1288628 - Intermittent devtools/client/netmonitor/test/browser_net_resend.js | new header added to sent request -

Status: RESOLVED WORKSFORME

(DevTools :: Netmonitor, defect, P3)

Description

[Treeherder Bug Filer]

Filed by: tomcat [at] mozilla.com

https://treeherder.mozilla.org/logviewer.html#?job_id=32403921&repo=mozilla-inbound

http://archive.mozilla.org/pub/firefox/tinderbox-builds/mozilla-inbound-win32/1469126541/mozilla-inbound_xp_ix_test-mochitest-devtools-chrome-7-bm110-tests1-windows-build46.txt.gz

Comment 1

[Thomas Wisniewski]

At first glance this would have been caused by a patchset I'm writing in bug 1285036, which was backed out. However, after running a couple of tries, I think something else may be going on.

First off, I did a try-run on central with my patchset applied: https://treeherder.mozilla.org/#/jobs?repo=try&revision=74b7c4156708

As expected, that triggered the dt7 failure, and the debugging info I added revealed that an extra header named "test-header" with value "true" was indeed there. So far so good.

However, I also ran a second try without my patchset applied, to compare the results... and it never even ran the test: https://treeherder.mozilla.org/#/jobs?repo=try&revision=c1734ca90eca

It marked dt7 as passing, but the job log shows that "browser_net_resend.js" was never even run.

This doesn't seem normal to me, so I thought I'd report my findings. I think I'll hop onto IRC to see if anyone can help me make sense of this...

Comment 2

[Thomas Wisniewski]

Thankfully it turns out that my patch is probably the cause of this failure, due to the tests apparently being over-sensitive to header capitalization. I'll continue my work on this in bug 1285036, but since I'm still not sure if the test is being run correctly in the normal case (or if it's just a weird logging quirk), I'm going to leave this bug open.

Comment 3

[BMO Automation]

Bulk assigning P3 to all open intermittent bugs without a priority set in Firefox components per bug 1298978.

Comment 4

[Thomas Wisniewski]

Given that this evidently hasn't triggered since my patch was backed out three weeks ago, I'd say this should be safe to mark as resolved. What do you think, tomcat?

Comment 5

[Carsten Book [:Tomcat]]

(In reply to Thomas Wisniewski from comment #4)
> Given that this evidently hasn't triggered since my patch was backed out
> three weeks ago, I'd say this should be safe to mark as resolved. What do
> you think, tomcat?

yeah will mark this as wfm now