Re: [Patch] to correctly sort the items with emphasis marks in a list

[Maxim Nikulin]

Hi Juan,

On 16/04/2021 01:21, you wrote:
> #+begin_src emacs-lisp :tangle list-var.el
>    (message "%S" (sort '("-\s\sv" "-\sv" "-\sa" "-\s\sa")
>                         #'string-collate-lessp))
> #+end_src
>
> #+begin_src sh
> exec 2>&1
> LC_ALL=en_US.UTF-8 emacs --batch -Q -l list-var.el
> #+end_src
>
> #+RESULTS:
> : -  a" "- a" "-  v" "- v
>
> #+begin_src sh
> exec 2>&1
> LC_ALL=es_ES.UTF-8 emacs --batch -Q -l list-var.el
> #+end_src
>
> #+RESULTS:
> : -  a" "-  v" "- a" "- v

You have managed to convince me that despite my first suspects the 
locale on your computer is correct. It is unexpectedly correct and it is 
more correct that most of locales in libc.

However I do not have opinion concerning you patch yet. I have not 
realized what is the proper way to sort list.

Space is significant. At least it may be. Only a few languages have got 
such fix, Spanish is among them 
https://sourceware.org/git/?p=glibc.git;a=blobdiff;f=localedata/locales/es_ES;h=aa919a26267fd6311b71d7aeb81655e55787b4df;hp=d17612f6726d0c098ac981e06f3702106540bb23;hb=159738548130d5ac4fe6178977e940ed5f8cfdc4;hpb=ce6636b06b67d6bb9b3d6927bf2a926b9b7478f5

Notice "collating-symbol <space>"

I have found example of sorting names in a language where woman surname 
usually have additional "a" in comparison to man surname.

printf "Ivanova Alla\nIvanov Adam\nIvanova Svetlana\n" \
  | LANG=pl_PL.UTF-8 sort
Ivanov Adam
Ivanova Alla
Ivanova Svetlana

es_ES behavior is just as the above example.

printf "Ivanova Alla\nIvanov Adam\nIvanova Svetlana\n" \
  | LANG=en_US.UTF-8 sort
Ivanova Alla
Ivanov Adam
Ivanova Svetlana

Ukrainian sort works better than Russian one with such example:

printf "Иванова Алла\nИванов Адам\nИванова Светлана\n" \
  | LANG=uk_UA.UTF-8 sort
Иванов Адам
Иванова Алла
Иванова Светлана

printf "Иванова Алла\nИванов Адам\nИванова Светлана\n" \
  | LANG=ru_RU.UTF-8 sort
Иванова Алла
Иванов Адам
Иванова Светлана

Man names are sorted first in such lists. Other cases might exist when 
significant space is undesired.

So sorting is tricky than I expected.