[MPlayer-dev-eng] [PATCH] support -mouse-movements on OS/2

[Reimar Döffinger]

On Sat, Mar 01, 2014 at 11:44:01PM +0900, KO Myung-Hun wrote:
> > But without panscan support it's not going to give the right results
> > e.g. when that feature is used.
> > Try for example "-panscanrange -1 -panscan -0.5" (this is supposed
> > to scale the video to half the window size), I guess your code will
> > produce mouse coordinates that are quite off.
> 
> No. It does not respond to this at all. ^^

Hm, I was quite convinced it would. No idea what I missed.

> > I don't know what options there are on OS/2, but most overlay based VOs
> > nowadays use calc_src_dst_rects to just handle all these options
> > (in particular panscan and border-pos-x/y).
> 
> Ok. I've used calc_src_dst_rects().

Oh, I didn't expect it to be that simple...

> BTW, border-pos-x/y should be used for OSD ? I cannot find any
> differences of a OS/2 build not using it from a Windows build using it.

Honestly the way these options interact with OSD is basically random.
The thing is, for most vos the OSD will be drawn right onto the video,
thus we really have no real choice on where it ends up (which means
these options will affect the OSD location).
-vo gl in particular is an exception, and I think there these options will
not effect the OSD there.
But I didn't test it, and I wouldn't say that the one or the other is
intended.
I would say ideally, the location of the OSD would be configurable, but
so for nobody really requested that (except for ASS subtitles, but there
it is the library deciding it on its own, based on the size of the
borders).

> Ah, the dvdnav white box cursor problem was solved. It was due to -slang
> option I'm having in config file.

Hm, I think slang is supposed to be ignored for menus, I think I
implemented something there.
But maybe it's not working properly.