[Git][ghc/ghc][wip/boxed-rep] 2 commits: Fix broken tests

[Ben Gamari]

Ben Gamari pushed to branch wip/boxed-rep at Glasgow Haskell Compiler / GHC


Commits:
7b1da914 by Andrew Martin at 2020-11-12T20:15:53-05:00
Fix broken tests

- - - - -
39caf4f1 by Andrew Martin at 2020-11-12T20:18:45-05:00
Remove incomplete comment

- - - - -


3 changed files:

- libraries/base/Data/Typeable/Internal.hs
- testsuite/tests/pmcheck/should_compile/T18249.hs
- testsuite/tests/rts/T8308/T8308.stdout


Changes:

=====================================
libraries/base/Data/Typeable/Internal.hs
=====================================
@@ -863,17 +863,7 @@ splitApps = go []
 -- even worse! We have to construct several different TyCons by hand
 -- so that we can build the fingerprint for TYPE ('BoxedRep 'LiftedRep).
 -- If we call `typeRep @('BoxedRep 'LiftedRep)` while trying to compute
--- the fingerprint of `TYPE ('BoxedRep 'LiftedRep)`, we get a loop since
--- that invocacion of `typeRep` leads to:
---
--- * fpTYPELiftedRep (starting point)
--- * typeRep @('BoxedRep 'LiftedRep)
--- * mkTrApp (applied to 'BoxedRep and 'LiftedRep, attempts TYPE shortcut)
--- * eqTypeRep (second argument is trTYPE, let's pursue this)
--- * sameTypeRep (second argument is trTYPE)
--- * typeRepFingerprint (argument is trTYPE, matches TrFun case)
--- * Dang, I have to think about this more tomorrow. The
---   loop is real, but it evades me right now.
+-- the fingerprint of `TYPE ('BoxedRep 'LiftedRep)`, we get a loop.
 --
 -- The ticket to find a better way to deal with this is
 -- #14480.


=====================================
testsuite/tests/pmcheck/should_compile/T18249.hs
=====================================
@@ -14,7 +14,7 @@ f :: Int# -> Int
 f !_ | False = 1
 f _ = 2
 
-newtype UVoid :: TYPE 'UnliftedRep where
+newtype UVoid :: TYPE ('BoxedRep 'Unlifted) where
   UVoid :: UVoid -> UVoid
 
 g :: UVoid -> Int


=====================================
testsuite/tests/rts/T8308/T8308.stdout
=====================================
@@ -1 +1 @@
-7
+1



View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/compare/fa16578b606480c7782c6bcb0845bae93840d376...39caf4f1edde204edd323596ba2b2b2890dce093

-- 
View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/compare/fa16578b606480c7782c6bcb0845bae93840d376...39caf4f1edde204edd323596ba2b2b2890dce093
You're receiving this email because of your account on gitlab.haskell.org.


-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.haskell.org/pipermail/ghc-commits/attachments/20201113/785ceddc/attachment-0001.html>