Re: Unicode typography in translations

[Ask Hjorth Larsen <asklarsen gmail com>]

2016-11-16 13:43 GMT+01:00 Rafael Fontenelle <rafaelff gnome org>:
> 2016-11-15 10:21 GMT-02:00 Ask Hjorth Larsen <asklarsen gmail com>:
> > Hi Rafael
> >
> > 2016-11-15 2:26 GMT+01:00 Rafael Fontenelle <rafaelff gnome org>:
> >
> > > It would be nice to have a script with regexp that could compare msgstr
> > > and
> > > msgid in a PO file, and report strings that are not in compliance with
> > > GNOME's HIG typography.  I don't have such scripting skill, but if
> > > someone
> > > has it, please consider do it.
> > >
> > > Regards,
> > > Rafael Fontenelle
> >
> > It is easy to recognize when the English string contains something,
> > and the translated string does not (e.g. to find a unicode ellipsis
> > that was translated to an ASCII ellipsis).  But if the English string
> > uses ASCII, it is not always easy.  For example recognizing exactly
> > when the en-dash could or should be used instead of an ASCII hyphen.
> >
> > It is probably a good assumption that any sequence of exactly three
> > dots should be a unicode ellipsis, no matter the context, but that's
> > the only trivial case.
> >
> > Best regards
> > Ask
>
>
> I agree that is not hard to recognize them while translating a PO file. I
> just would like to have a solution that allows a conformity checking that
> could be run anytime, as such Unicoded characters could be missed by the
> translator (myself included).
>
>
> Rafael Fontenelle

A simple conformity check:

  gtgrep -cn --msgstr '\.\.\.'   filename.po

I had a bit of a battle to get the regex escapes right in bash, but
this should weed out most false positives:

  gtgrep -cn --msgstr '(?<!\.)\.\.\.(?!\.)'   filename.po

However if the translation uses "...." or ".." when it should really
be using an ellipsis, the first form is better anyway, or maybe
(?<!\.)\.{2,4}(?!\.).  Probably it's best just to use the simple one
:)

(gtgrep comes from pyg3t)

Something similar could be done for the other characters, but of
course quotation marks vary a lot depending on language, so it would
not all be completely portable to all languages.

Best regards
Ask