[vlc-devel] Status of Win32 branch nightly

[Derk-Jan Hartman]

On 10 sep 2008, at 11:05, Marian Ďurkovič wrote:
> On Wed, Sep 10, 2008 at 10:30:33AM +0200, Derk-Jan Hartman wrote:
>> On 10 sep 2008, at 08:04, Marian Ďurkovič wrote:
>>> - we still have crash on AR change
>>
>> Did we have a crashlog of this ?
>> We should really see if there is not a way to fix this.
>
> Yes, there is - vout shouldn't be destroyed and recreated due to AR  
> change.
> What happens now is the following:
>
> Picture pointers before AR change - taken via
> msg_Err( p_vout, "DatePicture %p",p_pic ) in vout_DatePicture():
>
> [00000429] main video output error: DatePicture 0xaf61dcac
> [00000429] main video output error: DatePicture 0xaf61d94c
> [00000429] main video output error: DatePicture 0xaf61deec
> [00000429] main video output error: DatePicture 0xaf61da6c
>
> Picture pointers after AR change:
>
> [00000441] main video output error: DatePicture 0xb4a382ec
> [00000441] main video output error: DatePicture 0xb4a3852c
> [00000441] main video output error: DatePicture 0xb4a3864c
> [00000441] main video output error: DatePicture 0xb4a3840c
>
>
> In other words, after vout respawn, everything is at different
> places in memory.
>
> Now with libmpeg2 0.4.1, the library outputs STATE_INVALID_END
> and STATE_SEQUENCE during AR change, which means VLC doesn't try
> to fetch pictures from the old memory range.

We could just emulate those steps instead...

DJ