test Log


Commit Date  
[7be4db] by fwierzbicki@79b11f02-f210-0410-99aa-fb189b3d1ac4

Creating a branch for experimentation with invokedynamic (jsr 292).

2009-07-07 19:08:37 Tree
[ca8bf7] by zyasoft@79b11f02-f210-0410-99aa-fb189b3d1ac4

Created branch to rewrite PyString to use byte[] instead of String

2009-06-20 16:41:18 Tree
[a57fba] (unicodedata) by zyasoft@79b11f02-f210-0410-99aa-fb189b3d1ac4

Created branch for unicodedata implemented in Java

2009-06-20 16:40:29 Tree
[737ec4] by fwierzbicki@79b11f02-f210-0410-99aa-fb189b3d1ac4

Temporary home for 2.6 grammar + features while we concentrate on 2.5.1 in
trunk.

2009-06-17 21:09:27 Tree
[42c585] by zyasoft@79b11f02-f210-0410-99aa-fb189b3d1ac4

Creating branch for JSR223 dev

2009-05-01 03:08:07 Tree
[4391ec] by fijal@79b11f02-f210-0410-99aa-fb189b3d1ac4

Create a branch for experimenting with _rawffi implementation, that together
with _ctypes stolen from pypy gives ctypes

2009-04-29 02:40:41 Tree
[d89698] by zyasoft@79b11f02-f210-0410-99aa-fb189b3d1ac4

Creating newlist branch, to address #1026 (thread safety for lists)

2009-04-05 01:15:45 Tree
[04db96] (modjy) by fwierzbicki@79b11f02-f210-0410-99aa-fb189b3d1ac4

Merged revisions 5752,5756-5758,5761-5763,5765-5775,5778-5816,5818-5840,5842-5847,5854,5857-5859,5863-5865,5869,5875,5877,5884-5885,5889-5893,5900-5901,5903-5906,5908,5912-5918,5920,5923,5925,5927,5930-5932,5934-5947,5950-5951,5953-5957,5959,5961,5965-5966,5971-5976,5978-5979,5981-5983,5992,6001,6003,6006-6014,6023,6031-6037,6039-6040,6045-6047,6049-6052,6061-6065,6071 via svnmerge from
https://jython.svn.sourceforge.net/svnroot/jython/trunk/jython

2009-03-06 16:15:19 Tree
[23c38f] by zyasoft@79b11f02-f210-0410-99aa-fb189b3d1ac4

Python Bytecode VM branch

2009-01-20 13:47:20 Tree
[2432ec] by fwierzbicki@79b11f02-f210-0410-99aa-fb189b3d1ac4

Merged revisions 5900-5901,5903-5906,5908,5912-5918,5920,5923,5925,5927 via svnmerge from
https://jython.svn.sourceforge.net/svnroot/jython/trunk/jython

........
r5900 | fwierzbicki | 2009-01-09 10:16:08 -0500 (Fri, 09 Jan 2009) | 2 lines

Fix for wrong error line number when file ends with open paren.
........
r5901 | otmarhumbel | 2009-01-09 12:31:07 -0500 (Fri, 09 Jan 2009) | 1 line

reduce the size of the installer .jar by excluding external libraries we do not want to distribute
........
r5903 | fwierzbicki | 2009-01-09 15:40:11 -0500 (Fri, 09 Jan 2009) | 2 lines

Change release numbers in build.
........
r5904 | fwierzbicki | 2009-01-09 15:43:55 -0500 (Fri, 09 Jan 2009) | 2 lines

locked CPython to revision 68460.
........
r5905 | fwierzbicki | 2009-01-09 15:53:52 -0500 (Fri, 09 Jan 2009) | 2 lines

Update README
........
r5906 | pjenvey | 2009-01-09 19:49:02 -0500 (Fri, 09 Jan 2009) | 2 lines

fix broken import
........
r5908 | pedronis | 2009-01-10 07:39:15 -0500 (Sat, 10 Jan 2009) | 5 lines

yet another try at binop rule, with more exhaustive tests (verified against cpython) that now pass

see also PyPy issue 412
........
r5912 | cgroves | 2009-01-10 19:53:55 -0500 (Sat, 10 Jan 2009) | 6 lines

* Add xerces' version number to its file name for future explorers.
* Add a skip-brand property to build.xml that allows version branding to be
skipped, which keeps jython-dev.jar from being built even if no other files
change and shaves a couple seconds off a dev build.
........
r5913 | cgroves | 2009-01-10 20:03:33 -0500 (Sat, 10 Jan 2009) | 1 line

Compile Syspath.java for 1.5 instead of 1.6
........
r5914 | cgroves | 2009-01-10 20:12:16 -0500 (Sat, 10 Jan 2009) | 5 lines

Creating the empty.py file after the bytecode just meant that it had a more
recent modtime and would be preferred for importing, so it wasn't testing
anything.
........
r5915 | cgroves | 2009-01-10 22:21:25 -0500 (Sat, 10 Jan 2009) | 1 line

Cleanup
........
r5916 | cgroves | 2009-01-11 01:04:11 -0500 (Sun, 11 Jan 2009) | 6 lines

Convert to PyObject going into and Object coming out of java.util.Set
methods on PySet like PyDictionary does for Map methods and PyList does for
List methods. Without this, Java code can add non-PyObjects to the underlying
set causing the Python access to it to throw ClassCastExceptions.
........
r5917 | cgroves | 2009-01-11 01:24:45 -0500 (Sun, 11 Jan 2009) | 3 lines

Huh, I guess ECJ can handle inferring these generics, but javac sure can't.
........
r5918 | cgroves | 2009-01-11 05:35:10 -0500 (Sun, 11 Jan 2009) | 6 lines

In String compilation contexts, we can assume the bytes have already been
decoded and just pass them through when parsing instead of using ascii. This
gets test_doctest back to passing as it was before and keeps the pep 263
checks.
........
r5920 | cgroves | 2009-01-11 17:26:38 -0500 (Sun, 11 Jan 2009) | 6 lines

test320 - Moved to test_traceback_jy
test366,373 - Moved to test_java_integration
test391 - Moved to test_java_visibility
test395 - Tested by test_java_visibility
test396 - Tested by test_java_subclasses
........
r5923 | pjenvey | 2009-01-11 20:31:59 -0500 (Sun, 11 Jan 2009) | 1 line

bump copyright year
........
r5925 | cgroves | 2009-01-13 17:10:24 -0500 (Tue, 13 Jan 2009) | 1 line

Allow importation from bytecode in a zip file even if the source isn't present
........
r5927 | fwierzbicki | 2009-01-13 22:42:42 -0500 (Tue, 13 Jan 2009) | 4 lines

Small fix to grammar -- a "print" by itself should have it's NL attribute set
to true as CPython does. Because "false" doesn't make sense in this context
(it would be a no-op), the compiler doesn't check, so it passed unnoticed.
........

2009-01-14 03:58:51 Tree
[80a2bb] by fwierzbicki@79b11f02-f210-0410-99aa-fb189b3d1ac4

Merged revisions 5889-5893 via svnmerge from
https://jython.svn.sourceforge.net/svnroot/jython/trunk/jython

........
r5889 | pjenvey | 2009-01-08 16:23:29 -0500 (Thu, 08 Jan 2009) | 5 lines

jython-elementtree from: http://jython-elementtree.googlecode.com/svn/trunk@69
provides a partial emulation of pyexpat via xerces 2.9.1 and extra tests in
test_xml_etree_jy
from Sebastien Boisgerault
........
r5890 | pjenvey | 2009-01-08 16:26:19 -0500 (Thu, 08 Jan 2009) | 2 lines

xerces 2.9.1 for elementtree
........
r5891 | pjenvey | 2009-01-08 16:49:36 -0500 (Thu, 08 Jan 2009) | 2 lines

jarjar xerces into jython-complete and check for the mangled name
........
r5892 | pjenvey | 2009-01-08 17:05:28 -0500 (Thu, 08 Jan 2009) | 2 lines

unused import, coding standards
........
r5893 | otmarhumbel | 2009-01-08 17:54:04 -0500 (Thu, 08 Jan 2009) | 10 lines

rename the jython*.jar files:
jython.jar is now called jython-dev.jar
jython-complete.jar is now called jython.jar

the one we distribute is jython.jar,
containing the necessary external libraries, mangled

jython-dev.jar is intended for local developer usage,
containing no external libraries,
referencing the real external package names
........

2009-01-09 04:50:22 Tree
[fe5af4] by fwierzbicki@79b11f02-f210-0410-99aa-fb189b3d1ac4

Merged revisions 5842-5847,5854,5857-5859,5863-5865,5869 via svnmerge from
https://jython.svn.sourceforge.net/svnroot/jython/trunk/jython

........
r5842 | cgroves | 2009-01-04 17:15:09 -0500 (Sun, 04 Jan 2009) | 5 lines

Split java subclassing tests out of test_java_integration into
test_java_subclasses, move unduplicated bits of test_jsubclass into
test_java_subclasses.
........
r5843 | cgroves | 2009-01-04 18:06:27 -0500 (Sun, 04 Jan 2009) | 1 line

Add tests that have been failing or skipping for the past few months to the expected lists
........
r5844 | cgroves | 2009-01-04 18:20:34 -0500 (Sun, 04 Jan 2009) | 1 line

One more longstanding skip
........
r5845 | cgroves | 2009-01-04 22:56:25 -0500 (Sun, 04 Jan 2009) | 17 lines

test292 - Moved to test_java_integration
test293 - Importing no longer auto-imports submodules; deleted
test294 - Moved to test_java_subclasses
test295 - Moved to test_traceback
test296 - Moved to test_import_jy
test297,301,310,313,314,315,317 - Testing jythonc; deleted
test298 - Moved to test_import_jy
test299,300 - Tested by test_file
test303 - Moved to test_java_visibility
test307,308 - Moved to test_zipimport_jy
test309 - Tested by test_scope
test311 - Already disabled; deleted
test312 - Moved to test_sax_jy
test316 - Moved to test_java_integration
test318 - Tested by test_syntax
test319 - Moved to test_java_visibility
........
r5846 | cgroves | 2009-01-05 02:08:51 -0500 (Mon, 05 Jan 2009) | 1 line

From http://svn.python.org/projects/python/branches/release25-maint/Lib/test/test_doctest.py@59127
........
r5847 | cgroves | 2009-01-05 02:12:43 -0500 (Mon, 05 Jan 2009) | 10 lines

test302 - Move to test_pep263_jy

This adds a check for the encoding declared by the source file actually
matching up with what we were getting out of it, and switches the default
encoding to ascii to follow CPython. CPython assumes utf-8 in some single
compilation contexts, and this uses the same parser algorithm everywhere, so I
made a small change to test_doctest since Jython is throwing a SyntaxError on
utf-8 strings to a single compilation.
........
r5854 | amak | 2009-01-05 14:47:25 -0500 (Mon, 05 Jan 2009) | 2 lines

Fixed a bug where connect timeouts were not being honoured when set through socket.setdefaulttimeout()
http://bugs.jython.org/issue1218
........
r5857 | fwierzbicki | 2009-01-05 21:11:04 -0500 (Mon, 05 Jan 2009) | 2 lines

Remove some unused methods.
........
r5858 | fwierzbicki | 2009-01-06 15:57:41 -0500 (Tue, 06 Jan 2009) | 4 lines

Making the package names point to the raw original package names in source and
making the jarjar step process everything in one step so the package mangling
happens on our source as well as the included jars for jython-complete.jar.
........
r5859 | otmarhumbel | 2009-01-06 16:58:40 -0500 (Tue, 06 Jan 2009) | 2 lines

- using ${dist.dir} for both jython.jar and jython-complete.jar
- no unjar of jarjar.jar (since it does not exist any more)
........
r5863 | otmarhumbel | 2009-01-07 02:35:05 -0500 (Wed, 07 Jan 2009) | 2 lines

- added asm and constantine .jar files to the eclipse build path
- jarjar directory is not needed any more
........
r5864 | otmarhumbel | 2009-01-07 03:08:00 -0500 (Wed, 07 Jan 2009) | 5 lines

1) exlude jython.jar from the installation
this makes the installer autotests pass again
2) do not include the /Lib files into jython-complete.jar
once issue #1214 is fixed, 'java -jar jython-complete.jar' will run for all installation types
(for now, set the -Dpython.home variable if not in standalone mode)
........
r5865 | otmarhumbel | 2009-01-07 03:28:57 -0500 (Wed, 07 Jan 2009) | 1 line

preparation for an easy rename of the jython .jar files
........
r5869 | cgroves | 2009-01-07 13:12:46 -0500 (Wed, 07 Jan 2009) | 1 line

Use the path to the imported item as __file__, not the package path
........

2009-01-08 00:32:17 Tree
[bf1cde] by fwierzbicki@79b11f02-f210-0410-99aa-fb189b3d1ac4

Initial Jython 3000 branch -- don't get too excited everybody, the NetBeans
folks requested a Python 3000 grammar and I thought a branch would be the right
place to put it. I don't think it is quite time to get serious about a 3.0
just yet.

2009-01-04 20:36:10 Tree
[4bbd00] by fwierzbicki@79b11f02-f210-0410-99aa-fb189b3d1ac4

Branch to look at tighter integration of modjy.

2008-12-12 15:38:47 Tree
[645d3b] by fwierzbicki@79b11f02-f210-0410-99aa-fb189b3d1ac4

branch to create writable ast.

2008-11-17 21:20:25 Tree
[f6e4b8] by cgroves@79b11f02-f210-0410-99aa-fb189b3d1ac4

Branching for exposing Java classes as PyType and making proxies PyObject instead of PyInstance

2008-11-10 03:29:23 Tree
[b5df85] (advanced) by thobes@79b11f02-f210-0410-99aa-fb189b3d1ac4

Merged revisions 5287-5295,5297,5301-5305,5307-5308,5313-5317,5320-5321,5323-5325,5327-5329,5333,5339-5444,5446-5447,5452-5460,5462-5480,5482-5495 via svnmerge from
https://jython.svn.sourceforge.net/svnroot/jython/trunk/jython

........
r5287 | fwierzbicki | 2008-09-03 17:24:20 +0200 (Wed, 03 Sep 2008) | 3 lines

Small cleanup of compiler package. Mainly remove unused imports and rename
local vars that shadow member vars.
........
r5288 | fwierzbicki | 2008-09-04 18:57:30 +0200 (Thu, 04 Sep 2008) | 3 lines

Fix for some offset problems -- the biggest one being dedents now take on the
offset values of the previous token.
........
r5289 | fwierzbicki | 2008-09-05 05:36:04 +0200 (Fri, 05 Sep 2008) | 3 lines

Adjustment to offset fix. Overthought the offsets for simple
dedents.
........
r5290 | zyasoft | 2008-09-05 18:21:55 +0200 (Fri, 05 Sep 2008) | 4 lines

PySystemState.warnoptions was not initialized before adding to it in
the processing of the -W option.
........
r5291 | fwierzbicki | 2008-09-05 18:30:36 +0200 (Fri, 05 Sep 2008) | 2 lines

Broke test_traceback.py in my last commit. This fixes it again.
........
r5292 | fwierzbicki | 2008-09-06 06:28:11 +0200 (Sat, 06 Sep 2008) | 2 lines

Some tuning of node col and line positions to align better with CPython.
........
r5293 | fwierzbicki | 2008-09-06 20:31:14 +0200 (Sat, 06 Sep 2008) | 2 lines

Switch back installer25 -> installer. Leaked in from asm branch.
........
r5294 | fwierzbicki | 2008-09-06 20:58:02 +0200 (Sat, 06 Sep 2008) | 2 lines

increment version.
........
r5295 | fwierzbicki | 2008-09-06 21:34:07 +0200 (Sat, 06 Sep 2008) | 2 lines

Update README.
........
r5297 | fwierzbicki | 2008-09-06 22:57:46 +0200 (Sat, 06 Sep 2008) | 2 lines

Remove more leakage from asm branch into build.xml.
........
r5301 | fwierzbicki | 2008-09-07 07:57:26 +0200 (Sun, 07 Sep 2008) | 2 lines

Committing Nicholas Riley's patch after getting test_ast.py to work. See http://bugs.jython.org/issue1758279.
........
r5302 | otmarhumbel | 2008-09-08 16:41:41 +0200 (Mon, 08 Sep 2008) | 9 lines

Fix for issue #1123: Weird "unexpected at this time" error.

The tests if environment variables JAVA_HOME or JYTHON_HOME are set did not work:
- if the path contained a space
- if the variable really was not set

The passed (manual, so far) tests can be found in:
http://bugs.jython.org/msg3489
........
r5303 | leosoto | 2008-09-08 19:48:28 +0200 (Mon, 08 Sep 2008) | 1 line

SystemRandom implementation taken from CPython. It depends on urandom(), which seems supported now through jna-posix
........
r5304 | fwierzbicki | 2008-09-09 20:42:57 +0200 (Tue, 09 Sep 2008) | 2 lines

Fix some offset problems in DEDENT, INDENT, and EOF.
........
r5305 | fwierzbicki | 2008-09-09 21:06:17 +0200 (Tue, 09 Sep 2008) | 3 lines

from:
http://svn.python.org/projects/python/branches/release25-maint/Lib@66318
........
r5307 | fwierzbicki | 2008-09-09 23:15:54 +0200 (Tue, 09 Sep 2008) | 2 lines

Disable refcount test with is_jython check.
........
r5308 | otmarhumbel | 2008-09-10 00:34:55 +0200 (Wed, 10 Sep 2008) | 1 line

applied leosoto's patch from http://bugs.jython.org/issue1077 to avoid duplicate entries in jython-complete.jar
........
r5313 | fwierzbicki | 2008-09-10 20:36:16 +0200 (Wed, 10 Sep 2008) | 3 lines

Make it possible to get error nodes into the spots that are now
casting to exprType.
........
r5314 | otmarhumbel | 2008-09-11 00:45:07 +0200 (Thu, 11 Sep 2008) | 3 lines

since property cpythonlib.present is always true if jython is checked out correctly,
but is evaluated too early (long before checkout) during a full-build,
it has probably no usage any more
........
r5315 | otmarhumbel | 2008-09-11 01:06:04 +0200 (Thu, 11 Sep 2008) | 4 lines

reintroduced the possibility to suppress checkout in full-build
(setting do.checkout to 'false' does not prevent from checking out,
but not setting the property at all really does)
so full-build behaviour should remain the same
........
r5316 | fwierzbicki | 2008-09-11 02:37:30 +0200 (Thu, 11 Sep 2008) | 2 lines

update readme for Jython 2.5a3.
........
r5317 | fwierzbicki | 2008-09-11 02:39:43 +0200 (Thu, 11 Sep 2008) | 2 lines

Change defaults for alpha 3.
........
r5320 | fwierzbicki | 2008-09-11 21:48:35 +0200 (Thu, 11 Sep 2008) | 3 lines

Transform if/elif/else into the correct If nodes in the grammar file
instead of as a code post-processing step.
........
r5321 | fwierzbicki | 2008-09-11 22:18:43 +0200 (Thu, 11 Sep 2008) | 2 lines

Allow the possibility of using error nodes for bad stmtType.
........
r5323 | leosoto | 2008-09-13 19:04:58 +0200 (Sat, 13 Sep 2008) | 1 line

Test for the 'feature' described on #1089: cStringIO.StringIO.read always returns string objects
........
r5324 | amak | 2008-09-13 20:09:58 +0200 (Sat, 13 Sep 2008) | 11 lines

Checking in fixes for 3 bugs

http://bugs.jython.org/issue1119
- socket module has no attribute SO_ERROR

http://bugs.jython.org/issue1120
- invalid socket shutdown gives AssertionError, should be "transport endpoint not connected" socket.error

http://bugs.jython.org/issue1121
- listening socket shutdown expects the wrong kind of socket
........
r5325 | amak | 2008-09-13 20:31:07 +0200 (Sat, 13 Sep 2008) | 5 lines

Change of mind on behaviour when shutting down server sockets.

