Hi / re,

I'm not using tree-sitter personally, so it's a bit of a pain to know
whether an update brings runtime regressions.  If you care
about tree-sitter support it would be good that Laurent Cheylus
(and/or Edd) Cc both you and me for future tree-sitter updates.

On Wed, Apr 29, 2026 at 05:25:36PM +0200, Paco Esteban wrote:
> Paco Esteban <paco@e1e0.net> writes:
> 
> > [...]
> > Also, I'm not entirely sure about the changes 'make update-plist' made.  It
> > removes some folders, but there are entries for regular files inside those
> > folders, so I guess it's fine (?).

It should be fine but it's a bit of a pain.  There's zero problem for
a directory to be shipped by several packages, alas AFAIK make
update-plist tends to strip directories if they already are shipped by
a dep.  Right now I'm not sure which deps are shipping eg
share/emacs/site-lisp but anyway, it feels like needless churn for
update-plist to strip these entries.

Also it looks like update-plist moves %%pgtk%% to a mid-file location
because it appears right after a !%%no_x11%% entry.  Moving %%pgtk%%
before !%%no_x11%% and voila, make update-plist leaves it in place.
Just an update-plist quirk, I guess.  Anyway...

> Updated diff with the PLIST change removed, as per jca@'s request.

ok jca@ for post-unlock, also since 7.9 ships the same emacs and
tree-sitter versions a backport ought to be useful.

-- 
jca