On Tue, 1 Aug 2000, Marcus Meissner wrote:
> In article <NBBBJGOOMDFADJDGDCPHMEOHCHAA.law@sgi.com> you wrote:
> ...
> > I may be out in left field, but my idea of a thorough development
> > and release process would be to take all the tools one just built and
> > use those to produce a 3rd generation. Then take those tools and produce
> > a 4th generation. If 3rd!=4th (subtracting out time/date stamps) then
> > it would seem something is wrong -- or am I missing something?
>
> > Now that whole process would be automated and if the 3rd and 4th compare,
> > then kick off the automated test suites.
>
> > Maybe RedHat doesn't use a thorough source and release methodology
> > for their Beta's?
>
> Just FYI, we do exactly that at Caldera for the last years. (And we call it
> self-hosting ;)
GCC does something similar during its own installation: first it builds a
bare-bones compiler with the local cc, then recompiles itself to produce a
GCC-built GCC - then repeats that step, and compiles the two results.
There are a couple of issues with some platforms adding timestamps etc
within the binary, IIRC, but it's basically just a simple file compare.
Of course, it's not sufficient to detect the recursive Trojan cc approach,
but it's a pretty good consistency/sanity check.
James.
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/
This archive was generated by hypermail 2b29 : Mon Aug 07 2000 - 21:00:07 EST