Hey Jacob, Thank you so much for the fast turn around and the high quality patch!
I apologize for the delay. Turns out my CI was broken. You did not break any tests. "Jacob S. Gordon" <[email protected]> writes: > Hey Morgan, > > Thanks for your diligence! Attached is a patch that tries to address > the errors, but I don’t have Emacs 28 setup to test (one of these days > I’ll have to ask about your multi-Emacs CI setup). It's based on guix but it is an entire rabbit hole that seems to break more often then I'd like. I don't mind one bit doing the testing for you. I'd much rather you spent your time on more fruitful endeavors like adding this feature! > On 2026-05-18 20:40, Morgan Smith wrote: >> org-agenda.el:10303:1: Warning: variable `_' not left unused >> org-agenda.el:10303:1: Warning: variable `_' not left unused > > This seems similar to [bug#69108] fixed by ‘e680827e814’ for Emacs 30. > I’ve removed the ‘_’ bindings in case that’s it. > > [bug#69108]<https://lists.gnu.org/archive/html/bug-gnu-emacs/2024-02/msg00723.html> Yes that was exactly it! It looks like they might try to deprecate this syntax of `when-let*' as I can see in this quote from the docstring of `if-let*' #+begin_quote An older form for entries of VARLIST is also supported, where SYMBOL is omitted, i.e. (VALUEFORM). This means the same as (_ VALUEFORM). This form is not recommended because many Lisp programmers find it significantly less readable. A future release of Emacs may introduce a byte-compiler warning for uses of (VALUEFORM) in VARLIST. #+end_quote That being said, this patch is perfectly readable. I doubt anyone will assume you're assigning to `eq' or `not'. I started going down the rabbit hole of trying to rewrite it with `seq-reduce' (which seems like a useful function but I've never managed to find a place to use it) but I'm trying to commit to less rabbit holes so I purposefully gave up. >> In org-agenda-clock-goto: >> org-agenda.el:10312:12: Warning: sort called with 3 arguments, but accepts >> only 2 > > Oh interesting, I didn’t know ‘sort’ was different pre-Emacs 30 > (‘ae5f2c02bd2’): ‘(sort SEQ LESSP)’. I’ve adapted to the compatible > calling convention. Looks good to me! I've run it through CI Morgan with the following results: Tests pass on emacs 29, 30.2, and a recent emacs/master build. Tests pass with TZ set to UTC, Europe/Istanbul, and America/New_York. 4 test failures on Emacs 28 but that's just because CI Morgan is still a little broken. Applied, as commit 5261e65d5 Thank you very much!!
