Re: Mandrake RPMs rebuilt - Mailing list pgsql-hackers

From Thomas Lockhart
Subject Re: Mandrake RPMs rebuilt
Date
Msg-id 3C864567.A9896055@fourpalms.org
Whole thread Raw
In response to Mandrake RPMs rebuilt  (Thomas Lockhart <thomas@fourpalms.org>)
Responses Re: Mandrake RPMs rebuilt  (teg@redhat.com (Trond Eivind Glomsrød))
List pgsql-hackers
(dropped -general from the recipients)

> I try to nix RedHatisms from my spec -- it is not always possible since I
> only run RedHat on my Intel machines (I run SuSE 7.3 on my UltraSPARC, but
> have not had time to work on SuSEifying the 7.2 set yet), and I don't always
> know what the RedHatisms are.

So far, the only things I had to do were to generate an "mx" RPM for my
machine, since the eGenix modules are not apparently available from Mdk
sites.

Maybe this is one reason:
   WARNING: Using this file is only recommended if you really must   use it for some reason. It is not being actively
maintained!
 

(from the header of the mxDateTime module :(

> I'll look at your mx RPM's spec and compare to RedHat's -- if a finer grained
> dependency can be determined, then we can go that route, as the eGenix mx
> stuff goes by more names than just 'mx'.

Right. The RH spec seems to have called it "mx2" sometimes, and who
knows what else it is called once it is packaged (the tarballs have the
eGenix name in them).

The RH spec file did weird things like explicitly reference
"/usr/bin/python2.2" as the python executable rather than
/usr/bin/python, perhaps because of problems with their build of
python-2.2. Mandrake 8.1 has python-2.1, and I was able to use your
techniques for version detection from the Pg spec file instead.

Anyway, I'm not sure what the future of the mx package is, or whether
there is an alternative in the python package set. It is a fairly simple
build, so occasionally grabbing the RH package and rebuilding isn't a
big deal.
                      - Thomas


pgsql-hackers by date:

Previous
From: Lamar Owen
Date:
Subject: Re: Mandrake RPMs rebuilt
Next
From: Turbo Fredriksson
Date:
Subject: Re: Drop in performance for each INSERT/DELETE combo