Instead to raising an exception, best to let the failure pass silently, as cpython does.

http://bugs.jython.org/issue1121
........
r5327 | zyasoft | 2008-09-13 21:02:55 +0200 (Sat, 13 Sep 2008) | 2 lines

Tests that __module__ is available during class definition time, for #1022
........
r5328 | zyasoft | 2008-09-14 02:08:28 +0200 (Sun, 14 Sep 2008) | 4 lines

Adds tests: test_copy, test_profilehooks, test_random, test_syntax, test_trace that should be reasonable for Jython (possibly with modifications)

Provides coverage for #1738411
........
r5329 | pjenvey | 2008-09-14 02:27:02 +0200 (Sun, 14 Sep 2008) | 1 line

small cleanup
........
r5333 | fwierzbicki | 2008-09-16 01:57:05 +0200 (Tue, 16 Sep 2008) | 2 lines

make accept and traverse no-ops for error nodes.
........
r5339 | fwierzbicki | 2008-09-18 21:07:35 +0200 (Thu, 18 Sep 2008) | 3 lines

Fixed another eof/whitespace parsing issue from Django. Turned all of these
problems into a new test: test_eof_jy.py.
........
r5340 | fwierzbicki | 2008-09-18 21:37:19 +0200 (Thu, 18 Sep 2008) | 3 lines

Unicode ast node is a kludge from long ago whose time has passed. I forgot to
remove it when I unified str node handling in the parser.
........
r5341 | leosoto | 2008-09-20 04:08:34 +0200 (Sat, 20 Sep 2008) | 1 line

PyType now exposes __repr__ and __str__ as documented on <http://wiki.python.org/jython/JythonDeveloperGuide/ImplementingStrAndRepr>. This fixes #1131
........
r5342 | leosoto | 2008-09-21 02:20:23 +0200 (Sun, 21 Sep 2008) | 1 line

importing test_support from the test package on test_exceptions.py
........
r5343 | pjenvey | 2008-09-24 21:44:25 +0200 (Wed, 24 Sep 2008) | 4 lines

it's more correct for StdoutWrapper to write the object being printed to the
file instead of __str__'ing it
refs #1130
........
r5344 | pjenvey | 2008-09-24 21:56:28 +0200 (Wed, 24 Sep 2008) | 1 line

revert r5343, it's broken
........
r5345 | fwierzbicki | 2008-09-24 23:05:39 +0200 (Wed, 24 Sep 2008) | 4 lines

Switched PythonPartial.g from being a "grammar" to being a "parser grammar",
plus some additional related cleanup. This fixes all of the missing Token
warnings. Thanks to Terrance Parr for suggesting this fix.
........
r5346 | thobes | 2008-09-25 00:12:23 +0200 (Thu, 25 Sep 2008) | 4 lines

Fixed the bug of the __exit__ method not being invoked on the outer context
manager in the case of nested with statements.
The problem was that the exception handlers were not properly sorted.
........
r5347 | thobes | 2008-09-25 20:52:30 +0200 (Thu, 25 Sep 2008) | 7 lines

Fix of the bug with "def f(): lambda x=(yield): 1".
The problem was that the order of instructions were such that there was
elements on the stack before the yield-point that does not get restored.
I also found another problem in that the types for local variables in
CodeCompiler are not proper java types, this is in error in nearly all places,
and needs to be fixed in a later commit.
........
r5348 | pjenvey | 2008-09-25 23:12:28 +0200 (Thu, 25 Sep 2008) | 5 lines

bump tarfile from 2.5.1 to 2.5.2, from:
http://svn.python.org/projects/python/branches/release25-maint/Lib
tarfile.py@60730
test/test_tarfile.py@60730
........
r5349 | pjenvey | 2008-09-25 23:16:30 +0200 (Thu, 25 Sep 2008) | 6 lines

reapply previous tarfile modifications:
o don't assume a reference counting GC
o ensure all file handles are closed to avoid os.remove problems on windows
o always use the tempdir, even when testtar is a relative path (it is when ran
via the regrtest)
........
r5350 | fwierzbicki | 2008-09-26 16:41:55 +0200 (Fri, 26 Sep 2008) | 3 lines

Correct this for new __class__.__name__ package truncating behavior of Java
classes.
........
r5351 | thobes | 2008-09-26 17:37:19 +0200 (Fri, 26 Sep 2008) | 4 lines

Fix to allow the yield statement to apear in the first iterator of a generator
expression. This exposes another problem where the stack state is not persisted
from yielding to resuming.
........
r5352 | fwierzbicki | 2008-09-26 19:01:58 +0200 (Fri, 26 Sep 2008) | 3 lines

Fixes an infinite loop that occurs when there is an unterminated triple string
and ListErrorHandler is being used.
........
r5353 | fwierzbicki | 2008-10-01 04:01:33 +0200 (Wed, 01 Okt 2008) | 3 lines

Added castSlice and castSlices to allow optional error nodes for sliceType.
Also renamed makeExpr(s) and makeStmt(s) to castExpr(s) and castStmt(s).
........
r5354 | pjenvey | 2008-10-02 02:13:21 +0200 (Thu, 02 Okt 2008) | 2 lines

remove more remnants of org.python.antlr.ast.Unicode
........
r5355 | pjenvey | 2008-10-02 02:16:29 +0200 (Thu, 02 Okt 2008) | 1 line

whitespace
........
r5356 | pjenvey | 2008-10-02 04:31:03 +0200 (Thu, 02 Okt 2008) | 3 lines

from:
http://svn.python.org/projects/python/branches/release25-maint/Lib/test/test_asynchat.py@46906
........
r5357 | pjenvey | 2008-10-05 02:00:24 +0200 (Sun, 05 Okt 2008) | 3 lines

avoid test_asynchat on BSD where it deadlocks
refs #1064
........
r5358 | pjenvey | 2008-10-05 02:01:52 +0200 (Sun, 05 Okt 2008) | 1 line

fix mislabeled names
........
r5359 | zyasoft | 2008-10-06 05:02:29 +0200 (Mon, 06 Okt 2008) | 3 lines

From
http://svn.python.org/projects/python/branches/release25-maint/Lib/test/test_random.py@53510
........
r5360 | zyasoft | 2008-10-06 05:12:34 +0200 (Mon, 06 Okt 2008) | 6 lines

Removed older version of random.py in favor of CPythonLib's; added
_random.getrandbits (in PyRandom); modified test_random so that it
doesn't test implementation specific aspects of the random generator,
since we actually use java.util.Random, not the Mersenne Twister.
........
r5361 | zyasoft | 2008-10-06 20:57:58 +0200 (Mon, 06 Okt 2008) | 6 lines

Fixed test_quopri by encoding space or tab before newline and
resetting the column count upon a newline

Fixed #1144 so that -mMODULE works and remaining args are passed in sys.argv
........
r5362 | zyasoft | 2008-10-06 22:42:04 +0200 (Mon, 06 Okt 2008) | 3 lines

From
http://svn.python.org/projects/python/branches/release25-maint/Lib/test/test_trace.py@60582
........
r5363 | zyasoft | 2008-10-06 23:48:03 +0200 (Mon, 06 Okt 2008) | 6 lines

Removed test cases not applicable to Jython. Although it's not
possible to implement jump when compiled to Java bytecode, it's
possible that we can better align our trace with what is done in
CPython, so this should be revisited at some point.
........
r5364 | zyasoft | 2008-10-07 00:00:46 +0200 (Tue, 07 Okt 2008) | 3 lines

From
http://svn.python.org/projects/python/branches/release25-maint/Lib/test/test_profilehooks.py@35570
........
r5365 | zyasoft | 2008-10-07 00:08:13 +0200 (Tue, 07 Okt 2008) | 5 lines

Similar mass removal of tests as in test_trace, these all are related
to some fundamental differences in our VM, especially around exception
handling.

........
r5366 | pjenvey | 2008-10-10 01:23:05 +0200 (Fri, 10 Okt 2008) | 3 lines

fix lambdas within decorator args causing an NPE. we weren't initializing their
scopes
........
r5367 | fwierzbicki | 2008-10-10 20:12:58 +0200 (Fri, 10 Okt 2008) | 2 lines

Cleaning up dead code.
........
r5368 | zyasoft | 2008-10-11 04:24:42 +0200 (Sat, 11 Okt 2008) | 3 lines

From
http://svn.python.org/projects/python/branches/release25-maint/Lib/test/test_copy.py@42573
........
r5369 | zyasoft | 2008-10-11 04:38:55 +0200 (Sat, 11 Okt 2008) | 9 lines

Make "extended" builtin functions in __builtin__ (that is, functions
like max which use a class like MaxFunction for their definition)
copyable by copy.copy and return the correct type.

This fixes test_copy, along with enabling reflexive structures to be
compared, since we don't prohibit or otherwise detect them in the
copying anyway.
........
r5370 | zyasoft | 2008-10-11 05:37:39 +0200 (Sat, 11 Okt 2008) | 2 lines

Patch from #1075, thanks Matt Boersma!
........
r5371 | pjenvey | 2008-10-11 07:08:23 +0200 (Sat, 11 Okt 2008) | 3 lines

disable test_compiler as we don't intend to support the parser module compiler
relies on, with _ast being the way forward
........
r5372 | pjenvey | 2008-10-11 08:42:35 +0200 (Sat, 11 Okt 2008) | 3 lines

fix solid_base potentially resolving the wrong type when called before the mro
was initialized. caused best_base to blowup with a TypeError
........
r5373 | fwierzbicki | 2008-10-11 22:54:47 +0200 (Sat, 11 Okt 2008) | 3 lines

Throw a Python exception for two internal compiler errors that where just
printing a less useful error message and going on to an NPE.
........
r5374 | fwierzbicki | 2008-10-11 23:29:29 +0200 (Sat, 11 Okt 2008) | 11 lines

A very basic start to supporting pep 328 style relative imports like

from . import foo

This patch just passes the "level" (that is number of dots) information from
the compile to the import machinary without actually using it, and forbids

from . import *

As Python 2.5 does.
........
r5375 | pjenvey | 2008-10-12 01:16:54 +0200 (Sun, 12 Okt 2008) | 2 lines

fix unicodedata.lookup to return unicode instead of the codepoint int
........
r5376 | zyasoft | 2008-10-12 02:18:38 +0200 (Sun, 12 Okt 2008) | 3 lines

From
http://svn.python.org/projects/python/branches/release25-maint/Lib/test/test_struct.py@60892
........
r5377 | zyasoft | 2008-10-12 02:24:30 +0200 (Sun, 12 Okt 2008) | 2 lines

Implemented struct.Struct as well as other updates in the struct module for 2.5
........
r5378 | zyasoft | 2008-10-12 02:25:07 +0200 (Sun, 12 Okt 2008) | 2 lines

Missing module for previous commit!
........
r5379 | fwierzbicki | 2008-10-12 18:28:15 +0200 (Sun, 12 Okt 2008) | 7 lines

More groundwork for absolute_import. Passing level info into most of the
places it needs to go, also now sending the correct level information when
"from __future__ import absolute_import" is issued. Added a constant in
imp.java DEFAULT_LEVEL, so that it can be changed in one place when absolute
import becomes the default behavior.
........
r5380 | pjenvey | 2008-10-12 22:10:04 +0200 (Sun, 12 Okt 2008) | 5 lines

fix bad bytecode generated for:
a[[b for b, c in d]] = e
the store was losing track of the value (e) here because the list comp/tuple
unpack combo overwrote where e's location was held
........
r5381 | pjenvey | 2008-10-13 00:18:06 +0200 (Mon, 13 Okt 2008) | 2 lines

better unhashable TypeError messages
........
r5382 | pjenvey | 2008-10-13 00:37:53 +0200 (Mon, 13 Okt 2008) | 2 lines

update the deriveds for better unhashable TypeError messages
........
r5383 | pjenvey | 2008-10-13 01:00:34 +0200 (Mon, 13 Okt 2008) | 1 line

fix the renamed _collections
........
r5384 | fwierzbicki | 2008-10-13 19:19:47 +0200 (Mon, 13 Okt 2008) | 2 lines

Fixed issue http://bugs.jython.org/issue1111: keyword arguments not supported on __import__
........
r5385 | fwierzbicki | 2008-10-13 19:31:05 +0200 (Mon, 13 Okt 2008) | 4 lines

disable test_transformer for the same reason we disabled test_compiler: we
don't intend to support the parser module compiler relies on, with _ast being
the way forward. transformer.py is in the same area of code.
........
r5386 | fwierzbicki | 2008-10-13 20:50:06 +0200 (Mon, 13 Okt 2008) | 7 lines

Fixed "expected skips" that wheren't actually being skipped.

Also added test_peepholer and test_profile to expected failures since these are
platform specific tests. This is a work around, we should (most likely) go
back and force these expected failures to skip in Jython, but I want to do that
in one future sweep.
........
r5387 | fwierzbicki | 2008-10-13 21:41:45 +0200 (Mon, 13 Okt 2008) | 3 lines

from
http://svn.python.org/projects/python/branches/release25-maint/Lib/test/test_with.py@66619
........
r5388 | fwierzbicki | 2008-10-13 22:38:36 +0200 (Mon, 13 Okt 2008) | 4 lines

Disabling tests of "as" and "with" semi-keyword status, for now it's too hard
to get antlr to treat them as almost-keywords. I want to revisit, but this is
a low priority compared to other problems.
........
r5389 | cgroves | 2008-10-14 03:47:30 +0200 (Tue, 14 Okt 2008) | 1 line

Jarjar if antlr changed more recently than the jarjar'd jar. Jar.
........
r5390 | leosoto | 2008-10-14 05:20:10 +0200 (Tue, 14 Okt 2008) | 1 line

Now String is mapped to PyUnicode on ClassicPyObjectAdapter (instead of mapping it to PyString). Thus, for APIs which really need to return bytestrings (such as PyFile) we now have to explicitely declare the return type as PyString (and on code written in Python, use StringUtil.asPyString(String)). Fixes #1128
........
r5391 | fwierzbicki | 2008-10-14 18:43:46 +0200 (Tue, 14 Okt 2008) | 4 lines

Fix or skips for most remaining problems in test_syntax.py. All remaining
problems are due to the lack of enforcement on continuations not being
permitted in finally.
........
r5392 | fwierzbicki | 2008-10-14 21:28:28 +0200 (Tue, 14 Okt 2008) | 4 lines

Added inFinally parameter to suite rule, and a dynamically scoped parameter
continueIllegal to support disallowing continue in a finally clause. This
fixes test_syntax.
........
r5393 | pjenvey | 2008-10-15 01:08:45 +0200 (Wed, 15 Okt 2008) | 5 lines

o fix the socket hostname functions returning unicode hostnames instead of
strs. was causing test_wsgiref to fail on environments where
socket.getfqdn('localhost') != 'localhost'
o fix getaddrinfo default family and handling of bad familys
........
r5394 | pjenvey | 2008-10-15 02:46:02 +0200 (Wed, 15 Okt 2008) | 2 lines

fix solid_base considering slotted objects with a __dict__ as solid
........
r5395 | cgroves | 2008-10-15 06:42:15 +0200 (Wed, 15 Okt 2008) | 8 lines

Remove build/exposed from the classpath. Having it in there meant any source
navigation like F3 or Open Type would go to the exposed version rather than the
source version that you were really after. Not having it in there means you
need to add the exposed dir to the classpath of any run configurations in
Eclipse, but that felt like a lesser evil than my smiting my keyboard in
frustration every time Eclipse opened the wrong file.
........
r5396 | otmarhumbel | 2008-10-15 07:35:53 +0200 (Wed, 15 Okt 2008) | 1 line

fixed typo: _JAVA_HOME has to be JAVA_HOME
........
r5397 | otmarhumbel | 2008-10-15 10:10:37 +0200 (Wed, 15 Okt 2008) | 7 lines

safely test if JAVA_HOME or JYTHON_HOME is set

we can now handle both quoted and unquoted settings like:
JAVA_HOME=C:\Program Files\Java\jdk1.6.0_07
JAVA_HOME="C:\Program Files\Java\jdk1.6.0_07"

this should fix issue #1125
........
r5398 | otmarhumbel | 2008-10-15 17:56:58 +0200 (Wed, 15 Okt 2008) | 1 line

applied lsoto's PyString change (revision 5390) to this handler as well
........
r5399 | fwierzbicki | 2008-10-15 19:06:58 +0200 (Wed, 15 Okt 2008) | 6 lines

New dotted and absolute import support. I have local tests, but have not
figured out how to integrate them into the regression tests. Examining the
tests in CPython does not appear to give much guidance here, the testing of
these features looks pretty thin (mainly they test features around __package__
which is new in 2.6). Maybe I just haven't found the relevant tests?
........
r5400 | fwierzbicki | 2008-10-15 22:10:46 +0200 (Wed, 15 Okt 2008) | 2 lines

? was replaced by <module> in info.
........
r5401 | fwierzbicki | 2008-10-15 22:16:53 +0200 (Wed, 15 Okt 2008) | 5 lines

Guard the Java-specific top imports with an is_jython. This way some of the
bugtests can run in pure python to make it easier to tilt at this windmill:
http://code.google.com/p/google-highly-open-participation-psf/source/browse/trunk/submissions/jython-tests-26-50.txt?r=424
Windmill summary: Turn all bugtests into real unit tests or kill.
........
r5402 | fwierzbicki | 2008-10-15 22:22:02 +0200 (Wed, 15 Okt 2008) | 2 lines

Move bugtests/test076.py into regrtest.
........
r5403 | fwierzbicki | 2008-10-15 22:23:35 +0200 (Wed, 15 Okt 2008) | 2 lines

test077.py is well tested in our current unit tests.
........
r5404 | fwierzbicki | 2008-10-15 22:35:18 +0200 (Wed, 15 Okt 2008) | 2 lines

test078.py is outdated.
........
r5405 | fwierzbicki | 2008-10-15 22:37:54 +0200 (Wed, 15 Okt 2008) | 2 lines

"find" is well tested and doesn't need this extra test.
........
r5406 | fwierzbicki | 2008-10-15 22:41:18 +0200 (Wed, 15 Okt 2008) | 3 lines

test_pkgimport.py already does a reasonable job of checking multiple
imports of the same .py file.
........
r5407 | fwierzbicki | 2008-10-15 22:42:47 +0200 (Wed, 15 Okt 2008) | 2 lines

cleanup dead code.
........
r5408 | fwierzbicki | 2008-10-15 22:44:08 +0200 (Wed, 15 Okt 2008) | 2 lines

This basic re functionality is well tested in our existing unit tests.
........
r5409 | fwierzbicki | 2008-10-15 22:52:30 +0200 (Wed, 15 Okt 2008) | 4 lines

test089.py looks like it is very confused :)

Maybe string.replace changed oh so long ago? /me shrugs and deletes.
........
r5410 | fwierzbicki | 2008-10-15 22:55:46 +0200 (Wed, 15 Okt 2008) | 6 lines

Comment at the top

Does not work, will never work.

