Re: test-org-table/sort-lines: Failing test on macOS

[Michael Brand <michael.ch.brand@gmail.com>]

Hi all

On Tue, Nov 22, 2022 at 2:14 AM Ihor Radchenko <yantar92@posteo.net> wrote:
>
> Max Nikulin <manikulin@gmail.com> writes:
>
> >>> Concerning the test, I would split the current testcase into 2 parts
> >>> depending on WITH-CASE argument, check if caseless collation is
> >>> available and skip the related test otherwise.
> >>
> >> How can we check the availability?
> >
> > (string-collate-lessp "a" "B" "C" t)
>
> Thanks!

For this specific subject I can not see a follow up in this thread, or
an implementation on the branches bugfix or main. And the complaint of
the OP about the failing test is still valid on bugfix (on main ~make
test~ aborted before with "void-variable (file)" on "(delete-file
file)" quite early when loading testing/lisp/test-ob-tangle.el).

As far as I understand and agree, some Org functionality wants to
stick with ~string-collate-lessp~. Is it welcome if I finish working
on my local patch that plans to use

#+begin_src emacs-lisp
  (defun org-test-string-collate-lessp-ignore-case-supported-p
      (&optional locale)
    "Whether `string-collate-lessp' supports ignore case for LOCALE.
  According to the docstring of `string-collate-lessp' it does not
  implement ignore case for some locale on some operating
  systems (actually depending on libc of Emacs). E. g. on macOS
  `string-collate-lessp' does not ignore case for the locale C.

  See also https://debbugs.gnu.org/cgi/bugreport.cgi?bug=59275 and
  https://list.orgmode.org/orgmode/m2ilkwso8r.fsf@me.com"
    (let ((ignore-case t))
      (string-collate-lessp "a" "B" locale ignore-case)))
#+end_src

not to skip the test as suggested in the quote at the beginning of
this message but to bifurcate the expected result of the ERT to fix
the test for e. g. macOS on bugfix?