[Freeciv-Dev] (PR#15751) Original string bugs/typos, batch

["Daniel Markstedt" <himasaram@xxxxxxxx>]

<URL: http://bugs.freeciv.org/Ticket/Display.html?id=15751 >

> [sini.ruohomaa@xxxxxxxxxxxxxx - Mon Mar 06 14:31:34 2006]:
> 
> Hi,
> 
> Thanks for the reply, Daniel!
> 
> On Sun, 2006-03-05 at 15:15 -0800, Daniel Markstedt wrote:
> > On 2006-03-03, at 06.53, sini.ruohomaa@xxxxxx wrote:
> > > server/cityhand.c:387 and others suggest translating string
> > > "%s". 
> > I just suppose these kind of strings never have to be altered by 
> > translators. Such as the name of a nation etc.
> 
> Hmm. I had a warning light on but was confused about what it warned
> about. The important point I was missing is that a lone %s shouldn't
> just blindly be marked untranslateable again. Check if it's used wrong
> first.
> 

It definitley needs a TRANS comment, yes, as do all similar messages
really. In this case I couldn't even figure out what %s is supposed to
be by looking at the code.

> If the %s is used only for escaping a message string not under our
> control at all, like in Jason's example, I think it can't and should not
> be touched by translators - any repetitive markup for "incoming message"
> and so on should be in different translateable strings. It just looks
> like a new string, but isn't.
> 
> However, the fact that the string was marked translateable may mean that
> a string translated for one context is used in a new context, which is
> worse. The helpful coder may think: "If it's different for some
> language, they can modify the surrounding string." Of course oftentimes
> that's not enough to make "I had a thought" and "I thought" to work with
> the same "thought". Or fix the problem with "the English" used both as a
> nation name and an adjective. ;) 
> 

Haha. Yes this is a significant problem for Swedish as well (perhaps not
as much as for Finnish though. ;) )

There have been attempts on a solution that would give translators more
control of grammatical forms. See for example these discussions:

http://lists.complete.org/freeciv-dev@xxxxxxxxxxx/2000/08/msg00493.html.gz
http://lists.complete.org/freeciv-dev@xxxxxxxxxxx/2000/09/msg00018.html.gz


> > > server/citytools.c:894 seems to have unnecessary indenting left over
> > > from having "Game: " in the start. Looks funky.
> > The indenting is AFAICT intentional. Also for example:
> > #: server/plrhand.c:1684
> > #, c-format
> > msgid ""
> > "The capture of %s's capital and the destruction of the empire's "
> > "administrative\n"
> > "      structures have sparked a civil war.  Opportunists have flocked 
> > to the "
> > "rebel cause,\n"
> > "      and the upstart %s now holds power in %d rebel provinces."
> 
> I'm not sure if it's consistent in any way. To me it would be
> considerably nicer if cosmetic indentations and linebreaking were done
> by code instead of people, and if there's some silly reason that we need
> to get something to fit a tight space, it should be commented. I can't
> see from all server strings if they're supposed to be indented the same
> way, if the English version fits on a single line. (The Finnish often
> ends up longer, and if the English uses 80% of the first line I won't
> see if it has its second line indented. ;))
> 

I think everyone agree that the current solution here isn't ideal. These
indentions are still very rare though; AFAICT not much more than these
two examples.

> > > server/citytools.c:1017 string should probably end in "."
> > >
> > > server/citytools.c:1854 Don't put the space in the translateable
> > > substring, put it between the %s%s so we can do punctuation where it
> > > should be.
> > I didn't follow you here. :-)
> 
> I assume the first one is clear?
> In that particular case I think the string was 
> "Cannot complete <building>< from the worklist>"
> If the option for < from the worklist> is another explanation, it's
> nicer to have the space in the complete sentence:
> "Cannot complete <building> <from the worklist>"
> However, if you presume to put the empty string <> as the latter %s
> sometimes, I can see why you'd put them together that way (to avoid the
> extra space). In that case it's unfortunate but fine. ;)
> 

Ok, I follow you now. :) I can't see either why this "funky spacing"
should be. Fixed in updated patch.

> This kind of funky spacing and other places where there are no hint
> words between different strings like in "%s of the %s" are that
> magically lovely place for translator comments. Translating 3000 strings
> makes you very soon prefer those ones which won't force you to go check
> in code what each %s really stands for. We can't rely on the
> translateable sub-message being next to the full message either in the
> po file to give the context, as sometimes the strings get rather mixed
> for reasons I don't quite understand yet.
> 
> Cheers,
> 
> --Sini
> 

The translatable strings needs an overhaul, agreed. Especially since
we're trying to "win back" lots of translators for abandoned l10ns. IMHO
we should even concider removing a signficiant number of error strings
as tranlatable, to 'ease' the work for returning translators. Error
messages are seldom more helpful translated and may even make it harder
for users to get assistance when they either have to re-translate to
English or find a developer/advanced user who speaks their language.


Regards,
-Daniel

strfixes2.diff
Description: Binary data