[Git][ghc/ghc][wip/T20749] 29 commits: ghc-the-library: Retain cafs in both static in dynamic builds.
Sebastian Graf (@sgraf812)
gitlab at gitlab.haskell.org
Fri Feb 3 16:24:03 UTC 2023
Sebastian Graf pushed to branch wip/T20749 at Glasgow Haskell Compiler / GHC
Commits:
08ba8720 by Andreas Klebinger at 2023-01-30T21:18:45-05:00
ghc-the-library: Retain cafs in both static in dynamic builds.
We use keepCAFsForGHCi.c to force -fkeep-cafs behaviour by using a
__attribute__((constructor)) function.
This broke for static builds where the linker discarded the object file
since it was not reverenced from any exported code. We fix this by
asserting that the flag is enabled using a function in the same module
as the constructor. Which causes the object file to be retained by the
linker, which in turn causes the constructor the be run in static builds.
This changes nothing for dynamic builds using the ghc library. But causes
static to also retain CAFs (as we expect them to).
Fixes #22417.
-------------------------
Metric Decrease:
T21839r
-------------------------
- - - - -
20598ef6 by Ryan Scott at 2023-01-30T21:19:20-05:00
Handle `type data` properly in tyThingParent_maybe
Unlike most other data constructors, data constructors declared with `type data`
are represented in `TyThing`s as `ATyCon` rather than `ADataCon`. The `ATyCon`
case in `tyThingParent_maybe` previously did not consider the possibility of
the underlying `TyCon` being a promoted data constructor, which led to the
oddities observed in #22817. This patch adds a dedicated special case in
`tyThingParent_maybe`'s `ATyCon` case for `type data` data constructors to fix
these oddities.
Fixes #22817.
- - - - -
2f145052 by Ryan Scott at 2023-01-30T21:19:56-05:00
Fix two bugs in TypeData TH reification
This patch fixes two issues in the way that `type data` declarations were
reified with Template Haskell:
* `type data` data constructors are now properly reified using `DataConI`.
This is accomplished with a special case in `reifyTyCon`. Fixes #22818.
* `type data` type constructors are now reified in `reifyTyCon` using
`TypeDataD` instead of `DataD`. Fixes #22819.
- - - - -
d0f34f25 by Simon Peyton Jones at 2023-01-30T21:20:35-05:00
Take account of loop breakers in specLookupRule
The key change is that in GHC.Core.Opt.Specialise.specLookupRule
we were using realIdUnfolding, which ignores the loop-breaker
flag. When given a loop breaker, rule matching therefore
looped infinitely -- #22802.
In fixing this I refactored a bit.
* Define GHC.Core.InScopeEnv as a data type, and use it.
(Previously it was a pair: hard to grep for.)
* Put several functions returning an IdUnfoldingFun into
GHC.Types.Id, namely
idUnfolding
alwaysActiveUnfoldingFun,
whenActiveUnfoldingFun,
noUnfoldingFun
and use them. (The are all loop-breaker aware.)
- - - - -
de963cb6 by Matthew Pickering at 2023-01-30T21:21:11-05:00
ci: Remove FreeBSD job from release pipelines
We no longer attempt to build or distribute this release
- - - - -
f26d27ec by Matthew Pickering at 2023-01-30T21:21:11-05:00
rel_eng: Add check to make sure that release jobs are downloaded by fetch-gitlab
This check makes sure that if a job is a prefixed by "release-" then the
script downloads it and understands how to map the job name to the
platform.
- - - - -
7619c0b4 by Matthew Pickering at 2023-01-30T21:21:11-05:00
rel_eng: Fix the name of the ubuntu-* jobs
These were not uploaded for alpha1
Fixes #22844
- - - - -
68eb8877 by Matthew Pickering at 2023-01-30T21:21:11-05:00
gen_ci: Only consider release jobs for job metadata
In particular we do not have a release job for FreeBSD so the generation
of the platform mapping was failing.
- - - - -
b69461a0 by Jason Shipman at 2023-01-30T21:21:50-05:00
User's guide: Clarify overlapping instance candidate elimination
This commit updates the user's guide section on overlapping instance candidate
elimination to use "or" verbiage instead of "either/or" in regards to the
current pair of candidates' being overlappable or overlapping. "Either IX is
overlappable, or IY is overlapping" can cause confusion as it suggests "Either
IX is overlappable, or IY is overlapping, but not both".
