482073 - test_classifier.html issues resulting in intermittent failures

Status: RESOLVED FIXED

(Toolkit :: Safe Browsing, defect)

Description

[Jeff Walden [:Waldo]]

I think the iframe in this test needs to be *before* the doUpdate() call, because if the setTimeout isn't reached, the iframe hasn't been processed yet (so the test fails when getElementById returns null).

Second, and off-topic enough that I'd usually file a separate bug except my connection's horrible, I'd prefer that malware.com:80 be malware.example.com:80 (assuming the malware distinction can be added to subdomains without tainting parent domains) to avoid going outside the reserved domain names in RFC (something), but at least that's not a failure issue.

Comment 1

[Dave Camp (:dcamp)]

Attachment: fix

Just moved that update to onload.  Yeah, malware.com->malware.example.com makes sense.

Comment 2

[Jeff Walden [:Waldo]]

Comment on attachment 366446 [details] [diff] [review]
fix

Looks good except for this (not noticed until after filing the bug):

>diff -r be848dc25e6c toolkit/components/url-classifier/tests/mochitest/classifierFrame.html

> <!-- Try loading from a malware css URI
>-<link rel="stylesheet" type="text/css" href="http://malware.com/tests/toolkit/components/url-classifier/tests/mochitest/evil.css"></link>
>+<link rel="stylesheet" type="text/css" href="http://malware.example.com/tests/toolkit/components/url-classifier/tests/mochitest/evil.css"></link>

That comment's not closed, which messes with doing this test:
> 
> <!-- Try loading a marked-as-malware css through an @import from a clean URI -->
> <link rel="stylesheet" type="text/css" href="import.css"></link>

Comment 3

[Dave Camp (:dcamp)]

Attachment: comment fixed too

Comment 4

[Dave Camp (:dcamp)]

http://hg.mozilla.org/mozilla-central/rev/f4e4e522ebcd

Comment 6

[Dave Townsend [:mossop]]

Landed this on the branch: http://hg.mozilla.org/releases/mozilla-1.9.1/rev/c56f528026af