[mnet-devel] Help with mnet/fec/mmap/w32

rob kinninmont rob at evilrob.com
Wed Jul 2 18:36:15 BST 2003


Sorry for the delay in getting back to this: I put the mnet stuff on 
the back burner and focussed on other things I needed to get done for a 
while.

On Tuesday, June 24, 2003, at 02:29 PM, Myers W. Carpenter wrote:

> On Tue, Jun 24, 2003 at 01:06:30AM -0700, rob kinninmont wrote:
>> I spent a while delving into the code to figure out exactly where it
>> explodes (specifically it segfaults, sorry gets an 'access violation',
>> when the C code in fecmodule tries to turn around and call 'mmap.mmap'
>> i.e. "mmapobj = PyObject_Call(mmapfunc, mmap_args, mmap_kwargs);" in
>> the FileToMmap function)
>
> 1) can you get GDB or VC++'s debugger running on this?

In principle, yes.  I'm not sure what's involved though, having done no 
debugging on windows as yet.  Any tips & tricks would be appreciated. 
(I'm guessing that it's going to be the python.exe code I need to poke 
around in, and have no idea what's involved in getting that all loaded 
into the debugger etc.)

> 2) have you copied over the test_fec.py unit test into this source 
> tree of
> yours and run it?

I ran the tests, and got errors which look to my eye to be irrelevant 
to the actual issues I'm having:

----

[Tue Jun 24 19:10:57] yo$ PYTHONPATH=. python mnetlib/test/test_fec.py
.EEEE
======================================================================
ERROR: test_biger_expansion (__main__.EncodeTestCase)
----------------------------------------------------------------------
Traceback (most recent call last):
   File "mnetlib/test/test_fec.py", line 42, in tearDown
     os.unlink(ii)
OSError: [Errno 13] Permission denied: 
'c:\\DOCUME~1\\robk\\LOCALS~1\\Temp\\~216
4-2fec_test.random'

======================================================================
ERROR: test_raise_error_on_too_big_index (__main__.EncodeTestCase)
----------------------------------------------------------------------
Traceback (most recent call last):
   File "mnetlib/test/test_fec.py", line 42, in tearDown
     os.unlink(ii)
OSError: [Errno 13] Permission denied: 
'c:\\DOCUME~1\\robk\\LOCALS~1\\Temp\\~216
4-3fec_test.random'

======================================================================
ERROR: test_size_of_packets (__main__.EncodeTestCase)
----------------------------------------------------------------------
Traceback (most recent call last):
   File "mnetlib/test/test_fec.py", line 42, in tearDown
     os.unlink(ii)
OSError: [Errno 13] Permission denied: 
'c:\\DOCUME~1\\robk\\LOCALS~1\\Temp\\~216
4-4fec_test.random'

======================================================================
ERROR: test_small_k (__main__.EncodeTestCase)
----------------------------------------------------------------------
Traceback (most recent call last):
   File "mnetlib/test/test_fec.py", line 42, in tearDown
     os.unlink(ii)
OSError: [Errno 13] Permission denied: 
'c:\\DOCUME~1\\robk\\LOCALS~1\\Temp\\~216
4-5fec_test.random'

----------------------------------------------------------------------
Ran 5 tests in 28.611s

FAILED (errors=4)
[Tue Jun 24 19:11:35] yo$

----

Perhaps more interestingly though, I did spend a little time with the 
fec stuff, and it was pretty easy to put together some python wrapping 
it that successfully publishes (and restores) a file, dumping the 
blocks to local files.  So the fec stuff itself works fine, in my 
processes - it's just when I try and get mnet to publish it explodes...

I'll look into the debugging thing some more today.

Any tips appreciated.

cheers,
rob




-------------------------------------------------------
This SF.Net email sponsored by: Free pre-built ASP.NET sites including
Data Reports, E-commerce, Portals, and Forums are available now.
Download today and enter to win an XBOX or Visual Studio .NET.
http://aspnet.click-url.com/go/psa00100006ave/direct;at.asp_061203_01/01
_______________________________________________
mnet-devel mailing list
mnet-devel at lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/mnet-devel




More information about the Mnet-devel mailing list