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

Re: mutt/2560: Mutt chokes on invalid charset in UTF environment



The following reply was made to PR mutt/2560; it has been noted by GNATS.

From: Christian Ebert <blacktrash@xxxxxxx>
To: Mutt Developers <mutt-dev@xxxxxxxx>
Cc: bug-any@xxxxxxxxxxxxx
Subject: Re: mutt/2560: Mutt chokes on invalid charset in UTF environment
Date: Wed, 22 Nov 2006 08:28:16 +0000

 * Derek Martin on Wednesday, November 22, 2006 at 00:15:03 +0100:
 > On Tue, Nov 21, 2006 at 08:42:39PM +0100, Vincent Lefevre wrote:
 >> On 2006-11-21 15:05:02 +0100, Christian Ebert wrote:
 >>>  libiconv-1.10
 >> 
 >> You should try 1.11 (this is what I'm using, from MacPorts).
 
 I haven't got the time to do this ATM. It would probably mean to
 switch from fink to MacPorts, because loads of fink packages
 depend on fink's libiconv, or to compile it myself, which would
 mean to compile myself the stuff I have currently installed by
 fink as well etc. etc.
 
 Mind you, I am not complaining, and apologize if what I wrote
 sounded that way. I just want to point out what might happen to a
 naive luser like me.
 
 > Even if that works, it doesn't mean someone won't at a later date use
 > some encoding that somehow breaks iconv...  I missed the beginning of
 > the thread, so I apologize if I don't have a clear understanding of
 > what the submitter is seeing,
 
 Actually you have a very clear understanding of what you missed
 ;)
 
 I am seeing a freeze. The only way out is to Ctrl-Z to force
 quit.
 
 > but it sounds like mutt is silently failing and just does
 > nothing... 
 
 Exactly.
 
 > Doesn't it seem reasonable that mutt should more gracefully
 > handle failed conversions?
 
 You summed up the point I was trying to make.
 
 At the moment the situation from my POV looks like this:
 
 UTF-8 environment + libiconv > 1.10: no prob
 
 UTF-8 environment + libiconv <= 1.10:
 
 1. assumed-charset patch indispensable
    otherwise same symptom on messages with valid charset _names_
    but wrongly declared (bug 2425)
 
 2. mutt chokes on invalid charset name (when iconv fails)
 
 Solutions in order preferred by me ;) :
 
 1. Incorporate assumed-charset patch +
    give a warning when iconv fails, and the possibility to move,
    delete, or call the editor on the message from within Mutt.
 
 2. State in docs/installation instructions that vanilla Mutt
    requires libiconv >= 1.11 to work in an UTF-8 environment.
 
 I can live with the situation as it is, but eg. for someone new
 to Mutt, this might be an offputting experience.
 
 c
 -- 
 _B A U S T E L L E N_ lesen! --->> <http://www.blacktrash.org/baustellen.html>