[Bug 30843] New: Adobe Flash 10 (only 10.0.x.x) fails to redraw properly in fullscreen mode when hardware acceleration is enabled (affects NP plugin when playing Youtube movies and standalone flash-based apps like Machinarium)

wine-bugs at winehq.org wine-bugs at winehq.org
Mon Jun 4 09:01:49 CDT 2012


http://bugs.winehq.org/show_bug.cgi?id=30843

             Bug #: 30843
           Summary: Adobe Flash 10 (only 10.0.x.x) fails to redraw
                    properly in fullscreen mode when hardware acceleration
                    is enabled (affects NP plugin when playing Youtube
                    movies and standalone flash-based apps like
                    Machinarium)
           Product: Wine
           Version: 1.5.5
          Platform: x86
               URL: http://helpx.adobe.com/flash-player/kb/archived-flash-
                    player-versions.html
        OS/Version: Linux
            Status: UNCONFIRMED
          Severity: normal
          Priority: P2
         Component: directx-d3d
        AssignedTo: wine-bugs at winehq.org
        ReportedBy: mooroon2 at mail.ru
                CC: hverbeet at gmail.com, stefan at codeweavers.com
    Classification: Unclassified


This bug affects only Adobe Flash Player versions from 10.0 family (i.e. having
version numbers in 10.0.x.x range). I had tested it to affect almost each 10.0
release that had been released to public, namely: 10.0.12.36, 10.0.15.3,
10.0.22.87, 10.0.32.18, 10.0.42.34 and 10.0.45.2

Full screen hardware accelerated mode of Adobe Flash 10 had been failing hard
(hanging, crashing, e.t.c., check bug #25533 for example) on systems up until
the commit bd97589d. Starting with that commit and up to Wine 1.5.5 hangs no
longer happen but there's an issue with Flash player not being able to redraw
its window contents when in in fullscreen mode.

This problem only happen in case "Hardware acceleration" is enabled (relevant
checkbox is checked in "Properties->Display" of the dialog box that is
accessible by right clicking on flash player window contents and selecting
"Properties..." menu item). Inspecting CPU usage and wine trace logs makes me
believe that hardware-accelerated fullscreen mode differ from other operating
modes of Flash Player (non-fullscreen or fullscreen without hardware
acceleration) in that it uses direct3d to do rendering.

Easiest way to reproduce problem is by installing fresh version of Firefox into
the prefix, applying workaround for bug #17273 (check comment #11 in that bug),
installing any release of Adobe Flash 10 Player NP plugin and trying to
playback any clip from Youtube in fullscreen mode. In case prefix is configured
to run in non-virtual-desktop mode (i.e. Wine creates a separate X visual for
every window windows apps tries to create) - Flash Player fullscreen window
won't be redrawed properly (i.e. would display statically a frame that had been
playing at the moment of switching into fullscreen mode). In virtual desktop
mode active for prefix "fullscreen" Flash Player window would be redrawn but
there would be some rendering artifacts noticeable looking like "periodical
quick jumps backwards in time and than back". 

Trying to apply fixes like activating "AlwaysOffscreen" wined3d mode doesn't
help (at least on the workstations I used to test/reproduce this bug).

With standalone Adobe Flash 10 based appls like Machinarium which don't
auto-quit fullscreen mode upon focus loss another interesting observation could
be made: Flash Player fullscreen window contents seem to redraw properly as
soon as there's some other window overlapping player's window which moves back
and forth. Simply having some other window sitting still in front of Flash
Player fullscreen window doesn't helps, one would have to move it in order to
force Flash Player into redrawing its window contents.

I'm not attaching logs to this bug as it seems that it is pretty easily and
quickly reproducible by anyone wishing to debug it. As far as I can tell from
inspecting logs (and comparing then with ones captured with Flash Player 10.1 -
which doesn't experience this bug) there are no obviously suspicious lines
there for relevant debug channels (ones acquired by grepping for "DEBUG_CHA"
inside wined3d and winex11.drv source tree folders).

-- 
Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email
Do not reply to this email, post in Bugzilla using the
above URL to reply.
------- You are receiving this mail because: -------
You are watching all bug changes.



More information about the wine-bugs mailing list