Commit Graph

200 Commits

Author SHA1 Message Date
Thomas Heller dcea9a91c6 Add an optional 'offset' parameter to byref, defaultingto zero. 2008-06-10 14:02:46 +00:00
Gregory P. Smith dd96db63f6 This reverts r63675 based on the discussion in this thread:
http://mail.python.org/pipermail/python-dev/2008-June/079988.html

Python 2.6 should stick with PyString_* in its codebase.  The PyBytes_* names
in the spirit of 3.0 are available via a #define only.  See the email thread.
2008-06-09 04:58:54 +00:00
Travis E. Oliphant 0144f27602 Remove locking part of new buffer protocol. 2008-06-06 22:39:47 +00:00
Thomas Heller f89b04c164 Performance improvement: Use PyDict_Get/SetItem instead of
PyDict_Get/SetItemString.
2008-06-06 18:37:55 +00:00
Thomas Heller fbb9c0bf3c Issue #1798: Add ctypes calling convention that allows safe access of errno.
ctypes maintains thread-local storage that has space for two error
numbers: private copies of the system 'errno' value and, on Windows,
the system error code accessed by the GetLastError() and
SetLastError() api functions.

Foreign functions created with CDLL(..., use_errno=True), when called,
swap the system 'errno' value with the private copy just before the
actual function call, and swapped again immediately afterwards.  The
'use_errno' parameter defaults to False, in this case 'ctypes_errno'
is not touched.

On Windows, foreign functions created with CDLL(...,
use_last_error=True) or WinDLL(..., use_last_error=True) swap the
system LastError value with the ctypes private copy.

The values are also swapped immeditately before and after ctypes
callback functions are called, if the callbacks are constructed using
the new optional use_errno parameter set to True: CFUNCTYPE(...,
use_errno=TRUE) or WINFUNCTYPE(..., use_errno=True).

New ctypes functions are provided to access the ctypes private copies
from Python:

- ctypes.set_errno(value) and ctypes.set_last_error(value) store
  'value' in the private copy and returns the previous value.

- ctypes.get_errno() and ctypes.get_last_error() returns the current
  ctypes private copies value.
2008-06-06 08:33:46 +00:00
Thomas Heller 2e75c450d2 Backport from py3k: Implement the new buffer interface from pep3118
for ctypes instances.  Closes issue #2404.
2008-06-05 17:51:15 +00:00
Ronald Oussoren 5640ce2f1e MacOS X: Enable 4-way universal builds
This patch adds a new configure argument on OSX:
        --with-universal-archs=[32-bit|64-bit|all]

When used with the --enable-universalsdk option this controls which
CPU architectures are includes in the framework. The default is 32-bit,
meaning i386 and ppc. The most useful alternative is 'all', which includes
all 4 CPU architectures supported by MacOS X (i386, ppc, x86_64 and ppc64).

This includes limited support for the Carbon bindings in 64-bit mode as well,
limited because (a) I haven't done extensive testing and (b) a large portion
of the Carbon API's aren't available in 64-bit mode anyway.

I've also duplicated a feature of Apple's build of python: setting the
environment variable 'ARCHFLAGS' controls the '-arch' flags used for building
extensions using distutils.
2008-06-05 12:58:24 +00:00
Thomas Heller d5bb9215c9 Revert revisions 63943 and 63942 (Issue #1798: Add ctypes calling
convention that allows safe access to errno)

This code does not yet work on OS X (__thread storage specifier not
available), so i needs a configure check plus a more portable
solution.
2008-06-04 20:22:05 +00:00
Thomas Heller 8f22b88e28 Fix ctypes.set_errno for gcc. 2008-06-04 19:19:00 +00:00
Thomas Heller e70c3378c0 Issue #1798: Add ctypes calling convention that allows safe access to
errno (and LastError, on Windows).

ctypes maintains a module-global, but thread-local, variable that
contains an error number; called 'ctypes_errno' for this discussion.
This variable is a private copy of the systems 'errno' value; the copy
is swapped with the 'errno' variable on several occasions.

Foreign functions created with CDLL(..., use_errno=True), when called,
swap the values just before the actual function call, and swapped
again immediately afterwards.  The 'use_errno' parameter defaults to
False, in this case 'ctypes_errno' is not touched.

