| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
|
| |
* tests/winDde.test: for unique dde server names. Added
* win/tclWinDde.c: some more tests. Fixes [Bug 219293]
|
|
|
|
| |
r1.44 as 2003-04-11 optimizations broke Windows98 glob'ing.
|
| |
|
|
|
|
| |
closing a Windows serial port [Bug #718002] (schroedter)
|
| |
|
| |
|
| |
|
|
|
|
|
|
|
|
|
|
| |
the code to catch any errors returned by the windows functions
handling TLS ASAP instead of waiting to get some mysterious
crash later on due to bogus pointers. Patch provided by Joe
Mistachkin.
This is a stop-gap measure to deal with the low number of ?TLS
slots provided by some of the variants of Windows (60-80).
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
* generic/tclIO.c (Tcl_CutChannel, Tcl_SpliceChannel):
Invoke TclpCutSockChannel and TclpSpliceSockChannel.
* generic/tclInt.h: Declare TclpCutSockChannel and
TclpSpliceSockChannel.
* unix/tclUnixSock.c (TclpCutSockChannel, TclpSpliceSockChannel):
Dummy functions, on unix the sockets are _not_ handled
specially.
* mac/tclMacSock.c (TclpCutSockChannel, TclpSpliceSockChannel):
* win/tclWinSock.c (TclpCutSockChannel, TclpSpliceSockChannel):
New functions to handle socket specific cut/splice operations:
auto-initi of socket system for thread on splice, management of
the module internal per-thread list of sockets, management of
association of sockets with HWNDs for event notification.
* win/tclWinSock.c (NewSocketInfo): Extended initialization
assignments to cover all items of the structure. During
debugging of the new code mentioned above I found that two
fileds could contain bogus data.
* win/tclWinFile.c: Added #undef HAVE_NO_FINDEX_ENUMS before
definition because when compiling in debug mode the compiler
complains about a redefinition, and this warning is also treated
as an error.
|
|
|
|
|
| |
* win/tclWinInit.c: recognize Windows CE as a Win platform.
This just recognizes CE - full support will come later.
|
|
|
|
|
| |
* win/configure.in (SHELL): force it to /bin/sh as autoconf 2.5x
uses /bin/bash, which can fail to find exes in the path (ie: lib).
|
| |
|
|
|
|
|
| |
and tclWinPipe.c on 2003-04-14. Now passes TCL_PIPE_DLL in place of
TCL_DBGX.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
symbol for every compile. Instead, define
TCL_PIPE_DLL only when compiling tclWinPipe.c.
This will break other build systems, so
they will need to remove the TCL_DBGX define
and replace it with a define for TCL_PIPE_DLL.
* win/tclWinPipe.c (TclpCreateProcess):
Remove PREFIX_IDENT and DEBUG_IDENT from
top of file. Use TCL_PIPE_DLL passed in
from build env instead of trying to construct
the dll name from already defined symbols.
This approach is more flexible and better
in the long run.
|
|
|
|
| |
was broken by recent changes.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
* win/configure.in: Add check for FINDEX_INFO_LEVELS
from winbase.h, known to be a problem in VC++ 5.2.
Define HAVE_NO_FINDEX_ENUMS if the define does not
exist.
* win/tclWinFile.c: Put declarations for
FINDEX_INFO_LEVELS and FINDEX_SEARCH_OPS inside
a check for HAVE_NO_FINDEX_ENUMS so that these are
not declared twice. This fixes the Mingw build.
* win/tclWinTime.c: Rework the init of timeInfo
so that the number or initializers matches the
declaration. This was broken under Mingw. Add
cast to avoid compile warning when calling the
AccumulateSample function.
|
| |
|
|
|
|
|
|
| |
return values). Fixed Bug 710310 (duplicate test numbers in
clock.test). Made major changes to tclWinTime.c and related code
to improve loop filter stability.
|
| |
|
|
|
|
| |
declaration and stubs entries.)
|
|
|
|
|
|
|
|
|
|
|
| |
* win/configure.in: Set stub lib flag based
on new LIBFLAGSUFFIX variable.