This was initially discussed on this Discourse topic:
https://discourse.haskell.org/t/clarification-on-overlapping-instance-candidate-elimination/5677
- - - - -
7cbdaad0 by Matthew Pickering at 2023-01-31T07:53:53-05:00
Fixes for cabal-reinstall CI job
* Allow filepath to be reinstalled
* Bump some version bounds to allow newer versions of libraries
* Rework testing logic to avoid "install --lib" and package env files
Fixes #22344
- - - - -
fd8f32bf by Cheng Shao at 2023-01-31T07:54:29-05:00
rts: prevent potential divide-by-zero when tickInterval=0
This patch fixes a few places in RtsFlags.c that may result in
divide-by-zero error when tickInterval=0, which is the default on
wasm. Fixes #22603.
- - - - -
085a6db6 by Joachim Breitner at 2023-01-31T07:55:05-05:00
Update note at beginning of GHC.Builtin.NAmes
some things have been renamed since it was written, it seems.
- - - - -
7716cbe6 by Cheng Shao at 2023-01-31T07:55:41-05:00
testsuite: use tgamma for cg007
gamma is a glibc-only deprecated function, use tgamma instead. It's
required for fixing cg007 when testing the wasm unregisterised
codegen.
- - - - -
19c1fbcd by doyougnu at 2023-01-31T13:08:03-05:00
InfoTableProv: ShortText --> ShortByteString
- - - - -
765fab98 by doyougnu at 2023-01-31T13:08:03-05:00
FastString: add fastStringToShorText
- - - - -
a83c810d by Simon Peyton Jones at 2023-01-31T13:08:38-05:00
Improve exprOkForSpeculation for classops
This patch fixes #22745 and #15205, which are about GHC's
failure to discard unnecessary superclass selections that
yield coercions. See
GHC.Core.Utils Note [exprOkForSpeculation and type classes]
The main changes are:
* Write new Note [NON-BOTTOM_DICTS invariant] in GHC.Core, and
refer to it
* Define new function isTerminatingType, to identify those
guaranteed-terminating dictionary types.
* exprOkForSpeculation has a new (very simple) case for ClassOpId
* ClassOpId has a new field that says if the return type is
an unlifted type, or a terminating type.
This was surprisingly tricky to get right. In particular note
that unlifted types are not terminating types; you can write an
expression of unlifted type, that diverges. Not so for dictionaries
(or, more precisely, for the dictionaries that GHC constructs).
Metric Decrease:
LargeRecord
- - - - -
f83374f8 by Krzysztof Gogolewski at 2023-01-31T13:09:14-05:00
Support "unusable UNPACK pragma" warning with -O0
Fixes #11270
- - - - -
a2d814dc by Ben Gamari at 2023-01-31T13:09:50-05:00
configure: Always create the VERSION file
Teach the `configure` script to create the `VERSION` file.
This will serve as the stable interface to allow the user to determine
the version number of a working tree.
Fixes #22322.
- - - - -
5618fc21 by sheaf at 2023-01-31T15:51:06-05:00
Cmm: track the type of global registers
This patch tracks the type of Cmm global registers. This is needed
in order to lint uses of polymorphic registers, such as SIMD vector
registers that can be used both for floating-point and integer values.
This changes allows us to refactor VanillaReg to not store VGcPtr,
as that information is instead stored in the type of the usage of the
register.
Fixes #22297
- - - - -
78b99430 by sheaf at 2023-01-31T15:51:06-05:00
Revert "Cmm Lint: relax SIMD register assignment check"
This reverts commit 3be48877, which weakened a Cmm Lint check involving
SIMD vectors. Now that we keep track of the type a global register is
used at, we can restore the original stronger check.
- - - - -
be417a47 by Ben Gamari at 2023-01-31T15:51:45-05:00
nativeGen/AArch64: Fix debugging output
Previously various panics would rely on a half-written Show
instance, leading to very unhelpful errors. Fix this.
See #22798.
- - - - -
30989d13 by Ben Gamari at 2023-01-31T15:51:45-05:00
nativeGen: Teach graph-colouring allocator that x18 is unusable
Previously trivColourable for AArch64 claimed that at 18 registers were
trivially-colourable. This is incorrect as x18 is reserved by the platform on
AArch64/Darwin.
See #22798.
- - - - -
7566fd9d by Ben Gamari at 2023-01-31T15:51:45-05:00
nativeGen/AArch64: Fix graph-colouring allocator
Previously various `Instr` queries used by the graph-colouring allocator
failed to handle a few pseudo-instructions. This manifested in compiler
panicks while compiling `SHA`, which uses `-fregs-graph`.
Fixes #22798.
