At lines 110 of the startup script (installed with 1.6.2.13-r2) contains this code: else set -o pipefail ${NICE} /usr/sbin/asterisk ${OPTS} 2>&1 >/dev/null | logger -t asterisk_wrapper set +o pipefail result=$? fi The resetting of pipefail before grabbing the result in this case is causing the problem and set +o pipefail and result=$? just needs to be swapped about. Reproducible: Always
+*asterisk-1.6.2.14 (02 Dec 2010) + + 02 Dec 2010; <chainsaw@gentoo.org> +files/1.6.2/asterisk.initd3, + -files/1.6.2/asterisk.logrotate, -files/1.6.2/asterisk.logrotate2, + -asterisk-1.6.2.13.ebuild, -asterisk-1.6.2.13-r1.ebuild, + +asterisk-1.6.2.14.ebuild: + Consolidate patchset in a tarball on the mirrors, closes bug #344015 by + Marcin Baczyński. Drop logrotate USE-flag, always install snippet, closes + bug #344181 by Jeroen Roovers. Updated init script to fix restart bug by Jaco + Kroon, closes bug #345307. Version bump closes bug #346737 by Jesse Adelman, + with an honourable mention for Kerin Millar for pointing out some QA matters + which are now fixed. I also snuck this one into Kerin's 1.4.37 ebuild, that leaves just 1.8.0 for later.
This bug is not resolved in the current ebuilds (1.6.2.18.2,1.6.2.19,1.8.4.4). The version of /etc/init.d/asterisk with the fix is not used, but an old version that still has the error. The ebuilds are using the file net-misc/asterisk/files/asterisk.initd2 but the fix is in net-misc/asterisk/asterisk.initd3
Reopening, as per Erik Wallin this is not fixed.
+*asterisk-1.8.4.4-r1 (04 Jul 2011) +*asterisk-1.6.2.19-r1 (04 Jul 2011) + + 04 Jul 2011; Tony Vroon <chainsaw@gentoo.org> -asterisk-1.6.2.19.ebuild, + +asterisk-1.6.2.19-r1.ebuild, -asterisk-1.8.4.4.ebuild, + +asterisk-1.8.4.4-r1.ebuild: + Actually install the newest init script with Jaco Kroon's restart fixes; as + pointed out by Erik Wallin (reopened) bug #345307. Unfortunately I am not able to modify 1.6.2.18.2 at this time; it was fast-tracked to stable for a security bug.