[GRASS5] Re: [GRASS translations] Not yet gettexted?

(cc grass5)

On Thu, Feb 24, 2005 at 12:37:57AM -0500, Daniel Calvelo Aros wrote:

Hi all.

I did a quick find'n'grep over the beta2 sources, and it looks like there are
still several strings not marked for translation. I guess it's not an issue
that should be cleaned up for 6.0 final

Hi Daniel,

this is a known problem (see locale/README and the i18n web page).

, but I think it ought to be included
in the TODO. Maybe also scheduled for some point release.

Basically it's a lot of work, help is more than welcome.

It might also be a
good opportunity to write the paragraph on fprintf vs. G_message in SUBMITTING.

Good point.
Someone knowing better than me may give us a hint when to use G_message().

FYI, a "find . -name '*.c' -exec grep -HE 'G_(message|warning|fatal_error)'
\{\} \; | grep -v '_("' | cut -d: -f1 | sort |uniq | wc" gives 595 files with
unmarked G_* calls; the equivalent count for fprintf\(std(out|err) gives 332
files with strings not marked for translation.

Daniel.

-- Daniel Calvelo Aros

Markus

On Mon, Feb 28, 2005 at 12:37:52PM -0500, Daniel Calvelo wrote:

Markus Neteler <neteler@itc.it> on 28 Feb 2005 15:22:00 +0100 wrote:
>
> Basically it's a lot of work, help is more than welcome.

Ok. Have you got a rough idea of the priorities for translation? I reckon d.m,
g.*, r.* are more used than the rest of commands, but people with more grass
experience might have other ideas.

In a long morning I have indicated all r.* descriptions (> 1000).
All d.*, g.* etc are missing, many v.* as well. Just go ahead!

[me]
> It might also be a good opportunity to write the paragraph on fprintf vs.
> G_message in SUBMITTING.

[Markus]
> Good point.
> Someone knowing better than me may give us a hint when to use
> G_message().

locale/README states that both fprintf(stderr...) and fprintf(stdout...)
should be changed to G_message. I find that quite surprising: it seems to
erase the distinction between "normal" output and "warning-like" output.
Shouldn't fprintf(stderr...) become G_warning()?

You are right, it's badly explained.

Radim: what's the policy for G_message?

Markus

Markus Neteler wrote:

Good point. Someone knowing better than me may give us a hint when to use G_message().

locale/README states that both fprintf(stderr...) and fprintf(stdout...)
should be changed to G_message. I find that quite surprising: it seems to
erase the distinction between "normal" output and "warning-like" output.
Shouldn't fprintf(stderr...) become G_warning()?

You are right, it's badly explained.

Radim: what's the policy for G_message?

I think that G_message should be used for messages - informations about the process for user while fprintf(stdout...) for data output.
G_message output is not expected to be send to pipe or file. fprintf(stdout...) output is usually send to pipe or file.

It was discussed more times in the list, search the list.

Radim