Andrew Dunstan <andrew(at)dunslane(dot)net> writes:
> The patch seems reasonable on its face, but I doubt it's the issue.
> Rather I think what's happening here is that a test is hanging silently
> and lastcommand.log's mtime doesn't get updated, causing a misreporting
> of the run duration. So in addition to the above I have added some code
> to update that timestamp if the file exists (which should only be the
> case with a timeout).
Sounds reasonable. Thanks!
regards, tom lane