Re: doc-i18n-tool

[Jonathan Blandford <jrb redhat com>]

> > Yeah, this is something that we'll definately want to work on.  Does
> > anybody know if there's a systematic way to determine if these things
> > are "inline" versus "block level"?  (hint, hint, Dave?)
> 
>   Basically it could boil down to a list of elements whose text nodes
> don't need to be converted, in that case it's easy to exclude them.

Doc guys?  Can we get a list of these elements?

> > > Comment?
> > 
> > I just nagged a few people from #i18n into playing with this tonight,
> > and, at least on my test document, the .pot file isn't good.  Here's a
> > wee bit from it:
> > 
> > #: policies.sgml:4
> > msgid "\n"
> > "  "
> > msgstr ""
> > 
> > I poked Daniel about this on IRC (although as it turns out, I'm almost
> > too sleepy to ask constructive questions there), and he suggested that
> > "maybe the parser needs to be instructed to ignore whitespace".  I'll
> > stop right here, before I ramble too much further, and hopefully
> > somebody can take a look at this today.  Later,
> 
>   Hum, I will let Jonathan check this one.

My plan was to use xmllint for this.  In my example, I ran the doc
through xmllint --noblanks before passing it to doc-i18n-tool.  I don't
think doing anything else is worth doing.

-Jonathan