João Távora <joaotavora@gmail.com> writes:

> On Tue, Nov 1, 2022 at 10:48 AM Philip Kaludercic <philipk@posteo.net>
> wrote:
>
>>
>> BTW, if there are major objections to the language, I should point out
>> that the new `buffer-match-p' in Emacs 29 uses the same language and has
>> already found usage in a number of spots in core Emacs.  There would
>> still be time to address any issues you might have, and avoid a
>> long-term mistake.
>>
>
> For me, it looks like match-buffers is reinventing
> cl-remove-if-not and match-buffer-p is reinventing ... unary predicate
> function of a buffer?

You could say that, though I would argue it is an easier way to express
common predicates, while not making anything else more complicated.

E.g. `display-buffer-alist' makes use of it to associate display-buffer
rules with buffers.  Now you can add

      ((major-mode . help-mode) display-buffer-in-side-window)

instead of trying to match being a regular expression to catch all
*Help* buffer names of a function along the lines of

      (lambda (buf _alist)
        (with-current-buffer buf
          (derived-mode-p 'help-mode)))

> I'm not fond of these mini-languages because they're less expressive, they
> end up being only minimally less complicated and bug-prone, they can't
> automatically be byte-compiled for efficiency, and they can't automatically
> be byte-compiled for correctness/diagnostics.  If one makes a mistake,
> the backtrace is much more complicated.

I agree in principle, but this should be alleviated by using a lambda
function as a predicate.  The above check still works and can be used
anywhere you would use `buffer-match-p'.

> So these mini-languages may make sense to define filters in thunderbird or
> something, but throwing Elisp away here generally doesn't make sense to me.
>
> But there may be exceptions (although this project.el one doesn't seem one
> of them) so why don't you show examples of use of these new helpers and
> so we can compare side by side with the Elisp-only alternative.

I am biased, but I believe that the language could even find more use in
project.el, by having `project-buffers' just call `match-buffers' with a
special `buffer-match-p' predicate.  Here is a sketch of how that could
look like (I haven't tested it yet):