Gregory P. Smith [Tue, 14 Dec 2010 14:38:00 +0000 (14:38 +0000)]
Issue #1731717: Fixed the problem where subprocess.wait() could cause an
OSError exception when The OS had been told to ignore SIGCLD in our process
or otherwise not wait for exiting child processes.
R. David Murray [Tue, 14 Dec 2010 14:16:20 +0000 (14:16 +0000)]
#10695: use %s not %d so that a string 'port' does not cause a debug traceback
Passing the port as a string value works fine in regular mode, but
if you turned debug on it would throw an error trying to print the
port number, which is surprising and confusing.
Gregory P. Smith [Tue, 14 Dec 2010 13:43:30 +0000 (13:43 +0000)]
Issue #6559: fix the subprocess.Popen pass_fds implementation. Add a unittest.
Issue #7213: Change the close_fds default on Windows to better match the new
default on POSIX. True when possible (False if stdin/stdout/stderr are
supplied).
Update the documentation to reflect all of the above.
Barry Warsaw [Mon, 13 Dec 2010 18:04:23 +0000 (18:04 +0000)]
Issue 10687. When --without-pymalloc is given, $VERSION is the same as
$LDVERSION, which screws up the sym/hard-links. This avoids those games when
$VERSION == $LDVERSION.
Also, include a drive-by fix for an obvious syntax error.
Gregory P. Smith [Mon, 13 Dec 2010 07:59:39 +0000 (07:59 +0000)]
issue7213: Open the pipes used by subprocesses with the FD_CLOEXEC flag from
the C code, using pipe2() when available. Adds unittests for close_fds and
cloexec behaviors.
Gregory P. Smith [Mon, 13 Dec 2010 06:45:02 +0000 (06:45 +0000)]
Get rid of the close_fds DeprecationWarning. Changes the default on a per
platform basis. It remains False on Windows and changes to True on all
other platforms (POSIX). Based on python-dev discussion and
http://bugs.python.org/issue7213.
issue 10683
When the solution is converted to Visual Studio 2010, the command line to invoke make_buildinfo changes from:
$(SolutionDir)make_buildinfo.exe" Debug "$(IntDir)\"
to
$(SolutionDir)make_buildinfo.exe" Debug "$(IntDir)"
If the final backslash is omitted, the backslash in IntDir will escape the quote, thus passing the quote in as part of the path name.
This solution is a hack-fix to that problem by skipping any trailing quote from the path name. It works as long as we don't need any additional arguments to make_buildinfo.exe. This will help all those sould that are going to run this project through the visual studio autoconverter and get the same error.
R. David Murray [Sun, 12 Dec 2010 20:06:19 +0000 (20:06 +0000)]
#243654: only create a new MIME boundary if we don't already have one.
The rearranged code should do exactly what the old code did, but
the new code avoids a potentially costly re computation in the case
where a boundary already exists.
Barry Warsaw [Sat, 11 Dec 2010 21:32:01 +0000 (21:32 +0000)]
Create the hardlink between python-3.2m and python-3.2 in altbininstall target
instead of bininstall target so it shows up when you do 'make altinstall'.
Closes issue 10677.
Ronald Oussoren [Tue, 7 Dec 2010 14:41:05 +0000 (14:41 +0000)]
Two small changes to adjust framework builds to the new stable ABI
Both the Makefile and the script that is used on OSX to create the binary
installer refer to the directory containing the Makefile using the name
'config'. This name was changed with the new ABI (with default build flags
it is now named config-3.2m). This patch ensures that both files use the
correct name.
The build-installer.py script contains one other change: it now tests for the
Tcl/Tk framework version by looking at the 'Current' symlink in the framework
instead of runnning a script. This makes it possible to verify the version
that is in the SDK that's used during the build instead of the version that
is installed on the system.