Subject: Re: Multicast routing and unnumbered point-to-point interfaces
To: Hitoshi Asaeda <>
From: Jason R Thorpe <>
List: tech-net
Date: 10/06/2000 11:21:09
On Sat, Oct 07, 2000 at 02:57:04AM +0900, Hitoshi Asaeda wrote:

 > > It is also the intent to cause old source to fail to compile with the
 > > new kernel headers.
 > I'm afraid I can't understand why you insist to cause errors by
 > force.

A multicast routing daemon which compiles fine but then fails to
function isn't the greatest thing, IMO.  Preserving the binary
compatibility was a goal, but changing the API like that is really
the only way to force programs to get fixed (and, sometimes, even
notice that they're broken in the first place).

        -- Jason R. Thorpe <>