cython.git
14 years agosupport setting language level (2 or 3) for source code - currently configures string...
Stefan Behnel [Mon, 31 May 2010 20:30:08 +0000 (22:30 +0200)]
support setting language level (2 or 3) for source code - currently configures string literals and print()

14 years agoextended test case
Stefan Behnel [Mon, 31 May 2010 12:20:15 +0000 (14:20 +0200)]
extended test case

14 years agore-enable friendly exceptions to make map exception traces back to tree traces
Stefan Behnel [Fri, 28 May 2010 10:41:40 +0000 (12:41 +0200)]
re-enable friendly exceptions to make map exception traces back to tree traces

14 years agoremember constant values during new constant node creation
Stefan Behnel [Thu, 27 May 2010 16:19:34 +0000 (18:19 +0200)]
remember constant values during new constant node creation

14 years agoAnother IntNode fix.
Robert Bradshaw [Thu, 27 May 2010 04:41:40 +0000 (21:41 -0700)]
Another IntNode fix.

14 years agoFix types of marked integer literals.
Robert Bradshaw [Thu, 27 May 2010 04:21:16 +0000 (21:21 -0700)]
Fix types of marked integer literals.

14 years agoMore large literal as double fixes.
Robert Bradshaw [Thu, 27 May 2010 04:04:08 +0000 (21:04 -0700)]
More large literal as double fixes.

14 years agoFix large integer literals used in C double context.
Robert Bradshaw [Thu, 27 May 2010 03:45:55 +0000 (20:45 -0700)]
Fix large integer literals used in C double context.

14 years agoLarge literal int + double test.
Robert Bradshaw [Thu, 27 May 2010 03:29:53 +0000 (20:29 -0700)]
Large literal int + double test.

14 years agomove 'for from' syntax out of the way in Python file compilation
Stefan Behnel [Wed, 26 May 2010 14:18:32 +0000 (16:18 +0200)]
move 'for from' syntax out of the way in Python file compilation

14 years agofix some error reportings in parser
Stefan Behnel [Tue, 25 May 2010 19:09:36 +0000 (21:09 +0200)]
fix some error reportings in parser

14 years agoAllow longer int literals on 64-bit machines.
Robert Bradshaw [Mon, 24 May 2010 23:03:23 +0000 (16:03 -0700)]
Allow longer int literals on 64-bit machines.

14 years agofix indexing for inferred Py_UNICODE 'strings'
Stefan Behnel [Mon, 17 May 2010 14:43:08 +0000 (16:43 +0200)]
fix indexing for inferred Py_UNICODE 'strings'

14 years agoextended test case
Stefan Behnel [Mon, 17 May 2010 11:23:56 +0000 (13:23 +0200)]
extended test case

14 years agomake len(Py_UNICODE) efficient
Stefan Behnel [Mon, 17 May 2010 11:23:31 +0000 (13:23 +0200)]
make len(Py_UNICODE) efficient

14 years agofix test after fixing code
Stefan Behnel [Mon, 17 May 2010 09:25:57 +0000 (11:25 +0200)]
fix test after fixing code

14 years agodo not optimise len(str) as it doesn't map to a simple C-API function, small cleanup
Stefan Behnel [Mon, 17 May 2010 06:38:00 +0000 (08:38 +0200)]
do not optimise len(str) as it doesn't map to a simple C-API function, small cleanup

14 years agomerge
Stefan Behnel [Sun, 16 May 2010 08:10:06 +0000 (10:10 +0200)]
merge

14 years agooptimise predicates and methods for single unicode characters
Stefan Behnel [Sun, 16 May 2010 08:09:33 +0000 (10:09 +0200)]
optimise predicates and methods for single unicode characters

14 years agomerge
Robert Bradshaw [Sat, 15 May 2010 20:50:21 +0000 (13:50 -0700)]
merge

14 years agoAdd failing test to bugs list.
Robert Bradshaw [Sat, 15 May 2010 20:49:43 +0000 (13:49 -0700)]
Add failing test to bugs list.

