[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]

Re: rpms/lineakd/devel lineakd.spec, NONE, 1.1 .cvsignore, 1.1, 1.2 sources, 1.1, 1.2



On Fri, 08 Sep 2006 18:08:24 -0700, Chris Petersen wrote:

> Michael Schwendt wrote:
> > On Sun, 21 May 2006 12:27:50 -0700, Chris Petersen wrote:
> > 
> >> Author: xris
> >>
> >> Update of /cvs/extras/rpms/lineakd/devel
> > 
> >> %post
> >>
> >> /sbin/ldconfig
> >>
> >> ################################################################################
> >>
> >> %postun
> >>
> >> /sbin/ldconfig
> >>
> >> ################################################################################
> > 
> > This .spec file should not have been approved in a review. Inserting
> > such "#######..." lines in the middle of [and below] scriptlet sections
> > bears a big risk. They become part of the scriptlets itself. This causes
> > bad breakage in RPM transactions when /bin/sh is not the scriptlet interpreter.
> > Has happened before regularly, so avoid "######..." lines like the plague.
> > Use multiple empty lines instead.
> 
> I'm aware of the issues it creates, but no one could give me a good
> reason why it would be a problem to leave things as they are.

One good reason is you ought to prefer

%post -p /sbin/ldconfig
%postun -p /sbin/ldconfig

so rpm can optimise execution of such "scriptlets" and add automatic
dependencies on /sbin/ldconfig. This is especially useful in packages
where you will never execute more than ldconfig anyway.

As a side-effect, you don't see your long #### lines in the scriptlet
bodies ("rpm --query --package --scripts lineak"). It's a trade-off
between spec-file readability and scriptlet ugliness.


[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]