This might be related to our lack of nt.dup(). I filed a new bug (http://www.codeplex.com/IronPython/WorkItem/View.aspx?WorkItemId=7451) to track this.
Sorry, we missed the comment left by amecha in codeplex. If you see any bug has not been fixed, please do not hesitate to re-activate it or send email to this list. A temp solution is: do not close the socket even after calling makefile. Btw, with the current makefile, ~70% of test_socket.py pass. test_urllib.py and test_urllibnet.py pass with few-line change. Thanks for telling us know this. -----Original Message----- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf Of Sanghyeon Seo Sent: Thursday, January 18, 2007 10:48 PM To: Discussion of IronPython Subject: [IronPython] Socket's life after close() Issue #1368 is closed, but its primary use case is still broken. Have you tested with original poster's test case, urllib.urlopen()? amecha commented on Jan 10 about this closed issue stating that it is still broken and I can confirm. (btw, shouldn't the original poster be able to re-open the bug?) Here is a simplifed test case since urllib.urlopen() calls hosts of other functions that simply obscures the root cause. The test assumes that echo service is enabled on the localhost. # test.py import socket s = socket.socket() s.connect(('127.0.0.1', 7)) s.send('test') f = s.makefile() s.close() print f.read(4) Python can read fine after socket close(), but IronPython complains. -- Seo Sanghyeon _______________________________________________ users mailing list users@lists.ironpython.com http://lists.ironpython.com/listinfo.cgi/users-ironpython.com _______________________________________________ users mailing list users@lists.ironpython.com http://lists.ironpython.com/listinfo.cgi/users-ironpython.com