[vlc-devel] [PATCH] [GSoC] New VLC interface for Maemo

[Laurent Aimar]

On Mon, Nov 10, 2008, Antoine Lejeune wrote:
> Hi, I've corrected my patches according to the remarks I received on
> vlc-devel and IRC.
> 
> I've also add a patch that explains how to build VLC for Maemo (an
> INSTALL.maemo). The patches are still at the same address :
> http://phytos.dinauz.org/VideoLAN/patches/0001-A-new-GUI-for-Maemo-based-on-Hildon-framework.patch
> http://phytos.dinauz.org/VideoLAN/patches/0002-Maemo-Add-the-swscale_nokia770-library.patch
> http://phytos.dinauz.org/VideoLAN/patches/0003-Add-an-omap-framebuffer-video-output-module.patch
> http://phytos.dinauz.org/VideoLAN/patches/0004-Add-an-INSTALL.maemo-file.patch
> 
> Le lundi 10 novembre 2008 à 05:41:05, Rémi Denis-Courmont a écrit :
> > Weird that you need to link with the Bluetooth dialogs.
> 
> I have removed this unnecessary link.
> 
> > You can probably default this one to autodetect (!= no) rather than disabled 
> > (= yes), as I assume it applies to all linux-omap boards. I don't understand 
> 
> The configure only tests if there is an omapfb.h file. A lot of computer
> can have it but the module won't work on many of them.
> 
> > why you extend vout_thread_t - is this a significant performance 
> > optimization?
> 
> I extend vout_thread_t to disable the rendering when not needed (so that
> the vout thread won't write into the framebuffer). It is used when the
> user reduce the vlc window into the taskbar or when another window is
> over the vlc one : vlc has to stop writing into the framebuffer or the
> video will continue to be displayed over every thing.
> 
> It's the only way I found to do this.
 Why not moving it to the p_sys field ? You seems to be the one writing to
it *and* reading it.

-- 
fenrir