I looked in the browsable svn repo for parallel python at
http://parallelpython.googlecode.com/svn/trunk/
1) OverflowError: long int too large to convert to int
This is a bug in their code (comments are mine) in pytransport.py:
size_packed = self.r.read(struct.calcsize("!Q")) # reads in 8-
character string
msg_len = struct.unpack("!Q", size_packed)[0] # unpacks to a long
value, e.g. "12345678" -> 3544952156018063160L
msg = self.r.read(msg_len) #
3544952156018063160L is too long for file.read, which I presume must
be able to cast as an int
As an example:
>>> sys.stdin.read(3544952156018063160L)
Traceback (most recent call last):
File "<stdin>", line 1, in <module>
OverflowError: long int too large to convert to int
2) UnboundLocalError: local variable 'sresult' referenced before
assignment
This is another bug in their code. This is an error in pp.py (again,
comments are mine, some code omitted for brevity):
try:
# ...snip...
sresult = worker.t.receive() # sets sresult
except:
if self.__exiting:
return
else:
# at this point, __run doesn't return, but sresult was not defined
since the
# execution was unwound because, presumable, the first fatal you
got, the OverflowError
sys.excepthook(*sys.exc_info())
# ...snip...
job.finalize(sresult) # UnboundLocalError here from missing sresult
You can submit these as a bug report to the author if you want - feel
free to copy-paste or link to this post :-)
- zeph
--
http://mail.python.org/mailman/listinfo/python-list