* win/tcl.m4 (SC_CONFIG_CFLAGS): Set new
LIBFLAGSUFFIX that works like LIBSUFFIX,
it is used when creating library names.
The previous implementation would generate
-ltclstub85 instead of -ltclstub85s when
configured with --disable-shared.
|
|
|
|
|
| |
Mistachkin insists that he has a configuration that fails to launch
tcltest without it, and it appears harmless otherwise.
|
|
|
|
|
|
|
|
|
| |
such that giving only a topic name really returns all services
with that topic. [Bug 219155]
Prevent hangup caused by dde server applications failing to process
messages [Bug 707822]
* tests/winDde.test: Corrected labels and added a test for search
by topic name.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
* generic/tclStubInit.c (tclOriginalNotifier):
* mac/tclMacNotify.c (Tcl_SetTimer,Tcl_WaitForEvent):
* unix/tclUnixNotfy.c (Tcl_SetTimer,Tcl_WaitForEvent,
Tcl_CreateFileHandler,Tcl_DeleteFileHandler):
* win/tclWinNotify.c (Tcl_SetTimer,Tcl_WaitForEvent): Some linkers
apparently use a different representation for a pointer to a function
within the same compilation unit and a pointer to a function in a
different compilation unit. This causes checks like those in the
original notifier procedures to fall into infinite loops. The fix
is to store pointers to the original notifier procedures in a struct
defined in the same compilation unit as the stubs tables, and compare
against those values. [Bug 707174]
|
|
|
|
|
|
|
|
|
| |
* win/configure.in: Don't run the AC_CYGWIN
macro since it uses AC_CANONICAL_HOST under
autoconf 2.5X. Just check to see if __CYGWIN__
is defined by the compiler and set the
ac_cv_cygwin variable based on that.
[Bug 705912]
|
|
|
|
| |
fix that was made in tk instead of tcl sources.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
for a detailed explanation of why this is good.
This will no doubt break the build on some
platforms, let the flaming begin.
* tools/configure: Regen with autoconf 2.57.
* tools/configure.in: Require autoconf 2.57.
* unix/configure: Regen with autoconf 2.57.
* unix/configure.in: Require autoconf 2.57.
Apply AC_LIBOBJ changes from patch 529884.
* unix/tcl.m4: Ditto.
* win/configure: Regen with autoconf 2.57.
* win/configure.in: Require autoconf 2.57.
Don't subst LIBOBJS since this happens by
default, this avoids an autoconf error.
|
|
|
|
|
| |
* win/tclWinPipe.c: from tcl.h -- they are not part of Tcl's
public interface. Put them in win/tclWinPipe.c where they are used.
|
|
|
|
| |
panicking on Win9x.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
* generic/tcl.h: Tcl to 8.5a0.
* library.init.tcl:
* mac/README:
* macosx/Tcl.pbproj/project.pbxproc:
* tests/basic.test:
* tools/configure.in:
* tools/tcl.hpj.in:
* tools/tcl.wse.in:
* unix/configure.in:
* unix/tcl.spec:
* win/README:
* win/README.binary:
* win/configure.in:
* win/makefile.bc:
* win/makefile.vc:
* win/tcl.m4:
* tools/configure: autoconf
* unix/configure:
* win/configure:
|
|
|
|
| |
d to g in rules.vc.
|
|
|
|
|
|
|
|
|
| |
* win/tclWinDde.c: bundled release with Tcl 8.4.2
* library/reg/pkgIndex.tcl: registry bumped to version 1.1.1 for
* win/tclWinReg.c: bundled release with Tcl 8.4.2
* library/opt/pkgIndex.tcl: updated package index to version 0.4.4
|
|
|
|
|
|
| |
* win/configure.in: check for 'g' for debug build type, not 'd'.
