sheaf authored 6 years ago and Ben Gamari committed 6 years ago

Summary:
on windows, plugins are loaded via .a files,
but those paths were not being searched when loading plugins

Test Plan: ./validate

Reviewers: Phyx, bgamari

Reviewed By: Phyx

Subscribers: RyanGlScott, rwbarton, carter

GHC Trac Issues: #15700

Differential Revision: https://phabricator.haskell.org/D5253

(cherry picked from commit 70298db16c3f0ea4adb603ccb2b5e93eb9c7a556)

Summary:
Although these functions were marked as CONSTANT_FOLDED, they did
not have a corresponding builtinRule in PrelRules. The idea was
probably to add them eventually, but this hasn't manifested so
far.

The plan is to eventually add builtin rules for these functions
over Natural, so as a reminder we simply comment out the
CONSTANT_FOLDED  annotation instead of removing it completely.

Reviewers: hvr, bgamari

Reviewed By: bgamari

Subscribers: rwbarton, carter

Differential Revision: https://phabricator.haskell.org/D5267

(cherry picked from commit 3ec6fe8827956cc36b58cdf0bb1f5752eaa2a8ea)

Fangyi Zhou authored 6 years ago and Ben Gamari committed 6 years ago

Summary:
Ticket #15271 reports a case where 1e1000000000 is incorrectly
converted to 0.0. After some investigation, I discovered the number is
converted to rational correctly, but converting the ratio into a double
introduced an error.

Tracking down to how the conversion is done, I found the rts float
implementation uses `ldexp`, whose signature is
`double ldexp (double x, int exp);`
The callsite passes an `I_` to the second argument, which is
platform-dependent. On machines where `I_` is 64 bits and `int` is 32 bits, we
observe integer overflow behaviour.

Here is a mapping from rational to exponent with observations
1e646457008  -> 2147483645 (result = infinity, positive in int32)
1e646457009  -> 2147483648 (result = 0.0, overflow to negative in int32)
1e1000000000 -> 3321928042 (result = infinity, overflow to positive in int32)
1e1555550000 -> 5167425196 (result = 0.0, overflow to negative in int32)

We fix this issue by comparing STG_INT_MIN/MAX and INT_MIN/MAX and bound the
value appropriately.

Test Plan: New test cases

Reviewers: bgamari, erikd, simonmar

Reviewed By: bgamari

Subscribers: rwbarton, carter

GHC Trac Issues: #15271

Differential Revision: https://phabricator.haskell.org/D5271

(cherry picked from commit 311a63979cfa2c1e81be54b82205e681f6ec4f14)

Summary:
This patch fixes #15806, where we found that the `:k` command in GHCi
misses a validity checking for the type.

Missing validity checking causes `:k` to accept types that are not validated.
For example, `:k (Maybe (forall a. a -> a))` (incorrectly) returns `*`, while
impredictivity of type instantiation shouldn't be allowed.

Test Plan: ./validate

Reviewers: simonpj, goldfire, bgamari

Reviewed By: bgamari

Subscribers: rwbarton, carter

GHC Trac Issues: #15806

Differential Revision: https://phabricator.haskell.org/D5265

(cherry picked from commit 12cb5cf50b8b35394e2e2d57e1ac693c76f90833)

Ben Gamari authored 6 years ago and Ben Gamari committed 6 years ago

Summary:
Fixes #14784. Note that C++11 is quite conservative; we could likely accept
C++03 as well.

Test Plan:
```
$ cat >hi.c <<EOF
#include <Rts.h>
EOF
$ g++ -std=c++11 hi.c
```

Reviewers: simonmar, hvr

Subscribers: rwbarton, carter

GHC Trac Issues: #14784

Differential Revision: https://phabricator.haskell.org/D5244

(cherry picked from commit d3a1022fabb0ad337003fac774c6929f402ecb8b)

Zejun Wu authored 6 years ago and Ben Gamari committed 6 years ago

Summary:
Encountered assembly error due to undefined label `.LcaDcU_info_end` for
following code generated by `pprFrameProc`:

```
.Lsat_sa8fp{v}_info_fde_end:
  .long .Lblock{v caDcU}_info_fde_end-.Lblock{v caDcU}_info_fde
.Lblock{v caDcU}_info_fde:
  .long _nbHlD-.Lsection_frame
  .quad block{v caDcU}_info-1
  .quad .Lblock{v caDcU}_info_end-block{v caDcU}_info+1
  .byte 1
```

This diff fixed the error.

