[commit: ghc] wip/rae-new-coercible's head updated: Don't look under newtypes when unifying representationally. (2005d52)

[git at git.haskell.org]

Repository : ssh://git@git.haskell.org/ghc

Branch 'wip/rae-new-coercible' now includes:

     e74a9e9 Remove `inline` from integer_gmp_mpn_import1()
     7535c83 Comments only
     37c2ed4 Optimise partitionFunEqs for the 'false' case
     15a54be Improve the treatment of AppTy equalities
     d64e682 Comments and variable names only, in type checking of (e1 $ e2)
     1d44261 Revise the inert-set invariants again
     ac73d1a Revise flattening-notes
     5818378 RAE's response to SPJ's question in flattening-notes
     b06908b Fix #9871 by clarifying documentation.
     8688f6a Add doctest examples for Data.Maybe
     2515686 catch some recent typos
     c2c1888 Update `process` submodule to pull in Safe Haskell fixes
     e5974f8 Proposal for Backpack file format [skip ci]
     fc45f32 Implement -XStaticValues
     09b7943 fix misleading error message regarding function arity
     8b480d3 Document splitAt deviation from the Report
     df1307f Link pre-ARMv6 spinlocks into all RTS variants
     659ec2c Add proper expected output for T5435_dyn_asm on Darwin
     4d1c452 Only run subsections_via_symbols test when LLVM is available.
     7ca5bb0 compiler: fix trac issue #9817
     b006a1a Add Ord instances to TH
     13b0b46 Reorganise the work list, so that flattening goals are treated in the right order
     1496598 Get rid of TcMType.newWantedEvVar(s)
     3e234f7 Fix type-variable details naming (fixes misleading debug output)
     bcb967a When flattening, try reducing type-family applications eagerly
     832f8db Implement a fast path for new constraints looking like (a~b), namely unifyWanted
     37b3646 Testsuite wibbles from constraint-solver improvements
     fca85c9 Tests for Trac #9872
     a225c70 Comments only: move flattening notes to TcFlatten
     f0bb9e9 Merge remote-tracking branch 'origin/master' into wip/rae-new-coercible
     ad0e495 Sequelae from previous merge.
     418806d Flat constraint --> Simple constraint
     2005d52 Don't look under newtypes when unifying representationally.