guix-devel
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [PATCH] gnu: xterm: Accept $SHELL even if not in /etc/shells


From: John Darrington
Subject: Re: [PATCH] gnu: xterm: Accept $SHELL even if not in /etc/shells
Date: Thu, 13 Feb 2014 12:06:52 +0100
User-agent: Mutt/1.5.21 (2010-09-15)

On Thu, Feb 13, 2014 at 03:47:54AM -0500, Mark H Weaver wrote:
     John Darrington <address@hidden> writes:
     
     > On Thu, Feb 13, 2014 at 03:00:29AM -0500, Mark H Weaver wrote:
     >      This patch makes xterm honor $SHELL (or the shell in the user's 
password
     >      entry) even if it's not in /etc/shells.  WDYT?
     >      
     > It sounds like a good idea to me.  /etc/shells is supposed to be only a 
whitelist of
     > those shells which may be used for login.  Not an exhaustive list of 
shells which
     > may be used at all.
     >
     > However, I'm wondering if we are forking too many upstream packages.  We 
should
     > only patch software in order to allow it to build/install.
     
     Really?  Just enough to build/install?  Not enough to work properly?  I
     agree that we should stay as close as we reasonably can to upstream, but
     sometimes things have to be fixed to work with Guix, which after all is
     a rather unusual distro.

     FYI, xterm doesn't merely ignore your $SHELL setting if it's not in
     /etc/shells, it also *sets* $SHELL to "/bin/sh" for you in that case,
     and then proceeds runs it.

Then that to me, sounds like a bug in xterm and can potentially affect many 
OSes not only guix.
     
     IMO, it's not reasonable to have to add
     /home/<USER>/<PROFILE>/bin/<SHELL> for every combination of <USER>,
     <PROFILE>, and <SHELL> to /etc/shells, in order to prevent 'xterm' from
     overriding your $SHELL setting.

I don't disagree.
     
     > If they refuse the patch, then of course you can start your own
     > weavershell fork...
     
     Fork it to change two lines?
     
Two lines today, five tomorrow, twenty next week ...

It is true that Guix is somewhat unusual - and therefore it exposes bugs in 
packages which hitherto have gone unnoticed.  That doesn't change the fact that 
they are bugs in upstream.  Of course it might be difficult getting upstream to
accept a patch but we should try.

I'm just making the point that Guix is not a repository of bug fixes!
     
Just my $0.02

J'


-- 
PGP Public key ID: 1024D/2DE827B3 
fingerprint = 8797 A26D 0854 2EAB 0285  A290 8A67 719C 2DE8 27B3
See http://sks-keyservers.net or any PGP keyserver for public key.




reply via email to

[Prev in Thread] Current Thread [Next in Thread]