Ronald Oussoren [Thu, 6 Jul 2006 10:13:35 +0000 (10:13 +0000)]
This patch solves the problem Skip was seeing with zlib, this patch ensures that
configure uses similar compiler flags as setup.py when doing the zlib test.
Without this patch configure would use the first shared library on the linker
path, with this patch it uses the first shared or static library on that path
just like setup.py.
Thomas Heller [Thu, 6 Jul 2006 08:48:35 +0000 (08:48 +0000)]
Patch #1517790: It is now possible to use custom objects in the ctypes
foreign function argtypes sequence as long as they provide a
from_param method, no longer is it required that the object is a
ctypes type.
Thomas Heller [Thu, 6 Jul 2006 07:50:18 +0000 (07:50 +0000)]
Revert the change done in svn revision 47206:
Add a new function uses_seh() to the _ctypes extension module. This
will return True if Windows Structured Exception handling (SEH) is
used when calling functions, False otherwise.
Fred Drake [Thu, 6 Jul 2006 05:13:22 +0000 (05:13 +0000)]
- back out Expat change; the final fix to Expat will be different
- change the pyexpat wrapper to not be so sensitive to this detail of the
Expat implementation (the ex-crasher test still passes)
Thomas Wouters [Wed, 5 Jul 2006 11:03:49 +0000 (11:03 +0000)]
Fix bug in passing tuples to string.Template. All other values (with working
str() or repr()) would work, just not multi-value tuples. Probably not a
backport candidate, since it changes the behaviour of passing a
single-element tuple:
Ronald Oussoren [Tue, 4 Jul 2006 12:30:22 +0000 (12:30 +0000)]
Sync the darwin/x86 port libffi with the copy in PyObjC. This fixes a number
of bugs in that port. The most annoying ones were due to some subtle differences
between the document ABI and the actual implementation :-(
(there are no python unittests that fail without this patch, but without it
some of libffi's unittests fail).
Fix build problems with the platform SDK on windows. It is not sufficient to test for the C compiler version when determining if we have the secure CRT from microsoft. Must test with an undocumented macro, __STDC_SECURE_LIB__ too.
[Bug #1511911] Clarify description of optional arguments to sorted()
by improving the xref to the section on lists, and by
copying the explanations of the arguments (with a slight modification).
Thomas Heller [Mon, 3 Jul 2006 08:08:14 +0000 (08:08 +0000)]
Add a new function uses_seh() to the _ctypes extension module. This
will return True if Windows Structured Exception handling (SEH) is
used when calling functions, False otherwise.
Currently, only MSVC supports SEH.
Fix the test so that it doesn't crash when run with MingW compiled
_ctypes. Note that two tests are still failing when mingw is used, I
suspect structure layout differences and function calling conventions
between MSVC and MingW.
Thomas Heller [Mon, 3 Jul 2006 08:04:05 +0000 (08:04 +0000)]
Fixes so that _ctypes can be compiled with the MingW compiler.
It seems that the definition of '__attribute__(x)' was responsible for
the compiler ignoring the '__fastcall' attribute on the
ffi_closure_SYSV function in libffi_msvc/ffi.c, took me quite some
time to figure this out.
Gerhard Häring [Sun, 2 Jul 2006 17:48:30 +0000 (17:48 +0000)]
The sqlite3 module did cut off data from the SQLite database at the first null
character before sending it to a custom converter. This has been fixed now.
Fred Drake [Sat, 1 Jul 2006 16:28:20 +0000 (16:28 +0000)]
SF bug #1296433 (Expat bug #1515266): Unchecked calls to character data
handler would cause a segfault. This merges in Expat's lib/xmlparse.c
revisions 1.154 and 1.155, which fix this and a closely related problem
(the later does not affect Python).
Moved the crasher test to the tests for xml.parsers.expat.
Fred Drake [Fri, 30 Jun 2006 19:29:25 +0000 (19:29 +0000)]
- consistency nit: always include "()" in \function and \method
(*should* be done by the presentation, but that requires changes all over)
- avoid spreading the __name meme
Neal Norwitz [Thu, 29 Jun 2006 04:10:08 +0000 (04:10 +0000)]
Add new utility function, reap_children(), to test_support. This should
be called at the end of each test that spawns children (perhaps it
should be called from regrtest instead?). This will hopefully prevent
some of the unexplained failures in the buildbots (hppa and alpha)
during tests that spawn children. The problems were not reproducible.
There were many zombies that remained at the end of several tests.
