Re: Emacs i18n

[Eli Zaretskii <eliz@gnu.org>]

> Cc: juri@linkov.net, emacs-devel@gnu.org
> From: Paul Eggert <eggert@cs.ucla.edu>
> Date: Wed, 6 Mar 2019 11:47:23 -0800
> 
> On 3/6/19 10:15 AM, Eli Zaretskii wrote:
> 
> > how do we deal with strings that are computed by concatenation or
> > formatting?
> >
> The same way that other GNU packages deal with them: we redo calls, to
> make the strings easier to translate. For example, instead of this code
> (adapted from todo-mode.el):
> 
>   (message (concat "The highlighted item" (if (= count 1) " is " "s
> precedes ")
>                    "the timestamp %s.")
>            timestamp)
> 
> we do something like this:
> 
>   (nmessage count
>             "The highlighted item is not up to date."
>             "The highlighted items are not up to date."
>             timestamp)

That's the easy case.  This one is a bit tougher:

  (message "The program says: " (shell-command-to-string "foo"))

> It's inevitable that we'd need to redo Lisp code this way, as
> translators cannot be expected to be programming experts that understand
> arbitrary Lisp code involving 'concat' and whatnot. This is what other
> GNU packages have done, and Emacs can do something similar.

Except that Emacs is so much larger that doing this "like other
packages" might make the job infeasible.

Which is one reason why I think we should start from doc strings: they
are both easier and much more important.