On Fri, 13 Oct 2006 04:20:54 +0200 Roland Mainz wrote:
> Erm... my original patch (and request) was about the introduction of a
> hidden+private alias for "getconf" (called "astgetconf") which the test
> suite should use instead and then we could disable the "getconf" builtin
> command on our side (temporarily - which should give us a way to do our
> ksh93 putback with a working test suite and you would have enougth time
> to handle the "getconf" issue in peace) and not your side...

our concern is a temporary fork along with semantics we can't vouch for --
passing regression tests is the first step, but those tests were not
written to test the getconf builtin per-se

a secondary concern is we would have to build the temporary fork on our
side to test possible solaris specific problems on this list -- potentially
doubling our debugging and problem replication time

-- Glenn Fowler -- AT&T Research, Florham Park NJ --


Reply via email to