[Pkg-sysvinit-devel] Re: Fixing hwclock.sh

[Thomas Hood]

Henrique de Moraes Holschuh wrote:
> On Fri, 06 Jan 2006, Thomas Hood wrote:
>> > We *cannot* tolerate the whole boot crashing because of a sourced script
>> > screwing up, for example :(  Is there a way to source initscripts and still
>> > tolerate broken initscripts without it doing nasty things to /etc/init.d/rc
>> > like causing it to _die_ ?
>> 
>> There is no reasonable way.  If a sourced script does "exit" then rc exits,
>> with all the consequences of that.
> 
> I was more worried about flat out bugs when in -e mode or somesuch


I just checked.  Neither /etc/init.d/rc nor any .sh initscript on my system
does "set -e".  Which is good, but no guarantee that there aren't other
initscripts that do (or will in the future do) "set -e".


> Can either init itself or rc be teached to run sulogin and reboot if the
> regular rc runlevel scripts execution borks in any way?  One or two forks
> before we run the whole thing is not bad if it increases resiliency, even if
> that means one extra process sitting around.


I can imagine this being done.  But it may not be worth doing if what we are
defending against is programming errors.

We should get a check added to lintian.

   * Error if any /etc/init.d/*.sh script does an "exit" or "set -e"
-- 
Thomas