Hi Hans, On Tue, Oct 04, 2022 at 10:43:55AM +0200, Hans Verkuil wrote: > Hi Sakari, > > On 10/4/22 00:01, Sakari Ailus wrote: > > Hi Hans, > > > > On Mon, Oct 03, 2022 at 04:26:32PM +0200, Hans Verkuil wrote: > >>> +#define V4L2_SUBDEV_ROUTE_FL_SOURCE (1U << 1) > >> > >> Can we rename this to _FL_INTERNAL_SOURCE? Just 'SOURCE' is very confusing > >> IMHO. The name 'INTERNAL_SOURCE' makes it clear that it is generated internally, > >> and so does not come from an external sink-side endpoint. > >> > >> I also think that the documentation for this flag in patch 04/19 is very vague, > >> I'll comment on that patch as well. > > > > Having descriptive names is important but I think "SOURCE" as such is fine > > for the purpose. Adding "INTERNAL_" adds 9 characters to what is already a > > very long name, making the flag very clumsy to use in code. That's why I > > would prefer to keep it as-is. > > > > _FL_SOURCE is meaningless (at least to me): there are so many 'source' and 'sink' > references, that just plain 'SOURCE' doesn't help me understand what the flag > means. I did consider INT_SOURCE, but I thought 'INT' is too close to 'interrupt'. > I'm OK with that, though. > > Another alternative would be _FL_NO_SINK: that clearly conveys that 1) there is > no sink, and implies that 2) the source is internally generated. > > Or perhaps: _FL_SOURCE_ONLY? This appears as the best compromise IMO. NO_SINK is shorter but conveying the meaning through negation is what I don't like too much. -- Kind regards, Sakari Ailus
