[PATCH] Use cached XKB keymap when rules haven't changed

[Dan Nicholson]

On Sun, Nov 30, 2008 at 9:38 PM, Daniel Stone <daniel at fooishbar.org> wrote:
> On Mon, Dec 01, 2008 at 03:24:42PM +1000, Peter Hutterer wrote:
>> On Wed, Nov 26, 2008 at 06:28:58AM -0800, Dan Nicholson wrote:
>> >  So, it turns out that XkbCopyKeymap doesn't quite do what you'd expect,
>> >  failing to copy a few fields of the XkbDescRec. I'm not sure whether it's
>> >  appropriate to copy .device_spec from the cached keymap, but it didn't
>> >  seem to hurt in my testing.
>>
>> Tbh, I don't know either :) Daniel, any insights?
>
> I'm not sure why device_spec even _exists_ (except to be able to pass an
> XkbSrvInfoPtr instead of a DeviceIntPtr, but honestly, what's the point
> in doing that: just pass either a DeviceIntPtr or an XkbDescPtr), so
> that's on my xkb-atkins hitlist (ha!).  defined should almost certainly
> be copied, yeah.

`git grep device_spec' seems to show that it's set to XkbUseCoreKbd
during XkbAllocKeyboard() and never changed again anywhere else. So, I
don't think it matters to copy it in the server.

I was reading the xkb-atkins log a couple weeks ago. It looks like a
nice diet. :)

--
Dan