Wow. Deleted.
........
r5411 | fwierzbicki | 2008-10-16 03:48:49 +0200 (Thu, 16 Okt 2008) | 2 lines

another "will never work" test. Deleted.
........
r5412 | fwierzbicki | 2008-10-16 03:53:07 +0200 (Thu, 16 Okt 2008) | 3 lines

Just tests the existance of a __name__ and __doc__ for the exceptions module.
Not enough of a test to preserve.
........
r5413 | fwierzbicki | 2008-10-16 03:53:49 +0200 (Thu, 16 Okt 2008) | 2 lines

deleting completely empty test.
........
r5414 | fwierzbicki | 2008-10-16 04:21:51 +0200 (Thu, 16 Okt 2008) | 2 lines

Moved unique id test to test_jy_internals
........
r5415 | fwierzbicki | 2008-10-16 04:56:10 +0200 (Thu, 16 Okt 2008) | 2 lines

First get test102.py working again...
........
r5416 | fwierzbicki | 2008-10-16 05:11:28 +0200 (Thu, 16 Okt 2008) | 2 lines

test102 moved to test_jy_internals.py
........
r5417 | fwierzbicki | 2008-10-16 05:14:39 +0200 (Thu, 16 Okt 2008) | 2 lines

Another empty test.
........
r5418 | fwierzbicki | 2008-10-16 05:15:27 +0200 (Thu, 16 Okt 2008) | 2 lines

Another "does not work"
........
r5419 | pjenvey | 2008-10-16 05:28:52 +0200 (Thu, 16 Okt 2008) | 2 lines

another unicode/str regression: co_name and co_filename should be a str
........
r5420 | pjenvey | 2008-10-16 08:38:43 +0200 (Thu, 16 Okt 2008) | 3 lines

fix os.access not handling multiple modes, and make its X_OK work via
jna-posix's stat
........
r5421 | fwierzbicki | 2008-10-16 14:07:24 +0200 (Thu, 16 Okt 2008) | 3 lines

Move essential part of test106 to test_jy_compile.py (bare return not permitted
in exec).
........
r5422 | fwierzbicki | 2008-10-16 14:15:26 +0200 (Thu, 16 Okt 2008) | 2 lines

First make test107 pass in both python and jython
........
r5423 | fwierzbicki | 2008-10-16 14:42:06 +0200 (Thu, 16 Okt 2008) | 2 lines

move test107 to test_jy_internals.py
........
r5424 | fwierzbicki | 2008-10-16 16:26:23 +0200 (Thu, 16 Okt 2008) | 2 lines

Moved test108 to test_dict_jy
........
r5425 | fwierzbicki | 2008-10-16 16:27:55 +0200 (Thu, 16 Okt 2008) | 2 lines

This basic functionality is well tested.
........
r5426 | fwierzbicki | 2008-10-16 16:29:48 +0200 (Thu, 16 Okt 2008) | 2 lines

Basic dict functionality that is well tested in the unit tests.
........
r5427 | fwierzbicki | 2008-10-16 16:31:14 +0200 (Thu, 16 Okt 2008) | 3 lines

Pretty much just a test for lists being disallowed as a key which is already in
the unit tests.
........
r5428 | fwierzbicki | 2008-10-16 16:33:33 +0200 (Thu, 16 Okt 2008) | 2 lines

Basic slice ops that are already well tested.
........
r5429 | fwierzbicki | 2008-10-16 16:34:25 +0200 (Thu, 16 Okt 2008) | 2 lines

This test doesn't make much sense.
........
r5430 | fwierzbicki | 2008-10-16 16:36:02 +0200 (Thu, 16 Okt 2008) | 2 lines

basic test of int subclasses which is well covered by current tests.
........
r5431 | fwierzbicki | 2008-10-16 18:13:52 +0200 (Thu, 16 Okt 2008) | 2 lines

Leftover for long-gone test049.
........
r5432 | fwierzbicki | 2008-10-16 18:18:04 +0200 (Thu, 16 Okt 2008) | 2 lines

move test124 to test_jy_internals.
........
r5433 | pjenvey | 2008-10-16 18:54:06 +0200 (Thu, 16 Okt 2008) | 1 line

typo
........
r5434 | fwierzbicki | 2008-10-16 19:48:03 +0200 (Thu, 16 Okt 2008) | 3 lines

The reload builtin is tested in a number of places in the unit tests, so this
basic test of it can be deleted.
........
r5435 | fwierzbicki | 2008-10-16 19:48:56 +0200 (Thu, 16 Okt 2008) | 2 lines

Very basic test of assert that is no longer needed as assert is well tested.
........
r5436 | fwierzbicki | 2008-10-16 19:50:37 +0200 (Thu, 16 Okt 2008) | 2 lines

This test is already covered by a test in test_jy_internals.
........
r5437 | fwierzbicki | 2008-10-16 19:51:59 +0200 (Thu, 16 Okt 2008) | 2 lines

Covered in test_jy_internals already.
........
r5438 | fwierzbicki | 2008-10-16 20:56:04 +0200 (Thu, 16 Okt 2008) | 2 lines

moved test130 to test_cmp_jy.
........
r5439 | fwierzbicki | 2008-10-16 20:58:39 +0200 (Thu, 16 Okt 2008) | 2 lines

This is well tested in test_hexoct.py
........
r5440 | fwierzbicki | 2008-10-16 22:19:50 +0200 (Thu, 16 Okt 2008) | 2 lines

moved test131 into test_java_integration
........
r5441 | fwierzbicki | 2008-10-16 22:22:15 +0200 (Thu, 16 Okt 2008) | 2 lines

test_glob thoroughly tests glob.
........
r5442 | fwierzbicki | 2008-10-16 22:24:04 +0200 (Thu, 16 Okt 2008) | 3 lines

This really just tests that copy.copy exists. We run the real test_copy and
so have no need for this test.
........
r5443 | fwierzbicki | 2008-10-16 22:32:40 +0200 (Thu, 16 Okt 2008) | 2 lines

moved test136 to test_java_integration.py
........
r5444 | fwierzbicki | 2008-10-16 22:34:32 +0200 (Thu, 16 Okt 2008) | 2 lines

better name for test method.
........
r5446 | cgroves | 2008-10-16 22:59:53 +0200 (Thu, 16 Okt 2008) | 9 lines

Use assertRaises instead of
try:
code
except ExpectedException:
pass
else:
self.fail
........
r5447 | fwierzbicki | 2008-10-16 23:15:59 +0200 (Thu, 16 Okt 2008) | 3 lines

put in null guard for cflags call. cflags can definitely be null at this
point in the code.
........
r5452 | pjenvey | 2008-10-17 02:04:42 +0200 (Fri, 17 Okt 2008) | 2 lines

minor xrange optimization: use the small int cache
........
r5453 | cgroves | 2008-10-17 09:06:39 +0200 (Fri, 17 Okt 2008) | 1 line

A couple slices from Occam's razor, and speedups for items and keys by making arrays for PyList directly instead of making Lists and letting PyList turn them into arrays
........
r5454 | thobes | 2008-10-17 20:15:50 +0200 (Fri, 17 Okt 2008) | 3 lines

Modifying the generated bytecode so that yield as an expression does not empty
the stack.
........
r5455 | thobes | 2008-10-17 21:28:20 +0200 (Fri, 17 Okt 2008) | 2 lines

Removed some erronious code in visitYield.
........
r5456 | thobes | 2008-10-17 22:46:47 +0200 (Fri, 17 Okt 2008) | 4 lines

Improved the scope analysis for generator expressions.
Also fixed a bug in my recent commit where the fix for yield as an expression
accidentially reordered the value to be yielded on the stack.
........
r5457 | thobes | 2008-10-17 23:21:36 +0200 (Fri, 17 Okt 2008) | 4 lines

Added a check for return and yield in the same scope in ScopesCompiler,
There are cases that can be missed in CodeCompiler due to dead code
elimination, this gets around that.
........
r5458 | pjenvey | 2008-10-18 01:55:52 +0200 (Sat, 18 Okt 2008) | 3 lines

allow more heap for the regrtest to avoid the OutOfMemory errors. we don't seem
to actually be leaking -- we just have more test modules
........
r5459 | pjenvey | 2008-10-18 01:56:07 +0200 (Sat, 18 Okt 2008) | 2 lines

remove stray print statement that was preventing this from passing
........
r5460 | pjenvey | 2008-10-18 03:58:29 +0200 (Sat, 18 Okt 2008) | 4 lines

store the symbol name info in a LinkedHashMap to maintain the order they're
encountered in. fixes test_code
technically this should bump the bytecode magic but it's not that important
........
r5462 | pjenvey | 2008-10-18 22:23:01 +0200 (Sat, 18 Okt 2008) | 3 lines

fix os.access to never raise an exception with a valid mode, which I broke in
r5420
........
r5463 | pjenvey | 2008-10-19 02:01:20 +0200 (Sun, 19 Okt 2008) | 7 lines

fix forgetting to free the var/starargs of Calls, which produced bad byte for:
def gen():
if True:
# or b(**c)
a = b(*c)
yield d
........
r5464 | cgroves | 2008-10-19 05:04:44 +0200 (Sun, 19 Okt 2008) | 1 line

Some formatting and cleanup
........
r5465 | cgroves | 2008-10-19 07:14:22 +0200 (Sun, 19 Okt 2008) | 8 lines

Rename PyBuiltinFunction to PyBuiltinCallable, and remake PyBuiltinFunction as
a base for functions with the basic function stuff from PyBuiltinFunctionSet.
Extend that with PyBuiltinFunctionNarrow for functions with a fixed set of args
to match PyBuiltinMethodNarrow. Use the new function stuff to implement the
builtin functions that were using ExtendedBuiltinFunction as real functions
instead of PyObjects.
........
r5466 | cgroves | 2008-10-19 07:17:07 +0200 (Sun, 19 Okt 2008) | 1 line

Missed a sppot
........
r5467 | pjenvey | 2008-10-19 08:51:07 +0200 (Sun, 19 Okt 2008) | 3 lines

fix visitCompare leaving the left side on the stack, causing a minor short term
memory leak
........
r5468 | zyasoft | 2008-10-19 17:16:29 +0200 (Sun, 19 Okt 2008) | 3 lines

from
http://svn.python.org/projects/python/branches/release25-maint/Lib/test/test_codeccallbacks.py@46456
........
r5469 | pjenvey | 2008-10-19 22:51:56 +0200 (Sun, 19 Okt 2008) | 3 lines

move the last visitCompare fix to after the end label so we also avoid the
issue when a chained comparison fails fast
........
r5470 | zyasoft | 2008-10-20 00:04:51 +0200 (Mon, 20 Okt 2008) | 5 lines

Make error handling in codecs so that it is also surrogate-aware.
Fixes test_codeccallbacks

Skip tests from test_threading not applicable to Jython.
........
r5471 | cgroves | 2008-10-20 01:17:44 +0200 (Mon, 20 Okt 2008) | 1 line

Add a couple tests for subclass constructor visibility
........
r5472 | zyasoft | 2008-10-20 01:18:22 +0200 (Mon, 20 Okt 2008) | 7 lines

PEP 342 specifies that an exception raised by close during
finalization should be output to stderr, so we now do that.

Fixed a minor doctest output formatting issue so that the desired
syntax error is properly seen.
........
r5473 | cgroves | 2008-10-20 01:30:32 +0200 (Mon, 20 Okt 2008) | 1 line

Replace StringUtil.asPyString with Py.newString
........
r5474 | pjenvey | 2008-10-20 02:44:52 +0200 (Mon, 20 Okt 2008) | 2 lines

bump java's stack size to 768k on solaris, to fix test_cpickle there
........
r5475 | cgroves | 2008-10-20 03:01:39 +0200 (Mon, 20 Okt 2008) | 8 lines

Add an ant resource collection that's like union except it combines based on
the name of the resource instead of its full path. Use this to exclude the
unexposed version of classes from jython.jar to keep it quiet.

Resource collections didn't appear until Ant 1.7, but that came out nearly 2
years ago, so I'm hoping most people already have it.
........
r5476 | cgroves | 2008-10-20 03:40:25 +0200 (Mon, 20 Okt 2008) | 1 line

Remove newcompiler since it's not being used now and it's going to be replaced by the advanced compiler
........
r5477 | cgroves | 2008-10-20 03:41:32 +0200 (Mon, 20 Okt 2008) | 1 line

Move BaseTypeBuilder out of TypeExposer so the compile time bits of the exposing system can be excluded from jython.jar
........
r5478 | cgroves | 2008-10-20 04:03:44 +0200 (Mon, 20 Okt 2008) | 1 line

Cleanup, modernization
........
r5479 | cgroves | 2008-10-20 08:29:50 +0200 (Mon, 20 Okt 2008) | 9 lines

The pain of not having the generic inferring statics from Google collections
finally got to the point where I felt the need to whip up version for Jython.
Add Generic with static methods to make Lists, Maps and Sets that infer their
generic types from the generics of whatever they're being assigned to. I went
for brevity in naming the methods as I felt they'd be idiomatic enough in
Jython to read well, but could be convinced that more descriptive names -
newArrayList() instead of list() for example - would be better.
........
r5480 | thobes | 2008-10-20 13:13:38 +0200 (Mon, 20 Okt 2008) | 3 lines

Fixed the evaluation order bug made visual by test_generators.
Added a test file for evaluation order tests.
........
r5482 | thobes | 2008-10-20 19:19:57 +0200 (Mon, 20 Okt 2008) | 5 lines

Added support for class decorators.
The test file added for this (test_classdecorators) should be removed when we
import test_decorators from CPython 3.x, because that is where the caes come
from. This is also documented in that file.
........
r5483 | pjenvey | 2008-10-20 22:01:14 +0200 (Mon, 20 Okt 2008) | 2 lines

make sock.listen() explicitly require the backlog argument to match CPython
........
r5484 | pjenvey | 2008-10-20 22:05:02 +0200 (Mon, 20 Okt 2008) | 4 lines

revert r5474 and bump Java's stack size to 1024k, matching 64 bit Java's
default. 32 bit needs an increase to at least 512k to pass test_cpickle, but we
don't want to shrink 64 bit's default
........
r5485 | pjenvey | 2008-10-20 23:18:08 +0200 (Mon, 20 Okt 2008) | 5 lines

o delete references to test.test_ modules in the test package after running
them to conserve memory
o bump ant regrtest's heap to 144m which is about the minimum needed for
Apple's 64 bit Java 6 (server VM)
........
r5486 | fwierzbicki | 2008-10-20 23:37:06 +0200 (Mon, 20 Okt 2008) | 6 lines

PythonPartial.g should allow decorators without classes or functions so that

@foo

is a valid partial sentence.
........
r5487 | pjenvey | 2008-10-21 01:53:47 +0200 (Tue, 21 Okt 2008) | 2 lines

__iter__ is just another file_self call
........
r5488 | pjenvey | 2008-10-21 01:54:14 +0200 (Tue, 21 Okt 2008) | 3 lines

allow abuse of _fileobject by fixing its close() to act like CPython's when the
underlying _sock isn't a Jython socket. fixes test_urllib2net
........
r5489 | fwierzbicki | 2008-10-21 02:21:45 +0200 (Tue, 21 Okt 2008) | 2 lines

moved test138 to test_str_jy.
........
r5490 | fwierzbicki | 2008-10-21 02:24:00 +0200 (Tue, 21 Okt 2008) | 2 lines

spelling fix.
........
r5491 | fwierzbicki | 2008-10-21 02:35:49 +0200 (Tue, 21 Okt 2008) | 3 lines

Charlie Groves pushed me over the "maybe the test is just dumb?" barrier.
........
r5492 | fwierzbicki | 2008-10-21 02:38:12 +0200 (Tue, 21 Okt 2008) | 3 lines

This one looks like it is testing multiple inheritance of Java classes which
is long gone. Also it imports a module which shouldn't exist. Deleting.
........
r5493 | fwierzbicki | 2008-10-21 02:40:02 +0200 (Tue, 21 Okt 2008) | 2 lines

These three tests are well covered by existing tests.
........
r5494 | fwierzbicki | 2008-10-21 02:40:55 +0200 (Tue, 21 Okt 2008) | 2 lines

This is well tested by existing unit tests.
........
r5495 | fwierzbicki | 2008-10-21 02:47:13 +0200 (Tue, 21 Okt 2008) | 2 lines

Doesn't really test anything...
........

2008-10-21 10:53:16 Tree
[58f161] (jythonc) by fwierzbicki@79b11f02-f210-0410-99aa-fb189b3d1ac4

Creating an experimental branch to mess around with some jythonc ideas
in code.

2008-09-16 18:45:58 Tree
[4a62b9] by fwierzbicki@79b11f02-f210-0410-99aa-fb189b3d1ac4

Create branch to combine Python.g and PythonWalker.g

2008-08-19 21:17:36 Tree
[14bd16] by zyasoft@79b11f02-f210-0410-99aa-fb189b3d1ac4

Fixed PyUnicode.unicode_center, unicode_left, unicode_right. Merged revisions 4182,4185-4193,4195,4197-4218,4220-4221,4223-4233,4236,4242-4251,4264-4265,4268-4282,4287,4289-4290,4292,4301-4309,4312-4321,4324,4327-4371,4374,4379,4391,4393-4401,4404,4406,4409-4415,4419-4422,4426-4432,4435-4439,4442-4443,4445,4447,4449,4453-4454,4456,4458,4460,4462,4464,4467-4471,4473,4475-4486,4488-4489,4491-4492,4495,4497,4502-4503,4505-4510,4513-4520,4522,4525,4527-4528,4530,4534-4539,4541-4542,4544,4546,4550,4552-4564,4566-4576,4580,4583,4590-4595,4597-4598,4601-4602,4609-4612,4614-4627,4629-4631,4635-4639,4641,4649-4650,4663-4664,4666,4668-4671,4675,4681,4685,4687-4691,4696-4697,4699-4700,4703,4707,4713 via svnmerge from
https://jython.svn.sourceforge.net/svnroot/jython/trunk/jython

........
r4182 | pjenvey | 2008-02-26 16:04:15 -0700 (Tue, 26 Feb 2008) | 2 lines

include cgitb, it seems to work
........
r4185 | cgroves | 2008-02-28 09:55:33 -0700 (Thu, 28 Feb 2008) | 1 line

Tabs to spaces
........
r4186 | pjenvey | 2008-03-02 22:10:24 -0700 (Sun, 02 Mar 2008) | 1 line

fix os.lstat('/')
........
r4187 | pjenvey | 2008-03-04 16:49:26 -0700 (Tue, 04 Mar 2008) | 1 line

include pkgutil
........
r4188 | pjenvey | 2008-03-05 15:36:49 -0700 (Wed, 05 Mar 2008) | 4 lines

fix importFrom hiding nested ImportErrors by moving the PyModule impHook logic
into imp. also remove the magic that automatically imports a module when it's
looked up via its parent package's __getattr__
........
r4189 | pjenvey | 2008-03-05 19:45:45 -0700 (Wed, 05 Mar 2008) | 3 lines

make unsubscriptable[1] raise a TypeError instead of AttributeError and fix
deque's workaround for this
........
r4190 | pjenvey | 2008-03-05 20:57:51 -0700 (Wed, 05 Mar 2008) | 2 lines

handle overflows in lshift/rshift
........
r4191 | pjenvey | 2008-03-06 00:23:43 -0700 (Thu, 06 Mar 2008) | 3 lines

handle another overflow case in lshift/rshift. workaround the fact that lshift
no longer drops bits in marshal
........
r4192 | pjenvey | 2008-03-06 17:13:52 -0700 (Thu, 06 Mar 2008) | 4 lines

o rearrange getname/getglobal to NameError messages to CPython
o fix del of an undefined global raising a KeyError instead of NameError
o flatten getlocal(String) into getlocal(int)
........
r4193 | pjenvey | 2008-03-07 22:53:48 -0700 (Fri, 07 Mar 2008) | 3 lines

fix __import__ to actually handle the fromlist argument
fixes #1006
........
r4195 | amak | 2008-03-08 09:22:08 -0700 (Sat, 08 Mar 2008) | 5 lines

Checking in changes relating to two separate bugs.
1. UDP sockets implicitly created by the sendto() method are not configured correctly.
2. Should a recvfrom() call on an unbound UDP socket result in the implicit creation of a socket?
See bug # 1005 for further information.
http://bugs.jython.org/issue1005
........
r4197 | cgroves | 2008-03-08 12:41:40 -0700 (Sat, 08 Mar 2008) | 1 line

Add all the jars we can redistribute to extlibs so individual users don't have to fetch them to gain the optional functionality. Can't redistribute the oracle and informix jars, so we're still not completely set for making installers without some jar hunting
........
r4198 | pjenvey | 2008-03-10 16:05:34 -0600 (Mon, 10 Mar 2008) | 3 lines

rewrite call's arg parsing to match CPython's (most importantly its error
messages)
........
r4199 | pjenvey | 2008-03-10 18:15:02 -0600 (Mon, 10 Mar 2008) | 3 lines

load_module must strip the trailing __init__ (like py_compile/imp do) for the
correct module/java class name
........
r4200 | cgroves | 2008-03-14 09:21:46 -0600 (Fri, 14 Mar 2008) | 8 lines

Reorganize init targets so the full build specific initting just sets the
properties it wants to override for regular initting and then shares the rest
of the init.

