GitLab

Most notably this pulls in `MonadFail` instances
(see also #10751)

- add MonadFail instance for ContT
- re-order methods for consistency
- Add `MonadFail` instances
- Canonicalise Monad instances
- instance Bifunctor Constant

This one slipped through testing.

Adam Gundry authored Nov 26, 2015 and Ben Gamari committed Nov 26, 2015

This is a first draft of the user manual documentation for
DuplicateRecordFields. Feedback welcome. I've intentionally made this a
minimal change, but I wonder if it might make sense to pull out all the
subsections on record system extensions ("Traditional record syntax",
and "Record field disambiguation" through to "Record wildcards") into a
new section? That is, is it worth breaking down the rather monolithic
"Syntactic extensions" section?

Test Plan: n/a

Reviewers: simonpj, bgamari, austin

Reviewed By: bgamari, austin

Subscribers: thomie

Differential Revision: https://phabricator.haskell.org/D1522

Herbert Valerio Riedel authored Nov 26, 2015 and Ben Gamari committed Nov 26, 2015

Test Plan: IIAM

Reviewers: austin, bgamari, quchen

Reviewed By: bgamari, quchen

Subscribers: thomie

Differential Revision: https://phabricator.haskell.org/D1529

GHC Trac Issues: #10751

Ben Gamari authored Nov 26, 2015 and Ben Gamari committed Nov 26, 2015

Test Plan: Read it

Reviewers: austin, fryguybob

Reviewed By: austin

Subscribers: thomie, fryguybob

Differential Revision: https://phabricator.haskell.org/D1523

GHC Trac Issues: #10639, #367

This patch does some signficant refactoring to the treatment
of default methods in class declarations, and more generally
to the type checking of type/class decls.

Highlights:

* When the class has a generic-default method, such as
     class C a where
       op :: a -> a -> Bool
       default op :: Ord a => a -> a -> a
  the ClassOpItem records the type of the generic-default,
  in this case the type (Ord a => a -> a -> a)

* I killed off Class.DefMeth in favour of the very-similar
  BasicTypes.DefMethSpec.  However it turned out to be better
  to use a Maybe, thus
      Maybe (DefMethSpec Type)
  with Nothing meaning "no default method".

* In TcTyClsDecls.tcTyClGroup, we used to accumulate a [TyThing],
  but I found a way to make it much simpler, accumulating only
  a [TyCon].  Much less wrapping and unwrapping.

* On the way I also fixed Trac #10896 in a better way. Instead
  of killing off all ambiguity checks whenever there are any type
  errors (the fix in commit 8e8b9ed9), I instead recover in
  TcTyClsDecls.checkValidTyCl.

There was a lot of associated simplification all round

[skip ci]

Simplify and tidy up the MonadFail code.
See TcMatches.tcMonadFailOp

Less, code; and more robust.

This incidentally fixes a bug; see the change
to MonadFailErrors.stderr

Ben Gamari authored Nov 24, 2015 and Ben Gamari committed Nov 24, 2015

This should catch duplicate uniques in the future before Bad Things
happen.

Test Plan: Introduce a duplicate unique and validate

Reviewers: austin, hvr, thomie

Reviewed By: hvr, thomie

Subscribers: thomie

Differential Revision: https://phabricator.haskell.org/D1521

quchen authored Nov 24, 2015 and Ben Gamari committed Nov 24, 2015

Reviewers: hvr, austin, thomie, bgamari

Reviewed By: hvr, austin, thomie, bgamari

Subscribers: thomie

Differential Revision: https://phabricator.haskell.org/D1495

GHC Trac Issues: #11000

elaforge authored Nov 24, 2015 and Ben Gamari committed Nov 24, 2015

This puts the "Relevant bindings" section at the end.

It uses a TcErrors.Report Monoid to divide messages by importance and
then mappends them together.  This is not the most efficient way since
there are various intermediate Reports and list appends, but it probably
doesn't matter since error messages shouldn't get that large, and are
usually prepended.  In practice, everything is `important` except
`relevantBindings`, which is `supplementary`.

ErrMsg's errMsgShortDoc and errMsgExtraInfo were extracted into ErrDoc,
which has important, context, and suppelementary fields.  Each of those
three sections is marked with a bullet character, '•' on unicode
terminals and '*' on ascii terminals.  Since this breaks tons of tests,
I also modified testlib.normalise_errmsg to strip out '•'s.

--- Additional notes:

To avoid prepending * to an empty doc, I needed to filter empty docs.
This seemed less error-prone than trying to modify everyone who produces
SDoc to instead produce Maybe SDoc.  So I added `Outputable.isEmpty`.
Unfortunately it needs a DynFlags, which is kind of bogus, but otherwise
I think I'd need another Empty case for SDoc, and then it couldn't be a
newtype any more.

