Bug 98696 – Add tests for plugin snapshotting

• Status: RESOLVED
• Resolution: FIXED
• Component: WebKit Misc.

[Jon Lee]

Add layout tests to do basic functionality tests for plugin snapshotting. Might need to add support for WTR.

Tests include:
* Load plugin, click, verify running. Plugin outputs to log.
* Loading with display:none
* Loading normal, then switch to display:none, then go back. Should maintain whatever state it was in prior to style switch.

[Radar WebKit Bug Importer]

<rdar://problem/12456731>

[Dean Jackson]

*** Bug 112632 has been marked as a duplicate of this bug. ***

[Dean Jackson]

Created attachment 193669 [details]
Patch

[Dean Jackson]

This patch doesn't do the display:none test suggested, but it does cover the two simple cases.

I will wait for the bots to chew over it to make sure I got the test expectations correct.

[Dean Jackson]

Created attachment 193671 [details]
Patch

[Jon Lee]

Comment on attachment 193671 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=193671&action=review

> LayoutTests/ChangeLog:11
> +        and one that restarts a snapshotted plugin by clicking.

Are we confirming these tests by just looking for the existence of the label? Is that enough? Or should the plugin also create some output so we know that in the restart case it was started twice?

Is this reusing an existing plugin? What does that plugin do?

[Dean Jackson]

(In reply to comment #6)
> (From update of attachment 193671 [details])
> View in context: https://bugs.webkit.org/attachment.cgi?id=193671&action=review
> 
> > LayoutTests/ChangeLog:11
> > +        and one that restarts a snapshotted plugin by clicking.
> 
> Are we confirming these tests by just looking for the existence of the label? Is that enough? 

At the moment yes.

> Or should the plugin also create some output so we know that in the restart case it was started twice?

We should definitely do that. I'm not sure it needs to land in this first patch. Like I said, these are super simple tests at the moment.

> Is this reusing an existing plugin? What does that plugin do?

It's just the existing simple swf file. We should make a plugin that calls into the page.

[Jon Lee]

Ok. r=me.

(In reply to comment #7)
> (In reply to comment #6)
> > (From update of attachment 193671 [details] [details])
> > View in context: https://bugs.webkit.org/attachment.cgi?id=193671&action=review
> > 
> > > LayoutTests/ChangeLog:11
> > > +        and one that restarts a snapshotted plugin by clicking.
> > 
> > Are we confirming these tests by just looking for the existence of the label? Is that enough? 
> 
> At the moment yes.
> 
> > Or should the plugin also create some output so we know that in the restart case it was started twice?
> 
> We should definitely do that. I'm not sure it needs to land in this first patch. Like I said, these are super simple tests at the moment.
> 
> > Is this reusing an existing plugin? What does that plugin do?
> 
> It's just the existing simple swf file. We should make a plugin that calls into the page.

[Tim Horton]

Comment on attachment 193671 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=193671&action=review

> Source/WebKit2/ChangeLog:10
> +        (don't wait to get a screenshot).

screenshot?

[Dean Jackson]

Committed r146250: <http://trac.webkit.org/changeset/146250>