GitLab

This reverts commit 9e6e4796.

I reverted it because one of these two patches
  9e6e4796 Add purgeObj()
  b5e8b3b1 Make the linker API thread-safe

causes a seg-fault on Windows.  The seg-fault happens immediately
the linker is invoked, in ghci or in Template Haskell.

I believe that it is the "linker API thread-safe" commit that causes
the seg-fault; it happens even if the "purgeObj" commit alone is
reverted.  But since the two patches mess with the same code, to
revert the "linker API" patch I had revert both.

This replaces the note mentioning the lack of a `Natural`-type by a note
pointing to the new "Numeric.Natural" (#9818) module.

Don't pass empty string `[]` as "action-if-not-given" to `AC_ARG_ENABLE()`
macro, as this would otherwise lead to an empty else-block in the resulting
bash `configure` script. This bug was introduced via cb0a503a.

This issue was pointed out by @christiaanb

Reviewed By: christiaanb

Differential Revision: https://phabricator.haskell.org/D545

   {-# NOUNPACK #-}
   {-# NOUNPACK #-} !
were being parsed the same way.  The former was wrong.

Thanks to Alan Zimmerman for pointing this out

Fixed by e6a2050e
along with #9582, #9833

Summary:
Commit 37d64a51 removed the preprocessing step
for Parser.y.

Reviewers: rodlogic, austin

Reviewed By: austin

Subscribers: carter, thomie

Differential Revision: https://phabricator.haskell.org/D543

Sorry about this.  I somehow failed to include this one line in my patch.

This finally solves the issue of instance-method signatures that are
more polymorphic than the instanted class method.

See Note [Instance method signatures] in TcInstDcls.

A very nice fix for the two Trac tickets above.

Differential Revision: https://phabricator.haskell.org/D544

Reviewed By: austin

Differential Revision: https://phabricator.haskell.org/D541

This reverts commit 460eebec.

Thomas requested to revert the commit with the words:

> Please revert this commit, it is horribly wrong. I'll have a proper look
> later, but not supplying `-traditional` to the C preprocessor is the cause
> of #9828.

the reverted commit was related to #9094

Fwiw, this wasn't really a proper .lhs to begin with...

Reviewers: austin

Reviewed By: austin

Subscribers: thomie, carter

Differential Revision: https://phabricator.haskell.org/D540

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

Summary: This patch changes the error message as suggested in trac issue #8815 comments.

Reviewers: jstolarek, austin

Reviewed By: jstolarek, austin

Subscribers: jstolarek, thomie, carter

Differential Revision: https://phabricator.haskell.org/D533

GHC Trac Issues: #8815

Summary: Signed-off-by: Austin Seipp <austin@well-typed.com>

Test Plan: `./validate`

Reviewers: hvr

Subscribers: thomie, carter

Differential Revision: https://phabricator.haskell.org/D537

Summary:
In a statically linked GHCi symbol `environ` resolves to NULL when
called from a Haskell script.

When resolving symbols in a Haskell script we need to search the
executable program and its dependent (DT_NEEDED) shared libraries
first and then search the loaded libraries.

We want to be able to override functions in loaded libraries later.
Libraries must be opened with local scope (RTLD_LOCAL) and not global.
The latter adds all symbols to the executable program's symbols where
they are then searched in loading order. We want reverse loading order.

When libraries are loaded with local scope the dynamic linker
cannot use symbols in that library when resolving the dependencies
in another shared library. This changes the way files compiled to
object code must be linked into temporary shared libraries. We link
with the last temporary shared library created so far if it exists.
Since each temporary shared library is linked to the previous temporary
shared library the dynamic linker finds the latest definition of a
symbol by following the dependency chain.

See also Note [RTLD_LOCAL] for a summary of the problem and solution.

Cherry-picked commit 2f8b4c

Changed linker argument ordering

On some ELF systems GNU ld (and others?) default to
--as-needed and the order of libraries in the link
matters.

The last temporary shared library, must appear
before all other libraries. Switching the position
of extra_ld_inputs and lib_path_objs does that.

Fixes #8935 and #9186

Reviewers: austin, hvr, rwbarton, simonmar

Reviewed By: simonmar

Subscribers: thomie, carter, simonmar

Differential Revision: https://phabricator.haskell.org/D349

GHC Trac Issues: #8935, #9186, #9480

Summary:
The bash completion is simple but works both for ghc and ghci.
The README explains to the user what they have to do to get
it working (hopefully nothing).

Test Plan: Follow the README, then enjoy the cli completion in your terminal!

Reviewers: austin

Subscribers: thomie, carter, jstolarek

Differential Revision: https://phabricator.haskell.org/D536

GHC Trac Issues: #9005

Summary: Attaching semis to preceding AST element, not following

Test Plan: sh ./validate

Reviewers: hvr, austin

Reviewed By: austin

Subscribers: cactus, thomie, carter

Differential Revision: https://phabricator.haskell.org/D529

Summary:
Shorten long lines in DynFlags.
Describe --show-options in ghci usage guide.

Reviewers: jstolarek, austin

Reviewed By: jstolarek, austin

Subscribers: thomie, carter

Differential Revision: https://phabricator.haskell.org/D532

GHC Trac Issues: #9259

Summary:
Amazingly, the fix for this very old bug is quite simple: when type-checking,
maintain a set of "visible orphan modules" based on the orphans list of
modules which we explicitly imported.  When we import an instance and it
is an orphan, we check if it is in the visible modules set, and if not,
ignore it.  A little bit of refactoring for when orphan-hood is calculated
happens so that we always know if an instance is an orphan or not.

