Éric Araujo [Sat, 8 Oct 2011 02:09:15 +0000 (04:09 +0200)]
Fix packaging byte-compilation to comply with PEP 3147 (#11254).
I want to replace custom byte-compiling function with calls to
compileall before 3.3b1, but in the short term it’s good to have this
fixed.
Adapted from the distutils patch by Jeff Ramnani. I tested with -B, -O
and -OO; test_util and test_mixin2to3 fail in -O mode because lib2to3
doesn’t support it.
Éric Araujo [Fri, 7 Oct 2011 23:55:07 +0000 (01:55 +0200)]
Fix test_sysconfig when prefix != exec-prefix (#9100).
I tested this manually; it would be great to have buildbots using
installed Pythons, including Pythons configured with different prefix
and exec-prefix.
Antoine Pitrou [Fri, 7 Oct 2011 14:17:50 +0000 (16:17 +0200)]
Issue #13063: the Windows error ERROR_NO_DATA (numbered 232 and described
as "The pipe is being closed") is now mapped to POSIX errno EPIPE
(previously EINVAL).
Antoine Pitrou [Fri, 7 Oct 2011 14:16:31 +0000 (16:16 +0200)]
Issue #13063: the Windows error ERROR_NO_DATA (numbered 232 and described
as "The pipe is being closed") is now mapped to POSIX errno EPIPE
(previously EINVAL).
Éric Araujo [Thu, 6 Oct 2011 03:10:09 +0000 (05:10 +0200)]
Fix incorrect test.
The packaging.install.remove function (a.k.a. the uninstall feature)
takes a path argument to allow client code to use custom directories
instead of sys.path. The test used to give self.root_dir as path, which
corresponds to a prefix option, but prefix is not on sys.path, it’s only
the base directory used to compute the stdlib and site-packages
directory paths. The test now gives a valid site-packages path to the
function.
Éric Araujo [Thu, 6 Oct 2011 02:59:41 +0000 (04:59 +0200)]
Change one name in packaging’s test_uninstall to avoid confusion.
install_lib may be the name of a module, a command or an option, so I
find it clearer to use site_packages to refer to a string object
containing the path of the site-packages directory created in a
temporary directory during tests.
Victor Stinner [Wed, 5 Oct 2011 23:45:57 +0000 (01:45 +0200)]
Don't check for the maximum character when copying from unicodeobject.c
* Create copy_characters() function which doesn't check for the maximum
character in release mode
* _PyUnicode_CheckConsistency() is no more static to be able to use it
in _PyUnicode_FormatAdvanced() (in formatter_unicode.c)
* _PyUnicode_CheckConsistency() checks the string hash
Issue #13070: Fix a crash when a TextIOWrapper caught in a reference cycle
would be finalized after the reference to its underlying BufferedRWPair's
writer got cleared by the GC.
Éric Araujo [Tue, 4 Oct 2011 23:50:22 +0000 (01:50 +0200)]
Update skip message printed by test.support.get_attribute.
This helper was changed to work with any object instead of only modules
(or technically something with a __name__ attribute, see code in 3.2)
but the message stayed as is.
Éric Araujo [Tue, 4 Oct 2011 23:46:37 +0000 (01:46 +0200)]
Cosmetic fixes for whitespace and a regex in packaging.
The goal of the regex is to catch a (alpha), b (beta), c or rc
(release candidate), so the existing pattern puzzled me. Tests were
OK before and after the change.
Éric Araujo [Tue, 4 Oct 2011 23:28:24 +0000 (01:28 +0200)]
Fix markup used in the documentation of sys.prefix and sys.exec_prefix.
- Using the file role with {placeholders} is IMO clearer than fake
Python code.
- The fact that sys.version[:3] gives '3.2' is a CPython detail and
should not be advertised (see #9442), even if some stdlib modules
currently rely on that detail.
Éric Araujo [Tue, 4 Oct 2011 23:06:31 +0000 (01:06 +0200)]
Fix minor wording issue.
sys.maxunicode is not called and thus does not return anything; it *is*
something. (I checked the doc quickly to see if it tells that
expression return things but found nothing.)
I also removed markup that would just generate a useless link to the
enclosing section.