On Sat, May 17, 2008 at 12:19 PM, Bastian Blank <[EMAIL PROTECTED]> wrote: > On Sat, May 17, 2008 at 11:30:36AM -0400, Zack Weinberg wrote: >> Unfortunately, this is a "can't happen" condition, and there's not >> enough information for me to have any idea what's gone wrong, nor is >> there enough information in the full build log. > > What do you mean with "can't happen" condition?
"Malfunctioned" indicates that the test harness itself crashed. It was written very defensively, so I wasn't expecting ever to see that happen. > Why are the tests not loud enough to detect the real problem? Since this wasn't expected to happen, I didn't bother putting in verbose logging for that case. > Sure, there is a developer machine[1] available. I don't have access to developer machines. Would you mind doing it? zw -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
