[prev in list] [next in list] [prev in thread] [next in thread] 

List:       helix-video-dev
Subject:    [Video-dev] Re: HX_SURFACE_UPDATE again...
From:       "Greg Wright" <gwright () real ! com>
Date:       2004-04-27 17:41:14
Message-ID: 1d3d01c42c7e$d7046f30$ff8317ac () gwright5
[Download RAW message or body]

> I managed to get everything from the CVS again, got a working umake. I 
> encountered the same problems as before even with the Makefile generated 
> by umake. I finally managed to compile this clean capture_vframes that 
> came from the CVS and got the same problem as before: segfault on 
> HX_SURFACE_UPDATE. So I looked at the file you told me (fb_site, 

Are you running this in a debugger? Can you tell me the stack trace
of the core-dump or where it is actually crashing at? Or, are you
running your new plugin in a pre-built player?

> fb_surf) and checked what this frame buffer site did on ForceRedraw. 
> Then I saw it sent a pointer to HXxWindow instead of NULL (NULL is what 
> is sent in capture_vframes and make the whole thing crash). I searched 
> for the way this HXxWindow was created in the case of fb_site, and found 
> out a that trough the _create function of fb_site the HX window handler 
> was NULL or 0x42. I tried 0x42 in capture_vframes and I still get the 
> same segfault. I know this is not testing fb_site instead of 

It sounds like whatever renderer you are sending the surface update
to is using that window handle, and so it can't be NULL. What renderer is
this? You may need to recompile that renderer to not use X as well. I 
can't tell you for sure until I know what content you are playing back.


> nowindow_site but I guess it will behave the same way. I just want to 
> understand why passing NULL as argument for window handler in the 
> HX_SURFACE_UPDATE event makes the Linux version crash and doesn't bother 
> at all the Windows version.

What player and build are you using under linux (and how did you get
it) and what player/build did you use under windows?

--greg.

> 
> I hope there is a simple solution or my boss will probably kill me for 
> wasting so much time on porting this thing to Linux.
> 
> Thanks for the help,
> Kouiskas
> 



[prev in list] [next in list] [prev in thread] [next in thread]