In the worst case, this shouldn't cause any more problems,
though it may not help either. Time will tell.
Fred Drake [Thu, 29 Jun 2006 00:51:53 +0000 (00:51 +0000)]
SF bug #1504333: sgmlib should allow angle brackets in quoted values
(modified patch by Sam Ruby; changed to use separate REs for start and end
tags to reduce matching cost for end tags; extended tests; updated to avoid
breaking previous changes to support IPv6 addresses in unquoted attribute
values)
[Bug #1508766] Add docs for uuid module; this puts the module in the 'Internet Protocols' section. Arguably this module could also have gone in the chapters on strings or encodings, maybe even the crypto chapter. Fred, please move if you see fit.
Neal Norwitz [Wed, 28 Jun 2006 05:03:22 +0000 (05:03 +0000)]
According to the man pages on Gentoo Linux and Tru64, EACCES or EAGAIN
can be returned if fcntl (lockf) fails. This fixes the test failure
on Tru64 by checking for either error rather than just EAGAIN.
Ronald Oussoren [Tue, 27 Jun 2006 15:45:32 +0000 (15:45 +0000)]
macosx: Install a libpython2.5.a inside the framework as a symlink to the actual
dylib at the root of the framework, that way tools that expect a unix-like
install (python-config, but more importantly external products like
mod_python) work correctly.
Ronald Oussoren [Tue, 27 Jun 2006 12:53:52 +0000 (12:53 +0000)]
Use staticly build copies of zlib and bzip2 to build the OSX installer, that
way the resulting binaries have a better change of running on 10.3.
This patch also updates the search logic for sleepycat db3/4, without this
patch you cannot use a sleepycat build with a non-standard prefix; with this
you can (at least on OSX) if you add the prefix to CPPFLAGS/LDFLAGS at
configure-time. This change is needed to build the binary installer for OSX.
Neal Norwitz [Tue, 27 Jun 2006 04:09:13 +0000 (04:09 +0000)]
Instead of doing a make test, run the regression tests out of the installed
copy. This will hopefully catch problems where directories are added
under Lib/ but not to Makefile.pre.in. This breaks out the 2 runs
of the test suite with and without -O which is also nicer.
Attempt to fix build failure on OS X and Debian alpha; the symptom is
consistent with os.wait() returning immediately because some other
subprocess had previously exited; the test suite then immediately
tries to lock the mailbox and gets an error saying it's already
locked.
To fix this, do a waitpid() so the test suite only continues once
the intended child process has exited.
[Bug #1512163] Use one set of locking methods, lockf();
remove the flock() calls.
On FreeBSD, the two methods lockf() and flock() end up using the same
mechanism and the second one fails. A Linux man page claims that the
two methods are orthogonal (so locks acquired one way don't interact
with locks acquired the other way) but that clearly must be false.
Without this patch IDLE will get unresponsive when you open the debugger
window on OSX. This is both using the system Tcl/Tk on Tiger as the latest
universal download from tk-components.sf.net.
Hye-Shik Chang [Fri, 23 Jun 2006 21:16:18 +0000 (21:16 +0000)]
Bug #1511381: codec_getstreamcodec() in codec.c is corrected to
omit a default "error" argument for NULL pointer. This allows
the parser to take a codec from cjkcodecs again.
(Reported by Taewook Kang and reviewed by Walter Doerwald)
Fit Makefile for the Python doc environment better; this is a step toward
including the howtos in the build process.
* Put LaTeX output in ../paper-<whatever>/.
* Put HTML output in ../html/
* Explain some of the Makefile variables
* Remove some cruft dating to my environment (e.g. the 'web' target)
This makefile isn't currently invoked by the documentation build process,
so these changes won't destabilize anything.
Fred Drake [Fri, 23 Jun 2006 06:03:45 +0000 (06:03 +0000)]
- SF bug #853506: IP6 address parsing in sgmllib
('[' and ']' were not accepted in unquoted attribute values)
- cleaned up tests of character and entity reference decoding so the
tests cover the documented relationships among handle_charref,
handle_entityref, convert_charref, convert_codepoint, and
convert_entityref, without bringing up Unicode issues that sgmllib
cannot be involved in
Brett Cannon [Thu, 22 Jun 2006 16:49:14 +0000 (16:49 +0000)]
'warning's was improperly requiring that a command-line Warning category be
both a subclass of Warning and a subclass of types.ClassType. The latter is no
longer true thanks to new-style exceptions.