Bug#64437: PROPOSED] Must/Should/May in policy
On Thu, May 25, 2000 at 12:30:32AM +1000, Anthony Towns wrote:
> > > > > Every time you put more than one shell command (this
> > > > > includes using a loop) in a makefile command you
> > > > > - <em>must</em> make sure that errors are trapped. For
> > > > > + should make sure that errors are trapped.
> > > > This must remain a `must', not doing so usually results in incomplete or
> > > > unbuildable packages.
> > If half of the package is missing because of an undetected error in the
> > build process, it's a RC bug. It's better to leave this as is.
>
> If it actually causes problems, it's an RC bug, but that's because of
> the problems it causes; if it doesn't cause problems, it doesn't matter as
> much.
But how could it not not cause problems? (I hope you understand what I mean :)
I've never seen something like this produce something benign. Can you give
such example?
> It's still worth filing a bug about, but if it's not causing problems,
> then why throw the whole package out?
Well, I file RC bugs to get them fixed, not to get the package removed. (the
former should happen, the latter is just an ugly necessity if !former) :o)
> Making it a "must" doesn't make the bug any easier to detect, or any easier
> to fix,
And that matters?
> or, I'd hope, any less likely to be fixed in a package that's getting
> updated.
Well, people should generally pay more attention to RC bugs.
Anyway, never mind details like this now; can someone tell us when will the
whole thing enter Policy? Manoj, Julian?
--
Digital Electronic Being Intended for Assassination and Nullification
Reply to: