[Lldb-commits] [lldb] [lldb] Keep L1 memory cache chunks disjoint (PR #208347)

[Felipe de Azevedo Piovezan via lldb-commits]

https://github.com/felipepiovezan commented:

> I thought about it, it was simpler indeed. My intuition tells me that it's fragile to have a cache allowing overlap.

I pointed at the simplicity as a plus, but my real concern is the quadratic behaviour.
If we have N insertions, each of which has a tiny overlap with the one before, LLDB is going to be quite slow processing that. E.g. `[1000, 2100), [2000, 3100), [3000, 4100)...`.

Maybe the numbers here  are not big enough in practice, but it does concern me to have this kind of complexity in LLDB.

https://github.com/llvm/llvm-project/pull/208347