[vlc-devel] [PATCH 3/4] core: vout: pass an input_thread_t to vout_GetDevice()

Thomas Guillem thomas at gllm.fr
Fri Jan 17 15:13:52 CET 2020



On Fri, Jan 17, 2020, at 14:49, Rémi Denis-Courmont wrote:
> I don't care about nonexistent apps that use an as yet nonexistent input item option. And by the way, zero is not "a lot".

VLC for Android is not a nonexistent app, check the play store for usage stats.

> 
> There are a number of apps ignoring the documentation and setting avcodec-hw in libvlc_new(). They will break no matter what exactly as the documentation warned they could, and they will have to adjust their syntax.
> 
> I do care about VLC and LibVLC working correctly, so not reintroducing parenting bugs and/or module string matching bugs. 

The dec dev parent will always be the vout, even with my patches. 
> 
> I also care about performance and thus not recreating the decoder device at every ES selection

The dec dev won't be recreated except if the user configure a different one for each item, which is very unlikely. 

> 
> -1.
> 
> Le 17 janvier 2020 15:18:51 GMT+02:00, Thomas Guillem <thomas at gllm.fr> a écrit :
>>  A lot of LibVLC clients have a setting to enable/disable hw decoding. Changing the "avcodec-hw" string by "dec-dev" is quite easy. But having to destroy and create a new libvlc in order to update user preferences is a huge change and require a lot of adaptation in client apps.
>> 
>> 
>> On Fri, Jan 17, 2020, at 12:28, Rémi Denis-Courmont wrote:
>>> This "feature" broke anyway since avcodec-hw no longer exists. Users used to be able to set plenty of variables before sout-keep and then input resources were implemented. Moving settings to more global scope is nothing new.
>>> 
>>> Le 17 janvier 2020 10:30:34 GMT+02:00, Thomas Guillem <thomas at gllm.fr> a écrit :
>>>> It's OK to break gapless if the hw dec settings change between 2 inputs. 
>>>> 
>>>> VLC users were always able to configure hw dec settings via the input (via avcodec-hw and other vars). I thnk we should keep it that way.
>>>> 
>>>> Do you have an other solution in mind? 
>>>> 
>>>> On Thu, Jan 16, 2020, at 19:00, Rémi Denis-Courmont wrote:
>>>>> Le torstaina 16. tammikuuta 2020, 18.41.36 EET Thomas Guillem a écrit :
>>>>>> Optinal, will be used to get the input_thread_t "dec-dev" variable.
>>>>> We don't accept output options on input items for years for good reasons. It 
>>>>> was totally broken. And it prevents gap-less.
>>>>> 
>>>>> -- 
>>>>> Реми Дёни-Курмон
>>>>> http://www.remlab.net/vlc-devel mailing list
>>>>> To unsubscribe or modify your subscription options:
>>>>> https://mailman.videolan.org/listinfo/vlc-devel
>>>> vlc-devel mailing list
>>>> To unsubscribe or modify your subscription options:
>>>> https://mailman.videolan.org/listinfo/vlc-devel
>>> 
>>> -- 
>>> Envoyé de mon appareil Android avec Courriel K-9 Mail. Veuillez excuser ma brièveté. 
>>> _______________________________________________
>>> vlc-devel mailing list
>>> To unsubscribe or modify your subscription options:
>>> https://mailman.videolan.org/listinfo/vlc-devel
>> 
> 
> -- 
> Envoyé de mon appareil Android avec Courriel K-9 Mail. Veuillez excuser ma brièveté. 
> _______________________________________________
> vlc-devel mailing list
> To unsubscribe or modify your subscription options:
> https://mailman.videolan.org/listinfo/vlc-devel
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.videolan.org/pipermail/vlc-devel/attachments/20200117/ea2390ef/attachment.html>


More information about the vlc-devel mailing list