Last night I tried to reduce the time it takes to analyze matches for exhaustiveness and reachability using futures.
I'm using one future for each analysis of each match (thus, two futures per match),
and only for the part that doesn't interact with the rest of the compiler
(but -Ystatistics indicate that's a significant part of the time spent).
Unfortunately, compile times did not change ("significantly") on a 4-core i7.
It's hard to know for sure, of course, since benchmarking this properly is too time consuming,
so I was wondering whether I am using Future wrong.
thanks!
adriaan