bug#57150: 29.0.50; [PATCH] Add test coverage for overlay modification hooks

[Matt Armstrong <matt@rfc20.org>]

Ihor Radchenko <yantar92@gmail.com> writes:

> Matt Armstrong <matt@rfc20.org> writes:
>
>> I noticed there was no explicit test against the overlay modification
>> hook boundary conditions.  The attached patch adds some, which might be
>> nice to have if a person (possibly me) someday optimizes overlay
>> performance with a change in data structure.
>
> Just in case. Have you seen
> https://lists.gnu.org/archive/html/emacs-devel/2019-12/msg00323.htlm
> ?
>
> I am not sure how much progress have been made on the noverlay branch,
> but there was at least some existing test suitcase in there. Not sure if
> it has been merged or if your proposed patch is not covered.

Yes, thank you for pointing that out.  I have gone as far as importing
the entire history of emacs-devel into https://notmuchmail.org/ and read
as much history about the overlay problem as I could discover.

I have been in contact with both Andreas (the original feature/noverlay
author) and Vladimir (who later examined it).  Andreas had no plans to
return to looking at the problem and did not recall the precise state
the branch was left in.  Vladimir, despite his interest, was
unfortunately not able to get his employer to sign FSF papers.  Both
wished me luck.

Fortunately all of the test cases from the feature/noverlay branch were
merged to the master branch, I believe by Eli, shortly after they were
written years ago.

As for feature/noverlay itself, I have been exploring alternative
approaches that might be simpler or clearer than Andreas' approach, with
the clear advantage of seeing his implementation, of course.  Failing
any improvement I can always start right where he left off.