1390777 - Convert dom/html/test/test_fullscreen-api.html to comply with new data: URI inheritance model

Status: RESOLVED FIXED

(Core :: DOM: Security, defect)

Comment 1

[Christoph Kerschbaumer [:ckerschb]]

It seems that test is not loading any data: URIs but it's still failing on Windows:
https://treeherder.mozilla.org/#/jobs?repo=try&revision=37554efc3fb036d426fed8b8aca58148fcb8a4da&selectedJob=123487896

Comment 2

[Christoph Kerschbaumer [:ckerschb]]

Henry, I was able to reproduce the problem on MacOSX. The root of the problem is here [1]. One potential solution would be to rewrite the test to use a postMessagehandler. But totally up to you. Thanks for looking into this.

If you can't repro, let me know and I'll have a look at this one.

[1] https://dxr.mozilla.org/mozilla-central/source/dom/html/test/file_fullscreen-api.html#40

Comment 3

[Christoph Kerschbaumer [:ckerschb]]

It seems that particular subtest is disabled on Linux, hence I wasn't able to reproduce at first.

Comment 4

[Henry Chang [:hchang]]

Attachment: Bug 1390777 - Use srcdoc to replace src=data:xxx whenever possible.

Review commit: https://reviewboard.mozilla.org/r/169090/diff/#index_header
See other reviews: https://reviewboard.mozilla.org/r/169090/

Comment 5

[Henry Chang [:hchang]]

I have also spotted that use of data:uri (as well as another one in file_fullscreen-plugins.html)
and changed to srcdoc with a try run:

https://treeherder.mozilla.org/#/jobs?repo=try&revision=6e3e7eb862714ec50129aef1326d75b52fe435e9

They look all good so the only thing of concern to you is the use of 'srcdoc'.
Do you think if we should avoid using srcdoc as you mentioned in [1]?

https://bugzilla.mozilla.org/show_bug.cgi?id=1324406#c21

Thanks :)

Comment 6

[Christoph Kerschbaumer [:ckerschb]]

(In reply to Henry Chang [:hchang] from comment #5)
> They look all good so the only thing of concern to you is the use of
> 'srcdoc'.
> Do you think if we should avoid using srcdoc as you mentioned in [1]?

If srcdoc works we can obviously use srcdoc, happy to r+ that change if you flag me for review :-)

Comment 7

[Henry Chang [:hchang]]

https://treeherder.mozilla.org/#/jobs?repo=try&revision=6e3e7eb862714ec50129aef1326d75b52fe435e9

I triggered more tries and there are still 2 failures out of 60 runs. (on Windows 7 opt)
However, I wonder if those failures are relevant to test_fullscreen-api-race.html
since they always fail altogether. So, I tend to land this fix first once I make sure
this is a permanent failure. (Argument: if "data: to srcdoc" is the issue, I would expect
this should always fail and a cross-origin check in requestFullScreen implementation.)

[1] https://treeherder.mozilla.org/#/jobs?repo=try&revision=37554efc3fb036d426fed8b8aca58148fcb8a4da

Comment 8

[Henry Chang [:hchang]]

Comment on attachment 8897796 [details]
Bug 1390777 - Use srcdoc to replace src=data:xxx whenever possible.

Review request updated; see interdiff: https://reviewboard.mozilla.org/r/169090/diff/1-2/

Comment 9

[Christoph Kerschbaumer [:ckerschb]]

Comment on attachment 8897796 [details]
Bug 1390777 - Use srcdoc to replace src=data:xxx whenever possible.

https://reviewboard.mozilla.org/r/169090/#review174798

Those other failures on TRY definitely look like intermittents not related to the change here. r=ckerschb, thanks for fixing

Comment 10

[Pulsebot]

Pushed by [e-mail]:
https://hg.mozilla.org/integration/autoland/rev/11d08f128b4f
Use srcdoc to replace src=data:xxx whenever possible. r=ckerschb

Comment 11

[Carsten Book [:Tomcat]]

https://hg.mozilla.org/mozilla-central/rev/11d08f128b4f