Re: UNS: Re: pkg/39643 (security/keychain not working properly because of bash style scripting)

[Jochen Keil <Jochen.Keil%gmail.com@localhost>]

The following reply was made to PR pkg/39643; it has been noted by GNATS.

From: Jochen Keil <Jochen.Keil%gmail.com@localhost>
To: Martti Kuparinen <martti.kuparinen%iki.fi@localhost>
Cc: gnats-bugs%NetBSD.org@localhost, martti%NetBSD.org@localhost, 
pkg-manager%NetBSD.org@localhost, 
 pkgsrc-bugs%NetBSD.org@localhost, gnats-admin%NetBSD.org@localhost, 
obache%NetBSD.org@localhost, 
 j.keil%gmx.de@localhost
Subject: Re: UNS: Re: pkg/39643 (security/keychain not working properly because
 of bash style scripting)
Date: Wed, 01 Oct 2008 16:20:10 +0200

 Hello Martti,
 
  > Could you give more details about this because I have no problems with
  > keychain on NetBSD/amd64 4.0_STABLE or NetBSD/amd64 4.99.x.
 
 [
 ok..
 came across the solution now, i made some mistakes  myself i guess:
 i've set ENV=~/.kshrc in .profile having this line in .kshrc:
 eval `keychain -q --eval`
 now everytime sh is started it reads .kshrc (sic) due to $ENV which 
 leads to a call to keychain and so on..
 got to look for a solution now, i could either remove that line (having 
 no keychain available in non-login shells or use something like
 eval `cat ~/.keychain/<hostname>-sh`
 or (simple as it is and my solution of choice): unset ENV in .profile 
 cause ksh will read ~/.kshrc by default.
 ]
 
 Sorry for bugging you but i was honestly confident that it's got 
 something to do with bash cause of my quick fix.
 Anyway, your hint made me dig more into that problem and i even came 
 with the "right" solution.
 
 Thanks and best regards,
 Jochen