GitLab

Occurrences in terms are uses, in patterns they are not.
In this way we get unused-constructor warnings from modules like this

   module M( f, g, T ) where
     data T = T1 | T2 Bool

     f x = T2 x

     g T1 = True
     g (T2 x) = x

Here a T1 value cannot be constructed, so we can warn. The use
in a pattern doesn't count.  See Note [Patterns are not uses]
in RnPat.

Interestingly this change exposed three module in GHC itself
that had unused constructors, which I duly removed:
  * ghc/Main.hs
  * compiler/ghci/ByteCodeAsm
  * compiler/nativeGen/PPC/RegInfo
Their changes are in this patch.

There was a long thread on cvs-ghc which concluded that although it's
the Right Thing to deprecate Rank2Types and PolymorphicComponents, in
favour of RankNTypes, it would cause lots of busy-work for library
authors and the pain isn't worth the gain.

So this patch removes the deprecation, and documents the synonym-ity.

After disussing on cvs-ghc and cabal-devel, we decided, in the end,
that it was simpler to make Rank2Types and PolymorphicComponents into
synonyms for RankNTypes.

This reverts commit d277031d.

Conflicts:
	compiler/coreSyn/CoreLint.lhs

We were calling ranlib on the wrong path

Now that we can build dyn-only, we don't need to force
DYNAMIC_BY_DEFAULT to NO. We just set GhcLibWays appropriately.

This reverts commit c67b4313.

We no longer need to always force, as ghc-pkg is cleverer.

This makes life easier, as ghc.mk has access to HADDOCK_DOCS
and DYNAMIC_BY_DEFAULT, so knows the right way to build it.

In particular, this fixes it if we are using dynamic libraries by
default and don't build the vanilla way.

When dynamic-by-default, we may only have .dyn_hi files, and ghc-pkg
complains if there are no .hi files.

There are better dependencies for this in the DPH package, where all
the TH is.