[1974] in athena10
Re: [Debathena] #170: move elinks diversion to debathena-locker
daemon@ATHENA.MIT.EDU (Debathena Trac)
Tue Mar 31 15:27:13 2009
MIME-Version: 1.0
Content-Type: text/plain; charset="utf-8"
From: "Debathena Trac" <debathena@MIT.EDU>
Cc: debathena@mit.edu
To: broder@mit.edu, andersk@mit.edu
Date: Tue, 31 Mar 2009 19:25:15 -0000
Reply-To:
Message-ID: <052.59fe4e1220516cef4afa8e895df4dff7@mit.edu>
In-Reply-To: <043.a27c3625ddab579d7337cf51fefbc147@mit.edu>
Content-Transfer-Encoding: 8bit
#170: move elinks diversion to debathena-locker---------------------+------------------------------------------------------
Reporter: broder | Owner:
Type: defect | Status: new
Priority: major | Component: default
Resolution: | Keywords:
---------------------+------------------------------------------------------
Comment(by andersk):
{{{
debathena / testers / andersk 15:16 (Anders Kaseorg)
Uh, it is extremely dangerous to divert a binary without depending
on
it, because it breaks programs that test for the existence of that
binary.
debathena / testers / broder 15:18 (Evan Broder)
Huh?
debathena / testers / geofft 15:18 (forbidden-config-interpreter)
A stat will fail... are there really things that will lstat elinks?
debathena / testers / andersk 15:18 (Anders Kaseorg)
No, that’s not the point. The binary really _does_ exist, because
it
is now symlinked to the wrapper script.
debathena / testers / geofft 15:19 (select-without-choices)
Oh, right
debathena / testers / andersk 15:19 (Anders Kaseorg)
But when you run the wrapper script it will complain about the
missing .debathena-orig
debathena / testers / andersk 15:20 (Anders Kaseorg)
I think you guys are becoming way too comfortable to throw up
config-package-dev diversions on users’ machines without carefully
considering the consequences. c-p-d is not a magical black box
that
solves all your problems.
debathena / testers / andersk 15:20 (Anders Kaseorg)
We tried to make a good abstraction, but we simply don’t have the
tools available to prevent it from leaking in lots of ways.
debathena / testers / andersk 15:24 (Anders Kaseorg)
Also, this is not a -locker problem. This is only a problem with
AFS
_home_ directories, not with the existence lockers.
}}}
-- Ticket URL: <http://debathena.mit.edu/trac/ticket/170#comment:1>Debathena <http://debathena.mit.edu/>MIT Debian-Athena Project
