/tclIOUtil.c: Tcl_FSNewNativePath, Tcl_FSData
* generic/tclPathObj.c: Tcl_FSRegister, Tcl_FSUnregister
* generic/tclTest.c: Tcl_FSGetFileSystemForPath ...
This change complies with TIP #24.
***POTENTIAL INCOMPATIBILITY***
|
|
|
|
|
|
|
| |
empty string as path (TclEvalEx). In 8.4 the missing code caused
panics in the testsuite. It doesn't in 8.5. I am guessing that the
code path with the missing the incr-refcount is not invoked any
longer. Because the bug in itself is certainly the same.
|
| |
|
| |
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
* doc/ObjectType.3 Tcl_ObjType which all are supposed
* generic/tcl.decls to be a constant, but this was not
* generic/tcl.h reflected in the API:
* generic/tclDecls.h Tcl_ConvertToType
* generic/tclObj.c Tcl_GetObjType
* generic/tclCompCmds.c Tcl_RegisterObjType
* generic/tclOOMethod.c Introduced a CONST86_RETURN, so extensions which
* generic/tclTestobj.c use Tcl_ObjType directly can be modified to compile
against both Tcl 8.5 and Tcl 8.6
tclDecls.h is re-generated with "make genstubs"
This change complies with TIP #24
***POTENTIAL INCOMPATIBILITY***
|
|
|
|
| |
fixes bug [1605269].
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
of location for uplevel scripts. Resolved merge conflict on
info-37.0, switched !singleTestInterp constraint to glob matching
instead. Ditto info-22.8, removed constraint, more glob matching,
and reduced the depth of the stack we check. More is coming, right
now I want to commit the bug fixes.
* tests/oo.test: Updated oo-22.1 for expanded location tracking.
* generic/tclCompile.c (TclInitCompileEnv): Reorganized the
initialization of the #280 location information to match the flow
in TclEvalObjEx to get more absolute contexts.
* generic/tclBasic.c (TclEvalObjEx): Added missing cleanup of
extended location information.
|