* Rename CoAxiom.Eqn = Pair Type to TypeEqn,
  and use it for fundeps

* Use the FunDepEqn for injectivity, which lets us share a bit
  more code, and (more important) brain cells

* When generating fundeps, take the max depth of the two
  constraints.  This aimed at tackling the strange loop in
  Trac #12860, but there is more to come for that.

* Improve pretty-printing with -ddump-tc-trace

In this commit
   commit 6c0f10fa
   Author: Ben Gamari <bgamari.foss@gmail.com>
   Date:   Sun Nov 13 16:17:37 2016 -0500

    Kill Type pretty-printer

we switched to pretty-printing a type by converting it to an
IfaceType and pretty printing that.  Very good.

This patch fixes two things

* The new story is terrible for debug-printing with -ddump-tc-trace,
  because all the extra info in an open type was discarded ty the
  conversion to IfaceType.

  This patch adds IfaceTcTyVar to IfaceType, to carry a TcTyVar in
  debug situations.  Quite an easy change, happily.  These things
  never show up in interface files.

* Now that we are going via IfaceType, it's essential to tidy before
  converting; otherwise
     forall k_23 k_34. blah
  is printed as
     forall k k. blah
  which is very unhelpful.  Again this only shows up in debug
  printing.

I found that some TcTyVars were lurking in a PatSyn, because
tc_patsyn_finish was using the TcType -> TcType zonker rather
than the TcType -> Type zonker.  Eeek.

I fixing this I also tided up function naming a bit (still not
terrific), and removed the unused TcTyBinder type entirely.

En route to something else I discovered that TcInstDcls.addDFunPrags
was building a DFunUnfolding that had TcTyVars in it.  They should
never survive beyond type checking.  It was harmeless, but now affects
type pretty-printing.

This patch fixes it.

Just affects Id.mkTemplateLocal, and should make debug prints
a little less voluminous.  No user-visible change here.

IfaceTypes are really not well suited to finding free variables etc.
Nevertheless, there was quite a lot of code to do just that; but it
was only used to see if a kind is variable-free so as to decide
whether to print a forall binder.

This patch simplifies to deal with just that case, replacing all
the free-vars stuff with just ifTypeIsVarFree

Trac #12686 showed that we were allowing a term variable into
a type, by promotion.  I chose to squash this in the renamer.

I think this wave of commits just made these two a little better;
they must have been close to the threshold before.

The invariants for CTyEqCan mean that this check isn't needed.
Turn it into an ASSERT.

Up to now we've had a rule that a TyVar can't apppear in a type
seen by the type checker; they should all be TcTyVars.  But:

a) With -XTypeInType it becomes much harder to exclude them;
   see Note [TcTyVars in the typechecker] in TcType.

b) It's unnecessary to exculde them; instead we can just treat
   a TyVar just like vanillaSkolemTv.

This is what was causing an ASSERT error in
indexed-types/should_fail/T12041, reported in Trac #12826.

That patch allows a TyVar in a TcType.  The most significant
change is to make Var.tcTyVarDetails return vanillaSkolemTv.
In fact it already did, but (a) it was not documented, and
(b) we never exploited it.  Now we rely on it.

This patch takes further my refactoring of the constraint
solver, which I've been doing over the last couple of months
in consultation with Richard.

It fixes a number of tricky bugs that made the constraint
solver actually go into a loop, including

  Trac #12526
  Trac #12444
  Trac #12538

The main changes are these

* Flatten unification variables (fmvs/fuvs) appear on the LHS
  of a tvar/tyvar equality; thus
           fmv ~ alpha
  and not
           alpha ~ fmv

  See Note [Put flatten unification variables on the left]
  in TcUnify.  This is implemented by TcUnify.swapOverTyVars.

* Don't reduce a "loopy" CFunEqCan where the fsk appears on
  the LHS:
      F t1 .. tn ~ fsk
  where 'fsk' is free in t1..tn.
  See Note [FunEq occurs-check principle] in TcInteract

  This neatly stops some infinite loops that people reported;
  and it allows us to delete some crufty code in reduce_top_fun_eq.
  And it appears to be no loss whatsoever.

  As well as fixing loops, ContextStack2 and T5837 both terminate
  when they didn't before.

* Previously we generated "derived shadow" constraints from
  Wanteds, but we could (and sometimes did; Trac #xxxx) repeatedly
  generate a derived shadow from the same Wanted.

