| Re: [AD] 4.3 display update methods |
[ Thread Index |
Date Index
| More lists.liballeg.org/allegro-developers Archives
]
On Thursday 10 August 2006 13:39, Peter Hull wrote:
> On 8/10/06, Robert Ohannessian <ROhannessian@xxxxxxxxxx> wrote:
> > I was referring to NOT IMMEDIATE, FRONT BUFFER NOT RETAINED, BACK BUFFER
> > NOT RETAINED by default.
>
> So, are we agreed that this is the scheme that can be implemented at
> driver-level for all platforms?
I think this scheme should be implemented at the API level. If you can't
gaurantee that all drivers can do IMMEDIATE/FRONT BUFFER RETAINED/BACK BUFFER
RETAINED, then you shouldn't be able to say that the API can. Even if you run
across a driver that can do one or all of those things, you shouldn't expect
the API to.
> Do we
> need to mandate 'non-immediate' behaviour?
IMO, yes. In worst-case it can always be implemented as a memory
double-buffer. Although I do think there should be a mechanism to allow
immediate behavior (assuming all drivers have a method to do so), but it
should not ever be the default, and should be discouraged except for special
situations.