On Wed, 2012-09-26 at 17:42 +0200, Mikel Astiz wrote: > Hi Arun, > > On Wed, Sep 26, 2012 at 1:57 PM, Arun Raghavan [...] > > 18 and 19 are a no-go from me. I'm not okay with dependencies between > > unrelated modules. As mentioned on IRC a while back, I'd much rather > > signal an unwillingness to be suspended via sink/source flags. > > > > The alternative is we declare suspend-on-idle core, but tbh I think > > that's messier than just doing the flags. > > I will think of alternative approaches but keep in mind that the > "unwillingness" is idle-specific, so adding these flags also sounds to > me like declaring suspend-on-idle core. The unwillingness is not actually specific to module-suspend-on-idle. Imagine a hypothetical system policy module that implements something like wakelocks for audio. Applications explicitly inform the system when they want a device and when nothing needs it, the module suspends the device. This would also be an idle suspend, but not made by module-suspend-on-idle. While this is completely hypothetical, it's certainly feasible, which is why I want these sorts of underlying hardware properties to be exposed by the corresponding PA object. > >> - Patch 22 is experimental and should not be applied yet. > > > > Luiz mentioned the patch for that version is in, so if you could just > > update the version on the comment (or even just tell me on IRC), I'll > > update and push that one too. > > Let's discuss this with Luiz but I would rather leave it out. Even > though the patch is already applied in BlueZ, there has been no > release yet and chances are that the next release will take some time > and involve other changes. Sure, that sounds fine. Cheers, Arun
