Closed Bug 1063640 Opened 10 years ago Closed 7 years ago

Reftest analyzer hangs the browser, since Treeherder passes the full log

Categories

(Tree Management :: Treeherder, defect, P3)

defect

Tracking

(Not tracked)

RESOLVED WONTFIX

People

(Reporter: wlach, Unassigned)

Details

With bug 1059331 fixed, the reftest analyzer works again and is probably enough for an initial rollout. However, the current implementation has the following disadvantages:

1. You need to download the full log to see the reftest failures, which is excessive. 
2. We do a bunch of cross-origin stuff over http://, which is not really a best practice.

We should probably fix this by moving the reftest analyzer into treeherder and have it download a reftest failure log by id (and not URL) from the main treeherder site. 

There's some discussion in bug 1059331 on how to do this. See also this discussion on irc:

http://logs.glob.uno/?c=mozilla%23treeherder&s=2+Sep+2014&e=2+Sep+2014&h=reftest#c1590
OS: Linux → All
Priority: -- → P3
Hardware: x86_64 → All
Priority: P3 → P2
I've had to force-kill my browser due to this. I'd call that P1.
Priority: P2 → P1
Summary: Reftest analyzer functionality could be further improved → Reftest analyzer hangs the browser, since Treeherder passes the full log
I wonder if bug 1133617 will help with this too?
Priority: P1 → P2
Priority: P2 → P3
The proper fix here is either for (a) the test harness to put the reftest dumps in a separate smaller file, (b) the reftest analyser to be re-written.

For Treeherder to fix this, we'd have to scrape the log and store the reftest parts somewhere, which I don't think we want to do.
Status: NEW → RESOLVED
Closed: 7 years ago
Resolution: --- → WONTFIX
You need to log in before you can comment on or make changes to this bug.