  A big change in this patch is to have two kinds of Wanteds:
     [WD] behaves like a pair of a Wanted and a Derived
     [W]  behaves like a Wanted only
  See CtFlavour and ShadowInfo in TcRnTypes, and the ctev_nosh
  field of a Wanted.

  This turned out to be a lot simpler.  A [WD] gets split into a
  [W] and a [D] in TcSMonad.maybeEmitShaodow.

  See TcSMonad Note [The improvement story and derived shadows]

* Rather than have a separate inert_model in the InertCans, I've
  put the derived equalities back into inert_eqs.  We weren't
  gaining anything from a separate field.

* Previously we had a mode for the constraint solver in which it
  would more aggressively solve Derived constraints; it was used
  for simplifying the context of a 'deriving' clause, or a 'default'
  delcaration, for example.

  But the complexity wasn't worth it; now I just make proper Wanted
  constraints.  See TcMType.cloneWC

* Don't generate injectivity improvement for Givens; see
  Note [No FunEq improvement for Givens] in TcInteract

* solveSimpleWanteds leaves the insolubles in-place rather than
  returning them.  Simpler.

I also did lots of work on comments, including fixing Trac #12821.

I need this in a later patch

This bug meant that we were less eager about expanding
tuple superclasses than we should have been; i.e. we stopped
too soon.  That's not fatal, beause we expand more superclasses
later, but it's less efficient.

bollmann authored 8 years ago and Erik de Castro Lopo committed 8 years ago

Previously TH was incorrectly returning a `Dec` using a `ConT` instead
of `PromotedT`.

Test Plan: validate

Reviewers: mainland, jstolarek, osa1, goldfire, thomie, bollmann,
bgamari, RyanGlScott, austin

Reviewed By: RyanGlScott

Subscribers: erikd

Differential Revision: https://phabricator.haskell.org/D2188

GHC Trac Issues: #11629

This patch also removes the "catch all" pattern in the function and
explicitly lists constructors to get a warning in the future if a new
`HsType` was added.

Reviewers: bgamari, austin, simonpj

Reviewed By: simonpj

Subscribers: simonpj, thomie

Differential Revision: https://phabricator.haskell.org/D2753

GHC Trac Issues: #12711

Alexander Vershilov authored 8 years ago and Ben Gamari committed 8 years ago

Allow to unregister threadWaitReadSTM/threadWaitWriteSTM on
a non-threaded runtime. Previosly noop action was returned,
as a result it was not possible to unregister action, unless
data arrives to Fd or it's closed.

Fixes #12852.

Reviewers: simonmar, hvr, austin, bgamari, trofi

Reviewed By: bgamari, trofi

Subscribers: thomie

Differential Revision: https://phabricator.haskell.org/D2729

GHC Trac Issues: #12852

Ben Gamari authored 8 years ago and Ben Gamari committed 8 years ago

We need to ensure that we don't include Rts.h from bootstrap compiler.
See #12796.

Test Plan: Validate

Reviewers: austin

Subscribers: thomie

Differential Revision: https://phabricator.haskell.org/D2698

GHC Trac Issues: #12796

dobenour authored 8 years ago and Ben Gamari committed 8 years ago

Reviewers: bgamari, austin

Reviewed By: bgamari

Subscribers: thomie

Differential Revision: https://phabricator.haskell.org/D2733

Test Plan: none

Reviewers: austin, snowleopard, bgamari

Reviewed By: bgamari

Subscribers: thomie

Differential Revision: https://phabricator.haskell.org/D2722

Maciej Bielecki authored 8 years ago and Ben Gamari committed 8 years ago

This patch add new options `-Werror=...`, `-Wwarn=...` and
`-Wno-error=...` (synonym for `-Wwarn=...`).

Semantics:

  - `-Werror` marks all warnings as fatal, including those that don't
    have a warning flag, and CPP warnings.
  - `-Werror=...` enables a warning and marks it as fatal
  - `-Wwarn=...` marks a warning as non-fatal, but doesn't disable it

Test Plan: validate

Reviewers: austin, bgamari

Reviewed By: bgamari

Subscribers: mpickering, svenpanne, RyanGlScott, thomie

Differential Revision: https://phabricator.haskell.org/D2706

GHC Trac Issues: #11219

Reviewers: austin, simonmar, trofi, bgamari

Reviewed By: bgamari

Subscribers: mpickering, thomie

Differential Revision: https://phabricator.haskell.org/D2671

GHC Trac Issues: #12802

Ben Gamari authored 8 years ago and Ben Gamari committed 8 years ago

I thought that this would naturally resolve itself with the elimination
of the Type pretty-printer but somehow all of the stars in result
position of an arrow are still rendered in non-Unicode syntax. Quite
odd.

