RE: [AD] A small bugfix in allegro_exit

[Vincent Penquerc'h <Vincent.Penquerch@xxxxxxxxxx>]

Title: RE: [AD] A small bugfix in allegro_exit

> But isn't the destructor is readbmp.c supposed to delete what
> the constructor
> has added?

It does, but called after Fortify's atexit scheduled routine which
outputs the leaks is called. Which prompted me to wonder why the
atexit called bitmap loader unregistration function was calling the
registration again anyway, which is a bit dodgy, even if I see why
this was done that way :)
I just see with a slightly concerned face Allegro stuff being live
before allegro_init is called, and after allegro_exit is called.

--
Vincent Penquerc'h