> >     http://cr.opensolaris.org/~johansen/webrev-4381/
> >   
> I'm fine with this for now. I think in general, it's better to pass the 
> data to the exception directly, and have the str method build the 
> string. I couldn't find where FileListRetrievalError was defined though, 
> so I'm not certain if that's possible in this situation.

FileListRetrievalError is defined in filelist.  I was keeping with the
conventions that have been used for ManifestRetrievalError and
DatastreamRetrievalError.  The code for those exception handlers are all
essentially the same.  This exception does have a str method that prints
the data; however, since we don't always know what's going to cause a
retrieval error, I create some helpful message as the data, and then let
the handler actually print out that string.

> Can you test what happens to the Packagemanager when/if this new
> exception reaches client code? Actually, from what I can tell, it
> won't be any worse than what happened with the RuntimeException, but
> we should still check, and file a bug against the gui if the exception
> would take it down.

I thought we had discussed this during the I/O Errors review, but I'm
having a hard time finding any record of that.  To my knowledge, the GUI
wasn't catching any of the transport
[FileList|Manifest|Catalog|Datastream]RetrievalError exceptions, so I'm
not regressing existing behavior.

-j

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

Reply via email to