Re: Code quality of some -ts-mode major modes

[Eli Zaretskii <eliz@gnu.org>]

> Date: Fri, 17 Mar 2023 23:45:00 +0200
> Cc: ruijie@netyu.xyz, casouri@gmail.com, emacs-devel@gnu.org
> From: Dmitry Gutov <dgutov@yandex.ru>
> 
> On 17/03/2023 15:54, Eli Zaretskii wrote:
> >> I repeat my question from above, if we are ready to link to the
> >> grammars, wouldn't it make sense to populate the variable
> >> `treesit-language-source-alist' directly?
> > No, I don't want to do that, see above for the reasons.  (We had this
> > discussion about 2 months ago, when the command was added to Emacs.)
> 
> FWIW, I had no problems with that conclusion, but then I noticed that we 
> keep a separate list of sources inside admin/notes/tree-sitter, which 
> basically contains the same info, as well as all 6 known exceptions to 
> the general scenario (where in the usual case we can determine 
> everything just from the name of the language).
> 
> If we're going to keep it updated (and we apparently are), why not move 
> that info to treesit-language-source-alist instead.

My original plan was to remove that script from admin/notes before
releasing Emacs 29.  I decided to wait for a while, but I definitely
don't want to commit ourselves to maintaining the script in the future
versions.  The information is just one Internet search away, usually
the first hit is the one you want.

More generally, the tree-sitter stuff seems to be just starting to get
noticed by distros, and it is too early to make any firm decisions
related to availability of information and libraries.  We already did
in this case more than we usually do for optional features, and I
think it should be enough for now.  Further efforts that increase our
maintenance burden should wait until it is clear they are necessary.