[PATCH] shell: fix weston_output_mode_switch() usage

[Derek Foreman]

On 23/09/14 08:34 AM, Hardening wrote:
> Le 23/09/2014 09:37, Pekka Paalanen a écrit :
>> On Mon, 22 Sep 2014 11:01:10 -0500
>> Derek Foreman <derekf at osg.samsung.com> wrote:
>>
>>> On 22/09/14 06:41 AM, Pekka Paalanen wrote:
>>>> On Fri, 19 Sep 2014 14:43:23 -0500
>>>> Derek Foreman <derekf at osg.samsung.com> wrote:
>>>>
>>>>> Calling weston_output_mode_switch() with WESTON_MODE_SWITCH_RESTORE_NATIVE
>>>>> will result in the mode being set "back" to the passed in mode - so the
> 
> [...]
> 
>> No, its really all clients. Even the fullscreen client does not know if
>> the temporary mode change it requested actually happened.
>>
>>>> Therefore: pushed. Feel free to refactor the code in a follow-up,
>>>> IMHO. ;-)
>>>
>>> Ok, I'll try to get to that in the next couple of days.
>>
>> Thank you.
>>
>> Hardening, could you keep an eye here and make sure we don't break
>> anything for the RDP backend or such?
>>
> 
> Yes of course. IIRC at the time we had already suggested to split and
> rework the function.

Thanks.

>> I believe some of the details here are only to cater for RDP, like
>> non-enumerable modes.
>>
> I may be wrong but I remember a conversation where someone was telling
> that DRM also have arbitrary modes...

Actually, I think compositor-rdp.c does the right thing in this regard.
 It's the set_temporary callers that pass in pointers to stack local
structs.

Both of these usages are OK, as long as no future users of set_native
crop up that try to treat it like set temporary...