<<< Date Index >>>     <<< Thread Index >>>

Re: Charset issue?



On Thu, 10 May 2007 or thereabouts, Alain Bench came forth with:

>  On Friday, May 4, 2007 at 10:44:00 +1200, Roland Hill wrote:

> > I did this [export LANG=en_NZ] and the 'garbled' message is still
> > garbled (/264 etc)

>     So this can mean either that the en_NZ locale is broken or not
> installed (it must appear in "locale -a" list), or that Kyle is right
> (mislabelled mail). Could you please send me gzipped attached one
> iso-8859-1 mail, and one utf-8 from the ahum... garbling sender? ;-)


> > Kyle, I tried message-hooks as you suggested, 2 actually to deal with
> > utf-8 -> us-ascii and iso-8859-1 -> us-ascii, and all strange
> > characters are now question marks.

>     Those characters are all above 128 (\200 octal), and as such are not
> part of US-Ascii. You forced them to be treated as US-Ascii, so Mutt
> considers them invalid, and ?-masks them. That's of course not the
> wanted result, and is not what Kyle suggested. He suggested to alias
> utf-8 -> Latin-1, in the hope the content is really Latin-1 under the
> utf-8 label. The said characters do exist in Latin-1 (as in won´t and
> ¨reply¨).

Hi Alain,

My guess is that the messages are labelled incorrectly, as, after much
review, and can read all 3 charset's correctly when sent from 99% of
people. As this particular sender sometimes sends in us-ascii, I thought
maybe that was the 'real' charset of the 'bad mails'. I was playing - I
don't know this subject well.

File is attached as requested. 1 'bad' message in Latin-1 and 1 in utf-8.

Thanks for your assistance with this.

-- 
Regards,

Roland

PGP Key 0xDA39319B = BCF0 1214 BAE9 5A3D 46FC 21A6 360D 9398 DA39 319B

Attachment: badmail.gz
Description: GNU Zip compressed data

Attachment: signature.asc
Description: Digital signature