GitLab

Happily, the fix is simply deleting some old code. I love it when
that happens.

This may have performance implications.

We need to instantiate types in tuples. Quite straightforward.

When compiler_debugged only. This makes the Travis build green again.

With 399a5b46, the old `[$foo| ... |]`
syntax for quasi-quotes is no longer allowed.

This re-working of the typechecker algorithm is based on
the paper "Visible type application", by Richard Eisenberg,
Stephanie Weirich, and Hamidhasan Ahmed, to be published at
ESOP'16.

This patch introduces -XTypeApplications, which allows users
to say, for example `id @Int`, which has type `Int -> Int`. See
the changes to the user manual for details.

This patch addresses tickets #10619, #5296, #10589.

Given:
 * `file = "foo.a.b"`
 * `osuf = ".a.b"`  -- Note the initial dot.
 * `new_osuf = "c"`

Before (bad, the last character of the filename is dropped):
  `dropTail (length osuf + 1) file <.> new_osuf == "fo.c"`
After (good):
  `stripExtension osuf file <.> new_osuf` == "foo.c"

This regression was introduced in commit c489af73 (#5554). That commit
fixed a similar but different bug, and care has been taken to not
reintroduce it (using the the newly introduced
`System.Filepath.stripExtension`).

Given:
 * `file = "foo.a.b"`
 * `osuf = "a.b"`
 * `new_osuf = "c"`

Before c489af73 (bad, the full suffix should get replaced):
  `replaceExtension file new_osuf == "foo.a.c"`
After c489af73 (good):
  `dropTail (length osuf + 1) file <.> new_osuf == "foo.c"`
After this commit (still good):
  `stripExtension osuf file <.> new_osuf == "foo.c"`

Reviewed by: bgamari

Differential Revision: https://phabricator.haskell.org/D1692

GHC Trac Issues: #9760

Test Plan: run tests on powerpc and x86_64

Reviewers: hvr, austin, thomie, bgamari

Reviewed By: thomie, bgamari

Subscribers: thomie

Differential Revision: https://phabricator.haskell.org/D1694

This moves code around to more sensible places.

- Construction for CoAxiom is localised in FamInstEnv

- orphNamesOfxx moves to CoreFVs

- roughMatchTcs, instanceCantMatch moves to Unify

- mkNewTypeCo moves from Coercion to FamInstEnv, and is
  renamed mkNewTypeCoAxiom, which makes more sense

Instead of substituting, just augment the environment.
Less code, more efficient.

And the previous version had a bogus in-scope set which
triggered a WARNING

There were two problems here:
 - We were looking under a lambda without extending
   the in-scope env, which triggered a WARNING
   But there's no need to look under a lambda.

 - We were looking under a letrec without extending
   the in-scope env, which triggered the same WARNING
   Solution: extend the in-scope env

This is still needed by the Shake build system until things have been
reworked to use Cabal directly.

The Cortex A8 hardware apparently has a bug which ld.gold will try to
correct; however in order to do so it must have unstripped executables
lest we see warnings of the form (see #10376, #10464),

    /usr/bin/ld.gold: warning: cannot scan executable section 1 of ...
    for Cortex-A8 erratum because it has no mapping symbols.

Consequently we disabling stripping by default on this architecture.

A bit more discussion about this issue can be found in this [Android
issue](http://code.google.com/p/android/issues/detail?id=40794).

Test Plan: Try validating on ARM

Reviewers: erikd, austin, thomie

Reviewed By: austin, thomie

Differential Revision: https://phabricator.haskell.org/D1599

GHC Trac Issues: #10376, #10464

Python 3 support seems to have mildly bitrotten since #9184 was closed.
Luckily, only some minor tweaks seem necessary.

Since commit 03c4893e, ic_int_print is
retained from registered packages after `:reload`.

Fixes #11159.

Summary: A regression test for #10432, which seems to already be fixed.

Test Plan: ./validate

Reviewers: simonpj, austin, bgamari

Subscribers: thomie

Differential Revision: https://phabricator.haskell.org/D1691

GHC Trac Issues: #10432

A PatBind operates similarly to a FunBind, and so the RdrName and
infix/prefix state needs to be captured for use in API Annotations.

This introduces "freezing," an operation which prevents further
locations from being appended to a CallStack.  Library authors may want
to prevent CallStacks from exposing implementation details, as a matter
of hygiene. For example, in

```
head [] = error "head: empty list"

ghci> head []
*** Exception: head: empty list
CallStack (from implicit params):
  error, called at ...
```

including the call-site of `error` in `head` is not strictly necessary
as the error message already specifies clearly where the error came
from.

So we add a function `freezeCallStack` that wraps an existing CallStack,
preventing further call-sites from being pushed onto it. In other words,

```
pushCallStack callSite (freezeCallStack callStack) = freezeCallStack callStack
```

Now we can define `head` to not produce a CallStack at all

```
head [] =
  let ?callStack = freezeCallStack emptyCallStack
  in error "head: empty list"

ghci> head []
*** Exception: head: empty list
CallStack (from implicit params):
  error, called at ...
```

---

1. We add the `freezeCallStack` and `emptyCallStack` and update the
   definition of `CallStack` to support this functionality.