Make a separate jarjar dir so I can depend on the same jarjar'd files from
Eclipse as I do in ant.
........
r4201 | zyasoft | 2008-03-15 06:59:11 -0600 (Sat, 15 Mar 2008) | 1 line

PyDictionary now supports the ConcurrentMap interface and setdefault is now atomic, just like CPython (of course CPython is only atomic under weaker assumptions, for Jython we can take advantage of putIfAbsent semantics which are similar if not quite the same as setdefault)
........
r4202 | thobes | 2008-03-15 21:34:20 -0600 (Sat, 15 Mar 2008) | 1 line

Added support for braces.
........
r4203 | thobes | 2008-03-15 21:36:56 -0600 (Sat, 15 Mar 2008) | 3 lines

Added a global interpreter lock.
It is available as:
from __future__ import GIL
........
r4204 | cgroves | 2008-03-16 09:40:52 -0600 (Sun, 16 Mar 2008) | 1 line

Cleanup and Java 5isms for the proxy code
........
r4205 | pjenvey | 2008-03-16 12:44:50 -0600 (Sun, 16 Mar 2008) | 1 line

convert potential IllegalArgumentExceptions into OSErrors
........
r4206 | pjenvey | 2008-03-16 13:23:08 -0600 (Sun, 16 Mar 2008) | 2 lines

from
http://svn.python.org/projects/python/branches/release23-maint/Lib@61346
........
r4207 | pjenvey | 2008-03-16 13:27:37 -0600 (Sun, 16 Mar 2008) | 4 lines

merge r52803, r52824 from
http://svn.python.org/projects/python/branches/release23-maint/Lib
these were reverted from the 23 maintenance branch for the 2.3.7 security
release
........
r4208 | cgroves | 2008-03-16 14:25:25 -0600 (Sun, 16 Mar 2008) | 1 line

urllibnet passes on jython
........
r4209 | fwierzbicki | 2008-03-17 13:07:46 -0600 (Mon, 17 Mar 2008) | 2 lines

Initial commit of ast copy into trunk.
........
r4210 | pjenvey | 2008-03-17 13:58:29 -0600 (Mon, 17 Mar 2008) | 5 lines

add 'at' to object repr to more closely match CPython's. Mainly for doctests
that expect it
fixes #1009
patch from John Szakmeister
........
r4211 | cgroves | 2008-03-17 14:09:42 -0600 (Mon, 17 Mar 2008) | 1 line

Remove typo'd ClassWriter constant that was stripping out line numbers from exposed classes
........
r4212 | fwierzbicki | 2008-03-17 14:11:20 -0600 (Mon, 17 Mar 2008) | 2 lines

Added antlr libs for ast work.
........
r4213 | cgroves | 2008-03-17 14:28:32 -0600 (Mon, 17 Mar 2008) | 1 line

Fill in empty array for bases and remove a fix for null from the actual bases method that unifies the two
........
r4214 | pjenvey | 2008-03-17 14:45:01 -0600 (Mon, 17 Mar 2008) | 1 line

add an 'at' to file repr, more of the same of r4210
........
r4215 | cgroves | 2008-03-17 15:48:47 -0600 (Mon, 17 Mar 2008) | 1 line

Classic classes don't have __class__
........
r4216 | cgroves | 2008-03-17 16:17:05 -0600 (Mon, 17 Mar 2008) | 1 line

Rip out all the old school 'public static PyClass __class__' optimizations. They haven't been enabled since forever and new style types set their type explicitly and don't need the optimization
........
r4217 | cgroves | 2008-03-17 17:13:50 -0600 (Mon, 17 Mar 2008) | 1 line

Add some braindead tests for Java method and field visibility from Python
........
r4218 | cgroves | 2008-03-18 10:42:02 -0600 (Tue, 18 Mar 2008) | 1 line

Patch from njriley to disallow setting and deleting fields on builtin types.
........
r4220 | thobes | 2008-03-18 17:25:35 -0600 (Tue, 18 Mar 2008) | 2 lines

A few mistakes, references to old ClassFiles, that I have had locally on my system and so on.
Also forgot to upload the modifications to the AST-files.
........
r4221 | cgroves | 2008-03-18 17:43:14 -0600 (Tue, 18 Mar 2008) | 1 line

Rip out stale code and jythonc support from proxy code
........
r4223 | nriley | 2008-03-19 08:12:21 -0600 (Wed, 19 Mar 2008) | 1 line

Add heapq module.
........
r4224 | nriley | 2008-03-19 08:21:53 -0600 (Wed, 19 Mar 2008) | 1 line

Remove -m from usage message because it's not yet supported by Jython.
........
r4225 | nriley | 2008-03-19 12:02:52 -0600 (Wed, 19 Mar 2008) | 1 line

Fix new.classobj to create new-style classes if needed (matches CPython behavior); provide more useful exceptions if you try to create an old-style class deriving from a new-style class (doesn't match CPython behavior).
........
r4226 | pjenvey | 2008-03-19 13:40:06 -0600 (Wed, 19 Mar 2008) | 1 line

more 'at' added to __repr__
........
r4227 | pjenvey | 2008-03-19 13:43:50 -0600 (Wed, 19 Mar 2008) | 3 lines

use the builtin variable instead of underlying_class. a builtin type may
potentially lack an underlying java class (e.g. new style class Exceptions)
........
r4228 | fwierzbicki | 2008-03-19 14:30:15 -0600 (Wed, 19 Mar 2008) | 2 lines

New antlr parser moving into trunk (but w/o removing old parser -- some way to go there yet).
........
r4229 | fwierzbicki | 2008-03-19 16:01:51 -0600 (Wed, 19 Mar 2008) | 2 lines

Checked in antlr output in the interests of letting others get there stuff in.
........
r4230 | zyasoft | 2008-03-19 16:20:54 -0600 (Wed, 19 Mar 2008) | 1 line

threading now maps against java.util.concurrent more directly
........
r4231 | nriley | 2008-03-19 17:01:36 -0600 (Wed, 19 Mar 2008) | 1 line

Add convenience constructors for OSError, NotImplementedError and EnvironmentError.
........
r4232 | nriley | 2008-03-19 17:23:32 -0600 (Wed, 19 Mar 2008) | 1 line

Minimal implementation of gc module.
........
r4233 | fwierzbicki | 2008-03-19 18:07:50 -0600 (Wed, 19 Mar 2008) | 2 lines

Back out of ast and antlr in trunk. Probably better in a separate directory.
........
r4236 | zyasoft | 2008-03-20 01:00:58 -0600 (Thu, 20 Mar 2008) | 1 line

Copies from 2.5: The Queue module in 2.5 uses Condition, instead of its own wait logic. Given that join/task_done require an additional condition, which has to be on the same lock as the queue as a whole, it only makes sense to use this pure Python version, now that we have a different Java-based version in the threading module. test_bisect uses the internals of Queue.
........
r4242 | pjenvey | 2008-03-20 22:18:34 -0600 (Thu, 20 Mar 2008) | 2 lines

fix the new thread local
........
r4243 | pjenvey | 2008-03-21 16:10:09 -0600 (Fri, 21 Mar 2008) | 1 line

use our old thread._local instead
........
r4244 | pjenvey | 2008-03-21 16:46:20 -0600 (Fri, 21 Mar 2008) | 4 lines

o expose the abstract descriptor type and its __name__ and __objclass__ vars so
its subclasses provide them
o expose slot and methodescr to bring all descriptors into new style
........
r4245 | pjenvey | 2008-03-21 16:58:09 -0600 (Fri, 21 Mar 2008) | 1 line

match reprs to CPython
........
r4246 | zyasoft | 2008-03-21 19:25:20 -0600 (Fri, 21 Mar 2008) | 1 line

repr of __dict__ now uses repr for the keys even if they are String (as they almost always are)
........
r4247 | zyasoft | 2008-03-21 19:56:32 -0600 (Fri, 21 Mar 2008) | 1 line

Added Semaphore (inadvertently deleted, needs explicit testing) and __eq__ for thread equality
........
r4248 | zyasoft | 2008-03-21 21:37:25 -0600 (Fri, 21 Mar 2008) | 1 line

Semaphore.acquire now supports non-blocking operations, as it is supposed to (need better tests for refactoring this codebase)
........
r4249 | pjenvey | 2008-03-22 14:36:41 -0600 (Sat, 22 Mar 2008) | 4 lines

o add __ne__ to JavaThread
o fix RLock.acquire sometimes returning None and assert the lock is held in
release like CPython does
........
r4250 | pjenvey | 2008-03-22 16:16:30 -0600 (Sat, 22 Mar 2008) | 5 lines

o expose __name__ and __objclass__ explicitly on each PyDescriptor subclass.
PyDescriptor doesn't actually exist in Python (and in their mro's) so we have
to do it this way. r4244 doesn't work unless PyDescriptor is in the mro
o fix PyDataDescr's repr
........
r4251 | pjenvey | 2008-03-23 21:44:39 -0600 (Sun, 23 Mar 2008) | 1 line

fromlist is no longer needed here, small cleanup
........
r4264 | pjenvey | 2008-04-01 18:10:30 -0600 (Tue, 01 Apr 2008) | 1 line

use fully qualified type names
........
r4265 | pjenvey | 2008-04-01 18:32:10 -0600 (Tue, 01 Apr 2008) | 1 line

some cleanup: coding standards, whitespace
........
r4268 | pjenvey | 2008-04-01 19:29:16 -0600 (Tue, 01 Apr 2008) | 2 lines

o little more cleanup, rearrangement
o move decapitalize into StringUtil
........
r4269 | pjenvey | 2008-04-01 20:47:02 -0600 (Tue, 01 Apr 2008) | 1 line

test_gc and test_weakref no longer skip but fail
........
r4270 | pjenvey | 2008-04-01 21:19:56 -0600 (Tue, 01 Apr 2008) | 2 lines

wrap type __dict__s with a readonly PyDictProxy to prevent their modification
........
r4271 | pjenvey | 2008-04-02 00:02:12 -0600 (Wed, 02 Apr 2008) | 2 lines

fix slotted class variables -- don't overwrite them with a descriptor
........
r4272 | pjenvey | 2008-04-02 15:18:12 -0600 (Wed, 02 Apr 2008) | 3 lines

more tightening of get/set/del of PyType attributes
fixes #1019
........
r4273 | pjenvey | 2008-04-02 17:34:26 -0600 (Wed, 02 Apr 2008) | 2 lines

make xrange a reference to the class instead of a wrapper method
........
r4274 | pjenvey | 2008-04-02 17:54:38 -0600 (Wed, 02 Apr 2008) | 3 lines

types.py from
http://svn.python.org/projects/python/branches/release23-maint/Lib/types.py@31332
........
r4275 | pjenvey | 2008-04-02 17:58:02 -0600 (Wed, 02 Apr 2008) | 3 lines

kill our builtin types module for the pure Python version. it wasn't neccessary
and was referencing the type objects via PyType.fromClass, which can be bad
........
r4276 | pjenvey | 2008-04-02 21:14:16 -0600 (Wed, 02 Apr 2008) | 1 line

whitespace
........
r4277 | pjenvey | 2008-04-03 00:39:50 -0600 (Thu, 03 Apr 2008) | 1 line

move ExtraMath into util
........
r4278 | pjenvey | 2008-04-03 00:43:42 -0600 (Thu, 03 Apr 2008) | 1 line

fix import for moved ExtraMath
........
r4279 | pjenvey | 2008-04-03 13:41:56 -0600 (Thu, 03 Apr 2008) | 2 lines

add noKeywords, eensures no keyword args were passed.
equiv. of CPython's PyArg_NoKeywords
........
r4280 | pjenvey | 2008-04-03 15:25:21 -0600 (Thu, 03 Apr 2008) | 1 line

whitespace
........
r4281 | pjenvey | 2008-04-03 16:55:42 -0600 (Thu, 03 Apr 2008) | 1 line

add a helper for creating a tuple from an iterable
........
r4282 | pjenvey | 2008-04-03 23:16:25 -0600 (Thu, 03 Apr 2008) | 4 lines

o unify the descriptors' type checkers into PyDescriptor and make them use the
right error message for __get__
o remove PyDescriptor's unnecessary blurb
........
r4287 | pjenvey | 2008-04-04 16:45:08 -0600 (Fri, 04 Apr 2008) | 1 line

rearrange as per coding standards
........
r4289 | pjenvey | 2008-04-04 17:08:48 -0600 (Fri, 04 Apr 2008) | 1 line

add getPyObjectByType. raises an exception if the arg doesn't match the type
........
r4290 | pjenvey | 2008-04-04 17:11:57 -0600 (Fri, 04 Apr 2008) | 3 lines

Removed merge tracking for "svnmerge" for
https://jython.svn.sourceforge.net/svnroot/jython/branches/modern
........
r4292 | pjenvey | 2008-04-04 17:23:46 -0600 (Fri, 04 Apr 2008) | 1 line

another handy invoke overload
........
r4301 | pjenvey | 2008-04-05 14:10:55 -0600 (Sat, 05 Apr 2008) | 1 line

whitespace
........
r4302 | pjenvey | 2008-04-05 14:17:36 -0600 (Sat, 05 Apr 2008) | 1 line

actually plugin __cmp__ (though the implementation is still a bit broken at this point)
........
r4303 | pjenvey | 2008-04-05 15:10:08 -0600 (Sat, 05 Apr 2008) | 8 lines

get test_pickle passing:
o fix PyObject.slotnames incorrectly going up the __mro__ for __slotnames__
o completely disallow __class__ assignment involving builtins
o __getnewargs__ wasn't exposed on str/unicode
o base types' numSlots shouldn't have any effect on the need for userdict
o rename layoutAligns to compatibleForAssignment and put it in charge of
throwing TypeErrors (to match CPython)
........
r4304 | pjenvey | 2008-04-05 15:32:09 -0600 (Sat, 05 Apr 2008) | 1 line

fix builtin functions not supported in protocol 1
........
r4305 | pjenvey | 2008-04-05 17:55:24 -0600 (Sat, 05 Apr 2008) | 6 lines

get test_cpickle passing:
o treat class's with a custom metaclass as a regular class
o fix batch_setitems to work on a dict's iteritems
o make load_*unicode actually load unicode instead of str
o handle slotstates in protocol 2
........
r4306 | pjenvey | 2008-04-05 18:48:46 -0600 (Sat, 05 Apr 2008) | 4 lines

get test_xpickle passing:
fix protocol 2 longs being dumped/loaded in big endian (from BigInteger)
whereas pickle uses little endian
........
r4307 | pjenvey | 2008-04-05 22:27:07 -0600 (Sat, 05 Apr 2008) | 2 lines

ensure sane args to new.function to make test_new pass
........
r4308 | pjenvey | 2008-04-06 11:07:20 -0600 (Sun, 06 Apr 2008) | 1 line

translate a failed int conversion to a TypeError
........
r4309 | pjenvey | 2008-04-06 13:33:09 -0600 (Sun, 06 Apr 2008) | 1 line

make long() (no args) return a long instead of int
........
r4312 | pjenvey | 2008-04-06 21:56:47 -0600 (Sun, 06 Apr 2008) | 6 lines

o expose complex __coerce__ and handle coerce NotImplemented. patch #1783088
o throw a ValueError on complex overflows. patch #1779428
o special case complex(acomplex) to return acomplex
fixes #1758284, #1783088
patches from Keshav Upadhyaya
........
r4313 | pjenvey | 2008-04-06 23:01:27 -0600 (Sun, 06 Apr 2008) | 2 lines

cleanup complex __new__
closes out #1514533
........
r4314 | pjenvey | 2008-04-06 23:21:05 -0600 (Sun, 06 Apr 2008) | 1 line

oops, broke PyComplexDerived
........
r4315 | pjenvey | 2008-04-06 23:27:02 -0600 (Sun, 06 Apr 2008) | 2 lines

add __coerce__ to int/long/float, patch #1785415 from Keshav Upadhyaya
fixes #1777684
........
r4316 | pjenvey | 2008-04-07 13:47:42 -0600 (Mon, 07 Apr 2008) | 3 lines

expose __module__ on builtin functions. gives pickle less work to do when
saving them
........
r4317 | pjenvey | 2008-04-07 16:19:59 -0600 (Mon, 07 Apr 2008) | 4 lines

disable Exception.message until 2.5: Exception.__init__ must set it via
__setattr__ which has the potential for breaking < 2.5 Exception
subclasses. it's ready to go on the pep352 branch
........
r4318 | pjenvey | 2008-04-07 17:04:35 -0600 (Mon, 07 Apr 2008) | 1 line

s/CPython/Python/
........
r4319 | pjenvey | 2008-04-07 17:07:35 -0600 (Mon, 07 Apr 2008) | 1 line

more 'at' in reprs
........
r4320 | pjenvey | 2008-04-07 18:54:32 -0600 (Mon, 07 Apr 2008) | 4 lines

patch #1780153 from Rajesh Battala: fix %c formatting non chars. allows sync.
with CPythonLib's test_str
fixes #1768075
........
r4321 | pjenvey | 2008-04-08 14:06:31 -0600 (Tue, 08 Apr 2008) | 3 lines

pretty up exception tracebacks to look like CPython 2.5, including loading the
offending line numbers in PyTraceback
........
r4324 | pjenvey | 2008-04-08 23:29:31 -0600 (Tue, 08 Apr 2008) | 5 lines

o catch *any* exception in getattr and just return false
o disallow unicode names in getattr and hasattr (thanks ayeshaiqbal, re
#1775263)
fixes #1501932, #1768979
........
r4327 | pjenvey | 2008-04-09 00:59:40 -0600 (Wed, 09 Apr 2008) | 2 lines

really fix #1501932 this time, catch *any* exception in hasattr. with test
........
r4328 | fwierzbicki | 2008-04-09 07:59:40 -0600 (Wed, 09 Apr 2008) | 4 lines

Try #2 to get the new parser into trunk. Includes fix of antlr_gen task from Tobias
and a conditional antlr.notneeded that should only rebuild the grammar stuff if there
has been a change.
........
r4329 | fwierzbicki | 2008-04-09 08:08:31 -0600 (Wed, 09 Apr 2008) | 2 lines

Make the ast dir only generate the ast files (the grammar building is now in trunk).
........
r4330 | pjenvey | 2008-04-09 18:39:49 -0600 (Wed, 09 Apr 2008) | 1 line

NotImplemented is Trueish
........
r4331 | pjenvey | 2008-04-09 20:45:22 -0600 (Wed, 09 Apr 2008) | 1 line

use array.ArrayType
........
r4332 | fwierzbicki | 2008-04-10 18:09:27 -0600 (Thu, 10 Apr 2008) | 14 lines

The start of an _ast module. It relies on the new antlr parser which is now in trunk.
For now you must put antlr-runtime-3.0.1 (which is in the extlibs dir) in your
classpath to use it. Using it looks like this:

>>> import _ast
>>> compile("a=1", "<unknown>", "exec", _ast.PyCF_ONLY_AST)
tree: (Module (Assign (Target (Name a)) (Value (Num 1))))

Which outputs a tree representation for now.

