Bug#331553: Xsession could set +e when interpreting 3rd party scripts
Package: x11-common
Version: 6.8.2.dfsg.1-7
Severity: wishlist
Following the IRC discussion with Overfiend and later comments from
gravity, here is the complaint written down as bug report:
please consider no longer using "set -e" in the Xsession script or at
least set +e and -e later before/after the scripts in Xsession.d are
interpreted.
Reason: those scripts can be buggy and return with non-zero return
codes. Their maintainers may even be unaware of the outcome because they
are not testing it properly or don't handle all possible problems
because of the "works-for-me" paradigm. Or even worse, they read that
policy paragraph in their own way and interpret "or handle the exit
status" as a real OR (and not every maintainer is a good shell
programmer).
I know this sounds a bit constructed but there are cases where it fails.
And a simple error in some script should not make the whole startup
system fail. Even interpreting the policy as "other should fix their
crap, the bugs are there" is a bad excuse: the idea behind the policy
was IMHO the requirement to write failure-tolerant/correcting scripts,
while Xsession is deliberately designed to fail itself.
Eduard.
-- System Information:
Debian Release: testing/unstable
APT prefers unstable
APT policy: (990, 'unstable'), (500, 'testing'), (500, 'stable')
Architecture: i386 (i686)
Shell: /bin/sh linked to /bin/bash
Kernel: Linux 2.6.12
Locale: LANG=de_DE.UTF-8, LC_CTYPE=de_DE.UTF-8 (charmap=UTF-8)
Versions of packages x11-common depends on:
ii debconf [debconf-2.0] 1.4.58 Debian configuration management sy
ii debianutils 2.14.3 Miscellaneous utilities specific t
ii lsb-base 3.0-9 Linux Standard Base 3.0 init scrip
x11-common recommends no packages.
-- no debconf information
Reply to: