GitLab

This is patch that adds support for interruptible FFI calls in the form
of a new foreign import keyword 'interruptible', which can be used
instead of 'safe' or 'unsafe'.  Interruptible FFI calls act like safe
FFI calls, except that the worker thread they run on may be interrupted.

Internally, it replaces BlockedOnCCall_NoUnblockEx with
BlockedOnCCall_Interruptible, and changes the behavior of the RTS
to not modify the TSO_ flags on the event of an FFI call from
a thread that was interruptible.  It also modifies the bytecode
format for foreign call, adding an extra Word16 to indicate
interruptibility.

The semantics of interruption vary from platform to platform, but the
intent is that any blocking system calls are aborted with an error code.
This is most useful for making function calls to system library
functions that support interrupting.  There is no support for pre-Vista
Windows.

There is a partner testsuite patch which adds several tests for this
functionality.

We used to define these CPP symbols, but nothing on hackage uses them
and the first 2 are no longer correct (as we support multiple Haskell
versions).

The default language is now Haskell2010, so this was a little odd.
Also, --make is now on by default, so this was largely irrelevant.

This means most code doesn't get caught by monomorphic local bindings.

Instead of keeping a *set* of untouchable variables in each
implication contraints, we keep a *range* of uniques for the
*touchable* variables of an implication.  This are precisely
the ones we would call the "existentials" if we were French.

It turns out that the code is more efficient, and vastly easier
to get right, than the set-based approach.

Fixes Trac #4355 among others

This fixes Trac #4361.  In a rather delicate way, but
no more delicate than before.  A more remoseless typechecker
would reject #4361 altogether.

See Note [Avoid unecessary constraint simplification]

Note [Avoid unecessary constraint simplification] in TcSimplify

This fixes at least one case of performance regression in 7.0, and
is nice win on nofib:

        Program           Size    Allocs   Runtime   Elapsed
            Min          +0.3%    -63.0%    -38.5%    -38.7%
            Max          +1.2%     +0.2%     +0.9%     +0.9%
 Geometric Mean          +0.6%     -3.0%     -6.4%     -6.6%

The error cascade caused puzzling errors in T4093b, and
suppressing some seems like a good plan.  Very few test
outputs change.

This big-ish patch arranges that if an Id 'f' is 
  * Type-class overloaded 
       f :: Ord a => [a] -> [a]
  * Defined with an INLINABLE pragma
       {-# INLINABLE f #-}
  * Exported from its defining module 'D'

then in any module 'U' that imports D

1. Any call of 'f' at a fixed type will generate 
   (a) a specialised version of f in U
   (b) a RULE that rewrites unspecialised calls to the
       specialised on

  e.g. if the call is (f Int dOrdInt xs) then the 
  specialiser will generate
     $sfInt :: [Int] -> [Int]
     $sfInt = <code for f, imported from D, specialised>
     {-# RULE forall d.  f Int d = $sfInt #-}

2. In addition, you can give an explicit {-# SPECIALISE -#}
   pragma for the imported Id
     {-# SPECIALISE f :: [Bool] -> [Bool] #-}
   This too generates a local specialised definition, 
   and the corresponding RULE 

The new RULES are exported from module 'U', so that any module
importing U will see the specialised versions of 'f', and will
not re-specialise them.

There's a flag -fwarn-auto-orphan that warns you if the auto-generated
RULES are orphan rules. It's not in -Wall, mainly to avoid lots of
error messages with existing packages.

Main implementation changes

 - A new flag on a CoreRule to say if it was auto-generated.
   This is persisted across interface files, so there's a small
   change in interface file format.

 - Quite a bit of fiddling with plumbing, to get the 
   {-# SPECIALISE #-} pragmas for imported Ids.  In particular, a
   new field tgc_imp_specs in TcGblEnv, to keep the specialise
   pragmas for imported Ids between the typechecker and the desugarer.

 - Some new code (although surprisingly little) in Specialise,
   to deal with calls of imported Ids

I don't think the type distinction of declaring NameEnv with a newtype
(as it was) is really useful to us. Moreover, VarEnv is a UniqFM, and
I do sometimes want to build an envt with Ids and look up with Names.

This may not be the last word on the subject.

Previously it was rejecting the match

  Template: forall s t. map s t
  Actual:   map Int t

which should obviously be fine.  It turns out that this kind of match
comes up when specialising.  By freshening that t we could avoid the
difficulty, but morally the (forall t) binds t and the rule should
be alpha-equivalent regardless of the forall'd variables.

This patch makes it so, and incidentally makes matching a little
more efficient.  See Note [Eta expansion] in VarEnv.

This is another long-standing bug, in which there was a possibility
that a loop-breaker could lose its loop-breaker-hood OccInfo, 
and then the simplifer re-simplified the expression. Result, either
non-termination or, in the case of #4345, an unbound identifier.

The fix is very simple, in Id.transferPolyIdInfo. 
See Note [transferPolyIdInfo].

When adding specialisation for imported Ids, I noticed that the
Glorious Simplifier was repeatedly (and fruitlessly) simplifying the
same term.  It turned out to be easy to fix this, because I already
had a flag in the ApplyTo and Select constructors of SimplUtils.SimplCont.

See Note [Avoid redundant simplification]

This patch fixes a long-standing lurking bug, but it surfaced when I
was adding specialisation for imported Ids.

See Note [ImpRuleUsage], which explains the issue.   The solution
seems more complicated than the problem really deserves, but I
could not think of a simpler way, so I just bit the bullet and
wrote the code.  Improvements welcome.

(At least, I don't think there is anything else.)

I added a lot of comments too, to explain the preconditions;
esp Note [FamInstEnv]

The tag2Enum rule wasn't doing the right thing for
enumerations with a phantom type parameter, like
   data T a = A | B

Plus remember to zonk the free_tvs in TcUnify.newImplication