On Thu 17 Apr 2008 at 07:37PM, Danek Duvall wrote:
>     http://cr.opensolaris.org/~dduvall/pkg-misc3/
> 
> This is intended to fix
> 
>     669 Need method to print package licenses before installation
>     1144 "pkg list -as" fails when unable to write to the pkg directory
> 
> and ends up fixing part of
> 
>     208 pkg info, pkg list display no output for pkgs in the repository and 
> not installed
> 
> In particular, "pkg info" has grown a --license option (for 669), and a -r
> option (also for 669, but that's part of what 208 wants).  But I'm not
> adding "-s <server>" to "pkg info", and I'm not adding a remote version of
> what is now "pkg contents".
> 
> I'd like folks' thoughts on the loop I copied from image.py to client.py.
> That should be refactored, but given that it's slightly different in the

The layering/implicit knowledge in client.py that something is
going to be gzipped seems... sketchy IMO.

Test cases needed for each of the failure modes at:

client.py: 686 (info: -l and -r may not be combined)
client.py: 689 (nfo: must request remote info for specific packages)

client.py: I personally really like a couple of lines of vertical
whitespace between functions, esp. in python where there is no column
zero closing brace to signify function end.  Oh well.

More "get a list of things" sprawl.  Oh well.  We need to clean this up
post-May.

Can you post some sample output please?

Didn't totally understand the movement of gunzip stuff from generic to
misc....

client/image.py: Any deeper implications of this change?  Would
we want to warn the customer that they're fetching manifests but
they can't be stored locally?  For non-priv users, should we have
a stash in /tmp or something?  I could see this banging on the
server rather a lot...

client/image.py: I know you didn't write this but "parse" is a really
vague name and it took me some time to work out what this was doing.
Would this code be clearer, shorter and faster as follows?

        mfst = self.get_manifest(fmri, filtered = True)

        for depend in mfst.gen_actions_by_type("depend"):
                required, min_fmri, max_fmri = depend.parse()
                if required == False:
                        self.update_optional_dependency(min_fmri)

I couldn't work out why we needed a list comprehension.

        -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