* win/rules.vc (DBGX): correct to use 'g' for nmake win makefile
to match the cygwin makefile for debug builds. [Bug #635107]
|
|
|
|
| |
calling GetUserName
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
* generic/tcl.h:
* tools/tcl.wse.in:
* unix/configure:
* unix/configure.in:
* unix/tcl.m4:
* unix/tcl.spec:
* win/README.binary:
* win/configure:
* win/configure.in:
* macosx/Tcl.pbproj/project.pbxproj:
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
leaps of the performance counter. See the MSDN Knowledge Base article
Q274323 for the hardware problem that makes this necessary on certain
machines.
* tests/winTime.test: Revised winTime-2.1 - it had a tolerance of
thousands of seconds, rather than milliseconds. (What's six orders of
magnitude among friends?
Both the above changes are triggered by a problem reported at
http://aspn.activestate.com/ASPN/Mail/Message/ActiveTcl/1536811
although the developers find it difficult to believe that it accounts
for the observed behavior and suspect a fault in the RTC chip.
|
|
|
|
| |
that it works with non-ASCII7 user names. [Bug 685926]
|
|
|
|
|
|
|
|
|
|
|
|
| |
* win/configure.in: Generate error when attempting
to build under Cygwin. The Cygwin port of Tcl/Tk
does not build and people are filing bug reports
under the mistaken impression that someone is
actually maintaining the Cygwin port. A post to
comp.lang.tcl asking someone to volunteer as an
area maintainer has generated no results.
Closing bugs 680840, 630199, and 634772 and
marking as "Won't fix".
|
| |
|
| |
|
| |
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
TCL_DEBUG_IDENT, used only by TclpCreateProcess.
* unix/Makefile.in: Define TCL_DBGX.
* win/Makefile.in: Define TCL_DBGX.
* win/tclWinPipe.c (TclpCreateProcess):
Check that the Tcl pipe dll actually exists
in the Tcl bin directory and panic if it
is not found. Incorporate TCL_DBGX into
the Tcl pipe dll name. This fixes a really
mysterious error that would show up when
exec'ing a 16 bit application under Win95
or Win98 when Tcl was compiled with symbols.
The error seemed to indicate that the executable
could not be found, but it was actually the
Tcl pipe dll that could not be found.
|
|
|
|
| |
This version bundles gcc 3.
|
|
|
|
|
|
|
|
| |
* win/configure.in: Add test that checks to
see if the compiler can cast to a union type.
* win/tclWinTime.c: Squelch compiler warning
about union initializer by casting to union
type when compiling with gcc.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Invoke TclpCutFileChannel and TclpSpliceFileChannel.
* generic/tclInt.h: Declare TclpCutFileChannel
and TclpSpliceFileChannel.
* unix/tclUnixChan.c (FileCloseProc, TclpOpenFileChannel,
Tcl_MakeFileChannel, TclpCutFileChannel,
TclpSpliceFileChannel): Implement thread load data
cut and splice for file channels. This avoids
an invalid memory ref when compiled with -DDEPRECATED.
* win/tclWinChan.c (FileCloseProc, TclpCutFileChannel,
TclpSpliceFileChannel): Implement thread load data
cut and splice for file channels. This avoids
an invalid memory ref that was showing up in the
thread extension.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
* win/tclWinChan.c (Tcl_MakeFileChannel, squelch_warnings):
* win/tclWinFCmd.c (DoRenameFile, DoCopyFile, squelch_warnings):
Re-implement inline ASM SEH handlers for gcc.
The esp and ebp registers are now saved on the
stack instead of in global variables so that
the code is thread safe. Add additional checks
when TCL_MEM_DEBUG is defined to be sure the
values were recovered from the stack properly.
Remove squelch_warnings functions and add
a dummy call in the handler methods to squelch
compiler warnings.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
* win/configure.in: Define HAVE_ALLOCA_GCC_INLINE
when we detect that no alloca function is found
in malloc.h and we are compiling with GCC.
Remove HAVE_NO_ALLOC_DECL define.
* win/tclWin32Dll.c (TclpCheckStackSpace):
Don't define alloca as a cdecl function.
Doing this caused a tricky runtime bug because
the _alloca function expects the size argument
to be passed in a register and not on the stack.
To fix this problem, we use inline ASM when
compiling with gcc to invoke _alloca with the
size argument loaded into a register.
|