Skip to content

consistently use iSolveTC to solve typeclass goals

Ralf Jung requested to merge ralf/tc_control_full into gen_proofmode

This avoids the proofmode to be irritated by unrelated goals with a typeclass type.

LambdaRust timings (on my laptop, and I've been browsing the web but hopefully Firefox didn't use too much CPU...): Before

real	13m17,543s
user	50m32,616s
sys	0m33,451s

After

real	13m3,314s
user	49m7,732s
sys	0m33,080s

So, if anything, this seems to help with performance.

Merge request reports