Danek Duvall wrote:
> On Wed, Aug 20, 2008 at 05:21:27PM -0500, Shawn Walker wrote:
> 
>> http://cr.opensolaris.org/~swalker/pkg-1730/
> 
> So why isn't this done in the fmri constructor?

I don't know -- that's where I wanted to put it, until I saw that we had 
the valid_new_fmri method already in catalog.

I think it's odd that fmri doesn't throw an exception when it gets an 
invalid fmri.

> Or at least be a method on the fmri, rather than a method in the misc
> module?  Putting the test in the constructor would slow down object
> creation, and maybe not all parts of the system care, but it should at
> least be an operation on an fmri object.

I thought it would be useful to be able to check whether something was a 
valid pkg name without constructing an fmri.  If you think it should be 
a @staticmethod of the fmri class -- I'm ok with that.

I also agree with your concerns about putting the check in the fmri 
constructor.

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

Reply via email to