[osol-discuss] Re: rpm vs pkgadd (again!)

[UNIX admin]

> exactly which needed features are those that rpm has
> that pkgadd does not ?

That's really for Martin and not for me to say, since he seems to be the one preferring `rpm` to `pkgadd` and friends.

But since you're asking me, I'll tell you what I plan to do to `pkgadd`. I plan to extend it such that by setting SUNW_PKG_REVCHECK=true or similar in the pkginfo will make `pkgadd` revision aware.

I have to think about it some more, since I absolutely and without a doubt want my ehancement to be both backward and forward compatible with the present `pkgadd`.

There are principally several major features `pkgadd` is lacking, and those are:

if I need REV 1.0 or 2007.01.05.07 of some dependency,
and I have REV 3.2 or 2007.04.18.23 of the said software,
`pkgadd` should be able to recognize that I have satisfied the min REV required, and plough on instead of doing a string-like check and *insisting* that I need *exactly* REV 1.0 (or 2007.01.05.07, for example).

That right there is one of the biggest weaknesses of the System V packaging.

The second thing is how upgrades are handled; this, however, is an issue that can be addressed separately.

For example, on SGI's IRIX, `inst` is smart enough to note the difference between two same "subsystems" of a different revision: if a newer subsystem *does not* have files that the old one does, the now unnecessary files will be automatically and transparently removed by `inst`. This is an upgrade feature of `inst` that I find simply and without any doubt phenomenal!!!

Both of the aforementioned issues and scenarios is where `inst` excells. I have never worked with a UNIX/Linux software management subsystem as elegant and as intelligent as `inst`!

If SGI should tank tomorrow, *one* technology I would want to preserve from IRIX at all costs (even paying for it out of my own pocket!) would be `inst`. It is simply ingenious.

Message was edited by: 
        ux-admin
 
 
This message posted from opensolaris.org