Re: [AD] clipping line algorithm

[Sven Sandberg <svens@xxxxxxxxxx>]

Robert Jr Ohannessian wrote:
> My basic question is: should we allow the user to bypass the vtable?

I think we should, for the following reasons:
 - Backwards compatibility. bmp_read* and bmp_write* have been around 
forever, and it would be a pity to break programs using them.
 - Speed. Since we don't implement all possible functionality in 
Allegro, some of it has to be in add-ons, in user's programs, or in 
generic modules somewhere inbetween the two. Those would benefit from 
the additional speed of direct access rather than going through the vtable.

And I don't think it is difficult to support.

> If not, then it doesn't matter, because all the Allegro API functions 
> should route to some bitmap's vtable.

This is not how it works at the moment: polygon(), 
[scene_]polygon3d[_f](), draw_gouraud_sprite(), floodfill(), arc(), 
spline(), rect() and others are all outside the vtable (see 
include/allegro/gfx.h). Are you suggesting to move all such functions to 
the vtable? In any case, that's a different discussion, I'll reply on 
the discussion on how to improve the current API for low-level variants 
of getpixel/putpixel.

[snip]
> > particular situation where it is really significantly bad? As long as 
> > we have a version that uses the vtable, and BITMAP->id is able to 
> > indicate when the optimized version can be used instead, I think we 
> > should be fine?
>
> I like the idea, if only because people could check the ID before using 
> bmp_read/write on a bitmap.

Good! So how should this be implemented? I gave two suggestions, both 
based on flags in BITMAP->id, and you suggest a BITMAP->caps bitfield 
with the same functionality. Pros and cons in my opinion:
 - BITMAP->id seems like the easiest-to-implement solution. That's how 
it is currently done.
 - BITMAP->caps seems a bit more extendible in case we stumble on 
anything else that a bitmap needs to be queried for. BITMAP->id also has 
the drawback that the lowest bits (those not used by a flag) are used as 
an integer, to identify sub-bitmaps.

But I really don't mind which way we implement it, as long as it does 
the job.

Sven