On Wed, May 17, 2017 at 06:33:53PM +0100, James Hogarth wrote:
> On 17 May 2017 4:35 pm, "Peter Robinson" <pbrobinson@xxxxxxxxx> wrote:
>
> On Wed, May 17, 2017 at 3:48 PM, Matthew Miller
> <mattdm@xxxxxxxxxxxxxxxxx> wrote:
> > On Wed, May 17, 2017 at 09:15:51AM +0100, James Hogarth wrote:
> >> On 17 May 2017 at 08:46, Daniel P. Berrange <berrange@xxxxxxxxxx> wrote:
> >> > Converting apps from nettools to iproute is often non-trivial piece
> >> > of work. As such isn't really something Fedora package maintainers
> >> > should look to undertake as the risk of introducing regressions is
> >> > non-negligible. Bug reports really need to go the corresponding
> >> > upstream communities to get anything done.
> >> >
> >>
> >> That's a sensible position and one I can respect. I do wonder how much
> >> is upstream and how much is a result of packaging though, that itself
> >> might be an interesting investigation.
> >
> > I think this might be something that rises to the level of a
> > Change ("Officially Deprecate net-tools in Fedora"), and while working
> > with upstreams is going to be necessary, I think having a Fedora
> > tracker could be useful, if you're interested in putting in that
> > effort.
>
> Yes, there was already an effort to do this 6 years ago [1] which got
> some of the way there. I don't think the package itself will go away
> any time soon but it would be good to not need it in the core
> distribution. For a while we actually managed to do that but it's
> crept back in over time. I think a focus on getting things like
> cloud-init, vpnc-script, pki-server and similar packages that ship in
> core deliverables would be a good start.
>
> >> At the very least it may be worth checking for upstream bugs, filing
> >> them where they don't exist, and then filing a bugzilla (even with a
> >> tracking bug perhaps? Is this something for FPC to discuss maybe?)
> >> ticket linked to the upstream.
> >
> > Probably FESCo rather than FPC, unless we're going to ban depending on
> > net-tools or something like that.
>
> Yes, it would definitely be FESCo over FPC IMO
>
> [1] https://bugzilla.redhat.com/show_bug.cgi?id=687920
>
>
>
>
> Well I'm ready to put my virtual money where my virtual mouth is :)
>
> In the next couple of weeks I'll get the ticket and tracking bug in place
> with bugs on the relevant packages linked to it.
>
> If FESCo think it needs a Change as well I'll get that in place.
>
> We may not get all the way... But I think it'd be good to make a start and
> see just how far we can get.
>
> It was pointed out to me off list that bridge-utils is in a similar
> situation... But I think it's best to focus on just net-tools first and
> then we can take the next step separately when ready.
Here is a list of binary commands in net-tools:
/usr/bin/netstat
/usr/sbin/arp
/usr/sbin/ether-wake
/usr/sbin/ifconfig
/usr/sbin/ipmaddr
/usr/sbin/iptunnel
/usr/sbin/mii-diag
/usr/sbin/mii-tool
/usr/sbin/nameif
/usr/sbin/plipconfig
/usr/sbin/route
/usr/sbin/slattach
I don't mind removing dependencies on net-tools, as long as there
still exists these commands in the default install:
netstat
arp
ifconfig
route
I consider those to be a basic part of the user interface of any
Linux/UNIX system--there is too much historical precedent and
documentation to remove them IMO. It would be like trying to remove
"ls" just because there is a newer/better way to list files.
_______________________________________________
devel mailing list -- devel@xxxxxxxxxxxxxxxxxxxxxxx
To unsubscribe send an email to devel-leave@xxxxxxxxxxxxxxxxxxxxxxx
