GitLab

This patch affects libraries, and requires a submodule update.

Some other libraries, maintained by others, have redundant constraints,
namely:
           containers
           haskeline
           transformers
           binary

I have suppressed the redundant-constraint warnings by settings in
           validate-settings.mk
(in this commit)

The idea was promted by Trac #9939, but it was Christmas, so I did
some recreational programming that went much further.

The idea is to warn when a constraint in a user-supplied context is
redundant.  Everything is described in detail in
  Note [Tracking redundant constraints]
in TcSimplify.

Main changes:

 * The new ic_status field in an implication, of type ImplicStatus.
   It replaces ic_insol, and includes information about redundant
   constraints.

 * New function TcSimplify.setImplicationStatus sets the ic_status.

 * TcSigInfo has sig_report_redundant field to say whenther a
   redundant constraint should be reported; and similarly
   the FunSigCtxt constructor of UserTypeCtxt

 * EvBinds has a field eb_is_given, to record whether it is a given
   or wanted binding. Some consequential chagnes to creating an evidence
   binding (so that we record whether it is given or wanted).

 * AbsBinds field abs_ev_binds is now a *list* of TcEvBiinds;
   see Note [Typechecking plan for instance declarations] in
   TcInstDcls

 * Some significant changes to the type checking of instance
   declarations; Note [Typechecking plan for instance declarations]
   in TcInstDcls.

 * I found that TcErrors.relevantBindings was failing to zonk the
   origin of the constraint it was looking at, and hence failing to
   find some relevant bindings.  Easy to fix, and orthogonal to
   everything else, but hard to disentangle.

Some minor refactorig:

 * TcMType.newSimpleWanteds moves to Inst, renamed as newWanteds

 * TcClassDcl and TcInstDcls now have their own code for typechecking
   a method body, rather than sharing a single function. The shared
   function (ws TcClassDcl.tcInstanceMethodBody) didn't have much code
   and the differences were growing confusing.

 * Add new function TcRnMonad.pushLevelAndCaptureConstraints, and
   use it

 * Add new function Bag.catBagMaybes, and use it in TcSimplify

The main change is in TypeRep.pprTheta, so we print
       Eq a
for a singleton, but
      (Eq a, Show a)
for multiple constraints.

There are lots of trivial knock-on changes to error messages

When doing top-level defaulting, in TcSimplify.applyDefaultingRules, we
were temporarily making a unification variable equal to the default type
(Integer, say, or Float), as a 'given', and trying to solve. But this
relied on the unification variable being untouchable, which seems
complicated.  It's much simpler just to generate a new set of
constraints to solve, using newWantedEvVarNC in disambigGroup.

(I tripped over an ASSERT failure, and this solved it in a robust way.)

This fixes an ASSERT failure in TcBinds.  The problem was that we
were generating NoGen plan for a function with a partial type signature,
and that led to confusion and lost invariants.

See Note [Partial type signatures and generalisation] in TcBinds

I think the new implementation is a bit more efficient, because
it uses a work-list, rather than iterating over the entire set
every time

In particular
  In the type signature for:
     f :: Int -> Int
I added the colon

Also reword the "maybe you haven't applied a function to enough arguments?"
suggestion to make grammatical sense.

These tiny changes affect a lot of error messages.

Previously it was a SrcSpan, which can be an UnhelpulSrcSpan,
but actually for TcLclEnv and CtLoc we always know it is
a real source location, and it's good to make the types
reflect that fact.

There is a continuing slight awkwardness (not new with this
patch) about what "file name" to use for GHCi code.  Current
we say "<interactive>" which seems just about OK.

Signed-off-by: Edward Z. Yang <ezyang@cs.stanford.edu>

Summary:
Where we track timestamps of object files, also track timestamps
for interface files.  When -fno-code -fwrite-interface is enabled, use
the interface file timestamp as an extra check to see if the files are
up-to-date.  We had to apply this logic to one-shot and make modes.

This fix would be good to merge into 7.10; it makes using -fno-code
-fwrite-interface for flywheel type checking usable.
Signed-off-by: Edward Z. Yang <ezyang@cs.stanford.edu>

Test Plan: validate and new test cases

Reviewers: austin

Subscribers: carter, thomie

Differential Revision: https://phabricator.haskell.org/D596

GHC Trac Issues: #9243

Summary:
The isNothing maybe_old_linkable check predates
48bc81ad, which fixed #481 by requiring
recompilation information to be passed in as an argument to compileOne.
As a result, the check here is redundant: the client has already taken
a look at the object file to see if it is available or not.
Signed-off-by: Edward Z. Yang <ezyang@cs.stanford.edu>

Test Plan: validate

Reviewers: simonmar, austin

Subscribers: carter, thomie

Differential Revision: https://phabricator.haskell.org/D594

either one of the two recent commits (d8d00318, fd97d2a7) fixed it, or
there is some nondeterminism here. See #9938.

See Note [Replacement vs keeping].