If you do not use _ast or the PyCF_ONLY_AST flag, you should not need
antlr-runtime-3.0.1.jar in your path.
........
r4333 | pjenvey | 2008-04-11 17:23:41 -0600 (Fri, 11 Apr 2008) | 1 line

support the optional file arg to load_source
........
r4334 | pjenvey | 2008-04-12 14:56:15 -0600 (Sat, 12 Apr 2008) | 1 line

remove the now fully deprecated pieces of xrange, cleanup
........
r4335 | pjenvey | 2008-04-12 14:57:06 -0600 (Sat, 12 Apr 2008) | 1 line

more 'at' in repr
........
r4336 | pjenvey | 2008-04-12 15:48:49 -0600 (Sat, 12 Apr 2008) | 2 lines

make new types use a copy of their class dict instead of the actual dict
........
r4337 | pjenvey | 2008-04-12 15:51:42 -0600 (Sat, 12 Apr 2008) | 1 line

small cleanup
........
r4338 | pjenvey | 2008-04-12 18:11:50 -0600 (Sat, 12 Apr 2008) | 1 line

test_types_jy is testing classes not just types, rename to test_class_jy
........
r4339 | pjenvey | 2008-04-12 19:25:23 -0600 (Sat, 12 Apr 2008) | 23 lines

we use only PyFrame.getf_locals() to get at locals, whereas CPython in general
has two ways of getting them:

o the quicker way: grab the 'current' locals via f_locals directly: this is
what CPython's LOAD/STORE/DELETE_NAME and LOAD_LOCALS do
o the thorough way: first merge fast locals (cellvars and freevars) into
f_locals, then grab the updated f_locals: this is what locals() does, and also
what eval/exec/execfile etc. use

our LOAD_LOCALS equivalent doesn't use the quicker way and that can cause a
difference in the contents of class dicts passed to metaclasses'
__new__(M, name, bases, class_dict) (see the accompanying test)

so I've changed our LOAD_LOCALS to use the quick way, to fix this issue, and
also made our LOAD/STORE/DELETE_NAME equivalents use the quicker way too (which
means closures do a little less work)

Now we have:
o PyFrame.getf_locals -> quick way (and could be deprecated, only the compiler
uses it. just grab f_locals directly instead)
o new PyFrame.getLocals -> thorough way, used everywhere else outside of
PyFrame
........
r4340 | pjenvey | 2008-04-13 12:42:31 -0600 (Sun, 13 Apr 2008) | 1 line

whitespace
........
r4341 | pjenvey | 2008-04-13 12:51:14 -0600 (Sun, 13 Apr 2008) | 2 lines

py_compile from
http://svn.python.org/projects/python/branches/release25-maint/Lib/py_compile.py@38987
........
r4342 | pjenvey | 2008-04-13 12:55:39 -0600 (Sun, 13 Apr 2008) | 4 lines

rename our builtin py_compile to _py_compile, and call it from CPython's pure
python py_compile. adds support for compile's doraise option, which by default
swallows exceptions
........
r4343 | pjenvey | 2008-04-13 13:08:52 -0600 (Sun, 13 Apr 2008) | 3 lines

fix re.subx on empty strings with no match: CPython actually returns
pattern[0:0] in this case
........
r4344 | fwierzbicki | 2008-04-14 08:59:19 -0600 (Mon, 14 Apr 2008) | 2 lines

Added null check for compile_flags. Thanks to Georgy Berdyshev for pointing out the bug.
........
r4345 | pjenvey | 2008-04-14 18:07:42 -0600 (Mon, 14 Apr 2008) | 15 lines

o automatically define toString() in derived classes (instead of including it
from object.derived) but allow disabling of it. for PyString/UnicodeDerived;
they need toString() to pass through instead of calling __repr__ as that breaks
the expectation that PyString.toString() returns the real string

this means PyObject.toString() from java code always returns __repr__ *except*
for PyString/Unicode, which return the actual string
(thanks Mehendran)

o update the "should return a <type>" message in derived class descriptors to
match CPython 2.5's
o from the full gderived.py run: removed a useless constructor from
PyUnicodeDerived and updated the collections derives' getattr/setattr to use
interned strings
........
r4346 | pjenvey | 2008-04-14 19:01:18 -0600 (Mon, 14 Apr 2008) | 6 lines

o fix str.__repr__() to match repr(str) and separate it from toString
fixes #1510227, #1782565
mostly from Mehendran's patch in #1785475
o fix str.__str__ on str subclasses not returning a normal str
o rename unicode_toString to unicode___repr__ (that's what it really is)
........
r4347 | pjenvey | 2008-04-14 21:24:23 -0600 (Mon, 14 Apr 2008) | 2 lines

always escape \t \n \r, even for unicode, never escape \b \f
........
r4348 | pjenvey | 2008-04-14 22:11:00 -0600 (Mon, 14 Apr 2008) | 2 lines

fix single quotes not being escaped in unicode repr
........
r4349 | pjenvey | 2008-04-16 12:39:50 -0600 (Wed, 16 Apr 2008) | 9 lines

include the jarjar'd asm ClassReader in our jar and use it to get the
real class name of $py.class files before importing them. Allows us to
import misnamed/mislabeled $py.class files

the import statement has been hacking around this by recompiling the
.py (if it exists) when it happens. more importantly imp.load_module
(and load_compiled when we support it) don't and shouldn't workaround
this, instead load_module has been throwing NoClassDefFoundErrors
........
r4350 | pjenvey | 2008-04-16 17:22:27 -0600 (Wed, 16 Apr 2008) | 5 lines

add a special case to _binop_rule to satisfy CPython's handling of
__add__/__radd__ with str/unicode and str/unicode subclasses. replaces
the broken bandaid for this applied in r4020
Thanks Samuele Pedroni for the solution and test_basic
........
r4351 | pjenvey | 2008-04-17 00:53:27 -0600 (Thu, 17 Apr 2008) | 1 line

allow setDaemon(None) or anything else by filtering it through bool
........
r4352 | pjenvey | 2008-04-17 15:34:50 -0600 (Thu, 17 Apr 2008) | 3 lines

new style classes must define a __call__ to be callable, a __call__ provided by
__getattr__ doesn't count
........
r4353 | zyasoft | 2008-04-17 15:59:21 -0600 (Thu, 17 Apr 2008) | 1 line

Changes to enable dateutil on jython, plus a fix for a potential race on time.clock's initialization
........
r4354 | pjenvey | 2008-04-17 22:05:56 -0600 (Thu, 17 Apr 2008) | 2 lines

core.util is a more appropriate place for ByteSwapper and FileUtil
........
r4355 | fwierzbicki | 2008-04-18 13:03:31 -0600 (Fri, 18 Apr 2008) | 3 lines

Better coverage of _ast. Made AST an interface so that enums (like expr_contextType)
can pass instanceof(x, _ast.AST).
........
r4356 | pjenvey | 2008-04-18 14:45:13 -0600 (Fri, 18 Apr 2008) | 6 lines

o kill the annoying Java version of the os module. if it was actually ever
needed, it's not anymore
o store os.environ in sys so it can survive a reload of the os
module. this was causing a really vague new bug in test_javashell now
that test_import was actually able to reload os via imp.load_module
........
r4357 | pjenvey | 2008-04-18 16:03:25 -0600 (Fri, 18 Apr 2008) | 3 lines

avoid a ClassCastException when an object's __int__ surprises formatInteger by
returning a long
........
r4358 | pjenvey | 2008-04-18 17:40:39 -0600 (Fri, 18 Apr 2008) | 3 lines

utilize the str formatter to make float's repr/str resemble CPython's a little
more
........
r4359 | pjenvey | 2008-04-18 20:46:32 -0600 (Fri, 18 Apr 2008) | 2 lines

fix the default PyDictionary constructor unnecessarily creating two CHMs
........
r4360 | fwierzbicki | 2008-04-19 08:23:21 -0600 (Sat, 19 Apr 2008) | 3 lines

Expose __iter__ on tuple, list, and array types. Still need to do the
same on the Derived types.
........
r4361 | pjenvey | 2008-04-19 15:14:24 -0600 (Sat, 19 Apr 2008) | 7 lines

make the 'g' float formatter closer to CPython's output:
o format as exponent when -4 <= exponent < precision
o handle infinities
o pass the altFlag to the formatters
o fix precision of the exponential formatter's exponent
o don't need the altFlag padding when doing exponential format
........
r4362 | pjenvey | 2008-04-19 15:35:10 -0600 (Sat, 19 Apr 2008) | 3 lines

kill ExtraMath's hypot and log10 for their equivalents in Java 1.5's
java.lang.Math
........
r4363 | fwierzbicki | 2008-04-20 10:05:22 -0600 (Sun, 20 Apr 2008) | 2 lines

Added lineno and col_offset, and some debug info to PythonTree.
........
r4364 | fwierzbicki | 2008-04-20 12:19:34 -0600 (Sun, 20 Apr 2008) | 2 lines

remove debug statement.
........
r4365 | fwierzbicki | 2008-04-20 20:57:33 -0600 (Sun, 20 Apr 2008) | 2 lines

Fixed up to take advantage of Philip's recent improvements.
........
r4366 | fwierzbicki | 2008-04-21 15:58:41 -0600 (Mon, 21 Apr 2008) | 5 lines

took test_ast.py from CPython's release25-maint branch revision 62446. Modified a bit
to "shape" the results of __class__.__name__ since the results are different for Jython
for the moment (and may always differ, we'll see). About 50% of the tests work at the
moment. Also cleaned up the ast folder a bit.
........
r4367 | fwierzbicki | 2008-04-21 20:41:48 -0600 (Mon, 21 Apr 2008) | 3 lines

Better match line numbers and char positions with test_ast. Commented out the failures
for now. Also added support for eval and single_input.
........
r4368 | fwierzbicki | 2008-04-22 09:46:48 -0600 (Tue, 22 Apr 2008) | 2 lines

All but a couple of test_ast tests now pass.
........
r4369 | fwierzbicki | 2008-04-22 18:12:34 -0600 (Tue, 22 Apr 2008) | 2 lines

The rest of the modified test_ast now works.
........
r4370 | pjenvey | 2008-04-22 22:30:28 -0600 (Tue, 22 Apr 2008) | 1 line

expose __iter__ on sets
........
r4371 | pjenvey | 2008-04-24 22:22:43 -0600 (Thu, 24 Apr 2008) | 3 lines

use System.out and friends instead of FileDescriptor.out and friends, which
aren't correct under nailgun
........
r4374 | pjenvey | 2008-04-30 21:59:37 -0600 (Wed, 30 Apr 2008) | 2 lines

small rearrange per coding standards
........
r4379 | fwierzbicki | 2008-05-02 19:20:26 -0600 (Fri, 02 May 2008) | 3 lines

Add warning that Derived classes are generated. No warning added to Derived classes
in org/python/modules yet since they look like they where hand modified a bit.
........
r4391 | pjenvey | 2008-05-13 13:28:35 -0600 (Tue, 13 May 2008) | 1 line

whitespace
........
r4393 | pjenvey | 2008-05-13 15:15:18 -0600 (Tue, 13 May 2008) | 3 lines

update test_funcattrs to the latest 2.5 version, from:
http://svn.python.org/projects/python/branches/release25-maint/Lib/test/test_funcattrs.py@59171
........
r4394 | pjenvey | 2008-05-13 15:34:49 -0600 (Tue, 13 May 2008) | 8 lines

o cleanup PyFunction, use exposed annotations
o copy the new.function arg checking code into PyFunction __new__, so
types.FunctionType supports the same args. add test_func_jy with jek's test
cases
o add cell.cell_contents and an appropriate repr. use exposed annotations
fixes #1034
thanks Jason Kirtland
........
r4395 | pjenvey | 2008-05-13 16:18:05 -0600 (Tue, 13 May 2008) | 1 line

unused import
........
r4396 | pjenvey | 2008-05-13 16:30:47 -0600 (Tue, 13 May 2008) | 3 lines

new.py from:
http://svn.python.org/projects/python/branches/release25-maint/Lib/new.py@27241
........
r4397 | fwierzbicki | 2008-05-13 17:30:08 -0600 (Tue, 13 May 2008) | 3 lines

Extracted from astlib.py -- simple pretty printer that works for
both jython and python.
........
r4398 | fwierzbicki | 2008-05-13 17:31:56 -0600 (Tue, 13 May 2008) | 2 lines

change def name.
........
r4399 | pjenvey | 2008-05-13 19:13:52 -0600 (Tue, 13 May 2008) | 1 line

make types.InstanceType() handle None as second arg
........
r4400 | pjenvey | 2008-05-13 19:14:20 -0600 (Tue, 13 May 2008) | 6 lines

o match types.MethodType() (new PyMethod()) args to CPython's:
(func, self, type) instead of (self, func, type)
o kill most of newmodule for the simpler pure python version. we can't dump it
completely until a better workaround for the new.classobj hack is done; in the
meantime, keep it around at _new.classobj
........
r4401 | pjenvey | 2008-05-13 19:34:54 -0600 (Tue, 13 May 2008) | 1 line

ensure new.classobj can create new style classes
........
r4404 | fwierzbicki | 2008-05-15 13:13:41 -0600 (Thu, 15 May 2008) | 3 lines

Finally found the fix for empty modules. Empty Lexer output causes a null tree to
return. I've fixed by creating an empty Module node by hand for empty inuput.
........
r4406 | pjenvey | 2008-05-15 14:42:23 -0600 (Thu, 15 May 2008) | 3 lines

maintain func_defaults and func_closure as null instead of
Py.EmptyObjects/EmptyTuple, to more closely resemble CPython
........
r4409 | pjenvey | 2008-05-15 17:52:01 -0600 (Thu, 15 May 2008) | 4 lines

let null return values pass thru PyDataDescr get as None instead of triggering
AttributeErrors. usually what we want anyway and makes @ExposedGet
implementations easier
........
r4410 | pjenvey | 2008-05-15 20:03:31 -0600 (Thu, 15 May 2008) | 3 lines

fix isinstance(newstyle_obj, oldstyle_class) always failing when it can be True
via multiple inheritance. make isInstance more closely resemble CPython's
........
r4411 | pjenvey | 2008-05-16 11:55:45 -0600 (Fri, 16 May 2008) | 5 lines

fix sets incorrectly raising a TypeError when asked for superset/subset on
non-set iterables. also fixes a few other tests from 2.5 test_set
fixes #1035
patch from Jonathan Ellis
........
r4412 | pjenvey | 2008-05-16 14:27:52 -0600 (Fri, 16 May 2008) | 1 line

more ExposedGet now translating null for us
........
r4413 | pjenvey | 2008-05-16 15:19:03 -0600 (Fri, 16 May 2008) | 2 lines

cleanup PyMethod, use exposed annotations
........
r4414 | pjenvey | 2008-05-16 17:30:52 -0600 (Fri, 16 May 2008) | 1 line

more appropriate method name
........
r4415 | pjenvey | 2008-05-16 18:07:22 -0600 (Fri, 16 May 2008) | 7 lines

use our Java set implementation only for the builtin sets, and use CPythonLib's
sets module. so we can be more compatible with both versions compared to
CPython, as there's a few differences between the two. moves modules.sets into
core, PyImmutableSet -> PyFrozenSet, and causes test___all__ and CPythonLib's
test_sets to now pass
patch from Jonathan Ellis
........
r4419 | pjenvey | 2008-05-17 17:09:27 -0600 (Sat, 17 May 2008) | 1 line

match CPython's error message
........
r4420 | pjenvey | 2008-05-18 15:54:16 -0600 (Sun, 18 May 2008) | 1 line

use String.format
........
r4421 | pjenvey | 2008-05-18 16:51:09 -0600 (Sun, 18 May 2008) | 1 line

move the thread module into its own package
........
r4422 | pjenvey | 2008-05-18 17:00:22 -0600 (Sun, 18 May 2008) | 1 line

small cleanup
........
r4426 | pjenvey | 2008-05-20 12:24:57 -0600 (Tue, 20 May 2008) | 2 lines

test_set.py from:
http://svn.python.org/projects/python/branches/release25-maint/Lib/test/test_set.py@54954
........
r4427 | pjenvey | 2008-05-20 12:52:29 -0600 (Tue, 20 May 2008) | 4 lines

number of set fixes to get most of 2.5's test_set passing
fixes #1036
patch from Jonathan Ellis
........
r4428 | pjenvey | 2008-05-20 13:00:21 -0600 (Tue, 20 May 2008) | 1 line

make all baseset_ methods final, whitespace
........
r4429 | pjenvey | 2008-05-20 13:01:14 -0600 (Tue, 20 May 2008) | 2 lines

fix set pickling. we no longer need setsfactory
........
r4430 | pjenvey | 2008-05-20 13:14:37 -0600 (Tue, 20 May 2008) | 2 lines

special case symmetric_difference_update on yourself; just clear
avoids ConcurrentModificationExceptions
........
r4431 | fwierzbicki | 2008-05-20 13:19:52 -0600 (Tue, 20 May 2008) | 5 lines

Added a shell script to run Jython -- tested this in the asm branch a bit, but I think this
can safely go into trunk. This was adapted from a script from the JRuby project.
I also included the same support for The Java Interactive Profiler (JIP) that JRuby includes.
See http://jiprof.sourceforge.net/
........
r4432 | pjenvey | 2008-05-20 14:39:12 -0600 (Tue, 20 May 2008) | 2 lines

need to track line numbers for inplace ops
........
r4435 | pjenvey | 2008-05-20 17:20:43 -0600 (Tue, 20 May 2008) | 4 lines

o kill as_immutable, leftover from the sets module
implementation. cleanup asImmutable as a result (now asFrozen)
o fix cyclical repr problems via thread.enterRepr, and repr on set subclasses
........
r4436 | pjenvey | 2008-05-20 17:37:48 -0600 (Tue, 20 May 2008) | 1 line

remove unnecessary this., small cleanup
........
r4437 | pjenvey | 2008-05-20 17:39:54 -0600 (Tue, 20 May 2008) | 1 line

don't need deepcopy since our copy module is already from 2.5 and supports builtin sets
........
r4438 | pjenvey | 2008-05-20 17:54:02 -0600 (Tue, 20 May 2008) | 1 line

we also don't need union_update, nor does __nonzero__ need to be exposed
........
r4439 | fwierzbicki | 2008-05-21 18:10:14 -0600 (Wed, 21 May 2008) | 4 lines

Allow 'as' and 'exec' to be non-keywords for Java integration (These are Python keywords and
not Java keywords, so Java may use them as method names etc). Commented out start to the
complete list of these keywords in Python.g.
........
r4442 | fwierzbicki | 2008-05-24 08:56:35 -0600 (Sat, 24 May 2008) | 4 lines

Broke out antlr parsing of Expressions (aka: evals). Start of a
fix for evals with leading spaces. Extended PythonTreeWalker so
eval and interactive parsing is supported.
........
r4443 | pjenvey | 2008-05-24 17:46:09 -0600 (Sat, 24 May 2008) | 3 lines

fix function, method and slot descriptors always requiring 2 args
pointed out by lsoto
........
r4445 | fwierzbicki | 2008-05-25 06:47:37 -0600 (Sun, 25 May 2008) | 3 lines

Fail on first syntax error. Use antlr's RecognitionException
to handle syntax errors.
........
r4447 | cgroves | 2008-05-25 14:38:53 -0600 (Sun, 25 May 2008) | 10 lines

Add a basic project for Eclipse. It requires that ant be run first to make
gensrc and jarjar, though that could probably be fixed by some intrepid work
with the External Builder. It also requires an ANT_HOME variable to be defined
as the directory where ant is installed.

