GitLab

It's just a small cleanup. There should be no change in
behaviour.

Test Plan: ./validate --slow

Reviewers: austin, bgamari, simonpj

Reviewed By: simonpj

Subscribers: thomie, simonmar

Differential Revision: https://phabricator.haskell.org/D1901

This simple change fixes Trac #11563, #11520, #11516, #11399.
See esp the comments in #11520.

See Note [Fail fast on kind errors] in TcSimplify

Merge to 8.0 branch

Noticed as a sphinx warning:
    docs/users_guide/flags-warnings.gen.rst:97:
        WARNING: Inline interpreted text or phrase
        reference start-string without end-string.

Which pointed to broken table.

Before the patch table looked like:

    | :ghc-flag:`-Wno-unticked-promoted-constructors |
    | `                                              |

After the patch long link is on a single line:

    | :ghc-flag:`-Wno-unticked-promoted-constructors`    |
Signed-off-by: Sergei Trofimovich <siarheit@google.com>

Noticed when looked through sphinx warnings:
    docs/users_guide/safe_haskell.rst:4:
        SEVERE: Duplicate ID: "ghc-flag-Issue"
Signed-off-by: Sergei Trofimovich <siarheit@google.com>

On the GHC 8.0 RCs, multiple users reported a very strange error
whereby GHC would complain that the symbols names recorded in interface
files did not match the expected name.  The reason for this is
that they were using an old version of Cabal which chose symbol
names differently from the installed package ID ('id' field) which
the package was to be installed with; GHC 8.0 now mandates that
these coincides.

This change adds a test to ghc-pkg to make sure that 'id' and 'key'
(which is how Cabal previously reported what the symbol name
was supposed to be) match; if they don't match or key is missing,
we assume that the Cabal was too old.

Bikeshed points:

    - Should we offer more information about how to upgrade
      Cabal correctly (i.e. specify a version?)

    - Should we allow for a missing 'key'?  If we allow for
      'key' to be missing, we lose the ability to detect
      Cabal from GHC 7.8 or earlier being used.  If we
      require it to be specified, then it will not be possible
      for Cabal to deprecate the (unused) field and remove it
      without having BC for 8.0.
Signed-off-by: Edward Z. Yang <ezyang@cs.stanford.edu>

Test Plan: validate

Reviewers: austin, bgamari, hvr

Reviewed By: hvr

Subscribers: bergmark, thomie

Differential Revision: https://phabricator.haskell.org/D1892

GHC Trac Issues: #11558

Reviewers: austin, bgamari

Reviewed By: bgamari

Subscribers: thomie

Differential Revision: https://phabricator.haskell.org/D1899

The bug itself has already been fixed in #10734, so this
only adds another regression test (as given in the ticket).

Test Plan: ./validate

Reviewers: goldfire, austin, thomie, bgamari

Reviewed By: bgamari

Differential Revision: https://phabricator.haskell.org/D1898

GHC Trac Issues: #9022

GHC does not use passed paramaters anywhere
for this deprecated option.
Signed-off-by: Sergei Trofimovich <siarheit@google.com>

Previously "types" was inappropriately made plural instead of
"instance",

    instance Eq Ordering -- Defined in ‘GHC.Classes’
    ...plus 24 others
    ...plus 13 instance involving out-of-scope typess

GHC 7.10.2 and 7.10.3 had a derived `Eq` instance for `SrcLoc`, but it
seems to have been removed (see
6740d70d) during GHC 8.0 development.

Reviewers: hvr, austin, gridaphobe, bgamari

Reviewed By: gridaphobe, bgamari

Subscribers: thomie

Differential Revision: https://phabricator.haskell.org/D1894

Test Plan: GREP_OPTIONS=--blah ./validate

Reviewers: austin, thomie

Reviewed By: thomie

Differential Revision: https://phabricator.haskell.org/D1887

GHC Trac Issues: #11530

When CrossCompiling=YES or Stage1Only=YES, building the haddocks and the
User's Guide should be skipped, because haddock and mkUserGuidePart
depend on the GHC API.
See Note [No stage2 packages when CrossCompiling or Stage1Only] for
details.

There are several places in the build system where the variables
HADDOCK_DOCS and BUILD_SPHINX_* are checked. Instead of also checking
for the variables CrossCompiling or Stage1Only in all those places,
`make` will now exit with a nice error message when the user requests
the impossible.

Reviewers: rwbarton, austin, bgamari

Reviewed By: bgamari

Differential Revision: https://phabricator.haskell.org/D1882

The module CmmLayoutStack removes stack checks if a function does not
use stack space. However, it can only recognize checks of the form
Sp < SpLim. However, these checks get sometimes rewritten to
Sp >= SpLim (with both branches swapped), so we better recognize these
checks too.

