I sort of understand how to use IPS in a customer-facing "publication" 
environment.  However, I'm not clear on how to use it in day-to-day 
development.

In the SVR4 world, one would compile some or all of a source tree, 
producing some number of packages.  One would then install some or all 
of those packages on a test system, either first removing the old 
versions or applying the new on top of the old, as desired.  Since the 
built packages normally reside in the build workspace, each developer 
naturally has a personal set of packages, and can readily have several 
sets of packages for different releases, different build variations, et 
cetera.

How does this model map over to IPS?  Since IPS works only from a 
repository, it would appear that each developer must have a personal 
repository, and moreover may need to have multiple repositories when 
working on multiple releases or multiple variants in parallel.  I don't 
see any other way, but that seems like a very heavyweight answer.

We could just work with SVR4 packages most of the time, and only convert 
to IPS for official builds, but that seems wrong, both from a purism 
standpoint and because it means that developers would not be testing 
what's actually being deployed.

Does anybody have any thoughts on the subject?

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

Reply via email to