Re: patch for bug 312857

[Elijah Newren <newren gmail com>]

On 9/4/05, Harish Krishnaswamy <kharish novell com> wrote:
> > This late into the freeze I'd prefer to only approve absolutely
> > critical bugs as any change could (even if unlikely) accidentally
> > introduce a crasher.  Could you all answer a few questions?
> > >  How heavily has the patch been tested (by how many people and how
> > > much by them)?
> > >  - Are there any additional factors that might make the patch safer?
> > >
> Right. I understand your concern - considering the volume of requests
> from our team :-).
> These patches are mandatorily reviewed and tested by at least two
> developers and verified by the QA team independently. In fact, I have
> approved roughly just one-third of the freeze break requests that I have
> received last week.

Cool, that's good to know.  And I didn't know it before.  If such
information could just be added to all the requests (we can easily
forget stuff like that, we won't be able to tell if the process is
still the same, we occasionally get new members on the team, etc.),
it'd make things go much smoother.  We'd probably tend to just rubber
stamp requests...

> >  - How bad is the bug?  I first read it as something that makes the UI
> > look a little off (i.e. really
> >    annoying but not necessarily critical), but perhaps I misread it
> > and its a data-corruption issue?  I
> >    don't know...
> Yes. This is a nasty data corruption bug reported right after the freeze
> but fortunately before the release.

Yeah, details like this help, though perhaps it was just me misreading
the original.  *shrug*


Thanks for your patience with us,
Elijah