579235 - <object> with no size specified defaults to 0x0 even if the plug-in is playing media with some sort of intrinsic size

Status: RESOLVED INCOMPLETE

(Core Graveyard :: Plug-ins, defect)

Description

[Benoit Girard (:BenWa)]

I found the following Mozilla Quicktime tests: 
http://www-archive.mozilla.org/quality/browser/front-end/testcases/plugins/quicktimetest.html

This links to the following tests:
http://www.student.oulu.fi/~sairwas/object-test/video/

Several of these tests are not working and do not display the alternative text. These test are working under Webkit.

Comment 1

[Benoit Girard (:BenWa)]

This is also happening with Firefox 3.6.6.

Comment 2

[Benjamin Smedberg]

When did this break?

Comment 3

[Daniel Veditz [:dveditz]]

When you say "several" are not working which ones do you mean specifically? on 1.9.2.8pre they all seem to work except #2 and #4
  http://www.student.oulu.fi/~sairwas/object-test/video/qt2.html
  http://www.student.oulu.fi/~sairwas/object-test/video/qt4.html
Did those work prior to OOPP landing? Or are those just long-broken tests?

Comment 4

[Benoit Girard (:BenWa)]

The following test are broken:
AVI test 2&4
Quicktime Test 2&4
MPEG test #1 through 6

Plug-ins are not ran out-of-process on OS X on 3.6.x but could still be affected by some of the changes. We need to find when the regression occurred.

Comment 5

[Boris Zbarsky [:bzbarsky]]

Benoit, are you doing that regression search?  Or expecting someone else to do it?

Comment 6

[Benoit Girard (:BenWa)]

I'm working on another patch at the moment. If no one has a change to do this I'll come back to it soon.

Comment 7

[Daniel Veditz [:dveditz]]

please get back to us on the blocking nomination if this turns out to be an OOPP regression and/or affecting lots of people, but traffic in this bug isn't making this sound like a high priority issue.

Comment 8

[Benoit Girard (:BenWa)]

Updated the summary to clarify that Quicktime fails to display only under certain scenarios.

Can anyone verify the scenarios and make sure they are valid object definitions for Quicktime? I.e. defining an object with no width/height?

Comment 9

[Johnny Stenback (:jst)]

BenWa, did the failing cases work in 3.6 (pre OOPP) or 3.5? If so, they should work in newer builds as well.

Comment 10

[Benoit Girard (:BenWa)]

I've just tried the following UA and the test are also broken:

Mozilla/5.0 (Macintosh; U; Intel Mac OS X 10.6; en-US; rv:1.9.0.19) Gecko/2010031218 Firefox/3.0.19

This might not be the ideal test since it's a newer build from an older branch. Where can I find older builds from older release branches? 

I certainly wouldn't block on this issue until we are sure that the tests are valid and that this is a regression.

Comment 11

[Smokey Ardisson (offline for a while; not following bugs - do not email)]

(In reply to comment #10)
> Where can I find older builds from older release branches? 

ftp://ftp.mozilla.org/pub/mozilla.org/firefox/releases/

Comment 12

[Johnny Stenback (:jst)]

Not blocking based on the previous comments. Please renominate if more info appears that shows this is a regression.

Comment 13

[Boris Zbarsky [:bzbarsky]]

All that's happening here is that if no size is specified we size an nsObjectFrame for an <object> to 0x0.  The problem is that NPAPI doesn't have a way for us to ask the plug-in for its preferred size, last I checked.  I could have sworn we had an existing bug on this, but I can't find it right now....

Comment 14

[Benjamin Smedberg]

Resolving old bugs which are likely not relevant any more, since NPAPI plugins are deprecated.