In message <199407081359 .
mil>, "William C. Fenner" avows:
%--- Begin Cite ---%
> On Wed, 6 Jul 94 12:20:56 EDT Ed Strong wrote:
> > It looks like I'm expected to just blindly
> > install binary code from stanford in my kernel. I'd much prefer to st
> > from source code instead, if it's available.
> So why don't you? You have SunOS source, right? Just apply the diffs a
> rebuild your kernel. The multicast stuff from stanford is supplied as s
> code and kernel patches; the object files are simply supplied as a conve
> for those who don't have kernel source.
%--- End Cite ---%
Because what Sun (and probably many other vendors) ships as source code has
often come from a different source tree than that from which they build their binaries. In addition, they *never* ship patches to source, so you really run
the risk of opening up your system to old holes when you use source.
This is by no means meant to be a flame against Sun or any other vendor, in
fact one of our Sun contact is currently investigating the possiblity of
getting automatic source patches under our agreement. I just wanted to
point out a hazard in blithely depending on source code.
The topic has strayed though, followups should be directed to the
participants or a more apropriate group. Thanks.