| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
|
|
|
|
|
|
|
| |
* tools/tcl.wse.in: distinguish HEAD of CVS development
* unix/configure.in: from the recent 8.5a1 release.
* unix/tcl.spec:
* win/README.binary:
* win/configure.in:
* unix/configure: autoconf-2.57
* win/configure:
|
| |
|
|
|
|
|
|
|
|
|
|
| |
* generic/tclInt.h: but caused warnings related to
* generic/tclInt.decls: strict aliasing with GCC 3.3.
* generic/tclClock.c:
* generic/tclDate.c:
* generic/tclGetDate.y:
* win/tclWinTime.c:
* unix/tclUnixTime.c:
|
| |
|
|
|
|
|
|
|
|
| |
MsgWaitForMultipleObjectsEx for "a message is in the queue" wasn't
accurate. I removed the check on the case result==(WAIT_OBJECT_0 + 1)
This was having the error of falling into GetMessage and waiting
there by accident, which wasn't alertable through Tcl_AlertNotifier.
I'll do some more study on this and try to find-out why.
|
| |
|
| |
|
|
|
|
|
|
|
|
| |
* win/configure.in: Check for define of
MWMO_ALERTABLE in winuser.h.
* win/tclWinPort.h: If MWMO_ALERTABLE
is not defined in winuser.h then define it.
This is needed for Mingw.
|
|
|
|
|
| |
This provides a Windows test of the TIP #100 mechanism and
a sample to show how unloading an extension can be done.
|
| |
|
|
|
|
|
|
|
|
| |
"Asynchronous Procedure Calls" from its wait state. Only useful for extension
authors who decide they might want to try "completion routines" with
WriteFileEx(), as an example. From experience, I recommend that "completion
ports" should be used instead as the execution of the callbacks are more
managable.
|
|
|
|
| |
although vcvars32.bat sets both.
|
| |
|
|
|
|
|
| |
-Wconversion warning for GetGlobalAtomName. gcc is just wrong
about this.
|
|
|
|
|
|
|
| |
WaitForSingleObject returns a timeout. Tcl_Finalize called from
DllMain will pause all threads. Trust that the thread will get the
close notice at a later time if it does ever wake up before being
cleaned up by the system anyway.
|
|
|
|
|
|
|
|
| |
Use the tail component of the passed in
lib path instead of just blindly using
lib+4. That worked when lib was "lib/..."
but fails for other values. Thanks go to
Patrick Samson for pointing this out.
|
| |
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
* win/rules.vc:
* win/tcl.rc:
* win/tclsh.rc: Added an 'unchecked' option to the OPTS macro so a
core build with symbols can be linked to the non-debug enabled C
run-time. As per discussion with Kevin Kenny. Called like this:
nmake -af makefile.vc OPTS=unchecked,symbols
This clarifies the meaning of the 'g' naming suffix to mean only that
the binary requires the debug enabled C run-time. Whether the binary
contains symbols or not is a seperate condition.
|
| |
|
|
|
|
|
| |
* win/tclWinPipe.c: Fixed BuildCommandLine() to pass the new
cases.
|
|
|
|
|
|
|
|
| |
doing the "N backslashes followed a quote -> insert N * 2 + 1
backslashes then a quote" rule needed for the crt's
parse_cmdline().
* win/tclWinPipe.c: Fixed BuildCommandLine() to pass the new
cases.
|
|
|
|
|
| |
overflows and a couple envars upsetting invokations of cl.exe
and link.exe.
|
|
|
|
|
|
|
| |
* win/tclWinPipe.c (BuildCommandLine): Special case quoting for
'{' not required by the c-runtimes's _setargv().
* win/tclAppInit.c: Removed our custom setargv() in favor of
the one provided by the c-runtime. [Bug 672938]
|
|
|
|
|
|
| |
symbols. This is supposed to emulate the release build better to
avoid hiding problems that would only show themselves in a release
build.
|
| |
|
|
|
|
|
|
|
| |
'nul' so VC 5.2 doesn't try searching the path for it and failing with
a possible dialogbox popping up about having to add a CD to an
empty drive. Also added a SetErrorMode() call to disable any
dialogs. [Bug 885537]
|
| |
|
|
|
|
| |
variable names in
|
|
|
|
|
|
| |
process list. The delayed cut operation after the wait was going stale by
being outside the list lock. It now cuts within the lock and does a locked
splice for when it needs to instead. [Bug 859820]
|
|
|
|
| |
under VC++
|
| |
|
|
|
|
| |
-DTCL_USE_STATIC_PACKAGES is sent when compiling the shells
|
| |
|
| |
|
|
|
|
| |
and link when $(FULLWARNINGS) is set.
|
|
|
|
| |
macro; sets $(FULLWARNINGS).
|
|
|
|
|
|
|
|
| |
into the object file itself with #paragma comment (lib, ...)
when built with VC++. This will simplify linking for users of the
static library.
* win/makefile.vc: Removed 'advapi.lib' from $(baselibs).
|
|
|
|
|
| |
to see if there are waiters else uninitialized datum is manipulated.
[Bug 849007]
|
|
|
|
| |
nmake -af makefile.vc CHECKS=nodep
|
|
|
|
|
|
| |
blocks in place of __try and __except statements
to support gcc builds. This is needed after
David's changes on 2003-12-21. [Tcl patch 858493]
|
|
|
|
|
|
|
| |
to 'Tcl_Panic' (the function). The #define
of panic in tcl.h clearly states it is
deprecated in the comments.
[Patch 865264]
|
| |
|
| |
|
| |
|
|
|
|
|
| |
version strings. No more hard coding TCL_VERSION = 8.5 and having
to edit it when you swap cores.
|
|
|
|
|
|
|
|
| |
DllMain's DLL_PROCESS_DETACH. We can't 100% assured that Tcl is
being unloaded by the OS in a stable condition and we need to protect
the exit handlers should the stack be in a hosed state. AT&T style
assembly for SEH under MinGW has not been added yet. This is a first
part change for [Patch 858493]
|
| |
|
|
|
|
|
|
|
|
|
| |
with Borland.
* win/tclWinTime.c (TclpGetDate) : Borland's localtime() has
a slight behavioral difference.
From Helmut Giese <hgiese@ratiosoft.com> [Patch 758097].
|
|
|
|
|
| |
support Borland's rc tool. From Helmut Giese
<hgiese@ratiosoft.com>.
|