[vlc-devel] [PATCH] auhal setMute (refs #7329)

[Felix Paul Kühne]

Hello Rémi,

On 26.08.2012, at 18:12, "Rémi Denis-Courmont" <rem at videolan.org> wrote:

>   Hello,
> 
> On Sunday 26 August 2012, Felix Paul Kühne wrote:
>> On 26.08.2012, at 17:19, "Rémi Denis-Courmont" <remi at remlab.net> wrote:
>>>  Hello,
>>> 
>>> On Sunday 26 August 2012, Felix Paul Kühne wrote:
>>>> Well, it appears that the core no longer maintains the volume control
>>>> (see my patches, if you like).
>>>> 
>>>> Thus, every audio output needs to implement it itself as well as every
>>>> interface module. Sounds silly, adds lots of duplicated code and
>>>> work-arounds,
>>> 
>>> You know, this is insulting.
>> 
>> No, this is about the code level, not about you, your person or any of your
>> character's or body's qualities or characteristics. I didn't even say that
>> your work was BS.
> 
> Not explicitly but...

okay, I see what you mean and it wasn't intentional. I'm sorry about the somewhat harsh and arrogant words earlier.

>> Regrettably, ranting on this list seems to be the only way to get any
>> response from you, which is a shame.
> 
> This is a gross misrepresentation. You did not ask *me* anything. You just 
> assigned me two bugs that actually belong in the CoreAudio plugin.
I assigned them to core because the functionality needed was provided by the core in the past and still should be IMO. See below.


> I did not exhibit any activity on git, Trac, IRC or the ML; I was obviously 
> absent. In fact I was flying to a conference on *VideoLAN* sponsorship. So, 
> sorry but I am a normal individual: I do not have satellite Internet and I am 
> not always available within 48 hours notice.
Alright. In fact, after sending this previous email, I was wondering when you are going to San Diego. Honestly, I had the random thought "isn't Rémi going to the US this week, why isn't he there already  and when will he be"...

In fact, you usually respond within minutes or hours, which you obviously didn't while being on the plane. I forgot about this possibility and expected a "figure it out yourself"-like response, especially since you re-assigned the tickets without a comment and without responding to my previous mail including patches with work-arounds. This way, I expected you to have seen my mail long ago and decided not to care. I was wrong about this.

> However, my commit logs are on average the longest of all key developers. Last 
> time I checked, I also held the largest comments to code ratio (source: 
> Ohloh). You might not be bothered to read those comments, but that is your 
> problem, not mine.
> 
> If you thought some documentation was unclear, you were welcome to reply to 
> the commit log, or file a bug against the Documentation component, or whatever. 
> I don't recall of any such occurence.
Point taken.

> Where do you actually see duplication?
> Where do you actually see bloat?
Not within the code, but it's implications.

Why does every audio output module need to save its volume (if requested by the user) on its own? Why isn't there a generic cross-platform implementation for this? The bloat isn't huge, but looks like we will need "wasapi-volume", "auhal-volume", etc, while a single "volume" config would be sufficient like in 2.0.

> There is identical logic for software volume outputs (ALSA, JACK, KAI and 
> conditionally WaveOut). Because it is identical, it is factored (in volume.h). 
> So I really wonder where you saw duplicated code.
I missed this file mostly because auhal is no software volume output.

For hardware volume outputs, there seems to be no way to set the volume if no audio output is present. This sounds logical on a technical level, but isn't obvious for the user. People like to change the volume prior to starting a clip, so this value needs to go somewhere, so it is picked up when the aout is created. Of course, I could set "auhal-volume" to the desired value, but this way I'd abuse a config item for a variable and ignore the "volume-save" option. 

> As for bloat, the core for UI volume handling has become smaller and simpler. 
> Compare the sizes of src/audio_output/intf.c if you don't believe it... So 
> again, I doubt you actually saw what you claim to see.

I didn't want to say that you bloated the core. Quite the opposite. There is needed functionality missing now IMO.


Cheers,

Felix