[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Gnash-dev] XVideo Patch

From: Craig Kelley
Subject: Re: [Gnash-dev] XVideo Patch
Date: Thu, 22 May 2008 08:43:05 -0600

On Wed, May 21, 2008 at 5:41 PM, Bastiaan Jacques <address@hidden> wrote:
> On Tue, 20 May 2008, Craig Kelley wrote:
>> Hello Gnash developers;
>> Here's a first go at XVideo with the GTK+ gui:
> Looks pretty good. A few notes:
> * I see the patch includes some work on the GUI polling interface. Is
> this a dependency of the patch or an unrelated change?

That is unrelated.  I found that the GUI would get out sync with audio
tracks when it was using the simple g_timer callback.  This change
lowers the timer to 5ms, and then consults a timeline to see if a new
frame should be rendered.  I understand that there is ongoing work up
in the Gui object and in the backend to fix this for all GUIs.  This,
however, is a stopgap that seems to solve the problem with at least
the GTK+ guis.  The highest resolution it supports is 200fps, which
should be ample (I would imagine).

> * it would be nice to have other GUIs (and renderers) take advantage of
> XV and friends. I imagine we could make a X11Gui class from which X11
> based GUI (such as gtk) could inherit from. Or perhaps something
> entirely different. That said, this might be something for after the
> initial landing of your patch.

This is what the Ekiga project ended up doing.  They created a third
class, apart from Gui and Renderer that would handle hardware
acceleration.  It makes sense, and should be the case with Gnash as
well.  I didn't want to step on any toes by refactoring it all.
Unfortunately, to my knowledge, Qt does not have a nice way of
granting XVideo an X11 drawable.

Under Windows, it would be Direct X (which is the other class that
Ekiga implements).

> I'll do a closer review of the patch soon.

Thank you;  I'm particularly interested if it works with gst libmedia.
 I suspect that my local libraries are broken somehow, but gdb is of
little use because the floating point exception is thrown in some gst
thread that has no backtrace into Gnash.  :-/


-- finger address@hidden for PGP block

reply via email to

[Prev in Thread] Current Thread [Next in Thread]