Closed Siddhi-agg closed 5 months ago
It looks great! Out of curiosity, the test was indeed failing before the fix right?
Yes @NathanReb. The last test was failing before I implemented this fix.
I also noticed that the last test was only failing when we used two different cmi
. When I used the same variable two times, the test passed. So, it was indeed creating two signature types for type t
with different identifiers there.
When I used the same variable two times, the test passed. So, it was indeed creating two signature types for type t with different identifiers there.
yes that's expected, if you reuse the same signature without parsing it, the type are strictly identical. In practice that will never happen so our tests should always parse signatures separately, even if it means parsing the same string twice but we can't share the parsed cmi value.
Fixes #44.
I have modified the current signature to reflect the type equality for abstract types with the reference signature. I have also added an equality test case for abstract types and it works as required.