14 years agoC++ failures / tests
David Barnett [Sat, 15 May 2010 20:47:40 +0000 (13:47 -0700)]
C++ failures / tests

I noticed two failures to compile that I think are bugs:
 - "ctypedef Foo[int] Bar" gives a syntax error (i.e., there doesn't seem to be any way to typedef a templated type).
 - Defining a struct in a namespace and trying to convert it to a Python dict doesn't work. Instead you get nasty "__pyx_convert__to_py_THENAMESPACE" errors from the C++ compiler (it doesn't mangle the cname properly and tries to define __pyx_convert__to_py_THENAMESPACE::THESTRUCT).

I made a patch to *add tests* for the two bugs (not fix them).

14 years agoinfer plain object for bytes_string[i] and str/unicode for indexed unicode/str objects
Stefan Behnel [Sat, 15 May 2010 13:57:56 +0000 (15:57 +0200)]
infer plain object for bytes_string[i] and str/unicode for indexed unicode/str objects

14 years ago'fix' stupid test
Stefan Behnel [Fri, 14 May 2010 20:13:17 +0000 (22:13 +0200)]
'fix' stupid test

14 years agoinfer type of loop variable when for-in-looping over pointers, C arrays, unicode...
Stefan Behnel [Fri, 14 May 2010 20:03:16 +0000 (22:03 +0200)]
infer type of loop variable when for-in-looping over pointers, C arrays, unicode strings, etc.

14 years agotypo
Stefan Behnel [Thu, 13 May 2010 13:48:29 +0000 (15:48 +0200)]
typo

14 years agocomment fixes
Stefan Behnel [Thu, 13 May 2010 11:09:20 +0000 (13:09 +0200)]
comment fixes

14 years agofix comment
Stefan Behnel [Thu, 13 May 2010 05:18:08 +0000 (07:18 +0200)]
fix comment

14 years agoerror handling in __test__ dict creation code
Stefan Behnel [Wed, 12 May 2010 20:20:21 +0000 (22:20 +0200)]
error handling in __test__ dict creation code

14 years agofix star import utility code in Py3
Stefan Behnel [Wed, 12 May 2010 19:01:34 +0000 (21:01 +0200)]
fix star import utility code in Py3

14 years agoprevent redundant coercion from Py_UNICODE to a unicode string when the subscript...
Stefan Behnel [Wed, 12 May 2010 17:11:00 +0000 (19:11 +0200)]
prevent redundant coercion from Py_UNICODE to a unicode string when the subscript index is a Python object - this case is no longer optimised

14 years agofix doctest in Py3
Stefan Behnel [Wed, 12 May 2010 15:10:02 +0000 (17:10 +0200)]
fix doctest in Py3

14 years agofix copy+paste bug
Stefan Behnel [Wed, 12 May 2010 14:57:24 +0000 (16:57 +0200)]
fix copy+paste bug

14 years agofix test py_unicode_type, remove code from Optimize.py that became redundant when...
Stefan Behnel [Wed, 12 May 2010 14:45:19 +0000 (16:45 +0200)]
fix test py_unicode_type, remove code from Optimize.py that became redundant when Py_UNICODE started coercing to unicode by default

14 years agoprevent Py_UNICODE from becoming a plain integer during arithmetic operations
Stefan Behnel [Wed, 12 May 2010 14:24:29 +0000 (16:24 +0200)]
prevent Py_UNICODE from becoming a plain integer during arithmetic operations

14 years agoimplement ticket #535: fast index access into unicode strings
Stefan Behnel [Wed, 12 May 2010 13:48:13 +0000 (15:48 +0200)]
implement ticket #535: fast index access into unicode strings

14 years agofix ticket #525: let float values pass through the compiler literally
Stefan Behnel [Tue, 11 May 2010 17:31:27 +0000 (19:31 +0200)]
fix ticket #525: let float values pass through the compiler literally

14 years agoString literal identifiers ending in newlines.
Robert Bradshaw [Tue, 11 May 2010 04:18:46 +0000 (21:18 -0700)]
String literal identifiers ending in newlines.

14 years agobackout 63a493cdeee2 (usage of & is intentional)
Lisandro Dalcin [Mon, 10 May 2010 23:19:17 +0000 (20:19 -0300)]
backout 63a493cdeee2 (usage of & is intentional)

14 years agochange & to && in many places
Lisandro Dalcin [Mon, 10 May 2010 16:22:13 +0000 (13:22 -0300)]
change & to && in many places

14 years agoenable working test
Stefan Behnel [Sun, 9 May 2010 11:15:30 +0000 (13:15 +0200)]
enable working test

14 years agofix ticket #533: 'continue' in the 'else' block of a for-in-loop leaks a reference
Stefan Behnel [Sun, 9 May 2010 11:14:09 +0000 (13:14 +0200)]
fix ticket #533: 'continue' in the 'else' block of a for-in-loop leaks a reference

14 years agofix except+ for cppclass methods (with Denys Duchier)
Lisandro Dalcin [Sat, 8 May 2010 22:24:53 +0000 (19:24 -0300)]
fix except+ for cppclass methods (with Denys Duchier)

14 years agotest for ticket #533: reference leak in 'continue' statement
Stefan Behnel [Sat, 8 May 2010 20:35:17 +0000 (22:35 +0200)]
test for ticket #533: reference leak in 'continue' statement

14 years agoextended test case
Stefan Behnel [Thu, 6 May 2010 06:27:37 +0000 (08:27 +0200)]
extended test case

14 years agonew test for pointer operations
Stefan Behnel [Wed, 5 May 2010 19:53:25 +0000 (21:53 +0200)]
new test for pointer operations

14 years agofix last commit
Stefan Behnel [Wed, 5 May 2010 19:41:47 +0000 (21:41 +0200)]
fix last commit

14 years agofix ticket #145 also for CondExprNode, make "bint (+) non-bint -> object" a general...
Stefan Behnel [Wed, 5 May 2010 19:16:49 +0000 (21:16 +0200)]
fix ticket #145 also for CondExprNode, make "bint (+) non-bint -> object" a general rule for independent spanning type finding, merge duplicated type determination functionality from CondExprNode back into PyrexTypes.py

14 years agoadd some POSIX pxd's and a few tests
Lisandro Dalcin [Wed, 5 May 2010 17:51:56 +0000 (14:51 -0300)]
add some POSIX pxd's and a few tests

14 years agocollapse BoolBinopNode during constant folding, small fix for BoolNode folding
Stefan Behnel [Tue, 4 May 2010 19:11:33 +0000 (21:11 +0200)]
collapse BoolBinopNode during constant folding, small fix for BoolNode folding

14 years agofix ticket #145: the result of 'True or 5' must not be coerced into an integer
Stefan Behnel [Tue, 4 May 2010 19:00:24 +0000 (21:00 +0200)]
fix ticket #145: the result of 'True or 5' must not be coerced into an integer

14 years agoseveral test cases for ticket 145
Stefan Behnel [Tue, 4 May 2010 12:12:54 +0000 (14:12 +0200)]
several test cases for ticket 145

14 years agoextended test case based on ticket #145
Stefan Behnel [Tue, 4 May 2010 11:26:03 +0000 (13:26 +0200)]
extended test case based on ticket #145

14 years agonew test for ticket #354
Stefan Behnel [Tue, 4 May 2010 07:52:05 +0000 (09:52 +0200)]
new test for ticket #354

14 years agoenable .pop() optimisation also for typed lists
Stefan Behnel [Tue, 4 May 2010 05:55:20 +0000 (07:55 +0200)]
enable .pop() optimisation also for typed lists

14 years agoslight speed-up in unicode/bytes indexing
Stefan Behnel [Mon, 3 May 2010 16:13:59 +0000 (18:13 +0200)]
slight speed-up in unicode/bytes indexing

14 years agoMSVC: quick fix for int complex test failing in C++, do not use std::complex<int>
dalcinl [Mon, 3 May 2010 01:22:36 +0000 (22:22 -0300)]
MSVC: quick fix for int complex test failing in C++, do not use std::complex<int>

14 years agoDisable test for libc.stdint (missing stdint.h in MSVC)
dalcinl [Mon, 3 May 2010 00:43:09 +0000 (21:43 -0300)]
Disable test for libc.stdint (missing stdint.h in MSVC)

14 years agofix ssize_t tescase (ticket #399)
Lisandro Dalcin [Sun, 2 May 2010 21:46:17 +0000 (18:46 -0300)]
fix ssize_t tescase (ticket #399)

