Bug#207455: acknowledged by developer (Re: Bug#207455: packages.debian.org: HTML-encodes multi-byte characters as single bytes)

[Frank Lichtenheld <djpig@djpig.de>]

On Sat, Aug 30, 2003 at 03:23:27PM +0200, Josip Rodin wrote:
> On Sat, Aug 30, 2003 at 01:09:02PM +0200, Frank Lichtenheld wrote:
> > tags 207455 pending
> > thanks
> > 
> > In my current packages.d.o version, all Description: fields are 
> > handled as utf-8 (ddtp uses it and IMHO english descriptions should be
> > ascii).
> > 
> > Objections welcome.
> 
> I suppose you don't have the existing broken Latin1-using descriptions
> (if there are still any, there were some) break even further, do you?

[btw: sorry for my absence for at least the last week. But RL needed my
attention :)]

Hmm. I have a few possible strategies to handle descriptions (English
ones, the one from DDTP are all UTF-8):

1. Take them literaly and specify a charset=ascii for the page
2. Dito, but charset=iso-8859-1
3. Dito, but charset=utf-8
4. Use one of the three charsets but make a list of broken descriptions
that have to be converted

Currently we do (2) but I would prefer to go to (3). As long as policy
doesn't mandate one encoding for the description it's our decision
anyway and I would prefer to give everyone the same chance to break
something ;)

Gruesse,
-- 
Frank Lichtenheld <frank@lichtenheld.de>
www: http://www.djpig.de/