[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Bug#554194: ifupdown virtual package name and mass-filing (if accepted)



On Sat, Jan 16, 2010 at 06:09:19PM +0100, Bill Allombert wrote:
> On Sat, Jan 16, 2010 at 03:31:28PM +0100, Wouter Verhelst wrote:
> > On Thu, Nov 05, 2009 at 09:54:37AM +0100, martin f krafft wrote:
> > [...response that is not very relevant to this mail...]
> > 
> > There was no further discussion on this item since the above date. Since
> > I've recently uploaded ipcfg, I'd like this to be finalized. It
> > currently uses 'ifupdown' as the name to conflict/replace/provide, but I
> > don't consider that to be a particularly good idea.
> > 
> > I'm suggesting that the package name 'network-config-tool' be described
> > as a tool for a package providing 'ifup' and 'ifdown' binaries. These
> > should provide the following interface:
> > 
> > - support 'ifup <interface name>' or 'ifdown <interface name>' to bring
> >   an interface up or down, consistently with configuration, and exit
> >   with non-zero if either operation fails.
> 
> Is ifup "eth0=foo" supported ?

Should be fairly easy to add that to ipcfg, which might be a good idea.
Let's make that part of the interface, too.

> > - may provide a virtual interface name that does not map to an actual
> >   physical interface name, but instead uses internal logic to decide
> >   what to do.
> 
> Is not there a namespace issues wrt other interface that should be clarified ?

There are namespace issues, but I don't think they should be explained
in the interface specification; the tools should do so in their
documentation.

This is a "may" part of the specification, not a "should".

> > - ifup and ifdown should support a '-a' or '--all' option to configure
> >   or deconfigure 'all' interfaces. Here, 'all' is defined as 'all
> >   interfaces for which the tool's configuration defines that they should
> >   be brought up or down with the -a option'.
> 
> OK.
> 
> > - ifup and ifdown should support a '-v' or '--verbose' option to aid in
> >   debugging.
> 
> This requirement does not feel necessary.

If a tool calls ifup, it may wish to call it with -v to provide some
output to the user should bringing the interface up fail, which can be
useful. Perhaps it should be clarified that the output format of -v is
undefined.

> > - ifup and ifdown should support hook scripts in
> >   /etc/network/if-*.d:
> >   - the tool should provide a way for the user to set configuration
> >     values through environment variables, the name of which start with
> >     IF_
> >   - the tool should provide PHASE and MODE variables, describing what
> >     we're trying to do
> >   - (since I could not find a formal specification of the if-*.d hook
> >     script interface, I may have missed some things; if so, please let
> >     me know)

Obviously this also needs the IFACE= environment variable, defining the
interface on which we work.

-- 
The biometric identification system at the gates of the CIA headquarters
works because there's a guard with a large gun making sure no one is
trying to fool the system.
  http://www.schneier.com/blog/archives/2009/01/biometrics.html



Reply to: