Re: noverlay branch

[Matt Armstrong]

Stefan Monnier <monnier@iro.umontreal.ca> writes:

> I just updated the noverlay branch to the code in `master` (most of it
> was mechanical except for the fact that since the last commit on that
> branch we have gotten rid of Lisp_Misc and we moved to pdumper).

TLDR: things are looking pretty stable and good on the feature/noverlay
branch.  If anyone here tested it a month ago and found it lacking, give
it another whirl.

Stefan, unless I have a flash of insight I'm done with my differential
fuzz testing of the noverlay branch's ELisp functions that relate to
overlays.  I've sent you three patches (see bug#58703 and bug#58706) but
with those fixes my test harness now runs hundreds of thousands of
randomized operations against with no no difference detected between
Emacs master and the feature/noverlay branch.  Code at
https://git.sr.ht/~matta/emacs/tree/my/overfuzz/item/test/manual/noverlay/overfuzz.el,
though I haven't polished it up.

There are two calls to 'overlays_in' in buffer.c that I couldn't cover
from ELisp: 'disable_line_numbers_overlay_at_eob' and
'mouse_face_overlay_overlaps'.  They're both display related.

Trying to manually exercise 'disable_line_numbers_overlay_at_eob' I
couldn't find any ELisp that set `display-line-numbers-disable' on an
overlay, and only two that do it in a text property (vc/log-edit.el and
something in helm), so even if there were a bug here I'm not sure anyone
would discover it.

I haven't looked at mouse_face_overlay_overlaps, but as I move my mouse
around different things highlight as expected, and as I narrow the
buffer things behave the same on feature/noverlay as they do on master
(as far as I can tell).