Closed Bug 558434 Opened 14 years ago Closed 14 years ago

Winless, transparent, silverlight doesn't display correctly

Categories

(Core Graveyard :: Plug-ins, defect)

x86
Windows 7
defect
Not set
normal

Tracking

(blocking2.0 final+, blocking1.9.2 needed, status1.9.2 .4-fixed, status1.9.1 unaffected)

VERIFIED FIXED
Tracking Status
blocking2.0 --- final+
blocking1.9.2 --- needed
status1.9.2 --- .4-fixed
status1.9.1 --- unaffected

People

(Reporter: jimm, Assigned: jimm)

References

()

Details

(Keywords: verified1.9.2, verified1.9.3)

Attachments

(3 files, 1 obsolete file)

Attached image comparison
noticed this today while doing some testing.
Summary: [OOPP] Winess controls that leverage transparent regions don't display correctly in tab previews/aero peek → [OOPP] Winless controls that leverage transparent regions don't display correctly in tab previews/aero peek
Summary: [OOPP] Winless controls that leverage transparent regions don't display correctly in tab previews/aero peek → [OOPP] Winless controls that leverage transparent regions don't display correctly
(In reply to comment #0)
> Created an attachment (id=438159) [details]
> comparison
> 
> noticed this today while doing some testing.

Note in that comparison shot, the upper image shows a "black background" in the page w/silverlight, which should be transparent. The lower image with the white background is the preview. This bug is related to flash bug 552062.
No longer blocks: OOPP
This isn't an OOPP problem, I can reproduce this on trunk with oopp/disabled.
blocking2.0: --- → ?
Attached image screen cap
(In reply to comment #2)
> This isn't an OOPP problem, I can reproduce this on trunk with oopp/disabled.

I can also reproduce it on 3.6.4pre.
blocking1.9.2: --- → ?
What about 3.6.3? Regressions are a big deal, pre-existing bugs aren't so much.
(In reply to comment #5)
> What about 3.6.3? Regressions are a big deal, pre-existing bugs aren't so much.

It's a regression in 3.6.4, 3.6.3 doesn't exhibit the problem.
One unrelated thing I noticed while testing, I had "dom.ipc.plugins.enabled" set to false, with "dom.ipc.plugins.enabled.npctrl.dll" defaulted to true. The plugin container process for silverlight still launched. I had to disable "dom.ipc.plugins.enabled.npctrl.dll" before I was able to confirm this wasn't OOPP. Is that a bug in our prefs benjamin?
http://hg.mozilla.org/projects/firefox-lorentz/pushloghtml?startdate=2010-03-18&enddate=2010-03-23

Doesn't help much, but here's a regression range (3-18 -> 3-23) for the Lorentz nightlies.
Found it! Changes we made to nsObjectFrame::IsOpaque caused this. Silverlight doesn't set the transparent property to true in this case. Technically probably a bug in silverlight, but since older versions of fx didn't care, I suppose it partially our fault. Here are the changes:

http://hg.mozilla.org/releases/mozilla-1.9.2/rev/32a2e349d7cf

Either we back those changes out, or we quirk GetValue(nsPluginInstanceVariable_TransparentBool) to always return PR_TRUE for windowless silverlight version <= 3.0.
Attached patch fix (obsolete) — Splinter Review
Blocks: 526216
Attachment #438483 - Flags: review?(joshmoz)
Summary: [OOPP] Winless controls that leverage transparent regions don't display correctly → Winless, transparent, silverlight doesn't display correctly
Code freeze for 1.9.2.4 is midnight tonight.  Not blocking but if you can finish the review by then, we'd be happy to take this fix.
cc'ing Ashish @ MS. Ashish, in previous version of fx, we assumed windowless controls were transparent. In the next update, we've removed that assumption. It looks like silverlight relied on this. Silverlight should be calling set value for transparency, but doesn't appear to be doing so. We've fixed this by making the transparent property default true for silverlight windowless controls. To work around this, silverlight should always set the transparent property after setting the windowless property in future releases.
Attached patch updated patchSplinter Review
updated w/sl version info in the comment.
Attachment #438483 - Attachment is obsolete: true
Attachment #438552 - Flags: review?
Attachment #438483 - Flags: review?(joshmoz)
Attachment #438552 - Flags: review? → review?(joshmoz)
Attachment #438552 - Flags: review?(joshmoz) → review+
http://hg.mozilla.org/mozilla-central/rev/9614b3d45ff4
Status: NEW → RESOLVED
Closed: 14 years ago
Resolution: --- → FIXED
Attachment #438552 - Flags: approval1.9.2.4?
Comment on attachment 438552 [details] [diff] [review]
updated patch

a=LegNeato for 1.9.2.4
Attachment #438552 - Flags: approval1.9.2.4? → approval1.9.2.4+
blocking1.9.2: ? → needed
Verified

Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.9.2.4) Gecko/20100611 Firefox/3.6.4

Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.9.3a6pre) Gecko/20100619 Minefield/3.7a6pre
Status: RESOLVED → VERIFIED
blocking2.0: ? → final+
Product: Core → Core Graveyard
You need to log in before you can comment on or make changes to this bug.

Attachment

General

Created:
Updated:
Size: