On 03/12/2020 11:50, Michael Käppler wrote:
"IIUC, a patch that fails 'make check' ('fails' in the sense of 'errors
out') would have failed 'make test' in the old system, already."
The 'old system' was me doing the tests manually in the order of make,
make-testbaseline, make check, make doc.
In that order.
If any of them failed before make doc then I wouldn't bother doing make
doc. But it is technically feasible for make doc to fail but make check
to work (and vice versa).
This is why patches that have not started, not yet completed or have
failed make doc are filtered out when I run the useful 'countdown'
script Jonas did for us (me) that I use to generate the countdown email
and also to see which patches I needed to test.
James