Commit Graph

34504 Commits

Author SHA1 Message Date
Guido van Rossum 27ae52362c Fix test_uu.py by inserting an explicit conversion from bytes to str
into uu.py.  (This is *not* SF patch 1754339, which I rejected.)
2007-07-16 19:50:03 +00:00
Guido van Rossum 456fe5d3ca Fix a weird use of try/finally to close a file.
(There are more places that don't close 'f' at all if an error occurs,
but none have a bogus try/finally pattern.)
2007-07-16 19:42:05 +00:00
Guido van Rossum b358a2c423 SF patch# 1752647 by Joe Gregorio.
Fixes to make test_mmap.py pass unit tests.
Moves the interface from being string based to being bytes based.
2007-07-16 19:29:02 +00:00
Guido van Rossum 88079f472e Use the encoding specification when reading the source file. 2007-07-15 14:02:52 +00:00
Guido van Rossum 827bfd0742 Don't stop installing if compileall fails -- ever. 2007-07-15 13:12:42 +00:00
Guido van Rossum b7d3e65b3b Improve error handling; don't die from unicode errors or syntax errors. 2007-07-15 13:01:48 +00:00
Guido van Rossum 8ac004e698 Make chr() and ord() return/accept surrogate pairs in narrow builds.
The domain of chr() and the range of ord() are now always [0 ... 0x10FFFF].
2007-07-15 13:00:05 +00:00
Thomas Heller 49c12ac04e Fix ctypes tests on OS X. 2007-07-14 18:51:36 +00:00
Thomas Heller 3e6cbf178b For easier merging, revert ths change; it will be brought back in when
the trunk is merged.
2007-07-13 19:48:40 +00:00
Thomas Heller 7f16288134 atexit is a builtin module now, provide that on Windows. 2007-07-13 13:13:40 +00:00
Thomas Heller 19b52545df c_char, c_char_p objects and c_char array structure fields return
their value now as str, no longer str8.
2007-07-13 12:52:51 +00:00
Thomas Heller 27384da6e8 Do not accept str8 type in function calls any longer.
Accept bytes instead of str8 in the (unexposed in ctypes) BSTR type.
2007-07-13 12:07:59 +00:00
Thomas Heller ace05054d3 Remove all the Python 2.3 compatibility markers (the Python 3 ctypes
version is not compatible with Python 2.x anymore).
Remove backwards compatibility code.
2007-07-13 11:53:39 +00:00
Thomas Heller 84d66a0843 Remove all the Python 2.3 compatibility markers (the Python 3 ctypes
version is not compatible with Python 2.x anymore).
Remove backwards compatibility code.
2007-07-13 11:49:15 +00:00
Thomas Heller e5095e187b Structure fields of type c_char array or c_wchar array accept bytes or
(unicode) string.
2007-07-13 11:19:35 +00:00
Thomas Heller 745f5e2de7 No crashes on Linux in the ctypes test suite any longer, so we can
remove the 'crash-protection'.
2007-07-12 20:10:18 +00:00
Thomas Heller f4f771fe0f Fix segfault on Linux. 2007-07-12 20:06:41 +00:00
Thomas Heller 8750384dbc More easy fixes. The ctypes unittests pass now (on Windows). 2007-07-12 19:58:41 +00:00
Thomas Heller d660fabd24 binascii.hexlify returns a bytes object; we must convert it to str
before we can pass it to re.sub.
2007-07-12 19:56:28 +00:00
Thomas Heller 0d8a1fd78d Simple fixes. 2007-07-12 19:54:27 +00:00
Thomas Heller d68d131d29 Simple fixes. 2007-07-12 19:45:20 +00:00
Thomas Heller f0d8b6eee9 Simple fixes. 2007-07-12 19:42:59 +00:00
Thomas Heller 3821e31d19 c_void_p.from_param accepts bytes. Fix test_prototypes. 2007-07-12 19:38:33 +00:00
Thomas Heller 3d79dd9edf c_char_p.from_param accepts bytes. Fix test_parameters. 2007-07-12 19:33:24 +00:00
Thomas Heller 7a76e8f67c Fix another test. 2007-07-12 19:24:59 +00:00
Thomas Heller 9ccf468323 test_slicing does not segfault under linux anymore, so we can run this
test by default.
2007-07-12 19:22:10 +00:00
Thomas Heller 7775c716fc Accept bytes as parameter to foreign functions without prototype.
These are passed as byte strings (unicode strings are passed as wide
character strings).
2007-07-12 19:19:43 +00:00
Thomas Heller 60831316df Accept bytes for c_wchar_p instances and c_wchar array instances.
ctypes.create_unicode_buffer also accepts bytes now.

