I'm at the bottom of the IPS learning curve at the moment.  My area of 
interest is the fact that I maintain a couple of application packages 
which currently exist in SVR4 pkgadd format for Solaris (x86 and SPARC), 
and RPM format for Linux.  I'm trying to find out what's involved in 
migrating from pkgadd to IPS.

I found  
http://wiki.updatecenter.java.net/Wiki.jsp?page=UC20.Docs.Packaging#section-UC20.Docs.Packaging-IntermediateSVR4Packages
 
- which describes a way of managing existing SVR4 packages within IPS.  
That looked like a convenient stepping stone for me on the way to a 
native IPS solution.  However, I don't really understand what it does 
for me.  I followed the instructions, and ended up with a broken 
installation of my package: in fact, I'm not clear in what sense it is 
"installed" at all.  Certainly it doesn't exist in the pkgadd database, 
so it won't get very far if it calls pkgparam (as it does) to find out 
things like the name of its own installation directory.

My package has the usual entourage of pkgadd scripts (request, 
preinstall, postinstall), and I understand from various postings on 
pkg-discuss that these are considered works of the devil.  I have no 
intention to engage in any debate about that, but the bottom line is 
that my package does have these scripts, and its correct installation 
relies upon their being run in the way that happens with pkgadd.

So the question I am really asking is, is there any purpose in my 
continuing with the attempt to make my SVR4 packages behave correctly in 
IPS, or should I just forget it and concentrate on finding out how to do 
a proper IPS implementation?

--
Bill
_______________________________________________
pkg-discuss mailing list
[email protected]
http://mail.opensolaris.org/mailman/listinfo/pkg-discuss

Reply via email to