Re: ifupdown-extra upload to stable-proposed-updates

["Adam D. Barratt" <adam@adam-barratt.org.uk>]

On Thu, 16 Jun 2011 10:00:17 +0200, Javier Fernandez-Sanguino wrote:
> 2011/6/13 Adam D. Barratt <adam@adam-barratt.org.uk>:
> > On Sun, 2011-06-12 at 23:25 +0200, Javier Fernández-Sanguino Peña 
> > wrote:
> > The last of the changes in the list above has this changelog entry
> > associated with it:
> >
> > +  * if-up-scripts/static-routes:
> > +      - Fix typo that prevent the script from adding routes as it 
> > expected them
> > +        to have 'reject' when they shouldn't. Thanks to Mathieu 
> > Parent and
> > +        to Petru Ratiu for the patches. (Closes: #613632, #458395) 
> > (LP: #631533)
> >
> > but actually appears to both fix a bug (the inverted sense of the
> > "reject" test) and introduce new functionality relative to the 
> > current
> > version in stable, namely the adding of routes which _do_ have 
> > "reject"
> > associated with them.  Is that correct?
>
> You are right. I will strip off the new functionality and just fix 
> the bug.

Thanks.

> > The second query is more of a comment really.  I appreciate that 
> > this
> > isn't a regression from the previous matching code, but it seems to 
> > me
> > that this:
> >
> >  add_static_routes() {
> > -       cat $ROUTEFILE | egrep "${IFACE}$" |
> > +       cat $ROUTEFILE | egrep "^[^#].*${IFACE}$" |
> >
> > will match a line of "foo bareth0" in the route file where $IFACE
> > contains "eth0".  I'm not sure if this is an issue in practical use 
> > of
> > the package though.
>
> Hmmm.. .it is not fully a regression since it behaves similarly to 
> how
> it did previously, but the regular expression could be improved so
> that it only matched $IFACE when it is a full word and not part of 
> it.
> It is not that common to have similarly named interfaces, but I do
> agree that this, under some circumstances, could trigger a bug.
>
> I will ammend the regular expression in unstable and in the upload to
> proposed-updates.
>
> It is my understanding that fixing these two issues I could go ahead
> and do an unpload to 'stable'?

Assuming that the regular expression change is made in unstable 
_first_, yes that would be fine; thanks.  However, it's worth bearing in 
mind that the window for acceptance in to the 6.0.2 point release closes 
over the coming weekend.

Regards,

Adam