GitLab

Add new flag '-Wredundant-bang-patterns' that enables checks for "dead" bangs.
Dead bangs are the ones that under no circumstances can force a thunk that
wasn't already forced. Dead bangs are a form of redundant bangs. The new check
is performed in Pattern-Match Coverage Checker along with other checks (namely,
redundant and inaccessible RHSs). Given

    f :: Bool -> Int
    f True = 1
    f !x   = 2

we can detect dead bang patterns by checking whether @x ~ ⊥@ is satisfiable
where the PmBang appears in 'checkGrdTree'. If not, then clearly the bang is
dead. Such a dead bang is then indicated in the annotated pattern-match tree by
a 'RedundantSrcBang' wrapping. In 'redundantAndInaccessibles', we collect
all dead bangs to warn about.

Note that we don't want to warn for a dead bang that appears on a redundant
clause. That is because in that case, we recommend to delete the clause wholly,
including its leading pattern match.

Dead bang patterns are redundant. But there are bang patterns which are
redundant that aren't dead, for example

    f !() = 0

the bang still forces the match variable, before we attempt to match on (). But
it is redundant with the forcing done by the () match. We currently don't
detect redundant bangs that aren't dead.

Addresses #18422.

This is no longer necessary as there are now dedicated testsuite ways
which run tests with WinIO.

On CI (x86_64-linux-deb9-hadrian, compile_time/bytes_allocated):

    T10421     -1.8%    (threshold: +/- 1%)
    T10421a    -1.7%    (threshold: +/- 1%)
    T12150     -4.9%    (threshold: +/- 2%)
    T12227     -1.6     (threshold: +/- 1%)
    T12425     -1.5%    (threshold: +/- 1%)
    T12545     -3.8%    (threshold: +/- 1%)
    T12707     -3.0%    (threshold: +/- 1%)
    T13035     -3.0%    (threshold: +/- 1%)
    T14683     -10.3%   (threshold: +/- 2%)
    T3064      -6.9%    (threshold: +/- 2%)
    T4801      -4.3%    (threshold: +/- 2%)
    T5030      -2.6%    (threshold: +/- 2%)
    T5321FD    -3.6%    (threshold: +/- 2%)
    T5321Fun   -4.6%    (threshold: +/- 2%)
    T5631      -19.7%   (threshold: +/- 2%)
    T5642      -13.0%   (threshold: +/- 2%)
    T783       -2.7     (threshold: +/- 2%)
    T9020      -11.1    (threshold: +/- 2%)
    T9961      -3.4%    (threshold: +/- 2%)

    T1969 (compile_time/bytes_allocated)  -2.2%  (threshold: +/-1%)
    T1969 (compile_time/max_bytes_used)   +24.4% (threshold: +/-20%)

Additionally on other CIs:

    haddock.Cabal                  -10.0%   (threshold: +/- 5%)
    haddock.compiler               -9.5%    (threshold: +/- 5%)
    haddock.base (max bytes used)  +24.6%   (threshold: +/- 15%)
    T10370 (max bytes used, i386)  +18.4%   (threshold: +/- 15%)

Metric Decrease:
    T10421
    T10421a
    T12150
    T12227
    T12425
    T12545
    T12707
    T13035
    T14683
    T3064
    T4801
    T5030
    T5321FD
    T5321Fun
    T5631
    T5642
    T783
    T9020
    T9961
    haddock.Cabal
    haddock.compiler
Metric Decrease 'compile_time/bytes allocated':
    T1969
Metric Increase 'compile_time/max_bytes_used':
    T1969
    T10370
    haddock.base

Since Backpack the "home unit" is much more involved than what it was
before (just an identifier obtained with `-this-unit-id`). Now it is
used in conjunction with `-component-id` and `-instantiated-with` to
configure module instantiations and to detect if we are type-checking an
indefinite unit or compiling a definite one.

This patch introduces a new HomeUnit datatype which is much easier to
understand. Moreover to make GHC support several packages in the same
instances, we will need to handle several HomeUnits so having a
dedicated (documented) type is helpful.

Finally in #14335 we will also need to handle the case where we have no
HomeUnit at all because we are only loading existing interfaces for
plugins which live in a different space compared to units used to
produce target code. Several functions will have to be refactored to
accept "Maybe HomeUnit" parameters instead of implicitly querying the
HomeUnit fields in DynFlags. Having a dedicated type will make this
easier.

Bump haddock submodule

Where bindings can see evidence from the pattern match of the `GRHSs`
they belong to, but not from anything in any of the guards (which belong
to one of possibly many RHSs).

Before this patch, we did *not* consider said evidence, causing #18533,
where the lack of considering type information from a case pattern match
leads to failure to resolve the vanilla COMPLETE set of a data type.

Making available that information required a medium amount of
refactoring so that `checkMatches` can return a
`[(Deltas, NonEmpty Deltas)]`; one `(Deltas, NonEmpty Deltas)` for each
`GRHSs` of the match group. The first component of the pair is the
covered set of the pattern, the second component is one covered set per
RHS.

Fixes #18533.
Regression test case: T18533

As suggested in #18545.

The note has been rewritten by @simonpj in !3851

[skip ci]

T16916 (testing #16916) has been slightly fragile in CI due to its
reliance on CPU times. While it's hard to see how to eliminate
the time-dependence entirely, we can nevertheless make it more tolerant.

Fixes #16966.

