> Hi folks,
>
> I think the discussion, while very intersting, might
> be focussed a little
> too low level at the merits of package systems.
Yea, lets talk about HPMS intead of pros/cons of PMS.
I would rather work with existing PMS instead of trying
to replace them.
|--------------------------------------------------------|
HPMS | TWW HPMS=sbutils,pbutils and pkgutils |
|--------------------------------------------------------|
PMS | RPM | | PKGADD | | MSI | | SD-UX |
--------- ------------- ----------- --------------
| Linux | | Solaris | | Windows | | HP-UX |
OS | | | | | | | |
---------- ------------- -------------- -------------
from http://en.wikibooks.org/wiki/CPAM_with_TWW/User_Guide
>
> A distro will need to decide what it wants to support
> and how to do it.
>
> For example does your distro want to support Zones?
> If the answer is yes
> then you will need to implement the SVR4 package
> structure since thats what
> the zones creation and patching tools use. If you
> ou dont want to use SVR4
> and have a differnt package database, then you will
> need to re-implement
> the zones creation and update methodology for your
> chosen system. As
> someone who did a lot of patch testing on zones
> systems in s10 I can assure
> you that its not trivial to update zones. It's
> unfortunate that the patch
> and package tools are not open source right now, but
> you can build packages
> from the binaries as they stand if thats any
> consolation. Creating patches
> is another matter.
>
> Thats just one consideration. Even if you implement
> the ideal package
> management system for builders and users you still
> may face a lot of work
> getting the functionality you want throughout the OS.
>
> The main things wrong with the SVR4 that I'm seeiing
> in this discussion is
> a lack of features in the tools, rahter than a
> problem with the SVR4
> package and patch architecture.
pkgadd (SVR4 PMS) on soalris does has problem. no auto installation
of depended other packages is really a pain. no application depot infrasture
is another one. But this not its fault. these new feature were not in its
design spec. it need a technology update/improvement.
>
> This would lead me to the conclusion that the best
> thing to do would be to
> extend the current Sun Solaris tools as needed; for
> ease of use, updating,
> ease of package/patch creation. But can people wait
> the 9-12 months?
There are two approaches to get moden Package Management System on solaris.
1. After pkgadd is opensource and improve it from source code
level.
2. Write the modules to improve pkgadd externally without modfiy
the pkgadd source code.
This is already being done by TWW Inc. and released in GPL license.
So why spend the effort to modify the source code of pkgadd ?
tj
> Cheers,
> ~Al
>
> --
> Albert White - Patch System Test - Sun Ireland
> _______________________________________________
> opensolaris-discuss mailing list
> opensolaris-discuss@opensolaris.org
>
This message posted from opensolaris.org
_______________________________________________
opensolaris-discuss mailing list
opensolaris-discuss@opensolaris.org
