Re: [AD] new_api_branch: al_create_display and blit patch

[Evert Glebbeek <eglebbk@xxxxxxxxxx>]

On Monday 16 May 2005 17:33, Robert Ohannessian wrote:
> AGL uses flags for a completely different purpose than the Allegro API.
> The flags in Allegro specify the *operation* that should be done. In
> AGL, it specifies the *options* that should be used.
> 
> Granted, it's a semantic difference.
> 
> I feel more strongly about putting the flags in front for Allegro, but I
> won't argue for it. They go in either front or back.

I think Bob has a good point here. I think if the function prototype were 
written as
void al_blit(int how, source params, dest params) 
instead of
void al_blit(int flags, source params, dest params) 
we might not even have the discussion of wether flags go in back or in 
front.

My initial reaction, as I said, was also `wtf... flags should go in the 
back!', but when I thought more about it and played with the 
implementation and some samples of how it would be used, I actually did 
find `operation in front' more logical and easy to work with.

In other words, I propose to give it a chance. Afterall, the idea is to 
make a new API that is well-thought out. Personally, I also have no big 
objections to implementing it one way now and switching around if we 
decide we really can't stand it afterall while the code is this early in 
development.

Ever