GitLab

This controls whether or not the compiler warns if we're using an LLVM
version that's too old or too new. It's mostly useful when building the
compiler knowingly with an unsupported version, so you don't get a lot
of warnings in the build process.

There's no documentation for this since it's a flag only a few
developers would care about anyway.
Signed-off-by: Austin Seipp <mad.one@gmail.com>

Closes Trac #7572.
Signed-off-by: Austin Seipp <mad.one@gmail.com>

This adds some new functions: peekRunQueue, promoteInRunQueue,
singletonRunQueue and truncateRunQueue which help abstract away
manual linked list manipulation, making it easier to swap in
a new queue implementation.
Signed-off-by: Edward Z. Yang <ezyang@mit.edu>

Otherwise the configure script for e.g. base doesn't know that we're
cross-compiling, and fails trying to run an executable compiled by the
C cross-compiler.

The coercion optimizer was optimizing coercions inside of branched
axiom applications, sometimes invalidating the branch choice within
the axiom application. Now, we check to make sure we are not
invalidating this invariant before proceeding with the optimization.

It turned out that FunDeps.oclose was unused. So

* Remove oclose

* Rename oclose1 to oclose

* Move growThetaTyVars to FunDeps (from TcMType),
  because the comments treat it with oclose

* Move quantifyPred to TcSimplify (from TcMType),
  because it seemed orphaned

This was breaking tc226, following UNPACK-pragma reorg

This change gives a more helpful error message when the
user says    data T = MkT {-# UNPACK #-} Int
which should have a strictness '!' as well. Rather than
just a parse error, we get

  T7562.hs:3:14: Warning:
    UNPACK pragma lacks '!' on the first argument of `MkT'

Fixes Trac #7562

This fixes bug #1241 and #2247.  When UndecidableInstances are on,
we use the "Liberal Coverage Condition", which is what GHC used to do in
the past.  This is the gist of the check:

class C a b | a -> b
instance theta => C t1 t2

we check that `fvs t2` is a subset of `fd-closure(theta,fvs t1)`.

This is strictly more general than the coverage condition, while
it still guarantees consistency with the FDs of the class.  This
check is completely orthogonal to termination (it by no means guarantees
it).

I am not sure of the role of the "coverage condition" in termination---
the comments suggest that it is important.  This is why, for the moment,
we only use this check when UndecidableInstances are on.

All -fvia-C does nowadays is print a warning saying that it doesn't
do anything.

If we are not unregisterised then we skip the HCc phase.
Fixes #7563.

I don't actually know if suggesting -fllvm as a workaround is useful
advice, but -fvia-C certainly won't help as it doesn't do anything
any more.

We now just use checkCOrAsmOrLlvmOrInterp instead, which had
an identical definition.

bytestring fixes #7556.
terminfo fixes #7281.

hs-suffix-rules now calls hs-suffix-rules-srcdir, saving some duplication

Whether we check that .hi files have actually been created is now
controlled by $(ExtraMakefileSanityChecks) (defaults to NO).

Also updated comments about the .hi rule.

Moved some code from runPipeline' into runPipeline.

This will add the following preprocessor defines when Haskell source
files are compiled:

 * __SSE__ - If any version of SSE is enabled
 * __SSE2__ - If SSE2 or greater is enabled
 * __SSE4_2_ - If SSE4.2 is enabled

Note that SSE2 is enabled by default on x86-64.

We were adding "maintainer-clean : distclean" every time we call
build-package. All of these are redundant, as the same dependency
appears in the root ghc.mk.

We don't want the overhead of spawning a shell on Windows, but on other
platforms it's a useful sanity check.

make thought that it could make a .hi file for the C files in
libraries, which was causing problems when using dynamic-too.

It's now base + rts, rather than haskell98, that we link in by default.

e.g. narrow8Int# subsumes narrow16Int#, not the other way around.