[PATCH 1/2] Clarify the enum argument type

[Nils Chr. Brause]

On Thu, Oct 1, 2015 at 9:00 PM, Bill Spitzak <spitzak at gmail.com> wrote:
>
>
> On Thu, Oct 1, 2015 at 8:09 AM, Nils Chr. Brause <nilschrbrause at gmail.com>
> wrote:
>>
>> Hi,
>>
>> I would recommend to also include the name of the interface, where the
>> enum is defined, in the enum attribute, e.g:
>> <arg name="format" type="uint" enum="wl_shm.format"/>
>
>
> I believe popular opinion was that a "local" enum, belonging to the same
> object as the message or event, should not require this prefix. One big
> reason is so that the object can be renamed without having to edit it's
> contents.
>
> And currently only such "local" enums are being used in the protocols, so
> there is no need for parsers to support any such prefix yet, and no need to
> figure out it's rules. Whether the ability to get enums from other objects,
> or from some global space, does not have to be decided yet.

There are "non-local" enums, e.g.
wl_surface::set_buffer_transform::transform uses the enum
wl_output::transform, or am I missing something important here?