Re: scratch/igc fe90d556834: Make glyphs ambiguous roots

[Gerd Möllmann]

Helmut Eller <eller.helmut@gmail.com> writes:

> On Mon, Jul 29 2024, Gerd Möllmann wrote:
>
>>>> While the doc doesn't guarantee that all other (registered) threads are
>>>> suspended, it's hard to imagine what else MPS could do to stop other
>>>> threads from messing up the object.  (MPS must remove hardware barriers,
>>>> so that's not an option.)
>>>
>>> Good point, indeed.
>>>
>>> I've reverted that change. Thanks for checking!
>>
>> By the same argument, I think we could simplify face_cache and
>> image_cache which both use IGC_OBJ_PTR_VEC at the moment. 
>> IGC_OBJ_PTR_VEC could then go completely.
>
> You mean, you would use the regular xmalloc instead of igc_make_ptr_vec
> to allocate those vectors and trace the vectors in fix_face_cache.  Hm,
> yes that's indeed a similar situation.  As long as those caches aren't
> roots, I don't see a problem.

Yes, axactly. The bucket vectors and face_cache::faces_by_id and
image_cache::images could be xmalloc'd like in the non-MPS case,
and fix_face_cache and fix_image_cache could iterator over them and so
on. 

> I guess this would make the MPS and non-MPS code more alike.  We'd lose
> some details for igc-info, though that's a minor issue.

Yes, and things gets a bit simpler, which I'm personally a big fan of.
I'll put that on my todo list.

Thanks!