There's a bit further to go with this change (to report unused givens).
But it's already an improvement; see the latent bug described in the Note.

I always found calls to TcCanonical.xCtEvidence hard to grok; and I
found that we only had two left. This patch eliminates them, along
with xCtEvidence, its accompanying comments, and the auxiliary
XEvTerm type.

The two remaining calls were these:

 * One was in newSCWorkFromFlavored, where we'd already done
   case-splitting for given/wanted/derived.  So inlining the xCtEvidence
   made the code simpler, clearer, and faster.

 * The other was in canTuple; here all of xCtEvidence's functionality
   was needed, but inlining again made a net gain in code size and
   clarity.

This tidies up all the comments about recursive superclasses
and when to add superclasses.  Lots of duplicate and contradictory
comments removed!

I ended up introducing a new sub-section on instance termination.

By passing -O2, the bug appears depending on the order of clauses in
"solve", hence adding T9938B as the other variant.

Currently, T9938 is marked as broken, but maybe the bug is actually in
T9938B, where something (possibly inlining, as suggested by rwbarton)
affected the requirement to link against transformers.

Marked as known_broken

Summary: Signed-off-by: Edward Z. Yang <ezyang@cs.stanford.edu>

Test Plan: validate

Reviewers: austin, erikd

Reviewed By: erikd

Subscribers: erikd, carter, thomie

Differential Revision: https://phabricator.haskell.org/D588

Summary:
Flag `-l:` is GNU ld specific and not supported by the
Mac OS X link editor. So we create a temporary file name
lib<tmpname>.<so_ext> and link with the standard -l<tmpname>
option on Linux and OS X.

Fixes #9875

Test Plan: validate on Mac OS X

Reviewers: austin, hvr, ezyang

Reviewed By: ezyang

Subscribers: carter, thomie, ezyang

Differential Revision: https://phabricator.haskell.org/D579

GHC Trac Issues: #9875

This was introduced in 1617a10a (re #5364)

Summary:
* Throw an error when cross-compiling without a target definition.
  When cross compiling via LLVM, a target 'datalayout' and 'triple' must
  be defined or LLVM will generate code for the compile host instead of
  the compile target.

* Add aarch64-unknown-linux-gnu target.
  The datalayout and triple lines were found by using clang to compile a
  small C program and -emit-llvm to get the LLVM IR output.
Signed-off-by: Erik de Castro Lopo <erikd@mega-nerd.com>

Test Plan: validate

Reviewers: rwbarton, carter, hvr, bgamari, austin

Reviewed By: austin

Subscribers: carter, thomie, garious

Differential Revision: https://phabricator.haskell.org/D585

GHC Trac Issues: #9895

This involves recognizing lines starting with `"pattern "` as declarations,
keeping non-exported pattern synonyms in `deSugar`, and including
pattern synonyms in the result of `hscDeclsWithLocation`.

This drops a couple of `-fno-warn-*` which seem to have become obsolete
by now.

Moreover, with the cleaned up settings `./validate` passes with GHC
7.10.1 as bootstrap compiler.

Summary:
This makes it easier to see what is exported, and allows us to
add non-exported top-level names.

Reviewers: hvr, austin, ezyang

Reviewed By: ezyang

Subscribers: ezyang, carter, thomie

Projects: #ghc

Differential Revision: https://phabricator.haskell.org/D551

GHC Trac Issues: #9852

This removes compile warnings triggered by those modules due to redundant
imports and/or due to tabs

Summary:
This allows GHC HEAD to be bootstrapped using 7.10.
Signed-off-by: Edward Z. Yang <ezyang@cs.stanford.edu>

Test Plan: validate

Reviewers: austin

Subscribers: carter, thomie

Differential Revision: https://phabricator.haskell.org/D589

GHC Trac Issues: #9652

Signed-off-by: Edward Z. Yang <ezyang@cs.stanford.edu>

This should address #9924 as GHC's config.guess/sub versions need to be
up to date anyway.

First attempt via 7a2c9dde wasn't working properly. This attempt
should work better as it doesn't cause the makefile recipe to fail which
causes `make` to emit additional varying output.

Summary:
I'm not really happy about perpetuating the hackish fix for #8696,
but at least in the context of building with -fhpc, the performance
cost should be negligible.

I'm suspicious about PlainModuleInitLabel and the Windows stuff too,
but I don't know what it does / can't test it (respectively) so I'll
leave those alone for now.

Hopefully out-of-process TH will save us from these hacks some day.

The test is an adaptation of T8696. It's a bit more awkward since
I couldn't think of a way to get cross-module tickbox references
without optimizations (inlining), but ghci doesn't permit -O for
some reason.

Test Plan: harbormaster; validate

Reviewers: austin

Reviewed By: austin

Subscribers: carter, thomie

Differential Revision: https://phabricator.haskell.org/D583

GHC Trac Issues: #9762

Conflicts:
	testsuite/tests/ghci/scripts/all.T