Hi,

On Mon, Sep 06, 2021 at 10:02:11PM +0200, Mark Wielaard wrote:
> Still looking for explanations of the other FAILS (which I cannot
> reproduce locally).

I think I finally got it. The thing that all the tests that are
(sometimes) failing have in commen is that they use SIGUSR1 or SIGUSR2
to trigger a new scan or groom cycle. But they don't check whether the
initial scan or groom cycle has finished. So the attached patch
explicitly does that before the first SIGUSR1 or SIGUSR2 is fired at
the debuginfod server.

Pushed,

Mark