Status of internationalization

[Matt Mackall]

On Sat, Jun 11, 2005 at 12:55:21AM +0200, vseguip at gmail.com wrote:
>  Hello, I just started to evaluate mercurial and realized the
> repository web browserr didn't write the encoding in it's page so I 
> added a html charset line (patch attached) to set charset to UTF-8
> (default in my system) to the header template file. This allows me to
> visualize non ansi characters correctly.

Seems reasonable. Anyone have other opinions?

> Some questions on the issue:
> 
> - Does hg use UTF-8 always or is it system specific?

It was designed with UTF-8 in mind. File contents should be completely
8-bit clean, pathnames can contain any characters except nulls, and
most other things are fine with anything except linefeeds.

> - Does mercurial know about charsets at all? Should it know about
> them? (I think the best solution is to just convert the locale charset
> to UTF-8 before storing and convert from UTF-8 to the locale charset
> when retrieving, this way the repository is always portable)

I don't do any conversion at all, currently.

> - Are the basic blocks for i18n/l10n there? If not, what is the
> pythonish way to do this? (just starting to learn python and could
> give some help)

I'm afraid that I'm completely ignorant of i18n issues, Python or
otherwise. But I'm certainly open to learning.

-- 
Mathematics is the supreme nostalgia of our time.