Re: ommiting header causes multibyte errors

[Ariel]

On Thu, 7 Jun 2007, Bruno Haible wrote:

> > > This is normal. A POT or PO file that does not carry a character encoding
> > > specification in the header entry is assumed to be in ASCII. xgettext notices
> > > that its output would violate this rule and gives a warning and an error.

> > In that case I would have appreciated this being documented near the
> > option for it, it would have saved me quite a bit of time.
>
> OK, I'll mention this in the doc:

> -Don't write header with @samp{msgid ""} entry.
> +Don't write header with @samp{msgid ""} entry.  Note that using this option
> +will lead to an error if the resulting file would not entirely be in ASCII.

Thanks.

> > Actually, perhaps ever better: generate a minimal header whenever the
> > source is non-ascii, and the option is on, since in that case the header
> > is required rather then containing optional data.

> This is not really better, because if xgettext produces a minimal header
> despite of --omit-header, it is most likely not what the user wanted.

Maybe - after all why is the user asking for no header? I can think of two 
reasons, mine (not to have anything extra), in which case a minimal header 
with the encoding etc (but only if it's required), is fine - in fact I 
prefer it.

And for the reason of no variance for testing, in which case it's still 
fine.

I can see no reason why someone would want to use --omit-header with a 
non-ascii file, making it a never-used option, so giving the option 
something to do would not be causing any harm that I can see.

But, maybe rename it to --minimal-header

(And having --omit-header have the side effect of converting to ascii 
just seems wrong, especially when there isn't an official option to do 
that in xgettext.)

> > Also, aesthetically, it looks bad to have a field which is empty, but looks
> > like it should have something.

> When the format of PO files was specified 12 years ago, aesthetics did play
> a role (which is why breaking long strings into lines is possible), but
>  msgid ""
> was not considered ugly.

It's not the msgid "", it's the empty: "Report-Msgid-Bugs-To: \n".

> > I just don't want the extra fields to appear in the file. I need to send
> > it to someone very non-technical and I want the absolute minimum in the
> > file.
>
> Can you tell more: What is the non-technical person going to do with the
> file? Reading it? Reviewing it? Editing it? What kind of tool is he/she using
> for this purpose?

Editing it (adding the translations), and not with any tool written for 
.po files, just a text editor - probably notepad on windows, since it can 
handle UTF-8, and I know they have it.

And yes, I'm worried about them messing up the format, but I'll probably 
be able to fix it if that happens.

> I'm asking because it sounds like some other format might be better suited
> than the PO file format.

I didn't know there were any. (Besides .mo)

        -Ariel