[Rpm-maint] Thoughts on RPM revitalization, re: documentation

JP Vossen jp at jpsdomain.org
Tue Dec 19 18:17:31 UTC 2006


My main problem with RPM is/was a lack of a packaging standards and a 
really good "best practices for spec files" guide, with the latter 
mostly causing the former.  When I was co-maintainer of the Snort RPMs 
for a while (mid-2003 to mid-2004 or so), my biggest obstacle was 
learning the best way to do each thing that I needed to do.  I knew 
generally what I needed and wanted to do, but I was never able to find 
anything I really liked or that was accepted or definitive for *how* 
best to do it and what to do [1].

What I really wanted was:

1) A list of things that a spec file should do, and things it should not 
do (e.g. interactive prompt for user input).

2) A skeleton of a proper, best practices, etc. spec file.

3) Several simple example spec files.

4) Several moderate examples (Snort's package isn't very complex but it 
does have sub-packages).

5) Some really hairy examples.  (All of these examples should be trivial 
to do.  Just refactor existing FC specs to "best practices".)

I'm aware that I can disassemble existing SRPM packages to see how 
others did similar things.  And I did that.  It was an utter nightmare 
mess of different methods, styles, and so forth (mostly Red Hat packages 
(circa 2003)).

RPM technical resources and FC/RH package standards may have changed 
since then, but if not, that's my "feature request" list.  I should also 
note that for a variety of reasons I now use mostly Debian and Ubuntu, 
with some RHEL 3&4 thrown in, but no FC newer than 2 (don't ask).  So I 
haven't kept up on the RPM scene, unless you count using yum update 
periodically.

I've also had other issues like DB corruption (esp. RH8), as people on 
/. and elsewhere have mentioned.  I'm sure that'll be taken care of in 
the code cleanup and will say no more about it.  All that yammering 
aside, with the advent of yum there is nothing wrong with RPM that a 
little cleanup and documentation won't fix, so thanks to all of you are 
are going to go do it!

Hope this is useful,
JP

[1] Resources I used at the time as follows.  These are all good, but 
none provide (or did at the time) what I wanted.  I e-mailed Mr. 
Foster-Johnson about it but he was not in a position to be able to work 
on it then.

Red Hat RPM Guide by Foster-Johnson
http://www.rpm.org/RPM-HOWTO/build-it.html
http://www.rpm.org/max-rpm/
http://bent.latency.net/rpm-building

----------------------------|:::======|-------------------------------
JP Vossen, CISSP            |:::======|        jp{at}jpsdomain{dot}org
My Account, My Opinions     |=========|      http://www.jpsdomain.org/
----------------------------|=========|-------------------------------
Microsoft has single-handedly nullified Moore's Law.
Innate design flaws of Windows make a personal firewall, anti-virus
and anti-malware software mandatory. The resulting software arms race
has effectively flattened Moore's Law on hardware running Windows.



More information about the Rpm-maint mailing list