Danek Duvall wrote:
> On Tue, Oct 14, 2008 at 06:57:10PM -0500, Shawn Walker wrote:
> 
>> http://cr.opensolaris.org/~swalker/pkg-3704-3890
> 
> api_errors.py:
> 
>   - line 140: this seems like it'll be pretty verbose if there's more than
>     one fmri that isn't found.  Perhaps print the message once and then
>     each missing package on a line of its own?

I was following the convention we have for PlanCreationException.

> t_pkg_install.py:
> 
>   - line 1526: won't this fail to install because quux itself isn't in the
>     catalog?  This doesn't seem like an optional dependency failure.

No, because quux is already installed, so it shouldn't be a failure for 
quux -- it should be a failure for retrieving the optional dependency's 
information; see line 1521.

>   - I would also expect that if an optional dependency can't be found, it
>     should be ignored, and the install / update should succeed.  It may not
>     yet, but the test could fail until we fix the code.

I debated that, but then the problem is that we could get back to where 
we are now where an unhandled exception is thrown instead of an expected 
one.  I'd be more inclined to see the test changed once we change the 
behaviour.

Since these comments came after my putback, I'll open another bug for 
the items you spotted.

Thanks!

-- 
Shawn Walker
_______________________________________________
pkg-discuss mailing list
[email protected]
http://mail.opensolaris.org/mailman/listinfo/pkg-discuss

Reply via email to