Ryan Scott authored 7 years ago and David Feuer committed 7 years ago

GHC 8.4.1 is out, so now GHC's support window only extends
back to GHC 8.2. This means we can delete gobs of code that were
only used for GHC 8.0 support. Hooray!

Test Plan: ./validate

Reviewers: bgamari, erikd, dfeuer

Reviewed By: bgamari, dfeuer

Subscribers: alexbiehl, dfeuer, rwbarton, thomie, carter

Differential Revision: https://phabricator.haskell.org/D4492

[skip ci]

[skip ci]

T13615 needs multicore support from RTS:
    T13615: unknown RTS option: -N15

Signed-off-by: Sergei Trofimovich <slyfox@gentoo.org>

When interpreter is not profiled (see `interpreterProfiled` in
`DynFlags`) bytecode generator generates a NULL pointer as the cost
centre of a `BRK_FUN` instruction:

    let cc | interpreterProfiled dflags = cc_arr ! tick_no
           | otherwise = toRemotePtr nullPtr
    let breakInstr = BRK_FUN (fromIntegral tick_no) (getUnique this_mod) cc
    return $ breakInstr `consOL` code

We now take this into account when disassembling `BRK_FUN`.

Reviewers: bgamari, simonmar, erikd

Subscribers: rwbarton, thomie, carter

Differential Revision: https://phabricator.haskell.org/D4490

Unreg build failed as:

  $ ./configure --enable-unregisterised
  $ make

  HC [stage 1] libraries/ghc-prim/dist-install/build/GHC/PrimopWrappers.o
    ghc_1.hc: In function 'ghczmprim_GHCziPrimopWrappers_pdep8zh_entry':

    ghc_1.hc:1810:9: error:
     error: implicit declaration of function 'hs_pdep8'; did you mean 'hs_ctz8'?
       [-Werror=implicit-function-declaration]
     _c3jz = hs_pdep8(*Sp, Sp[1]);
             ^~~~~~~~
             hs_ctz8
       |
  1810 | _c3jz = hs_pdep8(*Sp, Sp[1]);
       |         ^

Signed-off-by: Sergei Trofimovich <slyfox@gentoo.org>

Test Plan: New unit test

Reviewers: andrewthad, niteria, bgamari, erikd

Subscribers: rwbarton, thomie, carter

GHC Trac Issues: #14900

Differential Revision: https://phabricator.haskell.org/D4485

Ben Gamari authored 7 years ago and Ben Gamari committed 7 years ago

Reviewers: hvr

Subscribers: rwbarton, thomie, erikd, carter

Differential Revision: https://phabricator.haskell.org/D4483

In commit 503219e3, we
stopped suppressing `-Wmissing-methods` warnings on class methods
whose names begin with an underscore. However, it seems the users'
guide documentation concerning this was never updated. Let's do so.

Test Plan: Read it

Reviewers: bgamari

Reviewed By: bgamari

Subscribers: rwbarton, thomie, carter

GHC Trac Issues: #12959

Differential Revision: https://phabricator.haskell.org/D4476

Ben Gamari authored 7 years ago and Ben Gamari committed 7 years ago

GCC throws this warning to inform us that __sync_fetch_and_nand's
behavior changed in GCC 4.4. However, this causes the build to fail when
-Werror is used.

Test Plan: Validate with -Werror

Subscribers: rwbarton, thomie, carter

Differential Revision: https://phabricator.haskell.org/D4481

Ben Gamari authored 7 years ago and Ben Gamari committed 7 years ago

As described in https://bugs.llvm.org/show_bug.cgi?id=8842, Clang
removed the __sync_fetch_and_nand builtins due to inconsistency in GCC's
behavior in 2010.  However, GCC has since clarified the behavior of
their builtins and consequently Clang re-added them in 2014.
Consequently this workaround should no longer be necessary.

Test Plan: Validate building with Clang

Subscribers: rwbarton, thomie, carter

Differential Revision: https://phabricator.haskell.org/D4480

- Show a more friendly error message when -fplugin is used with
  -fexternal-interpreter

- Add a few words to users guide about the interaction with -fplugin and
  -fexternal-interpreter

- Update test for #14335

Reviewers: bgamari

Reviewed By: bgamari

Subscribers: rwbarton, thomie, carter

GHC Trac Issues: #14335

Differential Revision: https://phabricator.haskell.org/D4456

This reverts commit 59d7ee53 and enables
the test for #14052.

(See #14052 for the discussion)

Reviewers: bgamari

Reviewed By: bgamari

Subscribers: rwbarton, thomie, carter

GHC Trac Issues: #14052

Differential Revision: https://phabricator.haskell.org/D4478

Summary:
The `-dynamic` flag does two things:

* In the code generator, it generates code designed to link against
  external shared libraries.  References outside of the current module
  go through platform-specific indirection tables (e.g. the GOT on ELF).

* It enables a "way", which changes which hi files we look
  for (`Foo.dyn_hi`) and which libraries we link against.

Some specialised applications want the first of these without the
second. (I could go into detail here but it's probably not all that
important).

This diff splits out the code-generation effects of `-dynamic` from the
"way" parts of its behaviour, via a new flag `-fexternal-dynamic-refs`.

Test Plan: validate

Reviewers: niteria, bgamari, erikd

Subscribers: rwbarton, thomie, carter

Differential Revision: https://phabricator.haskell.org/D4477

Simon Jakobi authored 7 years ago and Ben Gamari committed 7 years ago

Windows support was added in c93813d9

That image creates an unprivileged user to run the test suite under.

Simon Marlow authored 7 years ago and Ben Gamari committed 7 years ago

This was broken by D3746 and/or D3809, but unfortunately we didn't
notice because CI at the time wasn't building the profiling way.

