Re: [xine-devel] Re: 0.9.9 release plans?

[James Courtier-D. <Ja...@su...>]

Miguel Freitas wrote:

>On Sun, 2002-04-14 at 22:21, James Courtier-Dutton wrote:
>  
>
>>>I had not noticed the problem of seeking when plugin is blocked. I see 
>>>it now.
>>>I did see a problem with stopping or trying to exit xine when on a 
>>>still menu, I.E. input read_block function sleeping.
>>>I can fix the seeking on still menu problem as I fixed the stop/exit 
>>>problem as the issue is similar.
>>>
>>>
>>>      
>>>
>>It does not hang any more, but I have noticed that the seek function 
>>seeks to the wrong place.
>>I suspect the seek function in dvdnav, but will look at the problem later.
>>    
>>
>
>James, 
>
>You have not understood me: both of the suggestions i sent in my
>previous message had their drawbacks. The fix you commited to
>demux_mpeg_block actually _breaks_ seeking of mpeg2 streams, including
>when using the default dvd plugin.
>
>Since you only care about dvd and dvdnav, why can't you fix this issue
>using a patch similar to this one:
>
>--- input/input_dvdnav.c        13 Apr 2002 13:35:36 -0000      1.131
>+++ input/input_dvdnav.c        16 Apr 2002 03:09:38 -0000
>@@ -519,6 +519,9 @@
>        dvdnav_still_event_t *still_event =
>          (dvdnav_still_event_t*)(buf->mem);
> 
>+       buf->type = BUF_CONTROL_NOP;
>+       finished = 1;
>+
>        /* Xine's method of doing still-frames */
>        if (this->pause_timer == 0) {
>          dprint("dvdnav:input_dvdnav.c:Stillframe! (pause time =
>0x%02x)\n",
>
>dvdnav used to send BUF_VIDEO_FILL so the code to maintain the still
>duration across multiple read_block calls must be somewhere in cvs. you
>just have to change it to a NOP and none of input->stop() patches would
>be needed.
>
>
>regards,
>
>Miguel
>
No problem, that seems easy enough.
Cheers
James

>  
>