Re: The overlay branch (yet again)

[Vladimir Kazanov <vekazanov@gmail.com>]

Hi all again,

I went through the code written by Andreas Politz (in the
feature/noverlay branch) and all of the related discussions. He
completed an astonishing amount of work! Apart from changes to the
working code the branch includes 1300 LOC of manually written
overlay-related tests, 5700 LOC of automatically generated unit tests
and C-level interval tree tests. Check test/src/buffer-tests.el in the
feature/noverlay branch for reference.

It would be a waste to just throw all of that away. As a first step I
really want to put those tests to work.

Given that nobody replied to my original email I assume everybody is
busy with the new release. In the meantime I'd really like to try to
pick up the work already done on overlays. I have a following plan:

1. A separate patch containing all those tests.
2. A second patch, making overlay code independent from the underlying
implementation.
3. A third patch, providing a tree-based data structure using either
what Andreas built, or smth else.

The first patch should be fairy trivial: I'll extract overlay-specific
tests into a overlay-test.el (while attributing all of the changes to
Andreas, of course), clean up a bit and drop C-level tests. I believe
those tests are really useful the way they are right now and can be
merged separately.

Having those tests merged it will be much easier to go on with further
overlay-related changes (i.e. tweaking internal APIs, or underlying
data structures).

Is anybody interested in me doing this work?

On Tue, Nov 26, 2019 at 2:44 PM Vladimir Kazanov <vekazanov@gmail.com> wrote:
>
> Hi all,
>
> a few years ago I tried replacing the underlying overlay data
> structure with a tree-based one. Back then I ran out of spare time and
> had to delay the project.
>
> I'll have time on my hands this winter and would like to get back to
> the project again. This time it seems that most of the work was done
> by Andreas Politz in the feature/noverlay branch:
>
> https://lists.gnu.org/archive/html/emacs-devel/2018-03/msg00565.html
> https://lists.gnu.org/archive/html/emacs-devel/2019-09/msg00543.html
>
> It seems reasonable to pick up the branch and try to finish it, unless
> somebody is already working on it. So my questions are:
>
> is anybody working on the branch? why was it abandoned? Are there any
> known problems with it?
>
> Thank you
>
> --
> Yours sincerely,
>
>
> Vladimir Kazanov



-- 
Yours sincerely,


Vladimir Kazanov


--
С уважением,

Владимир Казанов