On Fri 18 Apr 2008 at 02:35PM, Danek Duvall wrote:
> Sure.  This is what's currently there.  The empty output from the first
> example was what I was asking about, but I'm sure the messaging could be
> improved for all the errors.  Here are the local infos:
> 
>     # In catalog, not installed (exit=0):
>     $ pkg info SUNWzsh

Hmm.  This one is the confusing one.  Should we print a stub of some
kind saying that the package is "known"  Isn't that what "known" is for?

> Here are the remote infos.  Note that the SUNWtcat output is the same as
> for the local case, because there's only one version to be had.  If there
> were more than one, and I didn't have the latest installed, it would get
> the information for the latest version from the server.  I believe that it
> would say "Not installed" in that case, though we should probably figure
> out how to make it say "Upgradable".

Yeah.  Or we should have a "Local Version" and "Remote Version"
thing in the output... or something.  Perhaps, file a bug, so we
have a record that we need to come back around.

>     # In catalog, not installed (exit=0):
>     $ pkg info -r SUNWzsh
>             Name: SUNWzsh
>          Summary: Z shell (zsh)
>            State: Not installed
>        Authority: opensolaris.org (preferred)
>          Version: 4.3.4
>      Build Release: 5.11
>           Branch: 0.84
>     Packaging Date: Wed Apr 16 22:12:11 2008
>             Size: 4.0 MB
>             FMRI: pkg:/[EMAIL PROTECTED],5.11-0.84:20080416T221211Z

I didn't think we had a state called "Not Installed"?  Stephen had
some very precise opinions about this, see bug #1168.

>     # In catalog, installed (exit=0):
>     $ pkg info -r SUNWtcat
>             Name: SUNWtcat
>          Summary: Tomcat Servlet/JSP Container
>            State: Installed
>        Authority: opensolaris.org (preferred)
>          Version: 5.5.25
>      Build Release: 5.11
>           Branch: 0.84
>     Packaging Date: Wed Apr 16 22:03:47 2008
>             Size: 23.7 MB
>             FMRI: pkg:/[EMAIL PROTECTED],5.11-0.84:20080416T220347Z

I like this output-- would it argue for -r to be the default?
Something to explore for later.

> Yeah, I dunno.  I'm pretty certain nothing will break because the manifest
> can't be written.  Anything where you wouldn't have perms to install will
> fail later now than it would have previously, but that's the only behavior
> change there.  As for banging on the server a lot, yeah, I suppose.
> Another wad?

Yeah-- maybe note it with a bug report.  I know you disabled:

        'pkg info -r' (no argument)

Should we (or did you?) do the same for "pkg list -as"?  That will have the
same effect of pulling down a lot of manifests too as best I can recall.

        -dp

-- 
Daniel Price - Solaris Kernel Engineering - [EMAIL PROTECTED] - blogs.sun.com/dp
_______________________________________________
pkg-discuss mailing list
[email protected]
http://mail.opensolaris.org/mailman/listinfo/pkg-discuss

Reply via email to