I really like your approach, thanks for sharing.
At first everything seemed to work.
However, I've found a problem with merging booleans.
See this truth table:
| case |
common |
specialized |
result |
correct? |
| 1 |
false |
true |
true |
yes |
| 2 |
true |
true |
true |
yes |
| 3 |
false |
false |
false |
yes |
| 4 |
true |
false |
true |
no |
Case no. 4 is incorrect.
It seems like the merge function uses or semantics.
However, the correct behaviour for our use case would be: Specialized value always wins.
Does anyone know how we could make this work?
I really like your approach, thanks for sharing.
At first everything seemed to work.
However, I've found a problem with merging booleans.
See this truth table:
falsetruetruetruetruetruefalsefalsefalsetruefalsetrueCase no. 4 is incorrect.
It seems like the merge function uses
orsemantics.However, the correct behaviour for our use case would be: Specialized value always wins.
Does anyone know how we could make this work?