The values are also swapped immeditately before and after ctypes
callback functions are called, if the callbacks are constructed using
the new optional use_errno parameter set to True: CFUNCTYPE(..., use_errno=TRUE)
or WINFUNCTYPE(..., use_errno=True).

Two new ctypes functions are provided to access the 'ctypes_errno'
value from Python:

- ctypes.set_errno(value) sets ctypes_errno to 'value', the previous
  ctypes_errno value is returned.

- ctypes.get_errno() returns the current ctypes_errno value.

---

On Windows, the same scheme is implemented for the error value which
is managed by the GetLastError() and SetLastError() windows api calls.

The ctypes functions are 'ctypes.set_last_error(value)' and
'ctypes.get_last_error()', the CDLL and WinDLL optional parameter is
named 'use_last_error', defaults to False.

---

On Windows, TlsSetValue and TlsGetValue calls are used to provide
thread local storage for the variables; ctypes compiled with __GNUC__
uses __thread variables.
2008-06-04 18:59:03 +00:00
Thomas Heller de2d78a153 Fix misspelled sys.platform name and misspelled filename. 2008-06-02 18:41:30 +00:00
Thomas Heller 9287acf83d ctypes NULL function pointers have a boolean False value now. 2008-05-29 19:42:34 +00:00
Thomas Heller a52b244cc1 Fix compiler warning. 2008-05-29 19:18:12 +00:00
Christian Heimes 593daf545b Renamed PyString to PyBytes 2008-05-26 12:51:38 +00:00
Martin v. Löwis 8c255e4173 Patch #1722225: Support QNX 6. 2008-05-23 15:06:50 +00:00
Thomas Heller 03b75ddf7c On HPUX, -fPIC must be used for linking. _ctypes now builds on HP-UX
IA64 and PA machines.

The ctypes unittests work fine on the IA64, but dump core in
test_qsort on the PA.
2008-05-20 19:53:47 +00:00
Georg Brandl a6b79f3b11 Add missing return type to dealloc. 2008-04-26 18:32:17 +00:00
Thomas Heller 6ad5fbb7ea Add from_buffer and from_buffer_copy class methods to ctypes types. 2008-04-25 15:44:16 +00:00
Thomas Heller bf027c48c8 Remove cyclic reference in CFuncPtr instances; see issue #2682.
Backport candidate for the release25-maint branch.
2008-04-24 18:14:19 +00:00
Thomas Heller 046e6a43ff Issue #2616: Implement ctypes.pointer() and ctypes.POINTER() in C for
better performance.
2008-04-14 16:10:07 +00:00
Thomas Heller d3ed492164 Performance improvements. 2008-04-11 14:20:26 +00:00
Thomas Heller 12c4e6478a Move backwards compatibility macro to the correct place;
PyIndex_Check() was introduced in Python 2.5.
2008-04-11 13:05:38 +00:00
Thomas Heller c4ef47611f More files from libffi 3.0.5.
Removed libffi.pc.in because it is not needed for ctypes.
2008-04-04 19:14:42 +00:00
Thomas Heller e1929d5133 Sync with files from the just released libffi 3.0.5 version. 2008-04-04 16:01:54 +00:00
Thomas Heller 8e8ba151dd Issue #2543: Make ctypes compatible (again) with Python 2.3, 2.4, and 2.5. 2008-04-04 08:35:44 +00:00
Thomas Heller fffd939056 Make _ctypes.c PY_SSIZE_T_CLEAN. 2008-03-25 21:18:39 +00:00
Thomas Heller 50a24d8bfd Include <alloca.h> on Solaris, see issue #1506.
It would probably be better to have a configure test for that, but
this is outside of my configure expertise.
2008-03-18 15:03:17 +00:00
Thomas Heller 41e22ff5cc Remove unneeded initializer. 2008-03-14 21:06:21 +00:00
Thomas Heller f3c0559b5e Issue 1872: Changed the struct module typecode from 't' to '?', for
compatibility with PEP3118.
2008-03-05 15:34:29 +00:00
Thomas Heller a23aede562 Try to fix the build for PY_LINUX. 2008-03-04 20:21:42 +00:00
Thomas Heller 8bdf81d2df Merged changes from libffi3-branch.
The bundled libffi copy is now in sync with the recently released
libffi3.0.4 version, apart from some small changes to
Modules/_ctypes/libffi/configure.ac.

