Hi Andy, On Sat, Oct 21, 2006 at 12:51:50PM -0700, Andy Gill wrote: > > 1. What is the best way of seeing what the expected failures are with > HEAD? > I'm seeing a large number with HEAD right now. Perhaps we could have > a script > that uploads the latest failures to an read-only wiki page?
The Linux HEAD nightly build e-mails to the cvs-ghc (or cvs-all) mailinglist include a list of the tests that failed. Currently there are a lot of failures due to a combination of factors (such as large post-6.6 merges, and the lack of nightly HEAD builds in the lead up to the release meaning we didn't get the nightly test reports). Hopefully we'll get this back under control soon. On a related note, something I've been thinking about: Should the expect_fail family of functions in the testsuite driver take a bug number as an argument? This would make it easy to close the relevant bug when fixing an expected-fail test. We would also probably want some sort of automated check that all referenced bugs are indeed open. I think it would be a nice requirement that for a test to be marked as expected-to-fail there must be an open bug about it. > 2. What are the checkin policies. I could not find anything on the wiki. I would say: * Try not to break anything * Discuss anything you think might be controversial before pushing it If you do end up breaking the build then it's not the end of the world, though, so don't sweat about it too much. History shows that even people called Simon are not immune from doing so! > 2(b). How does someone get a stable HEAD-ish build? Are the dates or > tags? I tend to wait for a night when the nightly builds go through cleanly and update then. Thanks Ian _______________________________________________ Cvs-ghc mailing list [EMAIL PROTECTED] http://www.haskell.org/mailman/listinfo/cvs-ghc
