Hi
On Mon, Nov 28, 2011 at 8:11 AM, Gabor Szabo <szab...@gmail.com> wrote:
> The spanish translator noticed that some of the strings are
> strangely broken up.
>
> The example was this one:
>
> #: lib/Padre/Wx/ActionLibrary.pm:1021
> msgid ""
> "Change the end of line character of the current document to that used on
> Mac "
> "Classic"
> msgstr ""
>
> In the original perl file it is a single string.
> "Change the end of line character of the current document to that used
> on Mac Classic"
>
> I checked and xgettext, the tool that generates the messages.pot file from
> the
> perl source has a --no-wrap flag.
>
> Using that I get this result:
>
> #: lib/Padre/Wx/ActionLibrary.pm:1021
> msgid "Change the end of line character of the current document to
> that used on Mac Classic"
> msgstr ""
>
> I wonder why none of our translators have complained about this?
> Is this because you are using tools that actually hide this breakage?
>
The tool I am using (OmegaT) handles well these multi-line segments.
But I noticed another issue that IMHO should be fixed:
Are you using the --sort-output flag?
The strings contained in messages.pot are sorted alphabetically, which is
great if you want to avoid duplicates (the only benefit is a smaller
resource file, since most CAT tools will propagate translations anyway),
but it also removes context information. It's better to see all the strings
of a dialog together to disambiguate when needed.
For instance, file VCS.pm contains the following strings:
# Subversion status codes
my %SVN_STATUS = (
' ' => { name => Wx::gettext('Normal') },
'A' => { name => Wx::gettext('Added') },
'D' => { name => Wx::gettext('Deleted') },
'M' => { name => Wx::gettext('Modified') },
'C' => { name => Wx::gettext('Conflicted') },
'I' => { name => Wx::gettext('Ignored') },
'?' => { name => Wx::gettext('Unversioned') },
'!' => { name => Wx::gettext('Missing') },
'~' => { name => Wx::gettext('Obstructed')
Most of these strings can have up to 4 different translations in Spanish,
depending on the gender and number. On the other hand, in different
dialogs, the translation can be different too (i.e., "Modified" could be
"Modificados" in one context and "Modificada" in a different context).
So:
- there is no point in avoiding duplication
- displaying strings "dialog-by-dialog" will help us determine the correct
translation for each case. We can always check the source code, but this
would make our lives easier.
What do you think?
Best,
Enrique
_______________________________________________
Padre-dev mailing list
Padre-dev@perlide.org
http://mail.perlide.org/mailman/listinfo/padre-dev
