James Carlson <james.d.carlson at Sun.COM> writes: > Mike Kupfer writes: >> It turns out that there's one more tool where the SCM Migration project >> could use some help figuring out what work is left, and that's rtichk. > > I explicitly (and separately) tested rtichk when I tested "wx pbchk." > It works, at least for the on-SWAN case.
Right, the xVM guys use it. Part of the problem is that while it works in the ideal case, it a least used to be incredibly (and uselessly) noisy when things went wrong. I think John may have fixed that. It also used to misreported errors in the off-SWAN case, such that it would tell me there was no RTI for my bugs, rather than that I couldn't reach webrti.sfbay > Obviously (?), it'll need another update and testing pass once we have > an OpenRTI tool and some associated new query mechanism. Yes, it'll need to be redone with a new method of making queries. I've been on the fence about when to approach the OpenRTI folks about that, given I'd expect they're still working on extracting it from the various SWAN dependencies. > The details from that message I sent: > > I manually verified "rtichk," running it by hand with various CR > numbers and checking the results against RTIs I looked up through the > web interface. The results all checked out. > > And, of course, our modified rtichk is being used in production _now_. > It's deployed in Burlington. And by everyone(?) working on the Xen part of the xVM stuff, since that is in mercurial, in production, now. I think the major cases work, and the corner cases are perhaps quite wrong. -- Rich