I gave up on using libffi3 files on os x.
Instead, static configuration with files from pyobjc is used.
2008-03-04 20:09:11 +00:00
Neal Norwitz ed0a593a10 Get ctypes working on the Alpha (Tru64). The problem was that there
were two module_methods and the one used depended on the order the
modules were loaded.  By making the test module_methods static,
it is not exported and the correct version is picked up.
2008-02-24 07:21:56 +00:00
Thomas Heller 981f31860b configure.ac: Remove the configure check for _Bool, it is already done in the
top-level Python configure script.
configure, fficonfig.h.in: regenerated.
2008-02-21 18:28:48 +00:00
Thomas Heller a06a1a88ee Add pickle support to ctypes types. 2008-02-13 20:21:53 +00:00
Brett Cannon 8352585909 Make sure a switch statement does not have repetitive case statements.
Error found through LLVM post-2.1 svn.
2008-02-07 22:27:10 +00:00
Thomas Heller 55b8c3e26f Fixed refcounts and error handling.
Should not be merged to py3k branch.
2008-02-06 20:29:17 +00:00
Thomas Heller 99120c4f7e Revert rev. 59925, it breaks comtypes (I need to further examine this). 2008-01-28 08:44:13 +00:00
Christian Heimes d7e1b2bd17 static PyObject* variables should use PyString_InternFromString() instead of PyObject_FromString() to store a python string in a function level static var. 2008-01-28 02:07:53 +00:00
Neal Norwitz b3c8243c47 Update to newer version of ffi. Fixes crashes and test failures of longdouble 2008-01-27 17:10:43 +00:00
Thomas Heller 5ca924038d Invert the checks in get_[u]long and get_[u]longlong. The intent was
to not accept float types; the result was that integer-like objects
were not accepted.

Ported from release25-maint.
2008-01-24 19:15:02 +00:00
Thomas Heller 71dba4ccee Replace Py_BuildValue with PyTuple_Pack because it is faster.
Also add a missing DECREF.
2008-01-24 18:54:12 +00:00
Thomas Heller e4c03e4520 Use a PyDictObject again for the array type cache; retrieving items
from the WeakValueDictionary was slower by nearly a factor of 3.

To avoid leaks, weakref proxies for the array types are put into the
cache dict, with weakref callbacks that removes the entries when the
type goes away.
2008-01-24 18:36:27 +00:00
Brett Cannon a35a8b11c3 Fix a function pointer declaration to silence the compiler. 2008-01-17 18:45:10 +00:00
Thomas Heller 0f75f9847a Use 'g' instead of 'D' as the ctypes typecode for c_longdouble, for
compliance with PEP 3118.
2008-01-16 20:34:37 +00:00
Thomas Heller 415c1e36a9 Raise a TypeError instead of a ValueError when too many initializers
are used in a Structure or Union constructor.
2008-01-16 19:45:51 +00:00
Thomas Heller 02ec289f3e Raise a TypeError if conflicting positional and named arguments are
passed to a Structure or Union constructor.
2008-01-16 19:37:33 +00:00
Thomas Heller 902d30752f Convert the internal ctypes array type cache to a WeakValueDict so
that array types do not live longer than needed.
2008-01-16 19:16:27 +00:00
Thomas Heller 7a66a1becc Issue 1821: configure libffi for amd64 on FreeeBSD. 2008-01-14 10:35:28 +00:00
Amaury Forgeot d'Arc 7cdf5f5c31 ?Why did my tests not notice this before?
Slots inheritance is very different from OO inheritance.
This code lead to infinite recursion on classes derived from StructType.
2008-01-14 01:07:27 +00:00
Amaury Forgeot d'Arc 08ccf202e6 As discussed in issue 1700288:
ctypes takes some liberties when creating python types: it modifies the types'
__dict__ directly, bypassing all the machinery of type objects which deal with
special methods.  And this broke recent optimisations of method lookup.
Now we try to modify the type with more "official" functions.
2008-01-14 00:22:44 +00:00