Commit 3211fa06 authored by Gabor Greif's avatar Gabor Greif

Spelling in comments [ci skip]

parent 258c7195
......@@ -145,7 +145,7 @@ becomes
fl
We still have recusion for non-overloaded functions which we
speciailise, but the recursive call should get specialised to the
specialise, but the recursive call should get specialised to the
same recursive version.
......
......@@ -988,7 +988,7 @@ tcSuperClasses :: DFunId -> Class -> [TcTyVar] -> [EvVar] -> [TcType]
-- $Ordp1 = /\a \(d:Ord a). dfunEqList a (sc_sel d)
--
-- See Note [Recursive superclasses] for why this is so hard!
-- In effect, be build a special-purpose solver for the first step
-- In effect, we build a special-purpose solver for the first step
-- of solving each superclass constraint
tcSuperClasses dfun_id cls tyvars dfun_evs inst_tys dfun_ev_binds sc_theta
= do { (ids, binds, implics) <- mapAndUnzip3M tc_super (zip sc_theta [fIRST_TAG..])
......
......@@ -797,7 +797,7 @@ Theorem [Stability under extension]
is an inert generalised substitution.
Conditions (T1-T3) are established by the canonicaliser
Conditions (K1-K3) are established by TcSMonad.kickOutRewriteable
Conditions (K1-K3) are established by TcSMonad.kickOutRewritable
The idea is that
* (T1-2) are guaranteed by exhaustively rewriting the work-item
......@@ -1077,7 +1077,7 @@ work?
* CDictCan (C tys) or CFunEqCan (F tys ~ fsk):
Yes if the inert set could rewrite tys to make the class constraint,
or type family, fire. That is, yes if the inert_eqs intersects
with the free vars of tys. For this test we use rewriteableTyVars
with the free vars of tys. For this test we use rewritableTyVars
which ignores casts and coercions in tys, because rewriting the
casts or coercions won't make the thing fire more often.
......@@ -1510,7 +1510,7 @@ kick_out_rewritable new_fr new_tv ics@(IC { inert_eqs = tv_eqs
-- Kick it out if the new CTyEqCan can rewrite the inert one
-- See Note [kickOutRewritable]
-- Used only on CFunEqCan, CDictCan, CIrredCan
-- hence no forallls in (ctEvPred ev), hence rewriteableTyVarsOfType ok
-- hence no foralls in (ctEvPred ev), hence rewritableTyVarsOfType ok
kick_out_ct ct | let ev = ctEvidence ct
= fr_may_rewrite (ctEvFlavourRole ev)
&& new_tv `elemVarSet` rewritableTyVarsOfType (ctEvPred ev)
......
......@@ -903,7 +903,7 @@ rewritableTyVarsOfType ty
go (TyConApp _ tys) = rewritableTyVarsOfTypes tys
go (AppTy fun arg) = go fun `unionVarSet` go arg
go (FunTy arg res) = go arg `unionVarSet` go res
go ty@(ForAllTy {}) = pprPanic "rewriteableTyVarOfType" (ppr ty)
go ty@(ForAllTy {}) = pprPanic "rewritableTyVarOfType" (ppr ty)
go (CastTy ty _co) = go ty
go (CoercionTy _co) = emptyVarSet
......
......@@ -78,7 +78,7 @@ extern void *adjustorCode;
* recover the writable address, we subtract 1 word from the executable
* address and fetch. This works because Linux kernel magic gives us two
* pointers with different addresses that refer to the same memory. Whatever
* you write into the writeable address can be read back at the executable
* you write into the writable address can be read back at the executable
* address. This method is very efficient.
*
* On iOS this breaks for two reasons: 1. the two pointers do not refer to
......
......@@ -55,8 +55,8 @@ else
# needs to make a choice which of the above two implicit rules to follow (it
# never runs 2 recipes for the same target, unless double colon rules are
# used, which we don't), logically it will choose the first rule: Lexer.o will
# depend on Lexer.hs in the build directory, that file will be build, and then
# Lexer.o can be build.
# depend on Lexer.hs in the build directory, that file will be built, and then
# Lexer.o can be built.
#
# In an sdist however, Lexer.hs is present in the source directory. It was
# copied there during the creation of the sdist by a rule in
......@@ -65,7 +65,7 @@ else
# building from sdist and 2) the sdist creation process renamed Lexer.x to
# Lexer.x.source. So normally make would now choose the second rule: Lexer.o
# will depend on Lexer.hs in the source directory, for which nothing needs to
# be done, and then Lexer.o can be build.
# be done, and then Lexer.o can be built.
#
# There is however another actor in play, a rule in sdist-ghc-file.mk, which
# after after instantiating with the same directory ($1=utils/genprimops) and
......
Markdown is supported
0% or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment