sheaf authored 1 year ago and Ben Gamari committed 1 year ago

This commit adds similar name suggestions when importing. For example

  module A where { spelling = 'o' }
  module B where { import B ( speling ) }

will give rise to the error message:

  Module ‘A’ does not export ‘speling’.
  Suggested fix: Perhaps use ‘spelling’

This also provides hints when users try to import record fields defined
with NoFieldSelectors.

(cherry picked from commit 1af2e773)

sheaf authored 1 year ago and Ben Gamari committed 1 year ago

In an import declaration such as

  import M ( var )

the import of the variable "var" should **not** bring into scope record
fields named "var" which are defined with NoFieldSelectors.
Doing so can cause spurious "unused import" warnings, as reported in
ticket #23557.

Fixes #23557

(cherry picked from commit c7bbad9a)

sheaf authored 1 year ago and Ben Gamari committed 1 year ago

This commit consolidates all the logic for looking up something in
the Global Reader Environment into the single function lookupGRE.
This allows us to declaratively specify all the different modes of
looking up in the GlobalRdrEnv, and avoids manually passing around
filtering functions as was the case in e.g. the function
GHC.Rename.Env.lookupSubBndrOcc_helper.

-------------------------
Metric Decrease:
    T8095
-------------------------
-------------------------
Metric Increase:
    T8095
-------------------------

(cherry picked from commit 7f0a86ed)

sheaf authored 1 year ago and Ben Gamari committed 1 year ago

These are simple helper functions that wrap the internal
field names gre_info, gre_par.

(cherry picked from commit 6fd8f566)

sheaf authored 1 year ago and Ben Gamari committed 1 year ago

Commit 3f374399 inadvertently broke the deprecation/warning mechanism
for record fields due to its introduction of record field namespaces.

This patch ensures that, when a top-level deprecation is applied to
an identifier, it applies to all the record fields as well.
This is achieved by refactoring GHC.Rename.Env.lookupLocalTcNames, and
GHC.Rename.Env.lookupBindGroupOcc, to not look up a fixed number of
NameSpaces but to look up all NameSpaces and filter out the irrelevant
ones.

(cherry picked from commit 6143838a)

The `'[]` case in `tc_infer_hs_type` is smart enough to handle arity-0 uses of
`'[]` (see the newly added `T23543` test case for an example), but the `'[]`
case in `tc_hs_type` was not. We fix this by changing the `tc_hs_type` case to
invoke `tc_infer_hs_type`, as prescribed in `Note [Future-proofing the type
checker]`.

There are some benign changes to test cases' expected output due to the new
code path using `forall a. [a]` as the kind of `'[]` rather than `[k]`.

Fixes #23543.

(cherry picked from commit c335fb7c)

Krzysztof Gogolewski authored 1 year ago and Ben Gamari committed 1 year ago

This implements GHC proposal
https://github.com/ghc-proposals/ghc-proposals/pull/596

Also add support for Int64# and Word64#; see testcase ShowPrim.

(cherry picked from commit 787bae96)

(cherry picked from commit b55a8ea7)

Vladislav Zavialov authored 1 year ago and Ben Gamari committed 1 year ago

Add the missing changelog.md entries and @since-annotations.

(cherry picked from commit 7f13acbf)

Krzysztof Gogolewski authored 1 year ago and Ben Gamari committed 1 year ago

Found by Simon in ghc/ghc#23567 (comment 507834)

The testcase isn't ideal because it doesn't detect the bug in master,
unless doNotUnbox is removed as in
ghc/ghc#23567 (comment 507692).
But I have confirmed that with that modification, it fails before
and passes afterwards.

(cherry picked from commit bf9b9de0)

Matthew Pickering authored 1 year ago and Ben Gamari committed 1 year ago

deb9 is now EOL so it's time to upgrade the i386 bindist to use deb10

Fixes #23585

(cherry picked from commit c39f279b)

sheaf authored 1 year ago and Ben Gamari committed 1 year ago

The function GHC.Tc.Errors.validHoleFits would end up panicking when
encountering a Given constraint. To fix this, it suffices to filter out
the Givens before continuing.

Fixes #22684

(cherry picked from commit 630e3026)

(cherry picked from commit 550af505)

sheaf authored 1 year ago and Ben Gamari committed 1 year ago

The GREInfo for a record field should include the collection of all
the data constructors of the parent TyCon that have this record field.
This information was being incorrectly computed in the tyThingLocalGREs
function for a DataCon, as we were not taking into account other
DataCons with the same parent TyCon.

Fixes #23546

(cherry picked from commit 61b1932e)

(cherry picked from commit 054261dd)

Matthew Pickering authored 1 year ago and Ben Gamari committed 1 year ago

There was an oversight in the driver which assumed that you would always
produce a `.o` file when compiling a .cmm file.

Fixes #23610

(cherry picked from commit 76983a0d)

sheaf authored 1 year ago and Ben Gamari committed 1 year ago

This extra bit of information was accidentally being discarded after
a refactoring of the way we reported problems when unifying a type
variable with another type. This patch rectifies that.

(cherry picked from commit 2b55cb5f)

Mario authored 1 year ago and Ben Gamari committed 1 year ago

(cherry picked from commit 4af7eac2)

Matthew Pickering authored 1 year ago and Ben Gamari committed 1 year ago

Polymorphic specialisation has led to a number of hard to diagnose
incorrect runtime result bugs (see #23469, #23109, #21229, #23445) so
this commit introduces a flag `-fpolymorhphic-specialisation` which
allows users to turn on this experimental optimisation if they are
willing to buy into things going very wrong.

Ticket #23469

(cherry picked from commit 9f01d14b)

Dave Barton authored 1 year ago and Ben Gamari committed 1 year ago

(cherry picked from commit 4457da2a)

(cherry picked from commit 78b2f3cc)

When the Win32 threaded IO manager shuts down, `ioManagerDie` sends an
`IO_MANAGER_DIE` event to the IO manager thread using the
`io_manager_event` event object. Finally, it will closes the event object,
and invalidate `io_manager_event`.

Previously, `readIOManagerEvent` would see that `io_manager_event` is
invalid and return `0`, suggesting that everything is right with the
world. This meant that if `ioManagerDie` invalidated the handle before
the event manager was blocked on the event we would end up in a
situation where the event manager would never realize it was asked to
shut down.

Fix this by ensuring that `readIOManagerEvent` instead returns
`IO_MANAGER_DIE` when we detect that the event object has been
invalidated by `ioManagerDie`.

Fixes #23691.

This test passes on dynamically-linked Alpine.

Due to spurious output changes described in #23648.

Metric Increase:
    T6048

To freeze emsdk, avoiding #23641.

x86-64/Darwin's toolchain inexplicably warns that collectFreshWeakPtrs
needs to be a prototype.

(cherry picked from commit bb0ed354)

Ackerman authored 1 year ago and Ben Gamari committed 1 year ago

GHC.Rename.Utils concrete helpers instead of wrapGenSpan + HS AST constructors

(cherry picked from commit dd782343)

Ackerman authored 1 year ago and Ben Gamari committed 1 year ago

(cherry picked from commit 3efe7f39)

Ackerman authored 1 year ago and Ben Gamari committed 1 year ago

This fixes #23492. The problem was that we used the real source span
of the field declaration for the generated catch-all case in the
selector function, in particular in the generated call to
`recSelError`, which meant it was included in the HIE output. Using
`generatedSrcSpan` instead means that it is not included.

(cherry picked from commit 356a2692)