Re: Exposing buffer text modifications to Lisp (was: Tree-sitter integration on feature/tree-sitter)

[Eli Zaretskii]

> From: Ihor Radchenko <yantar92@gmail.com>
> Cc: casouri@gmail.com,  emacs-devel@gnu.org
> Date: Sat, 18 Jun 2022 15:23:51 +0800
> 
> > Where you care about changes in buffer positions of the AST elements,
> > markers should take care of most, if not all, of them.  I presume you
> > already do use markers wherever possible? if not, why not?  Or what am
> > I missing?
> 
> Because lots of markers degrade Emacs regex search performance
> tremendously.
> 
> See https://list.orgmode.org/orgmode/scedec$2g0$1@ciao.gmane.io/
> and https://orgmode.org/list/87y21wkdwu.fsf@localhost

AFAIU, the right fix for this is to fix performance degradation when a
buffer has many markers, not avoiding the use of markers.

Here's one conclusion from this discussion that indicates changes
required to be done in core (other than a low-level modification hook
for buffer text) to take care of your AST implementation.

We already have a TODO item for making markers more efficient; any
takers?

> > Why would you miss significant changes if you base your implementation
> > on buffer-modification hooks?  If there are some situations where
> > buffer text is modified in ways that are significant for the update
> > of the AST, but buffer-modification hooks are NOT called, please
> > describe some of those situations, so we will have something concrete
> > to talk about.
> 
> The situation is third-party code doing bloody murder with
> 
> (with-silent-modifications
>  (insert "Some text not triggering modification hooks))
> 
> Another scenario is modifying text in indirect buffers created with
> make-indirect-buffer. (where there is no chance to install
> before/after-change-functions via clone-indirect-buffer-hook).

In at least the latter case the idea for a proper solution was
outlined by Stefan.

For other cases, I think a careful discussion on a case by case basis
will show the path towards solving each one of them.  It is possible
that some of them require further changes in core, but we won't know
until we discuss the details of each case.