ErrMsg's errMsgShortString is only used by the Show instance, which is
in turn only used by Show HscTypes.SourceError, which is in turn only
needed for the Exception instance.  So it's probably possible to get rid
of errMsgShortString, but that would a be an unrelated cleanup.

Fixes #11014.

Test Plan: see above

Reviewers: austin, simonpj, thomie, bgamari

Reviewed By: thomie, bgamari

Subscribers: simonpj, nomeata, thomie

Differential Revision: https://phabricator.haskell.org/D1427

GHC Trac Issues: #11014

Warn about incoherent/non-canonical 'Applicative'/'Monad' instance
declarations. Specifically the following invariants are checked:

In 'Monad' instances declarations warn if the any of the following
conditions does not hold:

 * If `return` is overridden it must be canonical (i.e. `return = pure`).
 * If `(>>)` is overridden it must be canonical (i.e. `(>>) = (*>)`).

In 'Applicative' instance declarations:

 * Warn if 'pure' is defined backwards (i.e. `pure = return`).
 * Warn if '(*>)' is defined backwards (i.e. `(*>) = (>>)`).

NB, this warning flag is not enabled via `-Wall` nor `-Wcompat`.

This addresses #11128

Reviewers: quchen, austin, bgamari

Reviewed By: bgamari

Differential Revision: https://phabricator.haskell.org/D1516

This is needed to allow subsequent patches to refer to `*>`.

While at it, this commit also groups together the `Applicative` definitions
to reduce confusion.

Reviewed By: austin

Differential Revision: https://phabricator.haskell.org/D1513

Ben Gamari authored Nov 23, 2015 and Ben Gamari committed Nov 23, 2015

Otherwise we'll get link time failures as `base` always builds
`GHC.ExecutionStack`.

Test Plan: Validate

Reviewers: austin

Subscribers: thomie

Differential Revision: https://phabricator.haskell.org/D1519

This responds to Trac #11112 by improving the error message
when the kind checker discovers something of kind Constraint
used when a type is expected

In order to accomplish this we need to ensure that emit DIEs for all
DebugBlocks, even those that have been optimized out, lest we end up
with undefined symbols of parents at link time.

Differential Revision: https://phabricator.haskell.org/D1279

Differential Revision: https://phabricator.haskell.org/D1387

We now only strip block information from DebugBlocks when compiling with
`-g1`, intended to be used when only minimal debug information is
desired. `-g2` is assumed when `-g` is passed without any integer
argument.

Differential Revision: https://phabricator.haskell.org/D1281

Differential Revision: https://phabricator.haskell.org/D1198#40948

Differential Revision: https://phabricator.haskell.org/D1198#40948

RIP-relative addressing isn't available on i386.

The casting here is a bit tricky since Dwarf_Addr is always 64-bits.
This means we first need to narrow to uintptr_t before casting to/from a
pointer for compatibility on 32-bit architectures.

31bcf9b6 changed the behavior of WERROR
such that it would only apply to the stage 2 build. The reason for this
was to avoid silly redundant imports and such breaking validation on
different bootstrap compiler versions.

Document the fact that WERROR now only applies to the stage 2 build.

niteria authored Nov 22, 2015 and Ben Gamari committed Nov 23, 2015

This fixes a non-determinism bug where depending on the order
of uniques allocated, the type variables would be in a different order
when abstracted for the purpose of lifting out an expression.

Test Plan:
I've added a new testcase that reproduces the problem
./validate

Reviewers: simonmar, austin, bgamari, simonpj

Reviewed By: simonpj

Subscribers: nomeata, thomie

Differential Revision: https://phabricator.haskell.org/D1504

GHC Trac Issues: #4012

Alan Zimmerman authored Nov 22, 2015 and Ben Gamari committed Nov 23, 2015

At the moment the API Annotations can only be used on the ParsedSource,
as there are changes made to the RenamedSource that prevent it from
being used to round trip source code.

It is possible to build a map from every Located Name in the
RenamedSource from its location to the Name, which can then be used when
resolved names are required when changing the ParsedSource.

However, there are instances where the identifier is not located,
specifically

  (GHC.VarPat name)
  (GHC.HsVar name)
  (GHC.UserTyVar name)
  (GHC.HsTyVar name)

Replace each of the name types above with (Located name)

Updates the haddock submodule.

Test Plan: ./validate

Reviewers: austin, goldfire, bgamari

Reviewed By: bgamari

Subscribers: goldfire, thomie, mpickering

Differential Revision: https://phabricator.haskell.org/D1512

GHC Trac Issues: #11019

Erik de Castro Lopo authored Nov 22, 2015 and Ben Gamari committed Nov 23, 2015

Test Plan: Test on Arm and PowerPc.

Reviewers: bgamari, thomie, austin

Differential Revision: https://phabricator.haskell.org/D1510

GHC Trac Issues: #11060