[issue7427] BadStatusLine is hell to debug

Martin Panter added the comment: This change was only made in 2.7a4, not 2.6 -- versions: +Python 2.7 -Python 2.6 ___ Python tracker ___ __

[issue7427] BadStatusLine is hell to debug

Martin Panter added the comment: As far as I can tell, the “line” attribute isn’t documented anyway. But Issue 8450 is opened about improving the exception when the connection is closed. -- nosy: +vadmium ___ Python tracker

[issue7427] BadStatusLine is hell to debug

Andrew Swan added the comment: I just got tripped up by this change, I wanted to catch the specific case of an http server closing a connection and assumed that the following would work: try: resp = conn.getresponse() except httplib.BadStatusLine, e: if len(e.line) == 0: # server closed

[issue7427] BadStatusLine is hell to debug

Jake McGuire added the comment: I think what's happening is that your connection is being closed due to inactivity, so the status line that comes back is empty. Printing repr(line) would probably make the emptiness clear, but maybe the httplib code should put in a more specific message in th

[issue7427] BadStatusLine is hell to debug

djc added the comment: Also, it might be useful here if it showed repr(line) instead of just line, but that'd just be icing on the cake. -- ___ Python tracker ___ ___

[issue7427] BadStatusLine is hell to debug

New submission from djc : For whatever reason, BadStatusLine tracebacks often don't show the line passed into them. Given the errr, heavy architecture of httplib, this makes it pretty bad to debug. It's not clear to me why this is: Traceback (most recent call last): File "/home/djc/src/couchdb