On 01-07-2022 23:32, Giacomo Leidi via Guix-patches via wrote:
+ (arguments
+ (list
+ ;; 3 failed, 47 passed, 21 skipped, 1 warning
+ #:tests? #f))
By disabling tests when they detect problems, we defeat the point of
having tests. A 'failing' test is a good test, as it detected a
problem (hopefully not a false positive ...) before the buggy result
is actually used for real.
It's like, say, firing a building inspector because they noticed that
the building would fall over in case of some wind from an
unanticipated direction, instead of making the building more sturdy,
changing the building plan or not proceeding with actually building
the building.
What problems did the tests detect?