Test Plan:
```
cd testsuite/test/profiling/should_run
make WAY=ghci-ext-prof
```

Reviewers: bgamari, michalt, hvr, erikd

Subscribers: rwbarton, thomie, carter

GHC Trac Issues: #14705

Differential Revision: https://phabricator.haskell.org/D4437

`accumArray` was lazier than documented. `accum` did not have
documented strictness. The extra laziness allowed thunks to build
up in the array. Force the results of applying the accumulating
function to resolve.

Reviewers: hvr, bgamari

Reviewed By: bgamari

Subscribers: alpmestan, rwbarton, thomie, carter

GHC Trac Issues: #14785

Differential Revision: https://phabricator.haskell.org/D4403

Michal Terepeta authored 7 years ago and Ben Gamari committed 7 years ago

This basically replaces all uses of `foldl` with `foldl'`. I've looked
at all the call sites and there doesn't seem to be any reason to prefer
the lazy version.

Signed-off-by: Michal Terepeta <michal.terepeta@gmail.com>

Test Plan: ./validate

Reviewers: bgamari, simonmar

Reviewed By: bgamari

Subscribers: rwbarton, thomie, carter

Differential Revision: https://phabricator.haskell.org/D4463

Since 7aaeaf81
`-g1` and `-g2` are actually different.
The tutorial below is correct.

Test Plan: harbormaster

Reviewers: bgamari, simonmar

Reviewed By: bgamari

Subscribers: rwbarton, thomie, carter

Differential Revision: https://phabricator.haskell.org/D4471

This reverts f5b275a2
and changes the places that looked for `Lit (MachStr _))`
to use `exprIsMbTickedLitString_maybe` to unwrap ticks as
necessary.
Also updated relevant comments.

Test Plan:
I added 3 new tests that previously reproduced.
GHC HEAD now builds with -g

Reviewers: simonpj, simonmar, bgamari, hvr, goldfire

Subscribers: rwbarton, thomie, carter

GHC Trac Issues: #14779

Differential Revision: https://phabricator.haskell.org/D4470

(the CPP guard is already wrapped with the same guard in line 1549)

Summary:
This came up when trying to build GHC HEAD with nix. We
do not set CCX for integer-gmp when running ./configure. We do
this however for libffi.

The result is, that if CCX is not set, we default to the system
one, of which there might be none (as in nixos's case).  This
will not show on a debian+nix or similar setup, where the system
`cxx` is still in place, and only shows up when the system tries
hard to sandbox everything (even cxx) as nixOS does.

We use `CXX`, which is set to either `clang` or `CC_STAGE1`, and
also usedfor `CC`, similar to what we do for libffi.c

Test Plan: ./validate

Reviewers: bgamari, hvr

Reviewed By: hvr

Subscribers: rwbarton, thomie, carter

Differential Revision: https://phabricator.haskell.org/D4473

Reviewers: trofi, bgamari

Reviewed By: trofi, bgamari

Subscribers: rwbarton, thomie, carter

Differential Revision: https://phabricator.haskell.org/D4467

This wasn't spotted rightaway in
8f0b2f5e
because the include-site deliberately ignored include-errors as
a Hack with the justification below:

> Hack. The file gmp/config.mk doesn't exist yet after running ./configure in
> the toplevel (ghc) directory. To let some toplevel make commands such as
> sdist go through, right after ./configure, don't consider this an error.

This may have contributed to #14891.

This patch is preparatory for the main fix for Trac #13324

Here, we simplify rnLHsInstType so that it does not try
to figure out the class name.  This turns out to have a good
(rather than bad) effect on error messages, and it prepares
the way for the main event.

Plus, less code!

Summary:
There were previously some situations where `(->)` would
not be desugared or reified as `ArrowT`, leading to various oddities
such as those observed in #14888. We now uniformly treat `(->)` as
`ArrowT` in Template Haskell–world by checking for any tycon that
has the same name as `(->)`, and converting that to `ArrowT`.

Test Plan: make test TEST=T14888

Reviewers: goldfire, bgamari, simonpj

Reviewed By: goldfire, simonpj

Subscribers: simonpj, rwbarton, thomie, carter

GHC Trac Issues: #14888

Differential Revision: https://phabricator.haskell.org/D4466

I tried to do this with

    commit 0a12d92a
    Author: Simon Peyton Jones <simonpj@microsoft.com>
    Date:   Wed Dec 13 12:53:26 2017 +0000

    Further improvements to well-kinded types

    The typechecker has the invariant that every type should be well-kinded
    as it stands, without zonking.  See Note [The well-kinded type invariant]
    in TcType.

    That invariant was not being upheld, which led to Trac #14174.  I fixed
    part of it, but T14174a showed that there was more.  This patch finishes
    the job.

But I didn't get it quite right as Trac #14873 showed.

This patch fixes the problem; although I am still a bit unhappy.
(See "A worry" in the HsApp case of tc_infer_hs_type.)

Comment:4 in Trac #14808 explains why I'm unhappy with the current
state of affairs -- at least the lack of documentation.

This smallpatch does nothing major:

* adds comments
* uses existing type synonyms more (notably FreeKiTyVarsWithDups)
* adds another test case to T14808

Moritz Angermann authored 7 years ago and Ben Gamari committed 7 years ago

Requires some ghc-cabal changes as well.

Test Plan: ./validate

Reviewers: hvr, bgamari

Reviewed By: bgamari

Subscribers: hsyl20, erikd, alpmestan, rwbarton, thomie, carter

Differential Revision: https://phabricator.haskell.org/D4453

This should have been part of 5f6fcf78.