On Wed, Sep 14, 2016 at 4:11 PM, Michael Catanzaro <mcatanzaro@gnome.org> wrote:

On Wed, 2016-09-14 at 20:50 +0100, Richard Hughes wrote:
> Three people gave the update positive
> karma and I can't believe all three did so without actually opening a
> JPEG-2000 image in any GTK-using or KDE-using app so there might be
> something more subtle going on.

I can believe it.

I repeat my call that packages should spend more time in testing. This
is very far from the first time we've had an update fly past without
sufficient time for testing. Serious proposal: +3 karma and the update
can be pushed after one week in testing, otherwise it has to wait two
weeks. Packages maintainers could still be able to push an update
through faster, but would be expected to do so only in exceptional
circumstances, like to respond to a serious regression.

This isn't a very extreme idea.

Michael

Updates to existing packages, perhaps, but I don't think this is a good idea for *new* packages. My experience is that new package updates rarely get tested (unless they're something extremely popular), and new packages have theoretically just been tested by both the maintainer (when packaging them) and the reviewer (when reviewing them), so there is likely less need for further testing than there would be for other updates. And also, it should be significantly less likely for a new package to break things than it would be for updates to existing packages.

Ben Rosser