Test Plan:
  ./validate

Also the case where we used to have assembly error is now fixed.
Unfortunately, I have limited insight here and cannot get a small enough repro
or test case for this.

Ben says:

> I think I see: Previously we only produced end symbols for the info
> tables of top-level procedures. However, blocks within a procedure may
> also have info tables, we will dutifully generate debug information for
> and consequently we get undefined symbols.

Reviewers: simonmar, scpmw, last_g, bgamari

Reviewed By: bgamari

Subscribers: rwbarton, carter

Differential Revision: https://phabricator.haskell.org/D5246

(cherry picked from commit cf961dcf5ebc26cbd960196ba387736334088303)

Kavon Farvardin authored 6 years ago and Ben Gamari committed 6 years ago

We now calculate the SSE register padding needed to fix the calling
convention in LLVM in a robust way: grouping them by whether
registers in that class overlap (with the same class overlapping
itself).

My prior patch assumed that no matter the platform, physical
register Fx aliases with Dx, etc, for our calling convention.

This is unfortunately not the case for any platform except x86-64.

Test Plan:
Only know how to test on x86-64, but it should be tested on ARM with:

`make test WAYS=llvm && make test WAYS=optllvm`

Reviewers: bgamari, angerman

Reviewed By: bgamari

Subscribers: rwbarton, carter

GHC Trac Issues: #15780, #14251, #15747

Differential Revision: https://phabricator.haskell.org/D5254

(cherry picked from commit c36a2b596a6ba9d7a0a80df01b3c041720c727ca)

Fixes #15404.

Richard Eisenberg authored 6 years ago and Ben Gamari committed 6 years ago

This was just a pretty-printer infelicity. Fixed now.

Test case: printer/T15761

(cherry picked from commit 38c28c1a)

We were recovering too eagerly from errors in pattern-synonym
type inference, leading to a cascade of confusing follow up errors
(Trac #15685, #15692).

The underlying issue is that a pattern synonym should have a closed,
fixed type, with no unification variables in it.  But it wasn't!

Fixing this made me change the interface to simplifyInfer slightly.
Instead of /emitting/ a residual implication constraint, it
now /returns/ it, so that the caller can decide what to do.

(cherry picked from commit 9ebfa03d)

I've added a check in my release script to ensure that this doesn't happen in
the future.

(cherry picked from commit 2605458930f2d79738fab4437f10793448d4232c)

See #15696 for more details. We now always enter dataToTag# argument (done in
generated Cmm, in StgCmmExpr). Any high-level optimisations on dataToTag#
applications are done by the simplifier. Looking at tag bits (instead of
reading the info table) for small types is left to another diff.

Incorrect test T14626 is removed. We no longer do this optimisation (see
comment:44, comment:45, comment:60).

Comments and notes about special cases around dataToTag# are removed. We no
longer have any special cases around it in Core.

Other changes related to evaluating primops (seq# and dataToTag#) will be
pursued in follow-up diffs.

Test Plan: Validates with three regression tests

Reviewers: simonpj, simonmar, hvr, bgamari, dfeuer

Reviewed By: simonmar

Subscribers: rwbarton, carter

GHC Trac Issues: #15696

Differential Revision: https://phabricator.haskell.org/D5201

(cherry picked from commit ac977688)

For years we have been marking CoVars as dead, becuase we
don't gather occurrence info from types.  This is obviously
wrong and caused Trac #15695.

See Note [Do not mark CoVars as dead] in OccurAnal.

(cherry picked from commit 02b303ee)

Simon Marlow authored 6 years ago and Ben Gamari committed 6 years ago

Summary:
When using -fexternal-interpreter, recover was not treating a Q
compuation that simply registered an error with addErrTc as failing.

Test Plan:
New unit tests:
* T15418 is the repro from in the ticket
* TH_recover_warns is a new test to ensure that we're keeping warnings when
  the body of recover succeeds.

Reviewers: bgamari, RyanGlScott, angerman, goldfire, erikd

Subscribers: rwbarton, carter

GHC Trac Issues: #15418

Differential Revision: https://phabricator.haskell.org/D5185

(cherry picked from commit d00c3086)

Alec Theriault authored 6 years ago and Ben Gamari committed 6 years ago

Summary: See the new T12005 test case for an example of this.