For GHCi, we preinitialize the visible modules set based on the list of
interactive imports which are active.

Future work: Cache the visible orphan modules set for GHCi, rather than
recomputing it every type-checking round.  (But it's tricky what to do when you
/remove/ a module: you need a data structure a little more complicated than
just a set of modules.)
Signed-off-by: Edward Z. Yang <ezyang@cs.stanford.edu>

Test Plan: new tests and validate

Reviewers: simonpj, austin

Subscribers: thomie, carter

Differential Revision: https://phabricator.haskell.org/D488

GHC Trac Issues: #2182

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

`gcdExtInteger` has been available since `integer-gmp-0.5.1`
(added via 71e29584)

This also exposes the following two type-specialised modular
exponentiation variants of `recipModInteger` useful for implementing a
`recipModNatural` operation.

  recipModBigNat :: BigNat -> BigNat -> BigNat
  recipModWord   :: Word#  -> Word#  -> Word#

`recipModInteger` has been available since `integer-gmp-0.5.1`
(added via 4d516855)

This makes use of the `powMod*` primitives provided by
`integer-gmp-1.0.0`. This is the `Natural`-version of the related
`GHC.Integer.GMP.Internals.powModInteger` operation.

The fallback implementation uses a square and multiply algorithm,
compared to which the optimized GMP-based implementation needs much less
allocations due to in-place mutation during the computation.

This also exposes the following type-specialised modular exponentiation
variants of `powModInteger` useful for implementing a `powModNatural`
operation.

  powModBigNat     :: BigNat -> BigNat -> BigNat -> BigNat
  powModBigNatWord :: BigNat -> BigNat -> Word#  -> Word#
  powModWord       :: Word#  -> Word#  -> Word#  -> Word#

`powModInteger` has been available since `integer-gmp-0.5.1`
(added via 4d516855)

A few instances were missed in 447f5926

Moreover, be more paranoid when testing for zero values, and try harder
to avoid passing denormalized zero `mpz_t`-values into GMP functions.

Summary:
Properly detect insoluble wanteds

This used to be correct, but was recently incorrectly refactored.

Reviewers: austin, hvr

Reviewed By: austin, hvr

Subscribers: thomie, carter

Differential Revision: https://phabricator.haskell.org/D531

Summary:
Add support for Partial Type Signatures, i.e. holes in types, see:
https://ghc.haskell.org/trac/ghc/wiki/PartialTypeSignatures

This requires an update to the Haddock submodule.

Test Plan: validate

Reviewers: austin, goldfire, simonpj

Reviewed By: simonpj

Subscribers: thomie, Iceland_jack, dominique.devriese, simonmar, carter, goldfire

Differential Revision: https://phabricator.haskell.org/D168

GHC Trac Issues: #9478

For ages NameSet has used different names,
  eg.   addOneToNameSet   rather than    extendNameSet
        nameSetToList     rather than    nameSetElems

etc.  Other set-like modules use uniform naming conventions.
This patch makes NameSet follow suit.

No change in behaviour; this is just renaming.

I'm doing this just before the fork so that merging is easier.

This patche fixes Trac #9574.

The previous Note [Renaming associated types] in RnTypes appears to me to be wrong;
it confused class and instance declarations.

I have:

 * Treated kind and type variables uniformly. Both must be bound on the LHS
   of an associated type instance.  Eg
       instance C ('KProxy :: KProxy o) where
          type F 'KProxy = NatTr (Proxy :: o -> *)
   is illegal because 'o' is not bound on the LHS of the instance.

 * Moved the Note to RnSource and fixed it up

This improves the error message from T7938. However it made the code in
T6118 incorrect. We had:
  instance SingE (a :: Maybe k) where
    type Demote a = Maybe (Demote (Any :: k))
and that is now rejected, rightly I think.

The `integerGmpInternals` test was disabled in
c774b28f as many of the primitives
tested in that test weren't available yet w/ `integer-gmp2`.

However, most operations have been reimplemented by now, with the
exception of

    recipModInteger  :: Integer -> Integer -> Integer
    gcdExtInteger    :: Integer -> Integer -> (Integer, Integer)
    powModSecInteger :: Integer -> Integer -> Integer -> Integer
    powModInteger    :: Integer -> Integer -> Integer -> Integer
    powInteger       :: Integer -> Word -> Integer

which are still missing, and will (time permitting) be reimplemented
over time.

This also adds `nextPrimeWord#` and `nextPrimeBigNat` predicates.

`nextPrimeInteger` has been available since `integer-gmp-0.5.1`
(added via f4973548).

This allows us to replace an object without actually unloading the old
object, which is necessary when we know we have references to the old
object so it can't be completely unloaded.  Using unloadObj() would
cause the GC (CheckUnload) to repeatedly and fruitlessly try to unload
the old object.

We used to be able to rely on the client to use the API in a
single-threaded way, but now that the GC calls into the linker to
unload objects this isn't a safe assumption.

This also adds `testPrimeWord#` and `testPrimeBigNat` predicates.

`testPrimeInteger` has been available since `integer-gmp-0.5.1`
(added via f4973548).
The `nextPrimeInteger` function is still missing though.

This is a knock-on from the -dump-to-file changes.
(I found that -ddump-cs-trace stuff wasn't coming out!)