Alan Shutko wrote:
> Will Yardley <[EMAIL PROTECTED]> writes:
 
>> The patch (to init.c) seems to be:
>>
>>    Tempdir = safe_strdup ((p = getenv ("TMPDIR")) ? p : "/tmp");
>> -  Editor = safe_strdup ((p = getenv ("EDITOR")) ? p : "vi");
>> +  Editor = safe_strdup ((p = getenv ("EDITOR")) ? p : "/usr/bin/editor");
>>    Visual = safe_strdup ((p = getenv ("VISUAL")) ? p : Editor);
> 
> Either I'm blind, or that should obey VISUAL if set, otherwise
> EDITOR, otherwise /usr/bin/editor....
> 
> [installs mutt, tries
> EDITOR="echo EDITOR" VISUAL="echo VISUAL" mutt [EMAIL PROTECTED]
> ]
 
> Hmm... the Debian unstable package seems to work as expected for me.
> Sure you don't have something weird in your environment or something?

As I indicated in the original message, I'm using Debian 3.0 (stable).
 
Marco d'Itri wrote:
> On Mar 19, Will Yardley <[EMAIL PROTECTED]> wrote:
 
> News->mail gating for linux.debian.user will be enabled right now.]

Sweet. Thanks.
 
>> Yes, but mutt (normally) obeys VISUAL if present - it's only the Debian
>> package which seems not to.

> Maybe you think this because the debian package is a 1.5 snaphost, and
> handling of $EDITOR and $VISUAL has changed since 1.4 (something else
> needed to be fixed, maybe this broken something too...).
> 
> I will *not* further discuss this with the mutt upstream unless
> somebody will provide reference to authoritative documentation about
> the correct semantics of $EDITOR and $VISUAL (I could not find any)
> and a patch implementing this.

The issue doesn't seem to be with stock mutt - only the Debian package.
Also, this isn't with 1.5.x, but with 1.3.28 (in -stable).
 
>> +  Editor = safe_strdup ((p = getenv ("EDITOR")) ? p : "/usr/bin/editor");
>>    Visual = safe_strdup ((p = getenv ("VISUAL")) ? p : Editor);
>> 
>> I'd think that VISUAL (if present) should override /usr/bin/editor (if
>> present), however this seems not to be the case.

> Read again the code: it does.
> If you think that some parts of mutt use Editor while it should use
> Visual instead please provide a patch with plenty of justifications
> (I do not like changing the default hardcoded behaviour of a package
> like mutt).

Ahh - I see - my other tests were done with 1.5.3 (built from source) -
I didn't realize that the behavior had changed since 1.3/1.4 branch.
This is probably the reason for my confusion.


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Reply via email to