441778 - Text zoom cancels when mousing over 'snapshot' links

Status: VERIFIED FIXED

(Firefox :: Disability Access, defect)

Description

[Marty Kay]

User-Agent:       Mozilla/5.0 (Windows; U; Windows NT 5.1; en-GB; rv:1.9) Gecko/2008052906 Firefox/3.0
Build Identifier: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-GB; rv:1.9) Gecko/2008052906 Firefox/3.0

I am disabled and have visual problems. I use the CTRL++ to zoom text. However while trying to access LiveJournal pages mousing over 'snap shot' links causes the text zoom to reset to very small fonts. It's very annoying as I then cannot read the webpage. 

This doesn't happen with other links, just those that lead to a 'snapshot' popup/preview of the link.



Reproducible: Always

Steps to Reproduce:
1.Zoom text (CTRL++ x4), 
2.Go to www.livejournal.com to your friends page or any community at random (eg http://community.livejournal.com/knitting/)
3.Mouse over the snapshot links, (links that have the little 'speech bubble' next to them)
Actual Results:  
The text shrinks to very small. I can't read it!!

Expected Results:  
Keep the font texts at the user setting (zoomed) while mousing over all snapshot links

I'm listing this as a major feature issue because for those of us with visual impairment this renders mozilla unreadable.

Comment 1

[Kevin Brosnan]

Confirmed - Mozilla/5.0 (Windows; U; Windows NT 6.0; en-US; rv:1.9) Gecko/2008052906 Firefox/3.0

For right now the best thing might be to disable the Snapshot bubble which can be done at http://www.snap.com/snapshots_faq.php | For Site Visitors | 3. How can I turn Snap Shots on or off?

Comment 2

[Marco Zehe (:MarcoZ)]

This really sounds serious enough to warrant a blocking request on 1.9.1, and if we can take a patch earlier, the better.

Comment 3

[Martijn Wargers (dead)]

Attachment: testcase

This is not really from minimizing the page in question, but I suspect it is the same issue as this testcase shows.

Comment 4

[Robert O'Callahan (:roc) (email my personal email if necessary)]

I think this is more likely to be the zoom pref stuff than layout.

Comment 5

[Myk Melez [:myk] [@mykmelez]]

I can't reproduce on Live Journal, but I can reproduce using the testcase.  We fixed a bug like this at one point, so this problem could be a regression from that fix, or it could be another bug that has the same symptom.  Looking into this...

Comment 6

[Myk Melez [:myk] [@mykmelez]]

Attachment: patch v1: fixes problem

The problem is that we are updating the zoom setting on location changes for subframes.  We should only be updating it when the top-level page's location changes.  Here's a patch that makes that change.

Note: I also remove the try-catch block that was added to compensate for bug 376222 now that bug 376222 has been fixed.

Comment 7

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

Comment on attachment 333049 [details] [diff] [review]
patch v1: fixes problem

It would be good to include more context in your diffs (especially given that bugzilla's diff viewer doesn't fully support git-style diffs).

Comment 8

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

It would also be nice to have an automated test. It should be relatively easy to write a browser chrome test for this (based on Martijn's testcase).

Comment 9

[Myk Melez [:myk] [@mykmelez]]

Attachment: automated test

(In reply to comment #8)
> It would also be nice to have an automated test. It should be relatively easy
> to write a browser chrome test for this (based on Martijn's testcase).

Yes, it certainly should be, but sadly, as with too many things Mozilla, it isn't.  This test should work but doesn't, as the sub-document load doesn't trigger a location change event when the test starts it programmatically.

The test works fine if you manually press the button, however (first jack up the setTimeout value to give yourself time to try that out).  I've tried a couple different approaches, but so far I haven't been able to figure out what the problem is.

Comment 10

[Myk Melez [:myk] [@mykmelez]]

Attachment: automated test that works

(In reply to comment #7)
> (From update of attachment 333049 [details] [diff] [review])
> It would be good to include more context in your diffs (especially given that
> bugzilla's diff viewer doesn't fully support git-style diffs).

Sorry about that.  I've upgraded my copy of Mercurial to a version that supports -U and made -U 8 the default.


(In reply to comment #9)
> This test should work but doesn't, as the sub-document load doesn't
> trigger a location change event when the test starts it programmatically.

Here's a test that works.  I had to bend over backwards to get the location change to happen, synthesizing an event whose handler initiates the sub-document load in a timeout, as initiating the sub-document load without the timeout (or directly from the test without the handler) successfully loads the sub-document but doesn't call location change handlers.

This version also listens for the location change itself and finishes up the test after it happens (rather than waiting for an arbitrary amount of time) for robustness and speed.

Comment 11

[Martijn Wargers (dead)]

Attachment: my_changed_automated_Test

Hi, I also started changing your automated test to make it working.
To make it work, I had to do the iframe loading after a timer. Not really sure, why.
finishTest is being done, after the iframe's load event.

Comment 12

[Myk Melez [:myk] [@mykmelez]]

Attachment: updated automated test

(In reply to comment #11)
> Hi, I also started changing your automated test to make it working.
> To make it work, I had to do the iframe loading after a timer.

Brilliant, thanks!  That simplifies things tremendously.  Here's a version that uses your technique.  It retains the progress listener from my patch, though, for reliability and performance.

Comment 13

[Myk Melez [:myk] [@mykmelez]]

changeset:   17074:15707e100c34
date:        Tue Aug 19 14:59:09 2008 -0700
summary:     bug 441778: don't update page zoom when a sub-document loads (f.e. in an iframe); r=gavin

Comment 14

[Tony Chung [:tchung]]

Verified fix on Mozilla/5.0 (Macintosh; U; Intel Mac OS X 10.5; en-US; rv:1.9.1b1pre) Gecko/20081006 Minefield/3.1b1pre.