asyncore does not check for EAGAIN and EPIPE errno

[socketpair]

BPO	11265
Nosy	@josiahcarlson, @giampaolo, @socketpair
Files	z.patch z.patch: patch with EPIPE included

^{Note: these values reflect the state of the issue at the time it was migrated and might not reflect the current state.}

Show more details

GitHub fields:

assignee = 'https://github.com/giampaolo'
closed_at = <Date 2011-03-03.14:26:22.378>
created_at = <Date 2011-02-21.05:18:50.977>
labels = ['type-bug', 'library']
title = 'asyncore does not check for EAGAIN and EPIPE errno'
updated_at = <Date 2011-03-03.18:30:30.965>
user = 'https://github.com/socketpair'

bugs.python.org fields:

activity = <Date 2011-03-03.18:30:30.965>
actor = 'socketpair'
assignee = 'giampaolo.rodola'
closed = True
closed_date = <Date 2011-03-03.14:26:22.378>
closer = 'giampaolo.rodola'
components = ['Library (Lib)']
creation = <Date 2011-02-21.05:18:50.977>
creator = 'socketpair'
dependencies = []
files = ['20817', '20964']
hgrepos = []
issue_num = 11265
keywords = ['patch']
message_count = 6.0
messages = ['128939', '129810', '129822', '129835', '129969', '129991']
nosy_count = 3.0
nosy_names = ['josiahcarlson', 'giampaolo.rodola', 'socketpair']
pr_nums = []
priority = 'normal'
resolution = 'fixed'
stage = None
status = 'closed'
superseder = None
type = 'behavior'
url = 'https://bugs.python.org/issue11265'
versions = ['Python 3.1', 'Python 2.7', 'Python 3.2', 'Python 3.3']

[socketpair]

According to man:
----------------------
ERRORS
EAGAIN or EWOULDBLOCK
The socket is marked nonblocking and no connections are present to be accepted. POSIX.1-2001 allows either error to be returned for this case, and does not require these constants to have the same value, so a portable application should check for both possibilities.
-----------------

patch included

[socketpair]

patch updated for handling EPIPE

[giampaolo]

Ok for EAGAIN.

I'm not completely sure about EPIPE on recv() as it is not listed here:
http://www.kernel.org/doc/man-pages/online/pages/man2/recv.2.html
...although I think it's ok to treat it as an alias for connection lost anyway.

In pyftpdlib I treat both EPIPE and EBADF as disconnect events, and this works with both select() and poll().

Your patches should be targeted for python 3.3 (and not 2.x).

[josiahcarlson]

Giampaolo pinged me over email...

These additional conditions look good, and should be targeted for 3.3 .

Thank you :)

[giampaolo]

Fixed in r88722.
I did not catch EAGAIN in connect() as it makes hang some tests.

[socketpair]

I did not catch EAGAIN in connect() as it makes hang some tests.
maybe incorrect tests? which test(s) failed ?

mmarkk@fad:/usr/include$ fgrep EWOULDB -r .
./asm-generic/errno.h:#define EWOULDBLOCK EAGAIN /* Operation would block */

So not catching EAGAIN ( at least in Linux) can not make some tests hang.