Typos, minor comment fixes

- Remove fstName, sndName, fstIdKey, sndIdKey - no longer used,
  removed from basicKnownKeyNames
- Remove breakpointId, breakpointCondId, opaqueTyCon, unknownTyCon -
  they were used in the old implementation of the GHCi debugger
- Fix typos in comments
- Remove outdated comment in Lint.hs
- Use 'LitRubbish' instead of 'RubbishLit' for consistency
- Remove comment about subkinding - superseded by
  Note [Kind Constraint and kind Type]
- Mention ticket ID in a linear types error message
- Fix formatting in using-warnings.rst and linear-types.rst
- Remove comment about 'Any' in Dynamic.hs - Dynamic
  now uses Typeable + existential instead of Any
- Remove codeGen/should_compile/T13233.hs
  This was added by accident, it is not used and T13233 is already in
  should_fail

3 files changed, 3 insertions, 3 deletions

diff --git a/compiler/GHC/HsToCore/Errors/Ppr.hs b/compiler/GHC/HsToCore/Errors/Ppr.hs
index 0d8dbcc9a9..b0d2b16409 100644
--- a/compiler/GHC/HsToCore/Errors/Ppr.hs
+++ b/compiler/GHC/HsToCore/Errors/Ppr.hs
@@ -88,7 +88,7 @@ instance Diagnostic DsMessage where
       -> mkSimpleDecorated $
           text "Ignoring useless SPECIALISE pragma for NOINLINE function:" <+> quotes (ppr poly_id)
     DsMultiplicityCoercionsNotSupported
-      -> mkSimpleDecorated $ text "Multiplicity coercions are currently not supported"
+      -> mkSimpleDecorated $ text "Multiplicity coercions are currently not supported (see GHC #19517)"
     DsOrphanRule rule
       -> mkSimpleDecorated $ text "Orphan rule:" <+> ppr rule
     DsRuleLhsTooComplicated orig_lhs lhs2
diff --git a/compiler/GHC/HsToCore/Expr.hs b/compiler/GHC/HsToCore/Expr.hs
index 6bf900da89..b79f5862b9 100644
--- a/compiler/GHC/HsToCore/Expr.hs
+++ b/compiler/GHC/HsToCore/Expr.hs
@@ -1130,7 +1130,7 @@ is what dsHsWrapped does.  Because we might have an HsVar without a
 wrapper, we check in dsHsVar as well. typecheck/should_fail/T17021
 triggers this case.
 
-Note that if `f :: forall r (a :: Type r). blah`, then
+Note that if `f :: forall r (a :: TYPE r). blah`, then
    const True f
 is absolutely fine.  Here `f` is a function, represented by a
 pointer, and we can pass it to `const` (or anything else).  (See
diff --git a/compiler/GHC/HsToCore/Utils.hs b/compiler/GHC/HsToCore/Utils.hs
index 020775c91c..b68cf061a0 100644
--- a/compiler/GHC/HsToCore/Utils.hs
+++ b/compiler/GHC/HsToCore/Utils.hs
@@ -529,7 +529,7 @@ There are a few subtleties in the desugaring of `seq`:
 
  3. (as described in #2409)
 
-    The isLocalId ensures that we don't turn
+    The isInternalName ensures that we don't turn
             True `seq` e
     into
             case True of True { ... }