Re: string-freeze exception request: add translated strings to GOK

[Bill Haneman <Bill Haneman Sun COM>]

On Wed, 2003-10-01 at 20:26, Christian Rose wrote:
> ons 2003-10-01 klockan 15.01 skrev Bill Haneman:
> > We'd like to add some translatable strings to GOK.  The additions fall
> > into two groups:
> > 
> > (1) mark g_warning strings for translation: we may be able to reduce
> > this to a single string and re-use, but the patch below uses two
> > different strings (more informative).
> 
> My personal opinion is that marking these g_warnings for translation
> could wait until after you've branched,

Sure, no problem.  I just don't want them to get forgotten :-)

> > (2) add strings to support internationalization of some types of GOK
> > keyboards which cannot currently be internationalized.  This is in
> > service of gok i18n bugs 122117 and 122112.
> >
> > The strings in #2 won't be exposed to users until 2.6 unless the users
> > customize their 'Compose' keyboards.  These strings are part of a set of
> > fixes to improve GOK's internationalization generally (see bug #107200
> > which these strings address, and GOK internationalization tracker bug
> > #122112).
> 
> This seems like a nasty fix to add on the stable branch -- it's not
> trivial and it seems it might open a whole can of bugs with issues that
> haven't been considered once we actually get some translations of these
> layouts, and thus it seems inappropriate to add it on the stable branch.

Perhaps.  Again, it depends on whether we want users to be able to use
GOK "alphabetical keyboards" in non-English locales in the 2.4
timeframe.  But no, this is not a regression.


> Why not an uppercase initial letter like in the other warnings?

Just a typo introduced in the revert/re-create patch flurry.

> 
> > +				g_warning (_("Your keyboard geometry cannot be determined"));
> [...]
> > +		g_warning (_("Your keyboard geometry cannot be determined!\n"));
> 
> I don't see why these messages need to be different.
Hmm, ok, the difference is meaningful for diagnostic purposed (one is
failure of the XKB extension generally, one is failure of a specific XKB
API).
> > -	/* TODO: these strings are useless without localization, but string freeze prevents inclusion */
> >  	gok_keyboard_add_keys_from_charstrings (pKeyboard, 
> > -						"level 0|abcdefghijklmnopqrstuvwxyz1234567890-=[];'#\\,./",
...

> This absolutely need a comment for translators(*), so that they know
> what on earth this is about and how it should be "translated"

There's a long note to the translators in the preceding comment.  Isn't
that supposed to get merged into the po files by intltool? i.e. here is
the comment from the code (sorry about the wrapped lines):

       /*
         * Note to Translators: the following strings should contain
your LANG/locale's
         * alphabet or, in the case of LANGs with a very large glyph
set, a set of
         * character-primitives which can be used to compose your
language's character set.
         * Each string below corresponds to the characters associated
with a particular
         * "shift level" in the XKB keyboard definition.
         * The prefix before the '|' character is just a context string
and need not be
         * translated.
         *
         * For languages where 'case' is used, the first string should
contain
         * the lowercase alphabet, the second string the uppercase
alphabet.
         * Note that unless your locale clearly requires that digits
and/or punctuation
         * precede alphabetic characters, we prefer that digits and
punctuation are placed
         * at the end of the string for GOK's purposes.
         * The third string is not used in the C locale but corresponds
to a shift
         * level of '2' which is often associated with the AltGr key on
physical keyboards.
         *
         * The order in which punctuation occurs is not critical, but if
there is a
         * "standard" layout for your locale's physical keyboards it is
nice to place
         * the punctuation in the order in which it appears on physical
keycaps.
         * For instance, since Shift-plus-digits for "1,2,3,4" on a US
keyboard
         * produce "!, \", 3, $", etc., it is convenient to list these
punctuation marks
         * in the same order in the "level 1" string as we do below for
the C locale.
         *
         * For languages
         * which use shift or level modifiers to select between
different glyphs, the meaning
         * of 'level' above, and as interpreted by the X server, may be
slightly different.
         * If you have questions or comments about the intent of this
string please contact
         * billh gnome org or david bolter utoronto ca for more
information.
         *
         */

Is there something else I need to do to get this comment pulled into the
PO file?

There's also a note in README-translators I believe, or one could
certainly be added.
 
> But to repeat myself again, I don't see any strong reason this needs to
> be added to the stable branch.

When will the GTP folks be looking at 2.5?  One motivation was to get
the string exposed to the translators ASAP.

- Bill
 
> (*)
> http://developer.gnome.org/doc/tutorials/gnome-i18n/developer.html#use-comments
I studied this before making the patch.  It doesn't explain how and
where the comments are extracted however, I had to get that from grep
and CVS.

> 
> 
> Christian
> 
>