bug#37755: Logic in init_fringe_bitmap should be moved to backends (maybe rif->define_fringe_bitmap)

[Carlos Pita]

> I don't see that call in the patch, nor any changes in gui_init_fringe
> that would modify its current effect.

Because nothing changed in gui_init_fringe itself. It did and does:

  for (bt = NO_FRINGE_BITMAP + 1; bt < MAX_STANDARD_FRINGE_BITMAPS; bt++)
      rif->define_fringe_bitmap (bt, fb->bits, fb->height, fb->width);
  for ( ; bt < max_used_fringe_bitmap; bt++)
    rif->define_fringe_bitmap (bt, fb->bits, fb->height, fb->width);

The change affects rif->define_fringe_bitmap instead. It now does:

- Create platform-dependent structures from platform-independent bitmaps.

Previously this was divided between init and define as:

- Init: manipulate platform-independent bitmaps in a platform-dependent way.
- Define: use this platform-dependently shuffled bitmaps to create
platform-dependent structures.

So the only thing that have moved down the initialization sequence is
the bit-shuffling gymnastics which, if any, are done in
gui_init_fringe now.

> Sure, but I said define-fringe-bitmap, which is the Lisp name of
> Fdefine_fringe_bitmap.

I meant to remark that they do quite different things not that you
mistake one for the other, sorry if I wasn't clear.

> > I suggest you take a look at the modified pseudo code I posted quite a
> > few message above.
>
> I will, but I'd like to see the full patch as well.

You have already seen it :)

> If RIF is the problem, we could make each terminal backend do this
> initialization unconditionally at dump time.

According to my rationale above, I don't see any problem at all. But,
as I have said, I ignore everything about the dumper. Yet, I find it
hard to believe that whatever this  dumper thing is, it needs the bits
to be in little-endian, 8-bit per row format, or any other
rif-specific pattern.

Hope it's clearer now.