Don't allocate a Ptr constructor per frame.

Here we add the unit ID to the info table provenance structure.

This allows us to shave a 64-bit word off of the packed IPE entry size.

Previously we would eagerly allocate `InfoTableEnt`s for each
info table registered in the info table provenance map. However, this
costs considerable memory and initialization time. Instead we now
lazily decode these tables. This allows us to use one-third the memory
*and* opens the door to taking advantage of sharing opportunities within
a module.

This required considerable reworking since lookupIPE now must be passed
its result buffer.

Such that we can add new helpers into GHC.InfoProv.Types without
breakage.

Vladislav Zavialov authored 1 year ago and Marge Bot committed 1 year ago

This should have been done in 0f0c53a5 but I missed it.

Ben Gamari authored 1 year ago and Marge Bot committed 1 year ago

There is nothing platform-dependent about our GOT implementation and
GOT support is needed by `T24171` on i386.

Cheng Shao authored 1 year ago and Marge Bot committed 1 year ago

See #24515 for details.

Ben Gamari authored 1 year ago and Marge Bot committed 1 year ago

This replicates the behavior on Windows, where `Hi.exe` will produce
profiling output named `Hi.prof` instead of `Hi.exe.prof`.

While in the area I also fixed the extension-stripping logic, which
incorrectly rewrote `Hi.exefoo` to `Hi.foo`.

Closes #24515.

Patrick authored 1 year ago and Marge Bot committed 1 year ago

The main purpose of this is to tuck the module name `xxx` in `module xxx where` into the hieAst.
It should fix #24493.

The following have been done:
1. Renamed and update the `tcg_doc_hdr :: Maybe (LHsDoc GhcRn)` to `tcg_hdr_info :: (Maybe (LHsDoc GhcRn), Maybe (XRec GhcRn ModuleName))`
   To store the located module name information.
2. update the `RenamedSource` and `RenamedStuff` with extra `Maybe (XRec GhcRn ModuleName)` located module name information.
3. add test `testsuite/tests/hiefile/should_compile/T24493.hs` to ensure the module name is added and update several relevent tests.
4. accompanied submodule haddoc test update MR in ghc/haddock!53

Jade authored 1 year ago and Marge Bot committed 1 year ago

This patch improves the error messages from invalid type signatures by
trying to guess what the user did and suggesting an appropriate fix.

Partially fixes: #17879

Ben Gamari authored 1 year ago and Marge Bot committed 1 year ago

As noted by #24344, `strerror` is not necessarily thread-safe.
Thankfully, POSIX.1-2001 has long offered `strerror_r`, which is
safe to use.

Fixes #24344.

CLC discussion: https://github.com/haskell/core-libraries-committee/issues/249

Ben Gamari authored 1 year ago and Marge Bot committed 1 year ago

This fixes a regression in retainer set profiling introduced by
b0293f78. Prior to that commit
the heap traversal word would be initialized by `SET_HDR` using
`LDV_RECORD_CREATE`. However, the commit added a `doingLDVProfiling`
check in `LDV_RECORD_CREATE`, meaning that this initialization no longer
happened.

Given that this initialization was awkwardly indirectly anyways, I have
fixed this by explicitly initializating the heap traversal word to
`NULL` in `SET_PROF_HDR`. This is equivalent to the previous behavior,
but much more direct.

Fixes #24513.

Ben Gamari authored 1 year ago and Marge Bot committed 1 year ago

This was simply reexporting things from `ghc-prim`. Instead reexport
these directly from `Data.Kind`. Also add build ordering dependency to
work around #23942.

Alan Zimmerman authored 1 year ago and Marge Bot committed 1 year ago

Also tweak rendering of SrcSpan to remove extra blank line.

Ben Gamari authored 1 year ago and Marge Bot committed 1 year ago

Most GhcExceptions are user-facing errors and therefore the
ExceptionContext has little value. Ideally we would enable
it in the DEBUG compiler but I am leaving this for future work.

