RE: Patch: do not convert all text parts to UTF-8

[<Dirk-Jan Binnema nokia com>]

Hi Thomas,
 
> -----Original Message-----
> From: ext Thomas Monjalon [mailto:thomas monjalon openwide fr] 
> Sent: Wednesday, August 13, 2008 12:01
> To: tinymail-devel-list gnome org; Binnema Dirk-Jan 
> (Nokia-OCTO/Helsinki)
> Subject: Re: Patch: do not convert all text parts to UTF-8
> 
> Hi,
> 
> Dirk-Jan Binnema nokia com wrote:
> > Hmmmm.... one comment, do we have some example message that fails
> > without this patch?
> 
> Yes.
> HTML emails with charset 8859-1 were converted to UTF-8 and 
> the header still 
> specified the original charset. The the HTML viewer was 
> trying to convert 
> characters from 8859-1 but it was already UTF-8. So my french 
> mails were 
> broken.

Ah... that is indeed broken.

> > And are there mails that worked before, but could break with this?
> 
> Mails that are nor plain text nor html ?
> 
> Parts like diff (text/x-diff) won't be converted and I'm sure 
> it's better.

Well, I was thinking about the case where an HTML mail is in non-UTF8,
but doesn't explicitely set the charset (or sets the wrong one).

Before, the result might not be very nice, but at least after 
UTF8-conversion we're not sending arbitrary bytes to the viewer.

With this change, we're feeding whatever is in the HTML part to the 
viewer, which may not like that very much.

So, while your patch seems like an improvement in many cases (thanks!), 
there is this corner case now. I don't oppose the patch, it's just
something
to keep in mind.

Best wishes,
Dirk.