14 years agoreverted last commit, cast must happen to the value, not its pointer
Stefan Behnel [Sun, 2 May 2010 16:10:39 +0000 (18:10 +0200)]
reverted last commit, cast must happen to the value, not its pointer

14 years agocode simplification
Stefan Behnel [Sun, 2 May 2010 15:59:06 +0000 (17:59 +0200)]
code simplification

14 years agofix 'cython.set' and 'cython.frozenset' when set/frozenset are declared names in...
Stefan Behnel [Sat, 1 May 2010 16:29:44 +0000 (18:29 +0200)]
fix 'cython.set' and 'cython.frozenset' when set/frozenset are declared names in the module

14 years agocode cleanup
Stefan Behnel [Sat, 1 May 2010 16:20:25 +0000 (18:20 +0200)]
code cleanup

14 years agocode cleanup
Stefan Behnel [Sat, 1 May 2010 16:10:42 +0000 (18:10 +0200)]
code cleanup

14 years agoavoid GIL error for optimised container bool tests that run in plain C
Stefan Behnel [Sat, 1 May 2010 15:38:23 +0000 (17:38 +0200)]
avoid GIL error for optimised container bool tests that run in plain C

14 years agofix error on def-nogil functions
Stefan Behnel [Sat, 1 May 2010 15:28:46 +0000 (17:28 +0200)]
fix error on def-nogil functions

14 years agoremoved dead copy of test file
Stefan Behnel [Sat, 1 May 2010 15:26:38 +0000 (17:26 +0200)]
removed dead copy of test file

14 years agossize_t in Cython now means Py_ssize_t in C ((ticket #399))
Lisandro Dalcin [Fri, 30 Apr 2010 18:35:39 +0000 (15:35 -0300)]
ssize_t in Cython now means Py_ssize_t in C ((ticket #399))

14 years agotest fix after extending dead code removal
Stefan Behnel [Thu, 29 Apr 2010 06:48:19 +0000 (08:48 +0200)]
test fix after extending dead code removal

14 years agoextended test case
Stefan Behnel [Thu, 29 Apr 2010 06:35:35 +0000 (08:35 +0200)]
extended test case

14 years agocode cleanup for if-const dead code removal
Stefan Behnel [Thu, 29 Apr 2010 06:26:06 +0000 (08:26 +0200)]
code cleanup for if-const dead code removal

14 years agoextended test case
Stefan Behnel [Thu, 29 Apr 2010 06:06:02 +0000 (08:06 +0200)]
extended test case

14 years agoextended test case
Stefan Behnel [Thu, 29 Apr 2010 06:01:34 +0000 (08:01 +0200)]
extended test case

14 years agocode cleanup for if-const dead code removal: move handling into ConstantFolding trans...
Stefan Behnel [Thu, 29 Apr 2010 06:00:37 +0000 (08:00 +0200)]
code cleanup for if-const dead code removal: move handling into ConstantFolding transform, where it can also replace the entire node

--HG--
rename : tests/run/consts.pyx => tests/run/if_const.pyx

14 years agonew test case
Stefan Behnel [Thu, 29 Apr 2010 05:49:46 +0000 (07:49 +0200)]
new test case

14 years agosizeof(long double)==sizeof(double) in MSVC conficts with numpy built with MinGW
dalcinl [Wed, 28 Apr 2010 23:30:58 +0000 (20:30 -0300)]
sizeof(long double)==sizeof(double) in MSVC conficts with numpy built with MinGW

14 years agofix testcase
Lisandro Dalcin [Wed, 28 Apr 2010 19:29:02 +0000 (16:29 -0300)]
fix testcase

14 years agointroduce CYTHON_UNUSED macro to annotate functions and parametes
Lisandro Dalcin [Wed, 28 Apr 2010 19:18:43 +0000 (16:18 -0300)]
introduce CYTHON_UNUSED macro to annotate functions and parametes
- Silent compiler warnings about unused function/parametes
- Defined to __attribute__((__unused__)) for GCC(>3.4) and ICC
- Applied to a __{get|release}buffer__ special methods
- Applied to a couple of (self,unused) parameter pairs

14 years agoPy3: __cmp__ is gone, __bool__ instead of __nonzero__
Lisandro Dalcin [Wed, 28 Apr 2010 15:04:13 +0000 (12:04 -0300)]
Py3: __cmp__ is gone, __bool__ instead of __nonzero__

