[Python-Dev] 2.1c1: 2nd test_asynchat fails on Solaris 8

[Guido van Rossum]

> SunOS asafoetida 5.8, gcc 2.95.2
> 
> "make test" fails on running test_asynchat.py for the second time. The
> traceback is:
> 
> Exception in thread Thread-1:
> Traceback (most recent call last):
>   File
> "/export/home/mark/src/python/CVS/python/dist/src/Lib/threading.py",
> line 378, in __bootstrap
>     self.run()
>   File "Lib/test/test_asynchat.py", line 12, in run
>     sock.bind((HOST, PORT))
> error: (125, 'Address already in use')
> 
> Traceback (most recent call last):
>   File "Lib/test/test_asynchat.py", line 56, in ?
>     main()
>   File "Lib/test/test_asynchat.py", line 51, in main
>     c = echo_client()
>   File "Lib/test/test_asynchat.py", line 32, in __init__
>     self.connect((HOST, PORT))
>   File
> "/export/home/mark/src/python/CVS/python/dist/src/Lib/asyncore.py", line
> 308, in connect
>     raise socket.error, why
> socket.error: (146, 'Connection refused')
> 
> Looks like Solaris takes a while to shut sockets down? (This is not a
> slow box, btw.) Or is there an option to not have the socket linger?

Dunno, but there *is* an option to allow reusing a socket.

> Also, test_sunaudiodev fails, since setup.py tests only whether the
> platform is a Sun, not whether there is a /dev/audio as well. Servers
> don't have a /dev/audio. This is clearly a minor nit - I did log a bug
> against this some time ago.

Compiling on a server doesn't mean you'll run on a server only.  But
the test should mark itself as "skipped" when /dev/audio doesn't
exist.  I don't know how easy that is.

--Guido van Rossum (home page: http://www.python.org/~guido/)