Test Plan: Validate

Reviewers: austin

Subscribers: thomie

Differential Revision: https://phabricator.haskell.org/D2704

GHC Trac Issues: #12550

Ben Gamari authored 8 years ago and Ben Gamari committed 8 years ago

Test Plan: Validate

Reviewers: austin

Subscribers: thomie

Differential Revision: https://phabricator.haskell.org/D2703

GHC Trac Issues: #12447

Ben Gamari authored 8 years ago and Ben Gamari committed 8 years ago

Test Plan: Validate

Reviewers: austin

Subscribers: thomie

Differential Revision: https://phabricator.haskell.org/D2702

GHC Trac Issues: #12024

Ben Gamari authored 8 years ago and Ben Gamari committed 8 years ago

Test Plan: Validate

Reviewers: austin

Subscribers: thomie

Differential Revision: https://phabricator.haskell.org/D2731

GHC Trac Issues: #12855

Andreas Klebinger authored 8 years ago and Matthew Pickering committed 8 years ago

Summary: Updated code comment regarding EquationInfo. Fixes T12856

Reviewers: austin, bgamari, mpickering

Reviewed By: mpickering

Subscribers: mpickering, thomie

Differential Revision: https://phabricator.haskell.org/D2730

GHC Trac Issues: #12856

We directly build up the correct AvailInfos rather than generating
lots of singleton instances and combining them with expensive calls to
unionLists.

There are two other small changes.

* Pushed the nubAvails call into the explicit export list
  branch as we construct them correctly and uniquely ourselves.
* fix_faminst only needs to check the first element of the export
  list as we maintain the (yucky) invariant that the parent is the
  first thing in it.

Reviewers: simonpj, austin, bgamari

Reviewed By: simonpj, bgamari

Subscribers: simonpj, thomie, niteria

Differential Revision: https://phabricator.haskell.org/D2657

GHC Trac Issues: #12754

Summary:
Fixes #12438. As discussed on the Haskell libraries mailing list here:
https://mail.haskell.org/pipermail/libraries/2016-November/027396.html

Reviewers: hvr, austin, bgamari

Reviewed By: bgamari

Subscribers: thomie

Differential Revision: https://phabricator.haskell.org/D2726

GHC Trac Issues: #12438

Summary:
Add some missing symbols that we
export from the public headers but forgot to include
in the runtime linker's symbol table.

This is a bit of a unsatifactory patch, since we have a bit of a
cat and mouse game going. We should find a way to automate this.

But I know of no good solutions at the moment that won't add all
rts symbols (including those we don't have an extern declaration for.).

So for now, just add the ones reported missing.

Test Plan: inplace/bin/ghc-stage2.exe --interactive

Reviewers: RyanGlScott, austin, erikd, simonmar, bgamari

Reviewed By: bgamari

Subscribers: thomie, #ghc_windows_task_force

Differential Revision: https://phabricator.haskell.org/D2727

GHC Trac Issues: #12846

Test Plan:
Build and test on x86_64/linux (perf-llvm) and armhf/linux.

Reviewers: austin, hvr, bgamari

Reviewed By: bgamari

Subscribers: RyanGlScott, thomie

Differential Revision: https://phabricator.haskell.org/D2719

Facundo Domínguez authored 8 years ago and Ben Gamari committed 8 years ago

It turned out that finalizers were run too early and information
resulting from simplifying constraints was not available.

This patch runs finalizers after a first call to simplifyTop, and
then calls simplifyTop a second time to deal with constraints
that could result from running the finalizers.

Fixes T12777

Test Plan: ./validate

Reviewers: goldfire, simonpj, bgamari, austin

Reviewed By: simonpj

Subscribers: mpickering, mboes, thomie

Differential Revision: https://phabricator.haskell.org/D2659

GHC Trac Issues: #12777

Both Windows and POSIX codepaths use System.Directory.canonicalizePath.

Ben Gamari authored 8 years ago and Ben Gamari committed 8 years ago

@Phyx is working on correctly fixing (pun intended) the underlying issue
that prompted this hack. It turns out that `timeout` it the culprit.
Moreover, this hack breaks on msys python builds, which don't export
`WindowsError`.

Test Plan: Validate on Windows with `msys` python.

Reviewers: Phyx, austin

Subscribers: thomie, Phyx

Differential Revision: https://phabricator.haskell.org/D2724

GHC Trac Issues: #12554