ports/173366: FIB in Apache RC script vs. rc.subr
Stefan Esser
se at FreeBSD.org
Sun Nov 4 21:40:01 UTC 2012
>Number: 173366
>Category: ports
>Synopsis: FIB in Apache RC script vs. rc.subr
>Confidential: no
>Severity: non-critical
>Priority: low
>Responsible: freebsd-ports-bugs
>State: open
>Quarter:
>Keywords:
>Date-Required:
>Class: sw-bug
>Submitter-Id: current-users
>Arrival-Date: Sun Nov 04 21:40:00 UTC 2012
>Closed-Date:
>Last-Modified:
>Originator: Stefan Esser
>Release: -CURRENT
>Organization:
FreeBSD.Org
>Environment:
FreeBSD x.y.z 10.0-CURRENT FreeBSD 10.0-CURRENT #24 r242481M: Fri Nov 2 17:23:22 CET 2012 root at x.y.z:/sys/amd64/compile/SE amd64
>Description:
r242184 added setfib(1) support to rc.subr and assumes, that any non-blank value is a FIB number.
But the RC script for apache sets "apache_fib" to "NO" if it does
not contain a FIB number.
This leads to a comamnd "setfib -F NO apache" and setfib complains
about the non-zero value.
I assume that the apache RC script should be modified to not pass
apache22_fib="NO" for the case of no FIB. OTOH, the code in rc.subr
could be changed to accept "NO" as equivalent to an empty value of
the FIB variable.
Setting "apache22_fib=0" just hides the problem. It does not harm,
but the default of not specifying a FIB should just work.
>How-To-Repeat:
On a -CURRENT system with Apache try to start it wothout setting a FIB:
# service apache22 start
>Fix:
The following patch lets rc.subr accept "NO" as equivalent to an empty xxx_fib variable. This patch is sufficient, but does not try to detect any permitted spelling of "NO" (vs. [Nn][Oo]).
--- /usr/src/etc/rc.subr 2012-10-27 23:15:46.554124610 +0200
+++ /etc/rc.subr 2012-11-04 21:46:29.109476560 +0100
@@ -645,6 +645,10 @@
_group=\$${name}_group _groups=\$${name}_groups \
_fib=\$${name}_fib
+ if [ "x$_fib" = "xNO" ]; then
+ _fib=
+ fi
+
if [ -n "$_user" ]; then # unset $_user if running as that user
if [ "$_user" = "$(eval $IDCMD)" ]; then
unset _user
>Release-Note:
>Audit-Trail:
>Unformatted:
More information about the freebsd-ports-bugs
mailing list