ftr: have made progress on the "flaky" `test` `@ru...
# development
w
ftr: have made progress on the "flaky"
test
@rules
issue. it's both a doozy, and hopefully also a small fix (fingers crossed). high ratio of "tests added" over "code fixed", i hope.
a
Ooh! Yay! Share? :)
w
sooo: we are non-deterministically ending up with a subtlely different rule graph... we're choosing to use the same rule, but for a different set of Params
meaning that the transitive deps of a rule are different.
i have two small fixes to turn that into an ambiguity error instead, but need to tweak them so that they still work when a rule cycles on itself
(ie, the non-determinism isn't the actual problem... the fact that there was two choices of rule was the problem)
a
Hurrah! Sounds like an adventure!