Test Plan: make TEST=T12005

Reviewers: bgamari, osa1

Reviewed By: osa1

Subscribers: osa1, rwbarton, carter

GHC Trac Issues: #12005

Differential Revision: https://phabricator.haskell.org/D5182

(cherry picked from commit 21efbc75)

Vladislav Zavialov authored 6 years ago and Ben Gamari committed 6 years ago

The -Wstar-is-type flag was added without documentation.
Now it has documentation.

Test Plan: Validate

Reviewers: bgamari

Reviewed By: bgamari

Subscribers: rwbarton, carter

Differential Revision: https://phabricator.haskell.org/D5203

(cherry picked from commit 07083fc4)

As #15571 reports, eager blackholing breaks sanity checks as we can't
zero the payload when eagerly blackholing (because we'll be using the
payload after blackholing), but by the time we blackhole a previously
eagerly blackholed object (in `threadPaused()`) we don't have the
correct size information for the object (because the object's type
becomes BLACKHOLE when we eagerly blackhole it) so can't properly zero
the slop.

This problem can be solved for AP_STACK eager blackholing (which unlike
eager blackholing in general, is not optional) by zeroing the payload
after entering the stack. This patch implements this idea.

Fixes #15571.

Test Plan:
Previously concprog001 when compiled and run with sanity checks

    ghc-stage2 Mult.hs -debug -rtsopts
    ./Mult +RTS -DS

was failing with

    Mult: internal error: checkClosure: stack frame
        (GHC version 8.7.20180821 for x86_64_unknown_linux)
        Please report this as a GHC bug:  http://www.haskell.org/ghc/reportabug

thic patch fixes this panic. The test still panics, but it runs for a while
before panicking (instead of directly panicking as before), and the new problem
seems unrelated:

    Mult: internal error: ASSERTION FAILED: file rts/sm/Sanity.c, line 296
        (GHC version 8.7.20180919 for x86_64_unknown_linux)
        Please report this as a GHC bug:  http://www.haskell.org/ghc/reportabug

The new problem will be fixed in another diff.

I also tried slow validate (which requires D5164): this does not introduce any
new failures.

Reviewers: simonmar, bgamari, erikd

Reviewed By: simonmar

Subscribers: rwbarton, carter

GHC Trac Issues: #15571

Differential Revision: https://phabricator.haskell.org/D5165

(cherry picked from commit 66c17293)

Summary: After a parse error in OPTIONS_GHC issue an error message instead of a compiler panic.

Test Plan: make test TEST=T15053

Reviewers: Phyx, thomie, bgamari, monoidal, osa1

Reviewed By: Phyx, monoidal, osa1

Subscribers: tdammers, osa1, rwbarton, carter

GHC Trac Issues: #15053

Differential Revision: https://phabricator.haskell.org/D5093

(cherry picked from commit df363a64)

Ben Gamari authored 6 years ago and Ben Gamari committed 6 years ago

Reviewers: Phyx

Reviewed By: Phyx

Subscribers: Phyx, rwbarton, thomie, carter

GHC Trac Issues: #15053

Differential Revision: https://phabricator.haskell.org/D4883

(cherry picked from commit f03f0d61)

Alec Theriault authored 6 years ago and Ben Gamari committed 6 years ago

Summary:
This means that 'GHC.Classes.(%,%)' is no longer mentioned in
error messages for things like

   class (a,b,c)  -- outside of 'GHC.Classes'
   class (a,Bool)

Test Plan: make TEST=T14907a && make TEST=T14907b

Reviewers: RyanGlScott, bgamari

Reviewed By: RyanGlScott

Subscribers: rwbarton, carter

GHC Trac Issues: #14907

Differential Revision: https://phabricator.haskell.org/D5172

(cherry picked from commit 9bfbc4e1)

After commit ef26182e,
the type variable binders in GADT constructor type signatures
are now quantified in toposorted order, instead of always having
all the universals before all the existentials. Unfortunately, that
commit forgot to update some code (which was assuming the latter
scenario) in `DsExpr` which desugars record updates. This wound
up being the cause of #15499.