14 years agoextended test case
Stefan Behnel [Wed, 28 Apr 2010 11:34:36 +0000 (13:34 +0200)]
extended test case

14 years agoextended test
Stefan Behnel [Tue, 27 Apr 2010 19:36:49 +0000 (21:36 +0200)]
extended test

14 years agomore dead code removal for conditionals
Stefan Behnel [Tue, 27 Apr 2010 19:31:20 +0000 (21:31 +0200)]
more dead code removal for conditionals

14 years agoavoid generating code for 'if' conditions that are known to be False
Stefan Behnel [Tue, 27 Apr 2010 18:37:32 +0000 (20:37 +0200)]
avoid generating code for 'if' conditions that are known to be False

14 years agofix constant folding in PrimaryCmpNode/CascadedCmpNode
Stefan Behnel [Tue, 27 Apr 2010 15:25:42 +0000 (17:25 +0200)]
fix constant folding in PrimaryCmpNode/CascadedCmpNode

14 years agosupport casting '<bytes>int_val' and coercing integers to bytes on assignment
Stefan Behnel [Tue, 27 Apr 2010 05:34:45 +0000 (07:34 +0200)]
support casting '<bytes>int_val' and coercing integers to bytes on assignment

14 years agoextended test case
Stefan Behnel [Mon, 26 Apr 2010 06:38:34 +0000 (08:38 +0200)]
extended test case

14 years agofix bytes indexing
Stefan Behnel [Mon, 26 Apr 2010 06:38:24 +0000 (08:38 +0200)]
fix bytes indexing

14 years agoPy3 test fix
Stefan Behnel [Mon, 26 Apr 2010 05:17:52 +0000 (07:17 +0200)]
Py3 test fix

14 years agokeep 'constant_result' set when injecting new constant nodes
Stefan Behnel [Mon, 26 Apr 2010 04:51:23 +0000 (06:51 +0200)]
keep 'constant_result' set when injecting new constant nodes

14 years agomerge
Stefan Behnel [Sun, 25 Apr 2010 19:55:39 +0000 (21:55 +0200)]
merge

14 years agooptimised char/Py_UNICODE indexing of bytes/unicode objects
Stefan Behnel [Sun, 25 Apr 2010 19:55:26 +0000 (21:55 +0200)]
optimised char/Py_UNICODE indexing of bytes/unicode objects

14 years agowhitespace
Stefan Behnel [Sun, 25 Apr 2010 18:31:19 +0000 (20:31 +0200)]
whitespace

14 years agocoerce Py_UNICODE to and from single character unicode strings by default
Stefan Behnel [Sun, 25 Apr 2010 16:14:58 +0000 (18:14 +0200)]
coerce Py_UNICODE to and from single character unicode strings by default

14 years agotiny cleanup
Robert Bradshaw [Sun, 25 Apr 2010 06:49:39 +0000 (23:49 -0700)]
tiny cleanup

14 years agoSome SCons fixups
Dag Sverre Seljebotn [Fri, 23 Apr 2010 17:49:29 +0000 (19:49 +0200)]
Some SCons fixups

14 years agosupport arbitrarily sized typedef integral types
Lisandro Dalcin [Fri, 23 Apr 2010 16:02:07 +0000 (13:02 -0300)]
support arbitrarily sized typedef integral types

14 years agoreverted last commit - only a problem for closure scope
Stefan Behnel [Fri, 23 Apr 2010 13:00:34 +0000 (15:00 +0200)]
reverted last commit - only a problem for closure scope

14 years agosort ext type fields in GC functions to get a more predictable C code result
Stefan Behnel [Fri, 23 Apr 2010 12:41:32 +0000 (14:41 +0200)]
sort ext type fields in GC functions to get a more predictable C code result

14 years agofix signatures in numpy.pxd
Stefan Behnel [Fri, 23 Apr 2010 10:50:45 +0000 (12:50 +0200)]
fix signatures in numpy.pxd

14 years agoPyrexTypes: reworked ranking of numeric types and cleanup
Lisandro Dalcin [Thu, 22 Apr 2010 17:10:16 +0000 (14:10 -0300)]
PyrexTypes: reworked ranking of numeric types and cleanup