It excludes the Informix and Oracle classes from the build since they aren't
included by default in extlibs, but they'll still be built by ant if you have
the jars.
........
r4449 | pjenvey | 2008-05-25 17:21:36 -0600 (Sun, 25 May 2008) | 2 lines

add CPython 2.3's special case for KeyError__str__
patch from lsoto, fixes #1040
........
r4453 | pjenvey | 2008-05-25 23:37:52 -0600 (Sun, 25 May 2008) | 1 line

r4449 synced us with 2.3 test_exceptions' output
........
r4454 | fwierzbicki | 2008-05-26 07:41:36 -0600 (Mon, 26 May 2008) | 2 lines

Tighten grammar around "print <<" statements. "print << a ," not allowed.
........
r4456 | fwierzbicki | 2008-05-26 11:26:16 -0600 (Mon, 26 May 2008) | 3 lines

Better exception handling in antlr parser (fail fast in PythonWalker,
and new class antlr/ParseException.
........
r4458 | fwierzbicki | 2008-05-27 06:08:02 -0600 (Tue, 27 May 2008) | 2 lines

Throw ParseException for non-default arg after default arg.
........
r4460 | fwierzbicki | 2008-05-27 10:56:34 -0600 (Tue, 27 May 2008) | 4 lines

Fail fast on lexer errors. Also correct lexing of
imaginary numbers (can start with zero, are never
octal or hex).
........
r4462 | fwierzbicki | 2008-05-27 12:48:24 -0600 (Tue, 27 May 2008) | 2 lines

Some cleanup of the parser driving classes.
........
r4464 | fwierzbicki | 2008-05-27 17:31:15 -0600 (Tue, 27 May 2008) | 4 lines

Very naive partial parsing in antlr grammar for interactive code
and doctests. Also renamed symbol->kind in InteractiveInterpreter
to be consistent with usage in parser code.
........
r4467 | pjenvey | 2008-05-28 16:40:27 -0600 (Wed, 28 May 2008) | 1 line

small cleanup
........
r4468 | pjenvey | 2008-05-28 21:54:32 -0600 (Wed, 28 May 2008) | 6 lines

fix test_socket's testConnectWithLocalBind locking up on subsequent test runs:
Not even SO_REUSEADDR allows reuse of an exact socket 4-tuple (localhost, PORT
- 1, localhost, PORT) stuck in TIME_WAIT (which happens after the first test
run), so catch the EADDRINUSE raised in that situation, decrement the local
PORT and try again
........
r4469 | pjenvey | 2008-05-28 23:29:20 -0600 (Wed, 28 May 2008) | 6 lines

o revert _map_exception back to creating new socket.errors every time. they're
mutable and unsafe as singletons
o map java exceptions to python in socket.shutdown
o handle all unmapped exceptions in one place and include the Java exception's
message in the translated socket.error's msg
........
r4470 | pjenvey | 2008-05-29 00:16:26 -0600 (Thu, 29 May 2008) | 5 lines

fix intermittent failures in testShutdown:
the server occasionally tearsDown before the client finishes, causing a socket
not connected error during client shutdown(). so block server tearDowns until
clients are ready to tearDown
........
r4471 | fwierzbicki | 2008-05-29 11:16:32 -0600 (Thu, 29 May 2008) | 2 lines

null checks for arrays in ast code.
........
r4473 | pjenvey | 2008-05-29 13:24:41 -0600 (Thu, 29 May 2008) | 1 line

don't consider tests skipped due to ResourceDenied as unexpected
........
r4475 | pjenvey | 2008-05-29 15:37:07 -0600 (Thu, 29 May 2008) | 6 lines

fix intermittent failures in test_select_new:
o allow socket.connect_ex to be called multiple times
o since we're using connect_ex for an asynchronous connects, don't assume it's
immediately ready in finish_connect
o same idea for server accept: don't assume it's immediately acceptable
........
r4476 | pjenvey | 2008-05-29 16:39:00 -0600 (Thu, 29 May 2008) | 1 line

revert part of r4409: Java 6 disallows the exposed method access to the private base
........
r4477 | pjenvey | 2008-05-29 16:51:31 -0600 (Thu, 29 May 2008) | 1 line

use String.format
........
r4478 | pjenvey | 2008-05-29 16:52:21 -0600 (Thu, 29 May 2008) | 1 line

be more efficient with memory in the generic readall
........
r4479 | pjenvey | 2008-05-29 16:56:22 -0600 (Thu, 29 May 2008) | 1 line

expect test_ast to fail for now, until we actually include antlr-runtime
........
r4480 | pjenvey | 2008-05-29 18:34:42 -0600 (Thu, 29 May 2008) | 3 lines

UserString.py from
http://svn.python.org/projects/python/branches/release25-maint/Lib/UserString.py@46876
........
r4481 | pjenvey | 2008-05-29 18:35:40 -0600 (Thu, 29 May 2008) | 4 lines

add str/unicode partition/rpartition
fixes #1796272, 1797751
thanks ukeshav
........
r4482 | nriley | 2008-05-29 20:47:57 -0600 (Thu, 29 May 2008) | 1 line

clean up start script, add test script, tested on Mac OS X/Linux/Solaris/Cygwin
........
r4483 | nriley | 2008-05-29 20:52:24 -0600 (Thu, 29 May 2008) | 1 line

remove debugging, interferes with help
........
r4484 | nriley | 2008-05-29 20:56:23 -0600 (Thu, 29 May 2008) | 1 line

fix help sort order
........
r4485 | fwierzbicki | 2008-05-29 21:13:31 -0600 (Thu, 29 May 2008) | 2 lines

Added line numbers and column to ast display.
........
r4486 | fwierzbicki | 2008-05-29 21:21:14 -0600 (Thu, 29 May 2008) | 2 lines

Fix lineno/col for yield.
........
r4488 | pjenvey | 2008-05-29 23:50:34 -0600 (Thu, 29 May 2008) | 1 line

invoke regrtest via the new dist/bin/jython script
........
r4489 | pjenvey | 2008-05-29 23:54:28 -0600 (Thu, 29 May 2008) | 3 lines

add a subprocess resource that's skipped by default (test_subprocess and a few
tests in test_chdir), because they take a long time
........
r4491 | pjenvey | 2008-05-30 12:31:03 -0600 (Fri, 30 May 2008) | 2 lines

merge r4441 from branches/Release_2_2maint/jython
........
r4492 | pjenvey | 2008-05-30 12:34:13 -0600 (Fri, 30 May 2008) | 3 lines

still not pure bourne shell compatible: we die on line 90 during substring
extraction. so use bash and fix aliases not being expanded in bash
........
r4495 | amak | 2008-06-01 18:02:17 -0600 (Sun, 01 Jun 2008) | 1 line

Merging changes from Release 2.2 branch, including new option support and some test refactoring.
........
r4497 | amak | 2008-06-01 18:12:57 -0600 (Sun, 01 Jun 2008) | 1 line

Submitted in an incomplete merge in the last checkin; fixing this now.
........
r4502 | pjenvey | 2008-06-01 21:18:56 -0600 (Sun, 01 Jun 2008) | 2 lines

add unicode __getitem/slice__ to fix a couple test_unicode failures,
make str's versions final
........
r4503 | pjenvey | 2008-06-01 21:20:23 -0600 (Sun, 01 Jun 2008) | 1 line

re-enable tests broken by #1758276 which was a while ago
........
r4505 | fwierzbicki | 2008-06-02 11:37:04 -0600 (Mon, 02 Jun 2008) | 2 lines

Fix start position for (Expr (Call))
........
r4506 | fwierzbicki | 2008-06-02 11:54:29 -0600 (Mon, 02 Jun 2008) | 2 lines

Improved Line/Col info for parens.
........
r4507 | fwierzbicki | 2008-06-02 12:03:33 -0600 (Mon, 02 Jun 2008) | 2 lines

Adjust line/col info for elif.
........
r4508 | fwierzbicki | 2008-06-02 12:26:07 -0600 (Mon, 02 Jun 2008) | 2 lines

Adjust line/col info for unary ops.
........
r4509 | fwierzbicki | 2008-06-02 12:40:37 -0600 (Mon, 02 Jun 2008) | 2 lines

dump astview output for a directory tree.
........
r4510 | fwierzbicki | 2008-06-02 13:46:09 -0600 (Mon, 02 Jun 2008) | 3 lines

Switching to antlr 3.1b in prep for moving to automatic AST shaping in
PythonWalker.g instead of the current "by hand" approach.
........
r4513 | pjenvey | 2008-06-02 15:57:01 -0600 (Mon, 02 Jun 2008) | 1 line

sort all test lists
........
r4514 | pjenvey | 2008-06-02 16:34:09 -0600 (Mon, 02 Jun 2008) | 3 lines

o fix __reduce__ for PyDequeDeriveds
o re-enable test_deque's usage of gc.collect now that we support it
........
r4515 | pjenvey | 2008-06-02 18:02:19 -0600 (Mon, 02 Jun 2008) | 1 line

move _weakref into its own dir
........
r4516 | pjenvey | 2008-06-02 19:56:41 -0600 (Mon, 02 Jun 2008) | 2 lines

fix isCallable to handle new style classes
........
r4517 | pjenvey | 2008-06-02 20:50:09 -0600 (Mon, 02 Jun 2008) | 5 lines

from 2.5 lib:
http://svn.python.org/projects/python/branches/release25-maint/Lib/weakref.py@45853
http://svn.python.org/projects/python/branches/release25-maint/Lib/test/test_weakref.py@53533
http://svn.python.org/projects/python/branches/release25-maint/Lib/test/mapping_tests.py@36025
........
r4518 | pjenvey | 2008-06-02 21:23:59 -0600 (Mon, 02 Jun 2008) | 1 line

make for_type public
........
r4519 | pjenvey | 2008-06-02 21:24:47 -0600 (Mon, 02 Jun 2008) | 6 lines

o use the 2.5 weakref module
o pull apart the _weakref types into their own class files, exposed with
annotations
o add more methods to ProxyType
o use the builtin ReferenceError instead of _weakref's own
........
r4520 | pjenvey | 2008-06-02 21:33:29 -0600 (Mon, 02 Jun 2008) | 1 line

revert mistaken commits in r4519
........
r4522 | pjenvey | 2008-06-02 22:36:28 -0600 (Mon, 02 Jun 2008) | 3 lines

allow udp's connect_ex to be called repeatedly, and make connect_ex return
EISCONN instead of 0 in non blocking mode (like CPython)
........
r4525 | pjenvey | 2008-06-03 17:59:32 -0600 (Tue, 03 Jun 2008) | 7 lines

get test_weakref passing (though it may need further tweaking to be reliable):
o avoid CPythonisms and force extra gc.collect()'ing all over the place
o dance around a potential Jython bug in test_getweakrefs
o fix proxy types, and ReferenceType.toString and __call__ which I
broke in r4519
o fix proxy() not reusing references in every situation
........
r4527 | nriley | 2008-06-03 22:07:27 -0600 (Tue, 03 Jun 2008) | 1 line

correct mistaken translation of CPython sort code; fixes #1835099. Thanks arnebef!
........
r4528 | pjenvey | 2008-06-04 11:46:25 -0600 (Wed, 04 Jun 2008) | 3 lines

fix ReferenceType __new/init__ not fully ignoring keyword args (allowing them
to pass through)
........
r4530 | pjenvey | 2008-06-04 11:52:48 -0600 (Wed, 04 Jun 2008) | 1 line

note a small fixme
........
r4534 | pjenvey | 2008-06-04 19:37:03 -0600 (Wed, 04 Jun 2008) | 3 lines

from
http://svn.python.org/projects/python/branches/release25-maint/Lib/test/test_long.py@52138
........
r4535 | pjenvey | 2008-06-04 19:58:17 -0600 (Wed, 04 Jun 2008) | 6 lines

o fix float __cmp__ to never trigger OverflowErrors
o fix float to long raising a java exception on overflow instead of
OverflowError
o bring in CPython 2.5's test_long now that we pass, minus the case of hex
strings. since we're now failing 2.3's
........
r4536 | pjenvey | 2008-06-04 22:43:32 -0600 (Wed, 04 Jun 2008) | 1 line

long(float('NaN')) -> 0
........
r4537 | pjenvey | 2008-06-05 12:14:18 -0600 (Thu, 05 Jun 2008) | 1 line

utilize Py.NoKeywords
........
r4538 | pjenvey | 2008-06-05 17:52:58 -0600 (Thu, 05 Jun 2008) | 1 line

we need the actual key, not its str, in the KeyError
........
r4539 | pjenvey | 2008-06-05 18:00:48 -0600 (Thu, 05 Jun 2008) | 1 line

fix missing comma which broke test___all__
........
r4541 | pjenvey | 2008-06-05 18:11:05 -0600 (Thu, 05 Jun 2008) | 1 line

match CPython's error message
........
r4542 | pjenvey | 2008-06-05 19:09:00 -0600 (Thu, 05 Jun 2008) | 10 lines

o add support for keyword args to dict and stringmap update
fixes #1814904 and 1814929
o add support for 2.5's dict __missing__
o disallow iteration when the size of the dict or stringmap has changed
o fix dict calling the __setitem__ instead of dict___setitem__ in some cases
o test_call (we no longer pass 2.3's) from:
http://svn.python.org/projects/python/branches/release25-maint/Lib/test/test_call.py@35376
and test_dict from:
http://svn.python.org/projects/python/branches/release25-maint/Lib/test/test_dict.py@54840
........
r4544 | pjenvey | 2008-06-05 22:05:30 -0600 (Thu, 05 Jun 2008) | 3 lines

raise a TypeError on invalid args to float
fixes #1037
patch from lsoto
........
r4546 | fwierzbicki | 2008-06-06 11:36:22 -0600 (Fri, 06 Jun 2008) | 2 lines

Fix and test for issue 1048 (List Comprehension in finally throws NPE).
........
r4550 | fwierzbicki | 2008-06-06 18:43:26 -0600 (Fri, 06 Jun 2008) | 3 lines

Looks like the site.py from CPython works just fine. Also enabled
help() though the __doc__s are definitely lacking :).
........
r4552 | pjenvey | 2008-06-07 13:56:15 -0600 (Sat, 07 Jun 2008) | 1 line

whitespace
........
r4553 | pjenvey | 2008-06-07 15:15:31 -0600 (Sat, 07 Jun 2008) | 7 lines

o have ant generate the version information into a prop file, which is loaded
by a Version class. includes the svn revision when it can get it (unix
platforms)
o support sys.subversion, including the branch info (not always the revision)
o utilize the build info in sys.version and in a CPythonish banner. hardcode
the copyright banner, now includes "help"
........
r4554 | pjenvey | 2008-06-07 16:23:00 -0600 (Sat, 07 Jun 2008) | 4 lines

match class AttributeError messages to CPython
fixes #1039
thanks lsoto
........
r4555 | pjenvey | 2008-06-07 16:38:12 -0600 (Sat, 07 Jun 2008) | 4 lines

fix derived __getitem__ swallowing custom LookupErrors
fixes #1038
patch by lsoto
........
r4556 | pjenvey | 2008-06-07 16:51:14 -0600 (Sat, 07 Jun 2008) | 3 lines

latest template updates. re-enable the random derived template as it hasn't
been updated in a while
........
r4557 | pjenvey | 2008-06-07 16:58:19 -0600 (Sat, 07 Jun 2008) | 1 line

not everyone's ant is supporting propertyfile out of the box, use echo instead
........
r4558 | pjenvey | 2008-06-07 17:04:34 -0600 (Sat, 07 Jun 2008) | 1 line

woops, ensure the dir first. thanks Ariane
........
r4559 | pjenvey | 2008-06-07 18:58:39 -0600 (Sat, 07 Jun 2008) | 2 lines

fix thread.exit from other threads exiting the Jython process
........
r4560 | pjenvey | 2008-06-07 19:01:51 -0600 (Sat, 07 Jun 2008) | 1 line

update the ticket referencing why this is disabled
........
r4561 | pjenvey | 2008-06-07 19:20:26 -0600 (Sat, 07 Jun 2008) | 2 lines

site.py from:
http://svn.python.org/projects/python/branches/release23-maint/Lib/site.py@35472
........
r4562 | pjenvey | 2008-06-07 19:21:23 -0600 (Sat, 07 Jun 2008) | 1 line

special case jython for a Lib/site-packages sitedir
........
r4563 | pjenvey | 2008-06-07 20:13:16 -0600 (Sat, 07 Jun 2008) | 1 line

support -? for help
........
r4564 | pjenvey | 2008-06-07 21:00:45 -0600 (Sat, 07 Jun 2008) | 2 lines

have BaseSet implement Set. not sure why this was commented out
fixes #1009477
........
r4566 | pjenvey | 2008-06-07 22:25:33 -0600 (Sat, 07 Jun 2008) | 3 lines

give method and builtin method custom hashCodes
fixes #515497
........
r4567 | pjenvey | 2008-06-07 22:36:23 -0600 (Sat, 07 Jun 2008) | 1 line

add a test for #654142 to close it out
........
r4568 | pjenvey | 2008-06-08 14:50:30 -0600 (Sun, 08 Jun 2008) | 2 lines

o more resilient parsing of headURL
o fix sys.version to not be preceeded with 'Jython'
........
r4569 | pjenvey | 2008-06-08 15:03:57 -0600 (Sun, 08 Jun 2008) | 2 lines

o cleanup, coding standards
o frame doesn't have a custom repr in CPython so removed ours
........
r4570 | pjenvey | 2008-06-08 15:34:14 -0600 (Sun, 08 Jun 2008) | 2 lines

o don't break the build if svnversion isn't found on unix
o workaround the lack of osfamily in exec in ant 1.6
........
r4571 | pjenvey | 2008-06-08 15:44:41 -0600 (Sun, 08 Jun 2008) | 1 line

only svnversion on unix
........
r4572 | pjenvey | 2008-06-08 16:55:28 -0600 (Sun, 08 Jun 2008) | 4 lines

fix types to handle __doc__ descriptors
fixes test_descr.docdescriptor, #1605006 and #1639663
thanks leouserz
........
r4573 | pjenvey | 2008-06-08 18:03:23 -0600 (Sun, 08 Jun 2008) | 3 lines

make 2.5's test_tuple pass: don't apply the repeat optimization to
tuple subclasses and strengthen tuple's hashCode
........
r4574 | pjenvey | 2008-06-08 18:24:08 -0600 (Sun, 08 Jun 2008) | 1 line

isCallable instead of __builtin__.callable, whitespace/coding standards
........
r4575 | pjenvey | 2008-06-08 18:24:41 -0600 (Sun, 08 Jun 2008) | 1 line

fix instancemethod() arg 3 not being optional
........
r4576 | pjenvey | 2008-06-08 19:51:15 -0600 (Sun, 08 Jun 2008) | 1 line

woops, restore to \!isCallable which I broke in r4574
........
r4580 | nriley | 2008-06-08 20:39:54 -0600 (Sun, 08 Jun 2008) | 1 line

fix regrtest ant target dependencies so it works from clean
........
r4583 | pjenvey | 2008-06-09 00:49:36 -0600 (Mon, 09 Jun 2008) | 5 lines

o add support for utf8 and utf16 decoders decoding incomplete input. from
CPython r37292 (the rest being in the 2.4 stdlib)
o fix _codecs.utf_16_decode supporting an optional byteorder arg when only
utf_16_ex_decode should
........
r4590 | pjenvey | 2008-06-10 15:23:43 -0600 (Tue, 10 Jun 2008) | 2 lines

match messages to CPython 2.5
........
r4591 | pjenvey | 2008-06-10 15:26:18 -0600 (Tue, 10 Jun 2008) | 3 lines

fix ArgPargser.getInt throwing a ClassCastException instead of OverflowError,
and support float arguments (albeit deprecated)
........
r4592 | pjenvey | 2008-06-10 15:47:26 -0600 (Tue, 10 Jun 2008) | 3 lines

change xrange's implementation to match CPython's and convert it to exposed
annotations
........
r4593 | pjenvey | 2008-06-10 16:40:39 -0600 (Tue, 10 Jun 2008) | 1 line

push BASE64_MAXBIN out to roughly CPython's value, to fix 2.4's test_base64
........
r4594 | pjenvey | 2008-06-11 12:16:34 -0600 (Wed, 11 Jun 2008) | 2 lines

from:
http://svn.python.org/projects/python/branches/release25-maint/Lib/test/test_dumbdbm.py@38966
........
r4595 | pjenvey | 2008-06-11 12:19:21 -0600 (Wed, 11 Jun 2008) | 1 line

explicitly close file handles to fix test_line_endings
........
r4597 | pjenvey | 2008-06-11 15:07:37 -0600 (Wed, 11 Jun 2008) | 1 line

expose __iter__
........
r4598 | pjenvey | 2008-06-11 15:07:54 -0600 (Wed, 11 Jun 2008) | 1 line

update docs per new BASE64_MAXBIN
........
r4601 | pjenvey | 2008-06-11 16:56:31 -0600 (Wed, 11 Jun 2008) | 2 lines

from:
http://svn.python.org/projects/python/branches/release25-maint/Lib/gzip.py@51032
........
r4602 | pjenvey | 2008-06-11 17:05:01 -0600 (Wed, 11 Jun 2008) | 1 line

workaround Java's lack of Z_SYNC_FLUSH
........
r4609 | pjenvey | 2008-06-12 19:34:13 -0600 (Thu, 12 Jun 2008) | 5 lines

the compiler shouldn't inherit the compiler flags from the current
frame on import as this causes futures to leak into different modules.
this was done for #1691096 which is actaully an invalid bug
fixes #1886758
........
r4610 | pjenvey | 2008-06-12 19:38:54 -0600 (Thu, 12 Jun 2008) | 1 line

re-enable test_email after r4609
........
r4611 | pjenvey | 2008-06-12 22:26:20 -0600 (Thu, 12 Jun 2008) | 1 line

a FIXME
........
r4612 | pjenvey | 2008-06-12 22:47:04 -0600 (Thu, 12 Jun 2008) | 1 line

fix test_descr.test_dir dir(None) == dir(Ellipsis)
........
r4614 | pjenvey | 2008-06-13 13:20:35 -0600 (Fri, 13 Jun 2008) | 2 lines

from:
http://svn.python.org/projects/python/branches/release25-maint/Lib/test/test_time.py@50696
........
r4615 | pjenvey | 2008-06-13 13:43:46 -0600 (Fri, 13 Jun 2008) | 5 lines

pass 2.5's test_time:
o support passing None to localtime, gmtime and ctime
o support 0 for month day and julian day, and 2 leap seconds
o disallow timestamps that overflow long. always validate the srftime tuple
........
r4616 | pjenvey | 2008-06-13 16:30:01 -0600 (Fri, 13 Jun 2008) | 2 lines

