Re: lynx-dev internationalization

[Nelson Henry Eric]

> again - that's what I'm doing. but there's a lot of strings that
> aren't #define'd in LYMessages_en.h -- JES's version finds many of
> them, and I'm adding a few, both to LYMessages_en.h and to the inline
> stuff (if I stop to #define every string I see, it'll take a week).

I see no problem with taking a week or a month.  Don't want to have to
keep coming back to it.

> Once done, it'll be simple to grep the result for 'gettext' and think
> about how best to proceed.

Rather than grep, it might be better to use the tools that the
translators will be using.  I'd recommend making your own lynx.pot file
because it will give you information that grep wouldn't.

Once you have the gettext package installed, it's pretty straight
forward.  I'm sure you can do it with more savvy, but simply something
like:

% ls ./WWW/Library/Implementation/*.[ch] ./src/*.[ch] > z-ls xgettext -f
% z-ls -o lynx.pot`date '+%m%d'`

will make a fresh *.pot file, tell you where there are errors, and tell
you where each string is used.  You might want the *.pot file around for
debugging, in addition to seeing how the thing works.

Just thoughts.

__Henry