Ben Gamari authored 1 year ago and Marge Bot committed 1 year ago

Here we introduce the `Backtraces` type and associated machinery for
attaching these via `ExceptionContext`. These has a few compile-time
regressions (`T15703` and `T9872d`) due to the additional dependencies
in the exception machinery.

As well, there is a surprisingly large regression in the
`size_hello_artifact` test. This appears to be due to various `Integer` and
`Read` bits now being reachable at link-time. I believe it should be
possible to avoid this but I have accepted the change for now to get the
feature merged.

CLC Proposal: https://github.com/haskell/core-libraries-committee/issues/199
GHC Proposal: https://github.com/ghc-proposals/ghc-proposals/pull/330

Metric Increase:
    T15703
    T9872d
    size_hello_artifact

Ben Gamari authored 1 year ago and Marge Bot committed 1 year ago

Here we introduce the `ExceptionContext` type and `ExceptionAnnotation`
class, allowing dynamically-typed user-defined annotations to be
attached to exceptions.

CLC Proposal: https://github.com/haskell/core-libraries-committee/issues/199
GHC Proposal: https://github.com/ghc-proposals/ghc-proposals/pull/330

* Main change: make the error message generated by mkFunTysMsg more
  accurate by changing "value arguments" to "visible arguments".

* Refactor: define a new type synonym VisArity and use it instead of
  Arity in a few places.

It might be the case that there other places in the compiler that should
talk about visible arguments rather than value arguments, but I haven't
tried to find them all, focusing only on the error message reported in
the ticket.

Simon Peyton Jones authored 1 year ago and Marge Bot committed 1 year ago

These were changes are all triggered by #24471.

1. Make GHC.Core.Opt.SetLevels.lvlMFE behave better when there are
   many free variables.  See Note [Large free-variable sets].

2. Make GHC.Core.Opt.Arity.floatIn a bit lazier in its Cost argument.
   This benefits the common case where the ArityType turns out to
   be nullary. See Note [Care with nested expressions]

3. Make GHC.CoreToStg.Prep.cpeArg behave for deeply-nested
   expressions.  See Note [Eta expansion of arguments in CorePrep]
   wrinkle (EA2).

Compile times go down by up to 4.5%, and much more in artificial
cases. (Geo mean of compiler/perf changes is -0.4%.)

Metric Decrease:
    CoOpt_Read
    T10421
    T12425

Ben Gamari authored 1 year ago and Marge Bot committed 1 year ago

This is a temporary measure to improve CI reliability until a proper
solution is developed.

Works around #23942.

Cheng Shao authored 1 year ago and Marge Bot committed 1 year ago

This commit adds an assertion to Bdescr() to assert the pointer is
indeed heap allocated. This is useful to rule out RTS bugs that
attempt to access non-existent block descriptor of a static closure, #24492
being one such example.

Cheng Shao authored 1 year ago and Marge Bot committed 1 year ago

This commit exposes HeapAlloc.h as a public header. The intention is
to expose HEAP_ALLOCED/HEAP_ALLOCED_GC, so they can be used in
assertions in other public headers, and they may also be useful for
user code.

Cheng Shao authored 1 year ago and Marge Bot committed 1 year ago

This patch adds a -xr RTS option to control the size of virtual memory
address space reserved by the two step allocator on a 64-bit platform,
see added documentation for explanation. Closes #24498.

Sylvain Henry authored 1 year ago and Marge Bot committed 1 year ago

Use faster implementations for the following primitives in the JS
backend by not using JavaScript's BigInt:
- plusInt64
- minusInt64
- minusWord64
- timesWord64
- timesInt64

Co-authored-by: Josh Meredith <joshmeredith2008@gmail.com>

Matthew Pickering authored 1 year ago and Marge Bot committed 1 year ago

* -he profiling mode
* -he profiling selector
* --automatic-era-increment

CLC proposal #254 - https://github.com/haskell/core-libraries-committee/issues/254