Re: [SSI-users] procps debian package

[Brian J. W. <Bri...@hp...>]

Everett Coleman II wrote:
> Okay,
> 
>   As i promised, here is the procps patch, i've made the *.deb's and
> tested them, I am currently using them on my system. =3D)
> 
>   Another question, with the nfs and portmap, i assume there is a
> reason why they're are sources for them in the redhat tree.  I will
> blindly create packages for them.  If somebody will explain to me what
> the ssi enhancements are for, that will be very well appreciated.

Every source tree in the repository is there because OpenSSI needs to 
modify/enhance it in some way. David Zafman did the work on nfs-utils 
and portmap, so he has a good understanding of what was changed in them. 
You can get a quick diff versus the base with `cvs diff -ur REDHAT -r 
OPENSSI-RH nfs-utils portmap_4`, although this will only show you files 
modified, not files added.

You might notice that there are nfs-utils.more/ and portmap_4.more/ 
directories. The reason for these is due to the way the RH version of 
these packages is generated. The specfiles don't use the copy of 
nfs.init or nfslock.init in the standard nfs-utils/ source tree, or the 
copy of portmap.init in portmap_4/. Instead, it references stand-alone 
copies of these files as source:

$ head specs/nfs-utils.spec
Summary: NFS utlilities and supporting daemons for the kernel NFS server.
Name: nfs-utils
Version: 1.0.1
Release: 3.9_ssi_7
Source0: http://prdownloads.sourceforge.net/nfs/nfs-utils-1.0.1.tar.gz
Source1: ftp://nfs.sourceforge.net/pub/nfs/nfs.doc.tar.gz
Source10: nfs.init
Source11: nfslock.init
Source12: SSInfs.init
Patch0: nfs-utils-0.3.3.drop-privs.patch

Therefore, I needed to store these files somewhere in the repository 
that was separated from the main source trees. Using the *.more naming 
convention seemed like a reasonable approach.

You may not need to do this for Debian, although storing the Debian and 
RH versions of a file in different locations does make syncing changes 
between them a bit trickier.

Hope this helps,

Brian