Philip Brown wrote:
Robert Milkowski wrote:
Martin Bochnig wrote:
On Fri, Aug 28, 2009 at 12:53 AM, Philip Brown<[email protected]> wrote:
....
SUNWxim does not depend on SUNWdtdte. Therefore, a patch for
SUNWxim, has no
business pulling in an implied dependancy on SUNWdtdte either.
But it does, and the patch fails without it.
That fairly tightly fits my own personal definition of
"broken patch creation policies/procedures".
So, how is switching to the "better technology" of IPS, going to
solve this
problem of broken patch creation process inside of Sun?
Hey Phil: Very easy! As no "--no-deps" is "supported", your scenario
would not even be possible anymore, not at all. That's the solution
(really).
>>
> As Shawn and Bart posted today --nodeps will be supported via 'pkgrecv
> --nodeps ...; pkg install' which is fair enough and does make sense.
But that is still inadequate.
Sometimes, there are VALID patch dependancies. Sometimes, you really
do NEED dependancies updated, when you update/patch one package,
otherwise, the patch breaks things.
Again, there are no patches - just packages.
I don't see an issue with new package version providing new dependencies
- will just work.
The problem with old patching system is that basically there is no
one-to-one relation ship with a patch and a package so patches tend to
update many packages. With IPS it is not the case so that specific
problem is gone by design.
If "[pkg update]" is taking the place of "patchadd", then there should
be a mode to pkg for, "install/update everything that is installed,
but dont add anything NEW".
Otherwise, the --no-deps is inadequate.
There is no such thing like 'pkg update' - if you want a new version
(with a bug fix or whatever) of package A you just install new version
of package A, that's it. In most cases you should be able to do it to
current OS image or to a cloned one if you want to be on a safe side.
Unless you meant 'pkg image-update --no-deps'. Well, if it is the case
what problem are you trying to solve? As now you lost even me being an
advocate for an easy --no-deps. Is there any commonly used packaging
system which allows something like this? What would be a real value of
something like this?
Philip - the answer is relatively easy - as there are no patches in IPS.
Not exactly true. there will still be patches, effectively speaking.
you're just not allowed to CALL them "patches" any more. they're just
called "updates" or whatever.. but there still needs to be the same QA
done as when you release a "patch", in terms of, "what does this
change affect, what are the side effects, what else needs to be updated?"
Actually you are mistaken here. There are just packages - no updates,
patches, etc.
If you want to upgrade a package you have to install a new version of it
which does actually solve the specific issue you described by design -
regardless of any processes.
One package won't be able to overwrite a files from another package
(unless infamous --force will be implemented) so you wan't be even able
to create a behavior of the old patching system.
And if sun doesnt improve the way they are handling patches right now,
there will eventually be the exact type of problem I described in my
original email:
Sun engineers will release a new "[patch/version/update]" of
something, for bugfix purposes. But if they dont properly scope the
update, it will either FAIL on deployment (because some truely
required dependancy was not updated as needed), or it will pull in
junk that isnt really needed on the system.
Again, it will have to be a new package and not a patch/package spanning
across different other packages...
Then sure, every large enough organization could be improve but I don't
think this is the right forum to address such concerns.
--
Robert Milkowski
http://milek.blogspot.com
_______________________________________________
pkg-discuss mailing list
[email protected]
http://mail.opensolaris.org/mailman/listinfo/pkg-discuss
