Re: Introducing Org-transclusion

[Noboru Ota <me@nobiot.com>]

Ihor Radchenko <yantar92@gmail.com> writes:

> I noticed that you are using overlays in text-clone. It may potentially
> cause slowdown in large Org buffers. Of course, it does not mean that
> org-transclusion should not be accepted. Just something that may be
> optimised.

Karl Voit (GitHub user name, novoid) gave me a comment in a similar vein
regarding potential performance issues as follows:

> As far as I understood: when you do have many overlays in deeply nested headings that are collapsed and you scroll over the lines with the collapsed headings, Emacs needs to handle those overlays and therefore is not able to perform a simple line change without issues.

Source:
https://github.com/nobiot/org-transclusion/issues/42#issuecomment-739570395

I personally have never come across such performance issue.  Are you
aware of some good test file or something to let other people reproduce
it?

For now, I am inclined to keeping the use of overlays as is. It's
partially because of my psychological inertia and partially because
text-clone makes uses of overlay's properties:

- "Evaporate" property to take care of when to finish text-clone
- Yanking does not copy overlay properties to avoid accidental cloning

If performance issues become more prevalent and it's something I can
address within my skill set, I will be happy to review the use of
overlays, of course.