summaryrefslogtreecommitdiff
path: root/Cython/Compiler/ModuleNode.py
Commit message (Collapse)AuthorAgeFilesLines
...
* Table of file names with absolute paths.Martin Raum2014-08-141-1/+1
| | | | | | | The table of file names written to the c or cpp file contains absolute file paths. This affects, in particular, profiling. The c method __pyx_TraceCall takes as one of its arguments a file name, which will be read from the table of file names.
* Support @staticmethod decorator for C++ classes.Robert Bradshaw2014-07-171-1/+3
|
* Don't generate wrappers for cimported enums.Robert Bradshaw2014-07-101-1/+3
|
* Allow extern cpdef enum to export values into Python-level namespace.Robert Bradshaw2014-07-101-0/+22
|
* use explicit relative imports everywhere and enable absolute imports by defaultStefan Behnel2014-06-171-17/+19
|
* Merge pull request #275 from cython/no-old-pythonscoder2014-03-221-2/+2
|\ | | | | Get rid of obsolete Python cruft.
| * Dissallow Python 2.4, 2.5, 3.0, and 3.1Robert Bradshaw2014-02-221-2/+2
| |
* | move embedded C main function code into external utility code fileStefan Behnel2014-03-151-193/+1
| |
* | restrict freelist usage to CPythonStefan Behnel2014-03-101-2/+2
| | | | | | | | | | --HG-- extra : amend_source : 9ae2e52935ae16af23c35b6329d53e559b7c39d0
* | apply freelist type check also when *adding* dead objects to the freelist to ↵Stefan Behnel2014-03-031-5/+12
| | | | | | | | make sure their eventual cleanup works correctly
* | fix crash by excluding heap types from the extension type freelistsStefan Behnel2014-03-031-1/+1
| |
* | use arbitrary but predictable C code order when cleaning up imported ↵Stefan Behnel2014-02-261-2/+2
| | | | | | | | extension types
* | use an actual set() for Scope.types_imported instead of a set-ishly used dictStefan Behnel2014-02-261-1/+1
|/
* Clang-compatible C++ nested destructor call.Robert Bradshaw2014-02-181-8/+1
|
* make C code order in 'from-import *' table deterministicStefan Behnel2014-01-231-1/+1
|
* fix test failureStefan Behnel2014-01-231-3/+10
|
* generate cdef classes and closure classes in a deterministic order to reduce ↵Stefan Behnel2014-01-191-16/+24
| | | | | | C code churn on regeneration also fixes a problem where duplicate entries (one used, one unused) could be selected somewhat arbitrarily during type sorting (apparently applied to but worked for memoryview/-slice)
* add safety check for now to avoid trying to write annotation files into ↵Stefan Behnel2014-01-181-3/+9
| | | | standard include directories etc.
* when includes are used, additionally generate one annotation file per ↵Stefan Behnel2014-01-181-2/+23
| | | | | | | included file --HG-- extra : amend_source : c2b02f8beedd17cdae81256283c30de6f5767e08
* close output file ASAPStefan Behnel2014-01-181-2/+2
|
* safety fixStefan Behnel2014-01-181-4/+6
|
* call object.__new__() only when we expect an error due to an abstract class ↵Stefan Behnel2014-01-121-3/+14
| | | | being instantiated
* revert dynamic tp_new() override because it breaks instantiating types with ↵Stefan Behnel2014-01-101-3/+2
| | | | arguments that object.__new__()/object.__init__() don't expect
* reuse object's tp_new() for simple extension types if possible to improve ↵Stefan Behnel2014-01-101-2/+3
| | | | support for multiple inheritance
* call PyBaseObject_Type.tp_new() in tp_new() functions instead of tp_alloc() ↵Stefan Behnel2014-01-101-2/+2
| | | | directly
* add guarding for special method docstringsR. Andrew Ohana2014-01-031-0/+6
|
* whitespaceRobert Bradshaw2013-12-271-4/+4
|
* avoid redundant NULL inits for extension type fields in tp_new()Stefan Behnel2013-12-221-8/+10
|
* fix C compiler warning about unused expression resultStefan Behnel2013-12-211-1/+1
|
* clean up lots of places where exceptions are being raised to reduce the ↵Stefan Behnel2013-12-061-4/+4
| | | | message building overhead
* support bytearray as auto encoding string typeStefan Behnel2013-11-061-4/+8
| | | | | | --HG-- rename : tests/run/str_ascii_auto_encoding.pyx => tests/run/bytearray_ascii_auto_encoding.pyx rename : tests/run/str_default_auto_encoding.pyx => tests/run/bytearray_default_auto_encoding.pyx
* fix commentStefan Behnel2013-11-021-1/+1
|
* spellingRobert Bradshaw2013-11-011-1/+1
|
* Get rid of 'virtual methods without virtual destructor' warning.Robert Bradshaw2013-11-011-0/+7
|
* Clear tp_print for cdef classes.Robert Bradshaw2013-11-011-0/+3
|
* disable unnecessary PyType_IS_GC() check in PyPy (not supported)Stefan Behnel2013-10-161-2/+4
|
* clear global reference to module dict in module cleanup codeStefan Behnel2013-10-131-0/+2
|
* Merge branch 'int-conversion'Robert Bradshaw2013-09-181-6/+5
|\
| * Python 2.4 fix.Robert Bradshaw2013-08-281-0/+1
| |
| * Fix some integer conversion tests.Robert Bradshaw2013-08-261-0/+4
| |
| * Remove old type conversion code.Robert Bradshaw2013-08-241-6/+0
| |
* | Set correct baseclass for heap allocated type subclasses.Robert Bradshaw2013-09-121-1/+5
| |
* | Cleanup type sharing.Robert Bradshaw2013-08-121-0/+3
|/
* avoid generating empty PyMethodDef array for types (especially closures)Stefan Behnel2013-08-101-0/+2
|
* fix code generated for __dealloc__() and __releasebuffer__() to call ↵Stefan Behnel2013-08-091-1/+0
| | | | WriteUnraisable() on exceptions instead of building a useless traceback
* be a tiny bit smarter about when to GC_Track for the base type in tp_dealloc()Stefan Behnel2013-08-091-2/+5
|
* clean up tp_dealloc() code and integrate call to tp_finalize() for Py3.4Stefan Behnel2013-08-091-40/+55
|
* make the NULL-instead-of-None change in tp_clear() optional and off by ↵Stefan Behnel2013-08-081-2/+15
| | | | default - breaks too much code for now
* Add decorator @gc_no_clear to disable tp_clear slot.Torsten Landschoff2013-08-021-1/+2
| | | | | | | | | | This can be useful if some external objects have to be cleaned up when the Python object dies. Otherwise the cyclic garbage collector may drop the references before the cleanup has any chance to run. Includes documentation and a simple unit test. Discussion: http://article.gmane.org/gmane.comp.python.cython.devel/14986
* Change tp_clear generation to clear to NULL.Torsten Landschoff2013-08-011-8/+1
| | | | | | | Better to crash early than to never know that the reference changed to None. See discussion here: http://article.gmane.org/gmane.comp.python.cython.devel/14833