581361 - Pause and Cancel button of download progress widget not functional

Status: VERIFIED FIXED

(Toolkit :: Add-ons Manager, defect)

Description

[Henrik Skupin [:whimboo][⌚️UTC+2]]

Mozilla/5.0 (Macintosh; Intel Mac OS X 10.6; rv:2.0b2) Gecko/20100720 Firefox/4.0b2

When you install an add-on and want to pause and cancel the download, it will not work.

Steps:
1. Open the add-ons manager
2. Search for an add-on which isn't installed yet
3. Install the add-on
4. Immediately try to hit the pause or cancel button on the download progress widget

There is no reaction when clicking the above mentioned buttons.

Comment 1

[Dave Townsend [:mossop]]

Pause support would require bug 553024 which we aren't going to get done for 4.0 so we should just remove the pause button. Cancelling ought to work though.

Comment 2

[Mike Beltzner [:beltzner, not reading bugmail]]

We ought to remove the pause button, then. :)

Comment 3

[Blair McBride [:Unfocused] (UNAVAILABLE)]

Attachment: Patch v1

Removes the pause button, and hooks up the cancel button. Spent ages unsuccessfully trying to automate a test for this, before eventually giving up - so hows about we use a litmus test instead? :)

Comment 4

[Henrik Skupin [:whimboo][⌚️UTC+2]]

(In reply to comment #3)
> Removes the pause button, and hooks up the cancel button. Spent ages
> unsuccessfully trying to automate a test for this, before eventually giving up
> - so hows about we use a litmus test instead? :)

Can you give more details where the problems are for a test?

Comment 6

[Dave Townsend [:mossop]]

Comment on attachment 501565 [details] [diff] [review]
Patch v1

Looks good but yeah I'd really like to see an automated test here if we can. What were the problems you were seeing?

Comment 7

[Blair McBride [:Unfocused] (UNAVAILABLE)]

Had trouble figuring out exactly when the cancel button could be clicked - essentially a timing issue, since httpd.js will serve the file so quickly. I could spend more time on it if you - but it would only be testing one simple line of code that's gluing things together (AFAIK, everything else is already tested).

Comment 8

[Henrik Skupin [:whimboo][⌚️UTC+2]]

To make httpd.js serving content slower, you will have to send the data on your own. I already gave that advice on another bug with the in-testsuite? flag set.

It's up to you. If we don't have an automated mochitest, we can create a Mozmill test.

Comment 9

[Dave Townsend [:mossop]]

Let's get this landed when the tree re-opens and at the least add a litmus test for it. If we have time we can try to add a full automated test before release.

Comment 10

[Blair McBride [:Unfocused] (UNAVAILABLE)]

Landed: https://hg.mozilla.org/mozilla-central/rev/bdc14e2b1308

Comment 11

[Henrik Skupin [:whimboo][⌚️UTC+2]]

Verified fixed with Mozilla/5.0 (Macintosh; Intel Mac OS X 10.6; rv:2.0b10pre) Gecko/20110112 Firefox/4.0b10pre ID:20110112033217

A Litmus and Mozmill test are possible.