999052 - Issue using Mozregression behind authenticating proxy? #7

Status: RESOLVED FIXED

(Testing :: mozregression, defect)

Description

[William Lachance (:wlach)]

[import from github]

edmorley:

As part of asking someone on BMO to find a reg range with MozRegression, they came across this issue:
https://bugzilla.mozilla.org/show_bug.cgi?id=640499#c16

Any ideas?

Thanks,

Ed

harth:

Ooh, that could be a problem with Python's httplib. I'm not familiar with that, and I'm not sure when I'll have time to look into it. Do you think it's a common problem?

edmorley:

It's the first I've seen of it - and the user in question just bisected manually in the end. Happy to just leave as is for now, just wondered if there was an easy answer, which there is presumably not :-)

mattbishop:

I want to add that this issue exists for non-authenticating proxies too.

ernstp:

Still an issue.. :-)

Comment 1

[Julien Pagès (:parkouss)]

I'm not sure this is still an issue since mozregression uses requests instead of httplib: https://github.com/mozilla/mozregression/commit/0c033d2ead4b4b5f71197957cf9e2f3a6163c777.

I remember that I could use mozregression behind a non-authenticating proxy (cntlm) without problems at my work a few month ago. The python requests package can be configured to use proxies with environment variables (http://docs.python-requests.org/en/latest/user/advanced/#proxies). The documentation indicates that HTTP_PROXY and HTTPS_PROXY could be used, but if I remember well it worked for me on a Linux OS by setting http_proxy and https_proxy.

I think that this bug is no more relevant. William, what do you think ?

Comment 2

[William Lachance (:wlach)]

Yes, what you say sounds plausible. Let's resolve this one and let people report bugs if they're still having problems.