On Thu, 2015-10-22 at 08:50 +0200, David Henningsson wrote: > > On 2015-10-21 17:26, Tanu Kaskinen wrote: > > So you think it's ok to have pa_device_port.preferred_profile, even if > > the field doesn't actually refer to a profile? I'm definitely not ok > > with that. > > The field refers to either: > Â 1) one or more profiles' input_name, or > Â 2) one or more profiles' output_name, or > Â 3) a profile's name > > Hence, in its current form, it does refer to one or more profiles. Given > that > "preferred_profile_name_or_profile_input_name_or_profile_output_name" > was too long, I shortened it to "preferred_profile". Do you have a > better name suggestion for the field name? "preferred_thing" maybe? :) I think that would be less misleading, but I don't care if you keep "preferred_profile", it will hopefully change anyway before too long. > > But if you think it's "too much effort for little gain" to > > make the code correct,it's probably the least hassle if we apply your > > patches, and I fix it up, since I have more motivation, assuming that > > you won't be blocking patches that add name reservation to the > > namereg etc. > > Ok, that sounds like a compromise I can agree on. > > (Even though I disagree with your wording - I don't think my code is > incorrect just because it doesn't improve the situation for multi-sink > profiles.) Right, there's nothing else incorrect than the name "preferred_profile", and changing that to "preferred_thing" wouldn't make me much happier, so complaining about incorrectness is pretty pointless. --Â Tanu