move Jython's custom test_descr tests into test_descr_jy
........
r4617 | pjenvey | 2008-06-13 18:56:25 -0600 (Fri, 13 Jun 2008) | 5 lines

fixes for test_descr:
o disallow empty separators to str.split
o don't expose complex.__cmp__
o be careful of kwargs in our file __new__ hack
........
r4618 | pjenvey | 2008-06-13 20:00:53 -0600 (Fri, 13 Jun 2008) | 2 lines

handle super subclasses in its __get__. fixes test_descr.supers
........
r4619 | pjenvey | 2008-06-13 21:37:27 -0600 (Fri, 13 Jun 2008) | 1 line

add toString, cleanup/coding standards
........
r4620 | pjenvey | 2008-06-13 21:51:58 -0600 (Fri, 13 Jun 2008) | 2 lines

allow super to be used with a proxy for obj
fixes test_descr.proxysuper
........
r4621 | pjenvey | 2008-06-13 23:37:36 -0600 (Fri, 13 Jun 2008) | 3 lines

fix PyClassMethodDescr.__get__ called with one arg, or None as one of
the args. convert to exposed annotations
........
r4622 | pjenvey | 2008-06-14 13:48:57 -0600 (Sat, 14 Jun 2008) | 4 lines

lame hack to make derived class generation handle package and import statements
for classes outside of org.python.core. a hacky solution because the java
generation code is really difficult about letting me do this correctly
........
r4623 | pjenvey | 2008-06-14 13:59:37 -0600 (Sat, 14 Jun 2008) | 2 lines

o __delslice__ step is optional
o fix ProxyType to use the canonical version of __setslice__
........
r4624 | pjenvey | 2008-06-14 14:25:11 -0600 (Sat, 14 Jun 2008) | 4 lines

fix test_descr.overloading: dervied classes weren't supporting
__set/delslice__. also fix __getslice__ eating LookupErrors while I'm here
(like __getitem__ used to)
........
r4625 | pjenvey | 2008-06-14 15:19:39 -0600 (Sat, 14 Jun 2008) | 2 lines

we pass the stock test_pow now
........
r4626 | pjenvey | 2008-06-14 16:27:30 -0600 (Sat, 14 Jun 2008) | 4 lines

fix ternary pow to always attempt __pow__ even when coercion fails (__pow__ on
'new style numbers' doesn't require that types match). fallback to binary pow
when modulo is None
........
r4627 | pjenvey | 2008-06-14 16:30:17 -0600 (Sat, 14 Jun 2008) | 3 lines

fix deriveds not supporting the modulo arg to __pow__. with r4626, fixes
test_descr.binopoverride
........
r4629 | pjenvey | 2008-06-14 17:52:25 -0600 (Sat, 14 Jun 2008) | 5 lines

o make property use fget's docstring if one wasn't specified. fixes 2.5
test_descr.properties
o cleanup. We don't need ExposedSets here because TypeErrors are thrown for us
without them. coding standards
........
r4630 | pjenvey | 2008-06-14 18:38:46 -0600 (Sat, 14 Jun 2008) | 1 line

more __get__ arg2 is optional
........
r4631 | pjenvey | 2008-06-14 18:42:42 -0600 (Sat, 14 Jun 2008) | 3 lines

o fix PyModule's lack of getDict breaking 2.5 test_descr.setdict
o match lack of __setitem__ exception to CPython
........
r4635 | pjenvey | 2008-06-14 21:11:06 -0600 (Sat, 14 Jun 2008) | 1 line

match message to CPython's
........
r4636 | pjenvey | 2008-06-14 21:12:12 -0600 (Sat, 14 Jun 2008) | 3 lines

raise a TypeError when __init__ returns anything other than None. Also prevents
__init__ from being a generator. fixes 2.5 test_descr.test_init
........
r4637 | pjenvey | 2008-06-14 21:13:49 -0600 (Sat, 14 Jun 2008) | 2 lines

update to:
http://svn.python.org/projects/python/branches/release25-maint/Lib/test/test_descr.py@60060
........
r4638 | pjenvey | 2008-06-14 21:17:48 -0600 (Sat, 14 Jun 2008) | 1 line

re-integrate the ability to run one test from the command line
........
r4639 | pjenvey | 2008-06-14 22:35:29 -0600 (Sat, 14 Jun 2008) | 4 lines

enable 2.5's test_descr in the regrtest with various tests disabled. give
regrtest more memory for now until slices are fixed (they're currently broken,
causing test_descr.slottrash to use a lot of memory)
........
r4641 | pjenvey | 2008-06-15 14:29:43 -0600 (Sun, 15 Jun 2008) | 2 lines

context manager-ize file
........
r4649 | fwierzbicki | 2008-06-16 07:17:01 -0600 (Mon, 16 Jun 2008) | 3 lines

Renamed test classes to clearly label them as testers. Eventually
these should be moved out of the core src.
........
r4650 | fwierzbicki | 2008-06-16 07:20:35 -0600 (Mon, 16 Jun 2008) | 2 lines

Oops some experimental code leaked into my last commit. Reverting.
........
r4663 | zyasoft | 2008-06-17 13:24:30 -0600 (Tue, 17 Jun 2008) | 1 line

Added 2.5 builtins all, any
........
r4664 | zyasoft | 2008-06-17 19:52:22 -0600 (Tue, 17 Jun 2008) | 1 line

min, max now take the keyword arg key, a 2.5 feature
........
r4666 | pjenvey | 2008-06-18 14:01:28 -0600 (Wed, 18 Jun 2008) | 3 lines

fix slotted objects always having a __dict__ which was broken in r4303. match
the CPython slot behavior in test_slots_jy. and fix the test_descr slots tests
........
r4668 | pjenvey | 2008-06-18 15:16:53 -0600 (Wed, 18 Jun 2008) | 2 lines

__findattr__ is unnecessary here
spotted by lsoto
........
r4669 | pjenvey | 2008-06-18 16:08:10 -0600 (Wed, 18 Jun 2008) | 2 lines

add local to the derived mappings and update it
noticed by lsoto
........
r4670 | pjenvey | 2008-06-18 17:21:51 -0600 (Wed, 18 Jun 2008) | 5 lines

o fix mro_internal to validate mro() results with solid_base
o rerarrange solid_base to search down __bases__ instead of __mro__ which may
not be defined yet
fixes #1056 and test_descr.altmro
........
r4671 | pjenvey | 2008-06-18 17:38:28 -0600 (Wed, 18 Jun 2008) | 1 line

make passable on CPython
........
r4675 | fwierzbicki | 2008-06-19 09:16:28 -0600 (Thu, 19 Jun 2008) | 3 lines

Added (C) at the request of a PhD. student using this and in
accordance with the PSF contributor agreement that I signed.
........
r4681 | pjenvey | 2008-06-19 18:02:52 -0600 (Thu, 19 Jun 2008) | 2 lines

apply -Qnew division to derived classes, not just builtins
........
r4685 | pjenvey | 2008-06-19 19:22:15 -0600 (Thu, 19 Jun 2008) | 1 line

forgot to expose __ior__
........
r4687 | nriley | 2008-06-19 20:54:06 -0600 (Thu, 19 Jun 2008) | 1 line

Don't install profile.properties for now; it breaks --profile.
........
r4688 | nriley | 2008-06-19 20:56:16 -0600 (Thu, 19 Jun 2008) | 1 line

Run Jython launcher test from ant.
........
r4689 | nriley | 2008-06-19 21:00:18 -0600 (Thu, 19 Jun 2008) | 1 line

Use correct variable name for classpath delimiter when profiling; test this behavior (related JRuby bug filed as JRUBY-2678).
........
r4690 | nriley | 2008-06-19 21:02:18 -0600 (Thu, 19 Jun 2008) | 1 line

Use the system class loader as a fallback if the current class loader is null. Supports additions to the classpath when Jython is loaded from the boot classpath (e.g., by the bin/jython launcher script.)
........
r4691 | pjenvey | 2008-06-19 23:20:10 -0600 (Thu, 19 Jun 2008) | 3 lines

update operator to 2.5 (minus the inplace operations). add attr/itemgetter and
nuke __is__ and __is_not__ which don't exist
........
r4696 | pjenvey | 2008-06-20 01:37:56 -0600 (Fri, 20 Jun 2008) | 5 lines

o refactor long __new__ to work like CPython's. eases the migration to allowing
__long__ to return PyIntegers
o don't support __tojava__ conversion here, PyObjects should just implement
__long__ anyway
........
r4697 | pjenvey | 2008-06-20 02:03:16 -0600 (Fri, 20 Jun 2008) | 1 line

update notimplemented failure info
........
r4699 | pjenvey | 2008-06-20 02:28:26 -0600 (Fri, 20 Jun 2008) | 2 lines

allow __long__ to return a PyInteger
........
r4700 | pjenvey | 2008-06-20 02:29:23 -0600 (Fri, 20 Jun 2008) | 1 line

allow deriveds __long__ to return a PyInteger
........
r4703 | pjenvey | 2008-06-20 11:08:28 -0600 (Fri, 20 Jun 2008) | 2 lines

only allow actual strings (not even their subclasses) to be interned
for 2.5's test_built.test_intern
........
r4707 | pjenvey | 2008-06-20 13:07:18 -0600 (Fri, 20 Jun 2008) | 1 line

allow old style __long__ to return a PyInteger
........
r4713 | pjenvey | 2008-06-22 15:15:16 -0600 (Sun, 22 Jun 2008) | 2 lines

throw a TypeError when __long__ isn't supported
fixes #1059
........

2008-06-23 20:19:38 Tree
[1c156d] by thobes@79b11f02-f210-0410-99aa-fb189b3d1ac4

Created a branch for the advanced compiler.

2008-06-19 08:32:01 Tree
[36659c] by pjenvey@79b11f02-f210-0410-99aa-fb189b3d1ac4

Merged revisions 4326-4415 via svnmerge from
https://jython.svn.sourceforge.net/svnroot/jython/trunk/jython

........
r4327 | pjenvey | 2008-04-08 23:59:40 -0700 (Tue, 08 Apr 2008) | 2 lines

really fix #1501932 this time, catch *any* exception in hasattr. with test
........
r4328 | fwierzbicki | 2008-04-09 06:59:40 -0700 (Wed, 09 Apr 2008) | 4 lines

Try #2 to get the new parser into trunk. Includes fix of antlr_gen task from Tobias
and a conditional antlr.notneeded that should only rebuild the grammar stuff if there
has been a change.
........
r4329 | fwierzbicki | 2008-04-09 07:08:31 -0700 (Wed, 09 Apr 2008) | 2 lines

Make the ast dir only generate the ast files (the grammar building is now in trunk).
........
r4330 | pjenvey | 2008-04-09 17:39:49 -0700 (Wed, 09 Apr 2008) | 1 line

NotImplemented is Trueish
........
r4331 | pjenvey | 2008-04-09 19:45:22 -0700 (Wed, 09 Apr 2008) | 1 line

use array.ArrayType
........
r4332 | fwierzbicki | 2008-04-10 17:09:27 -0700 (Thu, 10 Apr 2008) | 14 lines

The start of an _ast module. It relies on the new antlr parser which is now in trunk.
For now you must put antlr-runtime-3.0.1 (which is in the extlibs dir) in your
classpath to use it. Using it looks like this:

>>> import _ast
>>> compile("a=1", "<unknown>", "exec", _ast.PyCF_ONLY_AST)
tree: (Module (Assign (Target (Name a)) (Value (Num 1))))

Which outputs a tree representation for now.

If you do not use _ast or the PyCF_ONLY_AST flag, you should not need
antlr-runtime-3.0.1.jar in your path.
........
r4333 | pjenvey | 2008-04-11 16:23:41 -0700 (Fri, 11 Apr 2008) | 1 line

support the optional file arg to load_source
........
r4334 | pjenvey | 2008-04-12 13:56:15 -0700 (Sat, 12 Apr 2008) | 1 line

remove the now fully deprecated pieces of xrange, cleanup
........
r4335 | pjenvey | 2008-04-12 13:57:06 -0700 (Sat, 12 Apr 2008) | 1 line

more 'at' in repr
........
r4336 | pjenvey | 2008-04-12 14:48:49 -0700 (Sat, 12 Apr 2008) | 2 lines

make new types use a copy of their class dict instead of the actual dict
........
r4337 | pjenvey | 2008-04-12 14:51:42 -0700 (Sat, 12 Apr 2008) | 1 line

small cleanup
........
r4338 | pjenvey | 2008-04-12 17:11:50 -0700 (Sat, 12 Apr 2008) | 1 line

test_types_jy is testing classes not just types, rename to test_class_jy
........
r4339 | pjenvey | 2008-04-12 18:25:23 -0700 (Sat, 12 Apr 2008) | 23 lines

we use only PyFrame.getf_locals() to get at locals, whereas CPython in general
has two ways of getting them:

o the quicker way: grab the 'current' locals via f_locals directly: this is
what CPython's LOAD/STORE/DELETE_NAME and LOAD_LOCALS do
o the thorough way: first merge fast locals (cellvars and freevars) into
f_locals, then grab the updated f_locals: this is what locals() does, and also
what eval/exec/execfile etc. use

our LOAD_LOCALS equivalent doesn't use the quicker way and that can cause a
difference in the contents of class dicts passed to metaclasses'
__new__(M, name, bases, class_dict) (see the accompanying test)

so I've changed our LOAD_LOCALS to use the quick way, to fix this issue, and
also made our LOAD/STORE/DELETE_NAME equivalents use the quicker way too (which
means closures do a little less work)

Now we have:
o PyFrame.getf_locals -> quick way (and could be deprecated, only the compiler
uses it. just grab f_locals directly instead)
o new PyFrame.getLocals -> thorough way, used everywhere else outside of
PyFrame
........
r4340 | pjenvey | 2008-04-13 11:42:31 -0700 (Sun, 13 Apr 2008) | 1 line

whitespace
........
r4341 | pjenvey | 2008-04-13 11:51:14 -0700 (Sun, 13 Apr 2008) | 2 lines

py_compile from
http://svn.python.org/projects/python/branches/release25-maint/Lib/py_compile.py@38987
........
r4342 | pjenvey | 2008-04-13 11:55:39 -0700 (Sun, 13 Apr 2008) | 4 lines

rename our builtin py_compile to _py_compile, and call it from CPython's pure
python py_compile. adds support for compile's doraise option, which by default
swallows exceptions
........
r4343 | pjenvey | 2008-04-13 12:08:52 -0700 (Sun, 13 Apr 2008) | 3 lines

fix re.subx on empty strings with no match: CPython actually returns
pattern[0:0] in this case
........
r4344 | fwierzbicki | 2008-04-14 07:59:19 -0700 (Mon, 14 Apr 2008) | 2 lines

Added null check for compile_flags. Thanks to Georgy Berdyshev for pointing out the bug.
........
r4345 | pjenvey | 2008-04-14 17:07:42 -0700 (Mon, 14 Apr 2008) | 15 lines

o automatically define toString() in derived classes (instead of including it
from object.derived) but allow disabling of it. for PyString/UnicodeDerived;
they need toString() to pass through instead of calling __repr__ as that breaks
the expectation that PyString.toString() returns the real string

this means PyObject.toString() from java code always returns __repr__ *except*
for PyString/Unicode, which return the actual string
(thanks Mehendran)

o update the "should return a <type>" message in derived class descriptors to
match CPython 2.5's
o from the full gderived.py run: removed a useless constructor from
PyUnicodeDerived and updated the collections derives' getattr/setattr to use
interned strings
........
r4346 | pjenvey | 2008-04-14 18:01:18 -0700 (Mon, 14 Apr 2008) | 6 lines

o fix str.__repr__() to match repr(str) and separate it from toString
fixes #1510227, #1782565
mostly from Mehendran's patch in #1785475
o fix str.__str__ on str subclasses not returning a normal str
o rename unicode_toString to unicode___repr__ (that's what it really is)
........
r4347 | pjenvey | 2008-04-14 20:24:23 -0700 (Mon, 14 Apr 2008) | 2 lines

always escape \t \n \r, even for unicode, never escape \b \f
........
r4348 | pjenvey | 2008-04-14 21:11:00 -0700 (Mon, 14 Apr 2008) | 2 lines

fix single quotes not being escaped in unicode repr
........
r4349 | pjenvey | 2008-04-16 11:39:50 -0700 (Wed, 16 Apr 2008) | 9 lines

include the jarjar'd asm ClassReader in our jar and use it to get the
real class name of $py.class files before importing them. Allows us to
import misnamed/mislabeled $py.class files

the import statement has been hacking around this by recompiling the
.py (if it exists) when it happens. more importantly imp.load_module
(and load_compiled when we support it) don't and shouldn't workaround
this, instead load_module has been throwing NoClassDefFoundErrors
........
r4350 | pjenvey | 2008-04-16 16:22:27 -0700 (Wed, 16 Apr 2008) | 5 lines

add a special case to _binop_rule to satisfy CPython's handling of
__add__/__radd__ with str/unicode and str/unicode subclasses. replaces
the broken bandaid for this applied in r4020
Thanks Samuele Pedroni for the solution and test_basic
........
r4351 | pjenvey | 2008-04-16 23:53:27 -0700 (Wed, 16 Apr 2008) | 1 line

allow setDaemon(None) or anything else by filtering it through bool
........
r4352 | pjenvey | 2008-04-17 14:34:50 -0700 (Thu, 17 Apr 2008) | 3 lines

new style classes must define a __call__ to be callable, a __call__ provided by
__getattr__ doesn't count
........
r4353 | zyasoft | 2008-04-17 14:59:21 -0700 (Thu, 17 Apr 2008) | 1 line

Changes to enable dateutil on jython, plus a fix for a potential race on time.clock's initialization
........
r4354 | pjenvey | 2008-04-17 21:05:56 -0700 (Thu, 17 Apr 2008) | 2 lines

core.util is a more appropriate place for ByteSwapper and FileUtil
........
r4355 | fwierzbicki | 2008-04-18 12:03:31 -0700 (Fri, 18 Apr 2008) | 3 lines

Better coverage of _ast. Made AST an interface so that enums (like expr_contextType)
can pass instanceof(x, _ast.AST).
........
r4356 | pjenvey | 2008-04-18 13:45:13 -0700 (Fri, 18 Apr 2008) | 6 lines

o kill the annoying Java version of the os module. if it was actually ever
needed, it's not anymore
o store os.environ in sys so it can survive a reload of the os
module. this was causing a really vague new bug in test_javashell now
that test_import was actually able to reload os via imp.load_module
........
r4357 | pjenvey | 2008-04-18 15:03:25 -0700 (Fri, 18 Apr 2008) | 3 lines

avoid a ClassCastException when an object's __int__ surprises formatInteger by
returning a long
........
r4358 | pjenvey | 2008-04-18 16:40:39 -0700 (Fri, 18 Apr 2008) | 3 lines

utilize the str formatter to make float's repr/str resemble CPython's a little
more
........
r4359 | pjenvey | 2008-04-18 19:46:32 -0700 (Fri, 18 Apr 2008) | 2 lines

fix the default PyDictionary constructor unnecessarily creating two CHMs
........
r4360 | fwierzbicki | 2008-04-19 07:23:21 -0700 (Sat, 19 Apr 2008) | 3 lines

Expose __iter__ on tuple, list, and array types. Still need to do the
same on the Derived types.
........
r4361 | pjenvey | 2008-04-19 14:14:24 -0700 (Sat, 19 Apr 2008) | 7 lines

make the 'g' float formatter closer to CPython's output:
o format as exponent when -4 <= exponent < precision
o handle infinities
o pass the altFlag to the formatters
o fix precision of the exponential formatter's exponent
o don't need the altFlag padding when doing exponential format
........
r4362 | pjenvey | 2008-04-19 14:35:10 -0700 (Sat, 19 Apr 2008) | 3 lines

kill ExtraMath's hypot and log10 for their equivalents in Java 1.5's
java.lang.Math
........
r4363 | fwierzbicki | 2008-04-20 09:05:22 -0700 (Sun, 20 Apr 2008) | 2 lines