Test Plan: ./validate

Reviewers: austin, bgamari, simonpj

Reviewed By: simonpj

Subscribers: simonpj, thomie

Differential Revision: https://phabricator.haskell.org/D1881

GHC Trac Issues: #11533

This was largely fixed by the re-rework of the pattern match checker.
Resolves #5642.

Reviewers: austin, thomie

Subscribers: thomie

Differential Revision: https://phabricator.haskell.org/D1893

We also need to update `stgBindHasCafRefs` assertion with this change,
as we no longer have the pre-computed SRT, LiveVars etc. We rename it to
`topStgBindHasCafRefs` and implement it like this:

A non-updatable top-level binding may refer to a CAF by referring to a
top-level definition with CAFs. A top-level definition may have CAFs if
it's updatable. At this point (because this is done after TidyPgm)
top-level Ids (whether imported or defined in this module) are
GlobalIds, so the top-levelness test is easy. (see also comments in the
code)

Reviewers: bgamari, simonpj, austin

Reviewed By: simonpj

Subscribers: thomie

Differential Revision: https://phabricator.haskell.org/D1889

GHC Trac Issues: #11550

tyConRolesRepresentational is just a version of tyConRolesX, but
specialised for a Representational argument. Saves a bit of extra
argument passing and pattern matching, and tyConRolesX was often
called when we knew the argument role was Representational.

Rather to my surprise this made the compiler allocate 5% less
for tests T9872{b,c,d}.  At least I think it's this commit.
Good thing, regardless.

   mkTvSubst :: InScopeSet -> TvSubstEnv -> TCvSubst
produces a TCvSubst with an empty CvSubstEnv

..around newtype deriving instances.

See esp the new Note [Newtype-deriving instances]

No change in behaviour

This patch fixes Trac #11523.

* The basic problem was that TcRnTypes.superClassesMightHelp was
  returning True of a Derived constraint, and that led to us
  expanding Given superclasses, which produced the same Derived
  constraint again, and so on infinitely.  We really want to do
  this only if there are unsolve /Wanted/ contraints!

* On the way I made TcSMonad.getUnsolvedInerts a bit more
  discriminating about which Derived equalities it returns;
  see Note [Unsolved Derived equalities] in TcSMonad

* Lots of new comments in TcSMonad.

This is a small refactoring, no change in behaviour.

If we fail to typecheck by blowing the constraint simplifier
iteration limit, we want to see the limit-blowing meessage.
Previously it was being suppressed by the type /error/, which
suppress the iteration-limit /warning/.  Solution: make the
iteration-limit message into an error.

They were excessively verbose.  I've commented them out rather than
deleting so that they can easily be restored.

When investigating Trac #11523 I found that superclass
expansion was a little over-aggressive; we were sort of
unrolling each loop twice.

This patch corrects that, and adds explanatory comments.

Flag up the problem highlighted in Trac #11518 comment:15

This is a more stressful example of T11480.

I missed this on my previous commit, somehow
   2cf3ca Allow foralls in instance decls
Apologies.

This patch finally makes it possible to have explicit
foralls in an instance decl
   instance forall (a :: *). Eq a => Eq [a] where ...

This is useful to allow kind signatures or indeed
explicicit kind for-alls; see Trac #11519

I thought it would be really easy, because an instance
declaration already contains an actual HsSigType, so all
the syntactic baggage is there.  But in fact it turned
out that instance declarations were kind-checked a
little differently, because the body kind of the forall
is 'Constraint' rather than '*'.

So I fixed that.  There a slight kludge
(see Note [Body kind of a HsQualTy], but it's still a
significant improvement.

I also did the usual other round of refactoring,
improved a few error messages, tidied up comments etc.
The only significant aspect of all that was

  * Kill mkNakedSpecSigmaTy, mkNakedPhiTy, mkNakedFunTy
    These function names suggest that they do something
    complicated, but acutally they do nothing. So I
    killed them.

  * Swap the arg order of mkNamedBinder, just so that it is
    convenient to say 'map (mkNamedBinder Invisible) tvs'

  * I had to improve isPredTy, to deal with (illegal)
    types like
       (Eq a => Eq [a]) => blah
    See Note [isPeredTy complications] in Type.hs

Still to come: user manual documentation for the
instance-decl change.

..namely buildImplication.

Plus white space in TcDeriv

Noticed by uselex.rb:
    last_free_capability: [R]: exported from:
        ./rts/dist/build/Capability.o
    shutdownCapability: [R]: exported from:
        ./rts/dist/build/Capability.o
Signed-off-by: Sergei Trofimovich <siarheit@google.com>