This patch makes up for lost time by desugaring record updates in
a way such that the desugared expression applies type arguments to
the right-hand side constructor in the correct order—that is, the
order in which they were quantified by the user.

Test Plan: make test TEST=T15499

Reviewers: simonpj, bgamari

Reviewed By: simonpj

Subscribers: rwbarton, carter

GHC Trac Issues: #15499

Differential Revision: https://phabricator.haskell.org/D5060

(cherry picked from commit 63b6a1d4)

Tamar Christina authored 6 years ago and Ben Gamari committed 6 years ago

Summary:
The new filesystem code accidentally asks for write attributes
permissions when doing read-only access.

I believe this is what's causing the GHC 8.6.1 tarballs to fail
when installed to a privileged location.
I haven't been able to reproduce the issue yet, but this permission
bit is wrong anyway.

Test Plan: I'm still trying to workout how to test that this works,
changing the permissions on the folder doesn't seem to reproduce
the error on a tarball I made from before the change.

Reviewers: bgamari, tdammers

Reviewed By: bgamari

Subscribers: tdammers, monoidal, rwbarton, carter

GHC Trac Issues: #15667

Differential Revision: https://phabricator.haskell.org/D5177

(cherry picked from commit deceb21b)

Kavon Farvardin authored 6 years ago and Ben Gamari committed 6 years ago

- Fix for #13904 -- stop "trashing" callee-saved registers, since it is
  not actually doing anything useful.

- Fix for #14251 -- fixes the calling convention for functions passing
  raw SSE-register values by adding padding as needed to get the values
  in the right registers. This problem cropped up when some args were
  unused an dropped from the live list.

- Fixed a typo in 'readnone' attribute

- Added 'lower-expect' pass to level 0 LLVM optimization passes to
  improve block layout in LLVM for stack checks, etc.

Test Plan: `make test WAYS=optllvm` and `make test WAYS=llvm`

Reviewers: bgamari, simonmar, angerman

Reviewed By: angerman

Subscribers: rwbarton, carter

GHC Trac Issues: #13904, #14251

Differential Revision: https://phabricator.haskell.org/D5190

(cherry picked from commit adcb5fb4)

This breaks if LLVM is not available.

(cherry picked from commit d0d74842)

(cherry picked from commit ba086ca7)

The cabal update command appears to be timing out with no output after 10 minutes.

Apparently the override argument to add_directive_to_domain was added in sphinx
1.8.

(cherry picked from commit a257782f)

Zejun Wu authored 6 years ago and Ben Gamari committed 6 years ago

Encouter following error when `make`:

```
Extension error:
The 'ghc-flag' directive is already registered to domain std
```

as we register `ghc-flag` to `std` in `add_object_type` first and then
overtride it in `add_directive_to_domain`.

Test Plan:
  make -C utils/haddock/doc html SPHINX_BUILD=/usr/bin/sphinx-build

Reviewers: austin, bgamari, patrickdoc

Subscribers: rwbarton, carter

Differential Revision: https://phabricator.haskell.org/D5161

(cherry picked from commit 8c7d33a8)

It seems that both add_object_type and add_directive_to_domain both register a
directive. Previously sphinx didn't seem to mind this but as of Sphinx 1.8 it
crashes with an exception.

(cherry picked from commit 4eebc801)

Takenobu Tani authored 6 years ago and Ben Gamari committed 6 years ago

Summary:
Fix code-block layout for QuantifiedConstraints.

[ci skip]

Test Plan: build

Reviewers: bgamari, monoidal

Reviewed By: monoidal

Subscribers: rwbarton, carter

Differential Revision: https://phabricator.haskell.org/D5121

(cherry picked from commit 43967c0c)

Simon Marlow authored 6 years ago and Ben Gamari committed 6 years ago

Shortcutting the SRT for a static function can lead to resurrecting a
static object at runtime, which violates assumptions in the GC. See
comments for details.

Test Plan:
- manual testing (in progress)
- validate

Reviewers: osa1, bgamari, erikd

Reviewed By: bgamari

Subscribers: rwbarton, carter

GHC Trac Issues: #15544

Differential Revision: https://phabricator.haskell.org/D5145

(cherry picked from commit a324dfdf3b505ea30d134dc8183d7b4bb441ced4)

This reverts commit c15d44f8.