Undo dubious optimisation in resolveTC
Since hnf_compute doesn't do what it says, and it's causing some obscure
or complicated type class constraints not to resolve properly.
Fixes #1688
commit94be0eb612b0ac0ca54398edaf703b86a5fbd3f0
authorEdwin Brady <ecb10@st-andrews.ac.uk>
dateSat 20 Dec 2014, 23:28 (12 years ago)
parent4175a7f
src/Idris/ElabTerm.hs+1−1
@@ -1227,7 +1227,7 @@ resolveTC = resTC' []
12271227resTC' tcs def 0 topg fn ist = fail $ "Can't resolve type class"
12281228resTC' tcs def 1 topg fn ist = try' (trivial' ist) (resolveTC def 0 topg fn ist) True
12291229resTC' tcs defaultOn depth topg fn ist
1230− = do hnf_compute
1230+ = do compute
12311231 g <- goal
12321232 ptm <- get_term
12331233 ulog <- getUnifyLog