[Po4a-devel] Markup text alignment interoperability problem between versions 0.42 and 0.45 (debian)

[D. Barbier]

[Cc'ing Chals; whole message quoted so that Chals can read it, please
drop useless parts when replying]

On 2014-05-18 10:22 GMT+02:00 Martin Quinson wrote:
> Hello
>
> On Sat, May 17, 2014 at 10:26:05PM +0200, chals wrote:
>> On Sat, May 17, 2014 at 9:38 PM, D. Barbier <bouzim at gmail.com> wrote:
>>
>> > You are right, this is very annoying.  But this change comes from
>> > gettext (more precisely msgmerge), not po4a, and I do not know how to
>> > avoid it.
>> >
>>
>> Yeah, "annoying" is the exact right word to describe it. But I was
>> afraid to use it because I didn't want to be annoying myself.
>>
>> Thank you very much for the information. I somehow suspected it could
>> be an upstream thing. Your clarification is very valuable.
>>
>> Now I only wish there was a way to coordinate the versions of all the
>> contributors to the project, but it is easier said than done, not to
>> say impossible. Anyway, I am very happy to learn one more thing today.
>
> You may want to open a bug (or better, provide a patch) against the
> msgmerge program (package: gettext) so that the column at which it
> justifies the text can be configured from the command line. With that
> feature in msgmerge, we could do what you want in po4a.
>
> Thanks for the report, and your interest in po4a.
> Mt.

Hello Martin,

msmerge already provides the -w flag, but IIRC its behavior changed
with respect to non-alphanumeric characters.  See original report: on
an old system:
 "_* Zmienne dotyczące parametrów startowych live-config zaczynają się od #"
 "{LIVE_}#."
whereas a more recent system outputs:
 "_* Zmienne dotyczące parametrów startowych live-config zaczynają się od "
 "#{LIVE_}#."

It looks like msgmerge used to split words containing non-alphanumeric
characters, but now prefers not splitting.  This is just a wild guess,
this should be confirmed by testing.

Denis