Added lineno and col_offset, and some debug info to PythonTree.
........
r4364 | fwierzbicki | 2008-04-20 11:19:34 -0700 (Sun, 20 Apr 2008) | 2 lines

remove debug statement.
........
r4365 | fwierzbicki | 2008-04-20 19:57:33 -0700 (Sun, 20 Apr 2008) | 2 lines

Fixed up to take advantage of Philip's recent improvements.
........
r4366 | fwierzbicki | 2008-04-21 14:58:41 -0700 (Mon, 21 Apr 2008) | 5 lines

took test_ast.py from CPython's release25-maint branch revision 62446. Modified a bit
to "shape" the results of __class__.__name__ since the results are different for Jython
for the moment (and may always differ, we'll see). About 50% of the tests work at the
moment. Also cleaned up the ast folder a bit.
........
r4367 | fwierzbicki | 2008-04-21 19:41:48 -0700 (Mon, 21 Apr 2008) | 3 lines

Better match line numbers and char positions with test_ast. Commented out the failures
for now. Also added support for eval and single_input.
........
r4368 | fwierzbicki | 2008-04-22 08:46:48 -0700 (Tue, 22 Apr 2008) | 2 lines

All but a couple of test_ast tests now pass.
........
r4369 | fwierzbicki | 2008-04-22 17:12:34 -0700 (Tue, 22 Apr 2008) | 2 lines

The rest of the modified test_ast now works.
........
r4370 | pjenvey | 2008-04-22 21:30:28 -0700 (Tue, 22 Apr 2008) | 1 line

expose __iter__ on sets
........
r4371 | pjenvey | 2008-04-24 21:22:43 -0700 (Thu, 24 Apr 2008) | 3 lines

use System.out and friends instead of FileDescriptor.out and friends, which
aren't correct under nailgun
........
r4374 | pjenvey | 2008-04-30 20:59:37 -0700 (Wed, 30 Apr 2008) | 2 lines

small rearrange per coding standards
........
r4379 | fwierzbicki | 2008-05-02 18:20:26 -0700 (Fri, 02 May 2008) | 3 lines

Add warning that Derived classes are generated. No warning added to Derived classes
in org/python/modules yet since they look like they where hand modified a bit.
........
r4391 | pjenvey | 2008-05-13 12:28:35 -0700 (Tue, 13 May 2008) | 1 line

whitespace
........
r4393 | pjenvey | 2008-05-13 14:15:18 -0700 (Tue, 13 May 2008) | 3 lines

update test_funcattrs to the latest 2.5 version, from:
http://svn.python.org/projects/python/branches/release25-maint/Lib/test/test_funcattrs.py@59171
........
r4394 | pjenvey | 2008-05-13 14:34:49 -0700 (Tue, 13 May 2008) | 8 lines

o cleanup PyFunction, use exposed annotations
o copy the new.function arg checking code into PyFunction __new__, so
types.FunctionType supports the same args. add test_func_jy with jek's test
cases
o add cell.cell_contents and an appropriate repr. use exposed annotations
fixes #1034
thanks Jason Kirtland
........
r4395 | pjenvey | 2008-05-13 15:18:05 -0700 (Tue, 13 May 2008) | 1 line

unused import
........
r4396 | pjenvey | 2008-05-13 15:30:47 -0700 (Tue, 13 May 2008) | 3 lines

new.py from:
http://svn.python.org/projects/python/branches/release25-maint/Lib/new.py@27241
........
r4397 | fwierzbicki | 2008-05-13 16:30:08 -0700 (Tue, 13 May 2008) | 3 lines

Extracted from astlib.py -- simple pretty printer that works for
both jython and python.
........
r4398 | fwierzbicki | 2008-05-13 16:31:56 -0700 (Tue, 13 May 2008) | 2 lines

change def name.
........
r4399 | pjenvey | 2008-05-13 18:13:52 -0700 (Tue, 13 May 2008) | 1 line

make types.InstanceType() handle None as second arg
........
r4400 | pjenvey | 2008-05-13 18:14:20 -0700 (Tue, 13 May 2008) | 6 lines

o match types.MethodType() (new PyMethod()) args to CPython's:
(func, self, type) instead of (self, func, type)
o kill most of newmodule for the simpler pure python version. we can't dump it
completely until a better workaround for the new.classobj hack is done; in the
meantime, keep it around at _new.classobj
........
r4401 | pjenvey | 2008-05-13 18:34:54 -0700 (Tue, 13 May 2008) | 1 line

ensure new.classobj can create new style classes
........
r4404 | fwierzbicki | 2008-05-15 12:13:41 -0700 (Thu, 15 May 2008) | 3 lines

Finally found the fix for empty modules. Empty Lexer output causes a null tree to
return. I've fixed by creating an empty Module node by hand for empty inuput.
........
r4406 | pjenvey | 2008-05-15 13:42:23 -0700 (Thu, 15 May 2008) | 3 lines

maintain func_defaults and func_closure as null instead of
Py.EmptyObjects/EmptyTuple, to more closely resemble CPython
........
r4409 | pjenvey | 2008-05-15 16:52:01 -0700 (Thu, 15 May 2008) | 4 lines

let null return values pass thru PyDataDescr get as None instead of triggering
AttributeErrors. usually what we want anyway and makes @ExposedGet
implementations easier
........
r4410 | pjenvey | 2008-05-15 19:03:31 -0700 (Thu, 15 May 2008) | 3 lines

fix isinstance(newstyle_obj, oldstyle_class) always failing when it can be True
via multiple inheritance. make isInstance more closely resemble CPython's
........
r4411 | pjenvey | 2008-05-16 10:55:45 -0700 (Fri, 16 May 2008) | 5 lines

fix sets incorrectly raising a TypeError when asked for superset/subset on
non-set iterables. also fixes a few other tests from 2.5 test_set
fixes #1035
patch from Jonathan Ellis
........
r4412 | pjenvey | 2008-05-16 13:27:52 -0700 (Fri, 16 May 2008) | 1 line

more ExposedGet now translating null for us
........
r4413 | pjenvey | 2008-05-16 14:19:03 -0700 (Fri, 16 May 2008) | 2 lines

cleanup PyMethod, use exposed annotations
........
r4414 | pjenvey | 2008-05-16 16:30:52 -0700 (Fri, 16 May 2008) | 1 line

more appropriate method name
........
r4415 | pjenvey | 2008-05-16 17:07:22 -0700 (Fri, 16 May 2008) | 7 lines

use our Java set implementation only for the builtin sets, and use CPythonLib's
sets module. so we can be more compatible with both versions compared to
CPython, as there's a few differences between the two. moves modules.sets into
core, PyImmutableSet -> PyFrozenSet, and causes test___all__ and CPythonLib's
test_sets to now pass
patch from Jonathan Ellis
........

2008-05-17 02:23:25 Tree
[40cd71] by fwierzbicki@79b11f02-f210-0410-99aa-fb189b3d1ac4

Reviving the asm translation of the compiler to start a more gradual translation to the new compiler. This time for real.

2008-05-08 17:04:53 Tree
[9ddc4e] by pjenvey@79b11f02-f210-0410-99aa-fb189b3d1ac4

branch for PEP 352 (new style exceptions). this will have to stay on a branch
for the time being as it breaks a few things in the 2.3 stdlib

2008-04-04 22:39:09 Tree
[41d0f5] by zyasoft@79b11f02-f210-0410-99aa-fb189b3d1ac4

Merged revisions 4092,4096,4098-4114,4117-4120,4122-4132,4134-4135,4138,4140-4141,4143-4144,4146-4153,4156-4157,4159-4160,4162-4164 via svnmerge from
https://jython.svn.sourceforge.net/svnroot/jython/trunk/jython

........
r4092 | cgroves | 2008-02-03 14:37:51 -0700 (Sun, 03 Feb 2008) | 1 line

convert array, bool, classmethod, complex, enumerate and float from templates to annotations
........
r4096 | cgroves | 2008-02-03 23:27:28 -0700 (Sun, 03 Feb 2008) | 1 line

Migrate bugtests 51-75 into Lib/test from Jens Abraham's work on GHOP task 281: http://code.google.com/p/google-highly-open-participation-psf/issues/detail?id=281
........
r4098 | pjenvey | 2008-02-04 20:04:10 -0700 (Mon, 04 Feb 2008) | 1 line

from http://svn.python.org/projects/python/branches/release23-maint/Lib/test/test_logging.py@33658
........
r4099 | pjenvey | 2008-02-04 20:10:06 -0700 (Mon, 04 Feb 2008) | 1 line

use cpython_compatibile_select on jython
........
r4100 | pjenvey | 2008-02-04 22:05:41 -0700 (Mon, 04 Feb 2008) | 1 line

test_logging and test_pickletools now pass
........
r4101 | pjenvey | 2008-02-04 23:18:10 -0700 (Mon, 04 Feb 2008) | 1 line

test_iterlen currently fails, test_email fails due to bug #1886758
........
r4102 | pjenvey | 2008-02-04 23:29:37 -0700 (Mon, 04 Feb 2008) | 1 line

these will run now
........
r4103 | pjenvey | 2008-02-05 20:26:51 -0700 (Tue, 05 Feb 2008) | 3 lines

make addBuilder do more to fully accomplish what add/fillFromClass do, for
types initialized before their builders are
........
r4104 | pjenvey | 2008-02-06 18:01:51 -0700 (Wed, 06 Feb 2008) | 2 lines

organize imports, rearrange javadocs as per coding standards
........
r4105 | pjenvey | 2008-02-06 19:29:38 -0700 (Wed, 06 Feb 2008) | 1 line

convert deque to expose annotations
........
r4106 | pjenvey | 2008-02-07 13:28:53 -0700 (Thu, 07 Feb 2008) | 3 lines

kill safeRepr, which is really just an older, bogus way of getting the type
name. use getType().fastGetName() instead
........
r4107 | pjenvey | 2008-02-07 14:40:20 -0700 (Thu, 07 Feb 2008) | 6 lines

o change PyType's name attribute for builtin types; include the fully
qualified name of the type including its module, like CPython does. allows
builtin types in org.python.modules to have a valid __module__ value
o kill PyType's getFullName; it's only really needed in a couple places and
misused for fastGetName everywhere else
........
r4108 | pjenvey | 2008-02-07 15:05:29 -0700 (Thu, 07 Feb 2008) | 1 line

match error messages to CPython
........
r4109 | pjenvey | 2008-02-07 15:27:10 -0700 (Thu, 07 Feb 2008) | 3 lines

fallthrough method's __name__ and __doc__ to the underlying callable instead of
copying them. fixes the fact that we weren't always copying them
........
r4110 | pjenvey | 2008-02-07 16:29:55 -0700 (Thu, 07 Feb 2008) | 1 line

followup r4109: builtin functions need to expose getDoc too. fixes test_generators
........
r4111 | pjenvey | 2008-02-07 16:35:46 -0700 (Thu, 07 Feb 2008) | 4 lines

make Py.makeClass fallback to the builtins when looking for __name__ (for the
class's __module__), like CPython does -- a corner case that probably only
doctests care about. with tests
........
r4112 | pjenvey | 2008-02-07 16:36:46 -0700 (Thu, 07 Feb 2008) | 1 line

woops, don't need this exception
........
r4113 | pjenvey | 2008-02-08 13:39:38 -0700 (Fri, 08 Feb 2008) | 1 line

for changes to the compiler package, from http://svn.python.org/projects/python/branches/release23-maint/Lib@60660
........
r4114 | pjenvey | 2008-02-08 13:45:13 -0700 (Fri, 08 Feb 2008) | 4 lines

include a stripped version of the compiler package that avoids the parser
module and imp.get_magic, which we lack. mainly for nose which just uses
compiler.consts.CO_GENERATOR
........
r4117 | pjenvey | 2008-02-09 14:28:29 -0700 (Sat, 09 Feb 2008) | 5 lines

disable PyGenerator's finalize for now. for some reason the garbage collector
attempts to finalize generators before they're done with in some situations (in
particular seen when running the full nose test suite). finalize is for 2.5
style generators which we don't fully support now anyway
........
r4118 | pjenvey | 2008-02-09 14:40:55 -0700 (Sat, 09 Feb 2008) | 2 lines

handle $py.class as well as .pyc/o
........
r4119 | pjenvey | 2008-02-09 15:23:34 -0700 (Sat, 09 Feb 2008) | 1 line

make this more like the changeset for nose
........
r4120 | cgroves | 2008-02-09 19:24:14 -0700 (Sat, 09 Feb 2008) | 6 lines

Allow classmethods to be exposed on builtin types using @ExposedClassMethod and
expose dict with its fromkeys classmethod using that.

Expose defaultdict.
........
r4122 | cgroves | 2008-02-10 01:16:48 -0700 (Sun, 10 Feb 2008) | 1 line

respect verbose passed into test_main
........
r4123 | cgroves | 2008-02-10 01:17:36 -0700 (Sun, 10 Feb 2008) | 1 line

Convert remaining types in mappings to annotations
........
r4124 | cgroves | 2008-02-10 01:34:03 -0700 (Sun, 10 Feb 2008) | 1 line

Convert random which had an expose but was excluded from mappings for some reason and staticmethod which had generated code but no .expose. Should be the last of the gexpose code
........
r4125 | cgroves | 2008-02-10 01:46:19 -0700 (Sun, 10 Feb 2008) | 1 line

Clear out the gexpose newstyle setup
........
r4126 | pjenvey | 2008-02-10 16:30:29 -0700 (Sun, 10 Feb 2008) | 1 line

use a consistent docstring type
........
r4127 | cgroves | 2008-02-10 17:52:05 -0700 (Sun, 10 Feb 2008) | 6 lines

Make PyDataDescr a newstyle type and add setType on it so it can be given its
type by DescriptorExposer without requiring a PyType.fromClass.

Remove the reflection based descriptors.
........
r4128 | cgroves | 2008-02-10 18:25:46 -0700 (Sun, 10 Feb 2008) | 1 line

add TYPE to zipimporter so its builder will be setup before it's referenced in zipimport or ySystemState
........
r4129 | cgroves | 2008-02-10 18:28:20 -0700 (Sun, 10 Feb 2008) | 1 line

Track the types that will be initialized before their builders are assigned in Py.BOOTSTRAP_TYPES and complain if we bootstrap something that isn't in that set, or if something from that set isn't bootstrapped
........
r4130 | cgroves | 2008-02-13 02:52:10 -0700 (Wed, 13 Feb 2008) | 1 line

Get test_extcall to line up with jython. Leave single line test_extcall in output so it'll override the one from CPythonLib
........
r4131 | cgroves | 2008-02-13 02:52:28 -0700 (Wed, 13 Feb 2008) | 1 line

Remove unused field
........
r4132 | cgroves | 2008-02-16 00:34:24 -0700 (Sat, 16 Feb 2008) | 1 line

Make the module the interpreter is running __main__ not main
........
r4134 | cgroves | 2008-02-16 11:27:11 -0700 (Sat, 16 Feb 2008) | 1 line

Use SO_REUSEADDR so this can run repeatedly
........
r4135 | pjenvey | 2008-02-16 14:48:59 -0700 (Sat, 16 Feb 2008) | 3 lines

fallback to the class name of the IOException if its message is null (which is
typical of the nio IOExceptions)
........
r4138 | cgroves | 2008-02-16 18:11:37 -0700 (Sat, 16 Feb 2008) | 1 line

pull the jarjar stuff from asm_compiler over for use with pyasm
........
r4140 | cgroves | 2008-02-16 19:07:46 -0700 (Sat, 16 Feb 2008) | 1 line

Only jarjar if the underlying jars have changed so we don't update jython.jar if nothing has changed
........
r4141 | cgroves | 2008-02-16 19:09:56 -0700 (Sat, 16 Feb 2008) | 1 line

Actually include all the asm jars in checking if jarjar needs to run
........
r4143 | cgroves | 2008-02-16 19:59:09 -0700 (Sat, 16 Feb 2008) | 1 line

Actually use our jarjar'd asm
........
r4144 | cgroves | 2008-02-16 19:59:54 -0700 (Sat, 16 Feb 2008) | 1 line

Should've gone with the last commit
........
r4146 | pjenvey | 2008-02-17 22:31:45 -0700 (Sun, 17 Feb 2008) | 4 lines

flush stdout and stderr upon exit/cleanup of the interpreter, after calling the
exit funcs. fixes:
jython -c "import sys; sys.stdout.write('foo')"
........
r4147 | pjenvey | 2008-02-18 00:23:06 -0700 (Mon, 18 Feb 2008) | 4 lines

we need the absolute path of the shell exe when running a subprocess with a
custom environ that lacks PATH. cheat and default to /bin/sh for POSIX, as it's
worked well for CPython
........
r4148 | pjenvey | 2008-02-18 00:46:16 -0700 (Mon, 18 Feb 2008) | 4 lines

file doesn't need __nonzero__, it defaults to True since it lacks __len__. it
doesn't need __cmp__ either; having it actually caused a TypeError with:
tempfile.TemporaryFile() == 1
........
r4149 | pjenvey | 2008-02-18 12:01:57 -0700 (Mon, 18 Feb 2008) | 2 lines

subprocess from http://svn.python.org/projects/python/trunk@60896
........
r4150 | pjenvey | 2008-02-18 12:22:48 -0700 (Mon, 18 Feb 2008) | 6 lines

jython support for subprocess. PIPEs expose the Java Process' streams,
otherwise we sync to the Process' streams with a thread, like popen2.
stdin=None (share stdin with the parent) could block that coupling thread
forever (and stopping it with Thread.interrupt() would close the parent's
stdin), so we don't support syncing it
........
r4151 | pjenvey | 2008-02-18 19:06:39 -0700 (Mon, 18 Feb 2008) | 5 lines

o only escape command lists, leave the command strings for shell=True alone
(on windows)
o if starting the process fails for whatever reason just propagate its
IOException up, as an OSError
........
r4152 | pjenvey | 2008-02-18 19:13:55 -0700 (Mon, 18 Feb 2008) | 3 lines

check for EACCESS when opening files for writing and a FileNotFoundException is
thrown
........
r4153 | pjenvey | 2008-02-18 23:16:21 -0700 (Mon, 18 Feb 2008) | 3 lines

from
http://svn.python.org/projects/python/branches/release23-maint/Lib/distutils/command@38383
........
r4156 | pjenvey | 2008-02-19 12:39:23 -0700 (Tue, 19 Feb 2008) | 5 lines

o remove build_ext: I changed my mind on stopping the c compiler here, let's do
it in a JythonCompiler class instead
o bring in distutils.ccompiler from:
http://svn.python.org/projects/python/branches/release23-maint/Lib/distutils/ccompiler.py@37185
........
r4157 | pjenvey | 2008-02-19 13:06:53 -0700 (Tue, 19 Feb 2008) | 5 lines

o add JythonCompiler, which simply raises a CCompiler exception about
extensions not being supported. distributions can catch this and skip optional
extensions on Jython (simplejson and Genshi trunk already do this)
o stub out sysconfig vars that build_ext expects with os2's
........
r4159 | pjenvey | 2008-02-20 20:48:06 -0700 (Wed, 20 Feb 2008) | 4 lines

close the file handle: we might try deleting it before the garbage collector
does and cause this fatal error on Windows: 'The process cannot access the file
because it is being used by another process.'
........
r4160 | pjenvey | 2008-02-20 21:04:06 -0700 (Wed, 20 Feb 2008) | 6 lines

zipimporter fixes for windows:
o convert '/' to File.separator in ZipEntry names, and back to '/' in
getDataStream
o use File.separator in zip_searchorder
o always close zip files when finished
........
r4162 | pjenvey | 2008-02-21 20:19:09 -0700 (Thu, 21 Feb 2008) | 4 lines

o was using an invalid filename on windows instead of the actual testtar
filename
o ensure all file handles are closed to avoid os.remove problems on windows
........
r4163 | pjenvey | 2008-02-21 20:27:10 -0700 (Thu, 21 Feb 2008) | 1 line

close stray file handles
........
r4164 | pjenvey | 2008-02-21 22:55:07 -0700 (Thu, 21 Feb 2008) | 4 lines

o add a chdir test for subprocess
o fix the subprocess command line strings to work on windows
o close file handles
........

2008-02-23 22:31:27 Tree
Older >

Get latest updates about Open Source Projects, Conferences and News.

Sign up for the SourceForge newsletter:

JavaScript is required for this form.





No, thanks