[ANNOUNCE] FIAIF version 1.5.0-1

Anders Fugmann email hidden
Sat Jan 4 20:13:40 CET 2003

Bill Babcock wrote:
>    Bill Babcock wrote:
>    > 
>    > As an fyi, when I updated to 1.5.2-0pre3 via 'rpm -Uvh', after the
>    > installation went to 100%, the post install script stuff got hung in a
>    > pause - at least that's what strace showed.
>    > 
>    > I tried to erase and start again and the erase showed this:
>    > 
>    > [snake src]# rpm -e fiaif-1.5.2-0pre3
>    > /var/tmp/rpm-tmp.79027: line 1: fg: no job control
>    > error: %preun(fiaif-1.5.2-0pre3) scriptlet failed, exit status 1
>    Ok. I guess this has something to do with the '%_preun_service fiaif' 
>    line, as this is the biggest change from 1.4.5 -> 1.5.0. I have put up a 
>    new version (1.5.2-0pre4), which remove the usage of %_preun_service. 
>    Please test - I guess that you will still see the error on the old 
>    package when upgrading, but then all should be in order.
> I've found a thread in redhat's bugzilla that seems to point squarely
> at this problem:
> http://bugzilla.redhat.com/bugzilla/long_list.cgi?buglist=68056
Hmm. I'm not sure what to make of this, but It seems that the Mandrake 
specific entry made triggered a bug in RH's version of RPM.

> I've gone through some of the steps iterated there and then I had to
> go through them again after running the upgrade to 1.5.2-0pre4, but
> additional package testing seems to be ok.

> I would have to have a wider range of linux to test on to suggest the
> best solution, but under redhat, there is a file /etc/redhat-release
> which could be checked for, and then cased on.
FIAIF cannot support all kind's of distributions (different version of 
RH, Mandrake, Debian, Suse, PLD, etc.) in the makefile. It would simply 
be too hard to maintain. The Tests for directories seems to be the 
easiest and best.

> Additionally - make gets unhappy, in my experience, with comments w/in
> targets, and there were some shell things that needed to be corrected in
> that target. I made some changes that seemed to address the issue with
> that target - see the attached unidiff.
Ok. Comments have been removed, and the diff applied.
I wonder however why the two last tests are so equal. The first tests 
for existance of /etc/rc.d/init.d and the next tests for /etc/rc.d,
and they do the same. It seems that the 3. test is redundant, which is 
why I have removed it.

Anders Fugmann

More information about the fiaif mailing list