1062966 - Breakdown: Implement new handling of malware downloads in download panel

Status: RESOLVED FIXED

(Firefox :: Downloads Panel, defect)

Description

[Sevaan Franks [:sevaan]]

This is the implementation bug for the UX work carried out in Bug 1053890 regarding how downloads flagged as "malware", "potentially unwanted", or "uncommon" are handled in Firefox.

More information:
- https://bugzilla.mozilla.org/attachment.cgi?id=8481334
- https://bugzilla.mozilla.org/show_bug.cgi?id=1053890#c38

Comment 2

[Liz Henry (:lizzard) (relman/hg->git project)]

Release Note Request (optional, but appreciated)
[Why is this notable]: This is something I think we should note and document for users, maybe in the support knowledgebase, since it will affect potentially a large percentage of user downloads.
[Suggested wording]: User warnings defined for downloads that may be malware.
[Links (documentation, blog post, etc)]:  (Still needed)

Tyler I'm cc-ing you because I'm not sure if there's someone specific in SUMO who might want to do this.

Comment 5

[:Gavin Sharp [email: [e-mail]]]

Paolo: you had mentioned this being complicated. Does it need to be broken down?

Comment 6

[:Paolo Amadini]

(In reply to :Gavin Sharp [email: [e-mail]] from comment #5)
> Paolo: you had mentioned this being complicated. Does it need to be broken down?

I did the breakdown and added my own estimates directly on the bugs.

Comment 7

[Terrell Kelley]

What version is this fixed in? You gave it an "iteration," but that number doesn't correspond to any Firefox version, AFAIK

This need to be implemented quickly, to prevent users from turning off the feature entirely to get around a false positive. And Google's methodology of using multiple antiviruses is highly prone to false positives.

Comment 8

[Gian-Carlo Pascutto [:gcp]]

Marco, you morphed this bug from a metabug into a breakdown, but now there's no meta bug left to see when the feature is complete? On top of that it's very confusing as there were already loads of bugs duplicated to this one, which are now showing as RESOLVED FIXED, even though the bug isn't fixed yet.

(In reply to Terrell Kelley from comment #7)
> This need to be implemented quickly, to prevent users from turning off the
> feature entirely to get around a false positive. And Google's methodology of
> using multiple antiviruses is highly prone to false positives.

Note that we already partially disabled this feature in release Firefox to reduce the number of false positives.

Comment 9

[Marco Mucci [:MarcoM]]

Turned into a breakdown based on Comment #6.  Bug 1068664 and Bug 1068660 are currently assigned and being worked on in IT 36.1  Bug 1068656 is blocked by Bug 1068664.

(In reply to Gian-Carlo Pascutto [:gcp] from comment #8)
> Marco, you morphed this bug from a metabug into a breakdown, but now there's
> no meta bug left to see when the feature is complete? On top of that it's
> very confusing as there were already loads of bugs duplicated to this one,
> which are now showing as RESOLVED FIXED, even though the bug isn't fixed yet.
> 
> (In reply to Terrell Kelley from comment #7)
> > This need to be implemented quickly, to prevent users from turning off the
> > feature entirely to get around a false positive. And Google's methodology of
> > using multiple antiviruses is highly prone to false positives.
> 
> Note that we already partially disabled this feature in release Firefox to
> reduce the number of false positives.

Comment 10

[Lukas Blakk [:lsblakk] use ?needinfo]

Is there something landed in 35 that should get noted here?  I'm not clear on which of the bugs landed or in progress actually have user-facing impact worth noting.

Comment 11

[:Paolo Amadini]

(In reply to Lukas Blakk [:lsblakk] use ?needinfo from comment #10)
> Is there something landed in 35 that should get noted here?  I'm not clear
> on which of the bugs landed or in progress actually have user-facing impact
> worth noting.

No, bug 1068656 is the visible front-end work. I've added the "relnote" keyword there.

Comment 13

[Terrell Kelley]

Do you have any telemetry data on people disabling the feature in about:config? I thought maybe you might need to force the option back on upon release, if disabling was common enough.