Right now the qunit will log all results at the end.
This means that the runbot may wait for all qunit before detecting the
failure.
This also mean that all failure are in one ir.logging on runbot, making
the automated parsing difficult if multiple modules fails during the
same build.
We could also log all failure immediately, but grouping them my qunit
module will avoid duplicating logs for linked causes (one failure
leading to a `Expected %s assertions, but %s were run` message)
closesodoo/odoo#99912
Signed-off-by: Xavier Dollé (xdo) <xdo@odoo.com>
Xavier-Do
0a695ab6b2