[Stackless] Stackless Python 3000

[Jeff Senn]

Ditto on the "thanks for effort" !!

On OS-X there are chicken-egg problems because you (or some
automated process) converted the build script to py3k syntax...
which is a dilemma... because I'm not even sure what the
right thing to do is... :-)
e.g. should you need py3k to build py3k?  I guess one could carefully
re-write the script so that it does not depend on differences pre/post  
3k....

Anyway aside from that, building with configure seems to work.

[remember it's like this -- in case someone else wants to try:
export MACOSX_DEPLOYMENT_TARGET="10.4"
./configure --enable-framework --enable-stacklessfewerregisters -- 
enable-universalsdk
]

The multiprocessing module doesn't even build on OS-X --
and test_multiprocessing hangs. But you said you
might not care about that...

I get an occasional failure on test_socket (which I believe
has happened before in prior versions)

And a consistent error on test_asynchat with a bunch of output
like:

+ error: uncaptured python exception, closing channel  
<test.test_asynchat.echo_client at 0x16728d0> (<class  
'AttributeError'>:'socket' object has no attribute  
'handle_close_event' [/Users/senn/src/stackless/py3k/Lib/asyncore.py| 
readwrite|104] [/Users/senn/src/stackless/py3k/Lib/asyncore.py| 
__getattr__|390])

This looks like a general problem with asyncore... and not something
you are likely to have done in the port -- possibly poll() acts a
little differently on OS-X ... I'd have to carefully check diffs to
make sure...

"make teststackless" runs w/o any suspiciousness.

-Jas


On Jul 11, 2008, at 10:37 AM, Richard Tew wrote:

> Hi,
>
> CCP kindly let Kristjan Valur Jonsson and I spend our time this week
> enhancing Python 3000 with tasklets, channels and scheduling; that is,
> porting Stackless Python to it.  CCP will not be using Python 3000 for
> EVE any time in the near future, which makes this is a generous
> donation of our working time.  As of this morning, we are happy with
> the state of it, and can announce that it should be ready for use (as
> much as Python 3000 itself is).
>
> If you can take the time to download it and compile it, then run both
> the Python unit tests and the Stackless unit tests and reply with any
> unexpected failures you might experience, it would be much
> appreciated.
>
> You can check out the source code here:
>  http://svn.python.org/projects/stackless/branches/py3k/
>
> *nix instructions:
>  * Running the Python unit tests:
>      ./python Lib/test/regrtest.py
>  * Running the Stackless unit tests:
>      cd Stackless/unittests
>      ../../python runAll.py
>
> Windows instructions:
>  * Running the Python unit tests:
>      PCbuild\python_d.exe Lib\test\regrtest.py
>  * Running the Stackless unit tests:
>      cd Stackless\unittests
>      ..\..\PCbuild\python_d.exe runAll.py
>
> I am not interested in failures in the Python unit tests if it is this
> one (it happens in the non-Stackless version of the source code) in
> test_multiprocessing:
>
> Traceback (most recent call last):
>  File "Lib\test\test_multiprocessing.py", line 1795, in <module>
> main()
>  File "Lib\test\test_multiprocessing.py", line 1792, in main
> test_main(unittest.TextTestRunner(verbosity=2).run)
>  File "Lib\test\test_multiprocessing.py", line 1771, in test_main
> ManagerMixin.manager.start()
>  File "C:\SVN\Python\stackless\branches\py3k\lib\multiprocessing 
> \managers.py",
> line 499, in start
> self._process.start()
>  File "C:\SVN\Python\stackless\branches\py3k\lib\multiprocessing 
> \process.py",
> line 104, in start
> self._popen = Popen(self)
>  File "C:\SVN\Python\stackless\branches\py3k\lib\multiprocessing 
> \forking.py",
> line 194, in __init__
> dump(process_obj, to_child, HIGHEST_PROTOCOL)
>  File "C:\SVN\Python\stackless\branches\py3k\lib\pickle.py", line  
> 1350, in dump
> Pickler(file, protocol).dump(obj)
>  File "C:\SVN\Python\stackless\branches\py3k\lib\multiprocessing 
> \util.py",
> line 311, in _reduce_method
> if m.__self__ is None:
> AttributeError: 'function' object has no attribute '__self__'
>
> Cheers,
> Richard.
>
> _______________________________________________
> Stackless mailing list
> Stackless at stackless.com
> http://www.stackless.com/mailman/listinfo/stackless
>