please note that dgk and I aren't privy to the solaris regression tests so problems like getconf compatibility may take several rounds to resolve
I'd also appreciate holding off on terms like "we've reached a deadlock" until we've had a chance to have at the next round -- Glenn Fowler -- AT&T Research, Florham Park NJ -- On Sun, 12 Nov 2006 04:34:51 +0100 Roland Mainz wrote: > Mike Kupfer wrote: > > > > >>>>> "Roland" == Roland Mainz <roland.mainz at nrubsig.org> writes: > > > > Roland> - ksh93 has "getconf" enabled and bound to /bin as it is needed > > Roland> for the test suite. The new version of the builtin is fully > > Roland> compatible to the native version except that there are > > Roland> additional options > > > > I chatted briefly with Don about this. It's on his to-do list, but he > > probably won't get to it today. It looks like there are a couple issues > > that will need to be worked out before submitting the case. > Read http://bugs.grommit.com/show_bug.cgi?id=22#c3 - it seems the > problem isn't completely gone.. ;-( > ... and I am running out of options and ideas how to solve it. We can't > disable the "getconf" builtin since this would break the test suite (and > therefore we would ship a ksh93 which is no longer compatible to the > upstream versions) and the OS/Net rules forbid a putback with an > incompatible builtin enabled. > We've reached a deadlock.