Revert some tests in test_unicode: Since string literals are unicode
now, conversion takes place when byte literals are passed as unicode
parameters.
2007-07-12 19:06:25 +00:00
Thomas Heller d076228202 Change the StgDictObject's proto member to a PyUnicode_Object (was a
PyStringObject before).
2007-07-12 18:53:22 +00:00
Thomas Heller 2c5e96465f Accept bytes in c_char_p and c_wchar_p types. 2007-07-12 15:41:51 +00:00
Thomas Heller 3af4266d07 ctypes.c_char and ctypes.c_wchar now accept initialization from byte objects. 2007-07-12 14:58:32 +00:00
Thomas Heller f7c6d86832 Use a unicode string as unique_key instead of 8-bit string. 2007-07-12 13:55:37 +00:00
Walter Dörwald a1884664eb Fix typo in docstring. 2007-07-12 12:16:02 +00:00
Walter Dörwald 6720d91ee7 Rename PyTclObject_unicode to PyTclObject_str and use this
for the tp_str slot. Drop __unicode__ method.
2007-07-12 12:12:25 +00:00
Thomas Heller 7c1ebe33b8 Fix a ctypes test. 2007-07-12 12:05:40 +00:00
Thomas Heller 1f498ef63d Fix a ctypes test. 2007-07-12 11:59:52 +00:00
Thomas Heller b0f48abd87 Skip testing the special file "/dev/tty" on Windows. This test does
weird things if someone has a "\dev" directory on the current drive.
2007-07-12 11:29:02 +00:00
Thomas Heller fdeee3a6cd Second part of sf# 1752225: On windows, emulate ftruncate with Win32
api functions. Code from fileobject.c, patch by Amaury Forgeot d'Arc.

This patch also changes:

The return value of the native ftruncate function is checked for '!=
0' instead of '< 0' as before.

fileio_seekable returns bool now instead of an int.
2007-07-12 11:21:36 +00:00
Thomas Heller af2be262af First part of sf# 1752225: Support for wide filenames on Windows.
Patch by Amaury Forgeot d'Arc.
2007-07-12 11:03:13 +00:00
Thomas Heller 5fa3f0517b Py_USING_UNICODE is no longer defined on Windows (on posix it is still
in pyconfig.h).  Remove the workaround for the broken (in Python 2.3)
PyUnicode_(From|As)WideChar functions.
2007-07-12 11:00:22 +00:00
Guido van Rossum e409b97950 Make this a pure-ASCII file. This used to have a Latin-1 coding and
contain several string literals using non-ASCII characters.  Because
of the pain of displaying those characters, it's better to use \xXX
escapes for these instead of typing the accented characters.

Apologies if I missed something; it didn't look like there was anything
that depended on this file being encoded in Latin-1 or containing non-ASCII
characters.
2007-07-12 09:25:21 +00:00
Guido van Rossum fc0ffb7243 Merged revisions 56265-56300 via svnmerge from
svn+ssh://pythondev@svn.python.org/python/branches/p3yk

........
  r56274 | georg.brandl | 2007-07-11 08:10:46 -0700 (Wed, 11 Jul 2007) | 2 lines

  Fix #1751965, typo in isinstance.
........
2007-07-12 08:55:44 +00:00
Guido van Rossum a5d0c260ac Fix for tet_bigmem.py (the failure was caused by a spurious reference
to 'file', which no longer exists).
Fix inspired by Christian Heimes, SF# 1752195.
2007-07-12 08:11:23 +00:00
Guido van Rossum 652f446d42 Fixes for str/uni/bytes for gettext.py. test_gettext.py passes.
Fix by Christian Heimes, SF# 1751958, who writes:

  I tested the fixes with the Zope3 zope.app.locales packages.
  The mo files are loaded and parsed w/o any problem.
  The translation with gettext.gettext is working as expected.
2007-07-12 08:04:06 +00:00
Guido van Rossum 076da0957b Fix asynchat to use bytes instead of strings.
Fix by Alexandre Vassalotti, SF# 1752173.
2007-07-12 07:58:54 +00:00
Guido van Rossum e7a0d39978 Fixed a refcount leak in _PyUnicode_AsDefaultEncodedString().
Found by Amaury Forgeot d'Arc, SF# 1752317.
(Also removed a redundant check for !errors.)
2007-07-12 07:53:00 +00:00
Guido van Rossum dc09715d0b Use unicode instead of 8-bit strings.
Patch by Alexandre Vassalotti, SF# 1752229.
2007-07-12 07:44:15 +00:00
Guido van Rossum c3e7ffc670 obcheckin. 2007-07-11 20:56:15 +00:00
Guido van Rossum bdab7d3cb6 Fix test_poll.py. By Christian Heimes. 2007-07-11 20:43:16 +00:00
Thomas Heller 3ed0acceee Must create heaptypes with unicode names. 2007-07-11 20:02:35 +00:00