Re: Position of "latin" items in lists in RTL-locale contexts

[Marcus Lundblad <ml update uu se>]

tis 2017-05-16 klockan 09:01 +0200 skrev Claude Paroz:
> Le 15. 05. 17 à 23:10, Marcus Lundblad a écrit :
> > tis 2017-05-16 klockan 00:30 +0430 skrev mousavi arash gmail com:
> > > On Mon, May 15, 2017 at 11:45 PM, Marcus Lundblad <ml update uu.s
> > >
> > > Best Wishes,
> > > Arash
> > >
> > > [1] https://en.wikipedia.org/wiki/Right-to-left_mark
> >
> > Hi!
> >
> > Thanks for your answer Arash!
> > I attached a patch to the aforementioned bug which I think should
> > do
> > achieve the goal (embedding each component with the RLM or LRM
> > depending on the locale). From some quick testing it kinda looks
> > right
> > to me. But it would be great with more testing (I might very well
> > be
> > missing something obvious here).
>
> What about just offering the whole expression to translate: "%s →
> %s".
> Someone should check that the printf syntax ("%2$s ← %1$s") is
> properly 
> handled in JS code.
>
> Claude

I'm not sure that would work, since the strings could be names written
in either an LTR or RTL script, so depending on the first one, the
overrall direction of the rendered string would be affected, thus the
usage of the directional markers in the generated string.

If we assume the example above with the translation of the placeholder
string "%s → %s" to "%2$s ← %1$s" in an Arabic translation, this would
be rendered as expected (if I got it right) in case at least the second
string was a LTR one. But in case %2$s would be a name written in
Arabic letters, this would make the string being rendered from right to
left, so it would come out as:

<place 1> ← <place 2>

If I'm not completely mistaken.

//Marcus