2. We add `errorWithoutStackTrace`, a variant of `error` that does not
   produce a stack trace, using this feature. I think this is a sensible
   wrapper function to provide in case users want it.

3. We replace uses of `error` in base with `errorWithoutStackTrace`. The
   rationale is that base does not export any functions that use CallStacks
   (except for `error` and `undefined`) so there's no way for the stack
   traces (from Implicit CallStacks) to include user-defined functions.
   They'll only contain the call to `error` itself. As base already has a
   good habit of providing useful error messages that name the triggering
   function, the stack trace really just adds noise to the error. (I don't
   have a strong opinion on whether we should include this third commit,
   but the change was very mechanical so I thought I'd include it anyway in
   case there's interest)

4. Updates tests in `array` and `stm` submodules

Test Plan: ./validate, new test is T11049

Reviewers: simonpj, nomeata, goldfire, austin, hvr, bgamari

Reviewed By: simonpj

Subscribers: thomie

Projects: #ghc

Differential Revision: https://phabricator.haskell.org/D1628

GHC Trac Issues: #11049

This closes Trac #10897

Fixes Trac #11278

Summary:
This patch fixes gc_thread related compilation failure
on Solaris/i386 platform. It uses Linux way of __thread declared
gc_thread variable for register starving i386 from now.

Reviewers: bgamari, austin, erikd

Subscribers: thomie

Differential Revision: https://phabricator.haskell.org/D1688

Summary:
The *-ws-32 file were too difficult to keep up-to-date so @bgamari
updated the test outut normalization code in commit 786d528e
to make these files un-necessary.

Test Plan: test on PowerPC

Reviewers: hvr, austin, bgamari

Reviewed By: bgamari

Subscribers: thomie, bgamari

Differential Revision: https://phabricator.haskell.org/D1690

Added test and changed -Nmax to -maxN, -n was taken

Noticed strange -m behavoir and fixed -m from quietly
ignoring being passed invalid opts, e.g. "-msasd"

Reviewers: simonmar, hvr, austin, thomie, bgamari

Reviewed By: hvr, thomie, bgamari

Subscribers: bgamari, hvr, thomie, simonmar

Differential Revision: https://phabricator.haskell.org/D1677

GHC Trac Issues: #10728

I found it was possible to do this a bit more nicely
See Note [Family instance declaration binders] in HsDecls, and
    Note [Wildcards in family instances] in RnSource

The main change is to add PatSynPE to PromotionErr, so that
when we get an ill-staged use of a pattern synonym we get a
civilised error message.

We were already doing this in half-baked form in tcValBinds, but
this patch tidies up the impl (which previously used a hack rather
than APromotionErr), and does it in tcTyClsInstDecls too.

I'm reverting this until we agree a design.
See comment:5 in Trac #9793.

Incidentally the reference to Trac #9739 in the reverted
patch is bogus; it shold have said #9793.

This reverts commit 44640af7.

Summary:
This patch adds _DYNAMIC symbol to the list of OpenBSD symbols.
The patch fixes unknown _DYNAMIC symbol runtime linker error caused
by recent update of unix library.

Reviewers: bgamari, austin

Subscribers: thomie

Differential Revision: https://phabricator.haskell.org/D1689

Summary:
The basic idea is that we have a new set of "exposed modules"
which are /only/ used for plugins, i.e. -fplugin Foo and
--frontend Foo.  You can interact with this namespace
using the flags -plugin-package-id and -plugin-package.
By default, this namespace contains all modules in the
user namespace (as before), but you can toggle that using
-hide-all-plugin-packages.

There is one nasty hack: GhcMake respects -fplugin in
GHC_OPTIONS to make local plugins work correctly.  It also
bails out of you have an import of a module which doesn't
exist locally or in the package database.  The upshot is
that we need to be sure to check in the plugin modules
too, so we don't give a spurious failure when a plugin
is in the plugin namespace but not the main namespace.
A better way to fix this would be to distinguish between
plugin and normal dependencies in ModSummary.

I cheated a little and tweaked a few existing plugins
tests to exercise the new code paths.

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

Test Plan: validate

Reviewers: bgamari, austin, simonpj, duncan

Subscribers: thomie

Differential Revision: https://phabricator.haskell.org/D1661

GHC Trac Issues: #11244

Summary:
Previously, all package flags (-package, -trust-package,
-ignore-package) were bundled up into a single packageFlags
field in DynFlags, under a single type.  This commit separates
them based on what they do.

This is a nice improvement, because it means that Packages can
then be refactored so that a number of functions are "tighter":

    - We know longer have to partition PackageFlags into
      the ignore flag and other flags; ignore flags are just
      put into their own field.

    - Trust flags modify the package database, but exposed
      flags do not (they modify the visibility map); now
      applyPackageFlag and applyTrustFlag have tighter signatures
      which reflect this.

This patch was motivated by the need to have a separate visibility
map for plugin packages, which will be in a companion patch.
Signed-off-by: Edward Z. Yang <ezyang@cs.stanford.edu>

Test Plan: validate

Reviewers: austin, bgamari, duncan

Subscribers: thomie

Differential Revision: https://phabricator.haskell.org/D1659