- put panic related functions into GHC.Utils.Panic
- put trace related functions using DynFlags in GHC.Driver.Ppr

One step closer making Outputable fully independent of DynFlags.

Bump haddock submodule

Previously it collected everything, including "max bytes used". This is
problematic since the test makes no attempt to control for deviations in
GC timing, resulting in high variability. Fix this by only collecting
"bytes allocated".

This enables IDE support by haskell-language-server for ghc-heap.

Previously the desugarer would instead fall over when it realized that
there was no unfolding for an imported function with a SPECIALISE
pragma. We now rather drop the SPECIALISE pragma and throw a warning.

Fixes #18118.

Also fix its slightly wrong comment

Metric Decrease:
    T5030
    T12227
    T12545

As suspected by @simonpj in #18535, avoiding allocations in
`GHC.Utils.Misc.splitAtList` when there are no leftover arguments is
beneficial for performance:

   On CI validate-x86_64-linux-deb9-hadrian:
    T12227 -7%
    T12545 -12.3%
    T5030  -10%
    T9872a -2%
    T9872b -2.1%
    T9872c -2.5%

Metric Decrease:
    T12227
    T12545
    T5030
    T9872a
    T9872b
    T9872c

And prefix ~

(cherry picked from commit 8dbee2c578b1f642d45561be3f416119863e01eb)

We want to only run the check if ld is gold.

Fixes the fix to #17962.

Remove unused ApiAnns, add one for linear arrow.

Include API Annotations for trailing comma in export list.

Previously it failed as the `ghc` package was not visible.

Previously the code generator could produce corrupt C call sequences due
to register overlap between MachOp lowerings and the platform's calling
convention. We fix this using a hack described in Note [Evaluate C-call
arguments before placing in destination registers].

As mentioned in Note [Register parameter passing] the arguments of
foreign calls cannot refer to caller-saved registers.

as suggested by comments on !2330.

GHC 8.12.1 has been renamed to GHC 9.0.1.

See also:
  https://mail.haskell.org/pipermail/ghc-devs/2020-July/019083.html

[skip ci]

Close #18534.

See commentary in the patch.

Before this patch, this type:
  T :: forall k -> (k ~ k) => forall j -> k -> j -> Type
was printed incorrectly as:
  T :: forall k j -> (k ~ k) => k -> j -> Type

Before this change, GHC would
pretty-print   forall k. forall a -> ()
          as   forall @k a. ()
which isn't even valid Haskell.

This patch started as a small documentation change, an attempt to make
Note [Parser-Validator] and Note [Ambiguous syntactic categories]
more clear and up-to-date.

But it turned out that runECP_P/runECP_PV are weakly motivated,
and it's easier to remove them than to find a good rationale/explanation
for their existence.

As the result, there's a bit of refactoring in addition to
a documentation update.

Before this patch, we parsed types into a reversed sequence
of operators and operands. For example, (F x y + G a b * X)
would be parsed as [X, *, b, a, G, +, y, x, F],
using a simple grammar:

	tyapps
	  : tyapp
	  | tyapps tyapp

	tyapp
	  : atype
	  | PREFIX_AT atype
	  | tyop
	  | unpackedness

Then we used a hand-written state machine to assemble this
 either into a type,        using 'mergeOps',
     or into a constructor, using 'mergeDataCon'.

This is due to a syntactic ambiguity:

	data T1 a =          MkT1 a
	data T2 a = Ord a => MkT2 a

In T1, what follows after the = sign is a data/newtype constructor
declaration. However, in T2, what follows is a type (of kind
Constraint). We don't know which of the two we are parsing until we
encounter =>, and we cannot check for => without unlimited lookahead.

This poses a few issues when it comes to e.g. infix operators:

	data I1 = Int :+ Bool :+ Char          -- bad
	data I2 = Int :+ Bool :+ Char => MkI2  -- fine

By this issue alone we are forced into parsing into an intermediate
representation and doing a separate validation pass.

However, should that intermediate representation be as low-level as a
flat sequence of operators and operands?

Before GHC Proposal #229, the answer was Yes, due to some particularly
nasty corner cases:

	data T = ! A :+ ! B          -- used to be fine, hard to parse
	data T = ! A :+ ! B => MkT   -- bad

However, now the answer is No, as this corner case is gone:

	data T = ! A :+ ! B          -- bad
	data T = ! A :+ ! B => MkT   -- bad

This means we can write a proper grammar for types, overloading it in
the DisambECP style, see Note [Ambiguous syntactic categories].

With this patch, we introduce a new class, DisambTD. Just like
DisambECP is used to disambiguate between expressions, commands, and patterns,
DisambTD  is used to disambiguate between types and data/newtype constructors.

This way, we get a proper, declarative grammar for constructors and
types:

	infixtype
	  : ftype
	  | ftype tyop infixtype
	  | unpackedness infixtype

	ftype
	  : atype
	  | tyop
	  | ftype tyarg
	  | ftype PREFIX_AT tyarg

	tyarg
	  : atype
	  | unpackedness atype

And having a grammar for types means we are a step closer to using a
single grammar for types and expressions.

The patch is quite straightforward. The only tricky part is that
`Language.Haskell.TH.Lib.Internal` now must be `Trustworthy` instead
of `Safe` due to the `GHC.Exts` import (in order to import `TYPE`).

Since `CodeQ` has yet to appear in any released version of
`template-haskell`, I didn't bother mentioning the change to `CodeQ`
in the `template-haskell` release notes.

Fixes #18521.