[SLUG] Irritating code instead of normal characters

[Al Girling]

On Fri, Aug 06, 2004 at 10:21:14PM BST, Paul Teasdale wrote:
[...]
> Basically it may be Mutt that has started the issue you mention in your 
> original e-mail (irritating chars) and Fintan may be right in that it's a 
> Mutt thing. 
> 
> Here is what I have done. Use Mutt to send a normal e-mail to a user on my 
> network. Mutt chooses by default 'us-ascii' which is 7-bit so long as I don't 
> type any extended characters like the £ sign in the message (BTW if I do then 
> Mutt chooses an 8 bit 'unknown' character set - so it claims? I don't know 
> what the 'unknown' is about because I don't use Mutt normally).
> 
> Then I picked up the e-mail with KMail and added a reply to the message with 
> some £ signs in etc. All the time I was ensuring that KMail was set to reply 
> in the same character set the original e-mail was received in (ie: us-ascii).
> 
> Then I went back to Mutt and picked up the reply e-mail as sent by KMail. Low 
> and behold the pound signs did not show _however_ I got question marks 
> instead of the £ signs and not the =A3 business that you get. It maybe the 
> case that Outlook does something else with the mail hence maybe =A3 instead 
> of ?. I may try this at some point :)
> 
> What this does prove however is that if you send an e-mail in a specific 
> format and the recipient's e-mail client keeps that format and also the 
> recipient replies with some characters not supported by the original format 
> then they get replaced by 'irritating characters'.
> 
> So the long and short of it is that Mutt may have started the chain of events 
> that lead to the problem you mention.

OK.  I've had another think about this and done some more reading and
experimenting.

It seems to me that Mutt is using and defining the minimum content type
and transfer-encoding for the messages it creates.  However, when it
receives a message in say us-ascii/7bit encoding and you add characters
like £ or other more exotic ones like ¤ or ¥ Mutt alters the content and
format type appropriately.  Again this seems to be reasonable behaviour
as this means that if I add these characters then a mail client
receiving my message gets correct information about the sort of
characters it will be reading.

The only thing I've been unable to do is make Mutt replicate your "Mutt
chooses an 8 bit 'unknown' character set".  As here on this machine I
see;

Content-Type: text/plain; charset=iso-8859-1
Content-Transfer-Encoding: 8bit

for mail containing £ symbols.

I've also tried sending a message locally without extended characters
which is shown as us-ascii.  Replying to it adding the £ symbol which
changes it to iso-8859-1.  Replying to that message having removed the £
symbols which changes it back to us-ascii.

So.  I've commented out my latest additions regarding charset settings
in my .muttrc as I now think it was working correctly after all.

Toodle pip,

Al
-- 
Al Girling
Linux User: #290080            <http://counter.li.org>
Scarborough Linux User Group   <http://www.scarborough.lug.org.uk>