- - - - -
2cb500a5 by Ben Gamari at 2023-01-31T15:51:45-05:00
testsuite: Add regression test for #22798
- - - - -
03d693b2 by Ben Gamari at 2023-01-31T15:52:32-05:00
Revert "Hadrian: fix doc generation"
This is too large of a hammer.
This reverts commit 5640cb1d84d3cce4ce0a9e90d29b2b20d2b38c2f.
- - - - -
f838815c by Ben Gamari at 2023-01-31T15:52:32-05:00
hadrian: Sphinx docs require templated cabal files
The package-version discovery logic in
`doc/users_guide/package_versions.py` uses packages' cabal files to
determine package versions. Teach Sphinx about these dependencies in
cases where the cabal files are generated by templates.
- - - - -
2e48c19a by Ben Gamari at 2023-01-31T15:52:32-05:00
hadrian: Refactor templating logic
This refactors Hadrian's autoconf-style templating logic to be explicit
about which interpolation variables should be substituted in which
files. This clears the way to fix #22714 without incurring rule cycles.
- - - - -
93f0e3c4 by Ben Gamari at 2023-01-31T15:52:33-05:00
hadrian: Substitute LIBRARY_*_VERSION variables
This teaches Hadrian to substitute the `LIBRARY_*_VERSION` variables
in `libraries/prologue.txt`, fixing #22714.
Fixes #22714.
- - - - -
bd6c3038 by Sebastian Graf at 2023-02-03T17:23:56+01:00
Make DataCon workers strict in strict fields (#20749)
This patch tweaks `exprIsConApp_maybe`, `exprIsHNF` and frieds, and Demand
Analysis so that they exploit and maintain strictness of DataCon workers. See
`Note [Strict fields in Core]` for details.
Very little needed to change, and it puts field seq insertion done by Tag
Inference into a new perspective: That of *implementing* strict field semantics.
Before Tag Inference, DataCon workers are strict. Afterwards they are
effectively lazy and field seqs happen around use sites. History has shown
that there is no other way to guarantee taggedness and thus the STG Strict Field
Invariant.
Unfortunately, T21392 does no longer reproduce after this patch, so I marked it
as "not broken" in order to track whether we regress again in the future.
Fixes #20749, the satisfying conclusion of an annoying saga (cf. the ideas
in #21497 and #22475).
- - - - -
30 changed files:
- .gitlab-ci.yml
- .gitlab/gen_ci.hs
- .gitlab/jobs.yaml
- .gitlab/rel_eng/fetch-gitlab-artifacts/fetch_gitlab.py
- cabal.project-reinstall
- compiler/CodeGen.Platform.h
- compiler/GHC.hs
- compiler/GHC/Builtin/Names.hs
- compiler/GHC/Builtin/Types.hs
- compiler/GHC/Cmm/CallConv.hs
- compiler/GHC/Cmm/DebugBlock.hs
- compiler/GHC/Cmm/Expr.hs
- compiler/GHC/Cmm/Graph.hs
- compiler/GHC/Cmm/LayoutStack.hs
- compiler/GHC/Cmm/Lexer.x
- compiler/GHC/Cmm/Lint.hs
- compiler/GHC/Cmm/Node.hs
- compiler/GHC/Cmm/Opt.hs
- compiler/GHC/Cmm/Parser.y
- compiler/GHC/Cmm/Reg.hs
- compiler/GHC/Cmm/Sink.hs
- compiler/GHC/Cmm/ThreadSanitizer.hs
- compiler/GHC/Cmm/Type.hs
- compiler/GHC/Cmm/Utils.hs
- compiler/GHC/CmmToAsm.hs
- compiler/GHC/CmmToAsm/AArch64/CodeGen.hs
- compiler/GHC/CmmToAsm/AArch64/Instr.hs
- compiler/GHC/CmmToAsm/Dwarf.hs
- compiler/GHC/CmmToAsm/Dwarf/Types.hs
- compiler/GHC/CmmToAsm/PIC.hs
The diff was not included because it is too large.
View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/compare/07e6142fb15bac5a4059635c70639bc104e67d6b...bd6c3038f4bf3a48177c831535baf0928f503dfd
--
View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/compare/07e6142fb15bac5a4059635c70639bc104e67d6b...bd6c3038f4bf3a48177c831535baf0928f503dfd
You're receiving this email because of your account on gitlab.haskell.org.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.haskell.org/pipermail/ghc-commits/attachments/20230203/29de99ab/attachment-0001.html>
More information about the ghc-commits
mailing list