- Jul 14, 2019
-
-
These are unexploded minds as far as the linter is concerned. I don't want to hit in my MRs by mistake! I did this with `sed`, and then rolled back some changes in the docs, config.guess, and the linter itself.
-
- Jul 10, 2019
-
-
The simple optimiser was making an invalid transformation to join points -- yikes. The fix is easy. I also added some documentation about the fact that GHC uses a slightly more restrictive version of join points than does the paper. Fix #16918
-
- Jun 21, 2019
-
-
Due to #16799.
-
- Jun 16, 2019
-
-
The previous fix, !1095, didn't work as `--show-iface` ignores `-dsuppress-ticks`. Rework the test instead.
-
Previously there were a few cases where operations like `omit_ways` were incorrectly passed a single way (e.g. `omit_ways('threaded2')`). This won't work as the author expected.
-
- Jun 09, 2019
-
-
As noted in #16741, this test otherwise breaks when `base` is compiled with `-g`.
-
- May 31, 2019
-
-
Previously if we had f |> co where `f` had arity type `ABot N` and `co` had arity M and M < N, `arityType` would return `ABot M` which is wrong, because `f` is only known to diverge when applied to `N` args, as described in Note [ArityType]: If at = ABot n, then (f x1..xn) definitely diverges. Partial applications to fewer than n args may *or may not* diverge. This caused incorrect eta expansion in the simplifier, causing #16066. We now return `ATop M` for the same expression so the simplifier can't assume partial applications of `f |> co` is divergent. A regression test T16066 is also added.
-
- May 26, 2019
-
-
Following the discussion under #16473, this change allows the specializer to work on any dicts in a lambda, not just those that occur at the beginning. For example, if you use data types which contain dictionaries and higher-rank functions then once these are erased by the optimiser you end up with functions such as: ``` go_s4K9 Int# -> forall (m :: * -> *). Monad m => (forall x. Union '[State (Sum Int)] x -> m x) -> m () ``` The dictionary argument is after the Int# value argument, this patch allows `go` to be specialised.
-
- Mar 22, 2019
-
-
As per https://prime.haskell.org/wiki/Libraries/Proposals/MonadFail Coauthored-by:
Ben Gamari <ben@well-typed.com>
-
- Mar 20, 2019
-
-
- Mar 15, 2019
-
-
This moves all URL references to Trac tickets to their corresponding GitLab counterparts.
-
- Mar 08, 2019
-
-
Trac #10069 revealed that small NOINLINE functions didn't get split into worker and wrapper. This was due to `certainlyWillInline` saying that any unfoldings with a guidance of `UnfWhen` inline unconditionally. That isn't the case for NOINLINE functions, so we catch this case earlier now. Nofib results: -------------------------------------------------------------------------------- Program Allocs Instrs -------------------------------------------------------------------------------- fannkuch-redux -0.3% 0.0% gg +0.0% +0.1% maillist -0.2% -0.2% minimax 0.0% -0.8% -------------------------------------------------------------------------------- Min -0.3% -0.8% Max +0.0% +0.1% Geometric Mean -0.0% -0.0% Fixes #10069. ------------------------- Metric Increase: T9233 -------------------------
-
- Mar 05, 2019
- Mar 01, 2019
-
-
Ben Gamari authored
Otherwise these tests break spuriously when core libraries are compiled with source notes.
-
- Feb 22, 2019
-
-
In this commit commit 7833cf40 Date: Thu Jan 24 17:58:50 2019 +0100 Look through newtype wrappers (Trac #16254) we made exprIsConApp_maybe quite a bit cleverer. But I had not paid enough attention to keeping exactly the correct substitution and in-scope set, which led to Trac #16348. There were several buglets (like applying the substitution twice in exprIsConApp_maybe, but the proximate source of the bug was that we were calling addNewInScopeIds, which deleted things from the substitution as well as adding them to the in-scope set. That's usually right, but not here! This was quite tricky to track down. But it is nicer now.
-
- Feb 19, 2019
-
-
exprIsConApp_maybe could detect that I# 10 is a constructor application, but not that Size (I# 10) is, because it was an application with a nontrivial argument.
-
For case-of-known constructor to continue triggering early, exprIsConApp_maybe is now capable of looking through lets and cases. See #15840
-
-
- Feb 01, 2019
-
-
Sebastian Graf authored
-
- Jan 30, 2019
-
-
Ben Gamari authored
This eliminates most uses of run_command in the testsuite in favor of the more structured makefile_test.
-
Ben Gamari authored
This reverts commit 76c8fd67.
-
Ben Gamari authored
-
- Dec 25, 2018
-
-
Ben Gamari authored
* Mark arith011 as broken with integer-simple As noted in #16091, arith011 fails when run against integer-simple with a "divide by zero" exception. This suggests that integer-gmp and integer-simple are handling division by zero differently. * This also fixes broken_without_gmp; the lack of types made the previous failure silent, sadly. Improves situation of #16043. * Mark several tests implicitly depending upon integer-gmp as broken with integer-simple. These expect to see Core coming from integer-gmp, which breaks with integer-simple. * Increase runtime timeout multiplier of T11627a with integer-simple I previously saw that T11627a timed out in all profiling ways when run against integer-simple. I suspect this is due to integer-simple's rather verbose heap representation. Let's see whether increasing the runtime timeout helps. Fixes test for #11627. This is all in service of fixing #16043.
-
- Dec 24, 2018
-
-
Ömer Sinan Ağacan authored
Currently we duplicate top-level binder ids for no reason: $fEqHsExpr_$c/= $fEqHsExpr_$c/= = \ @ id_a27U $dEq_a27V eta_B2 eta1_B1 -> case $fEqHsExpr_$c== $dEq_a27V eta_B2 eta1_B1 of { False -> True; True -> False } with this patch we drop the first line when type signatures are not printed (-dsuppress-type-signatures, which is implied by -dsuppress-all) Reviewers: simonpj, bgamari Reviewed By: simonpj Subscribers: rwbarton, carter Differential Revision: https://phabricator.haskell.org/D5472
-
- Dec 22, 2018
-
-
Ömer Sinan Ağacan authored
-
- Dec 12, 2018
-
-
Simon Peyton Jones authored
This patch collects a few improvements triggered by Trac #15696, and fixing Trac #16029 * Stop making toCleanDmd behave specially for unlifted types. This special case was the cause of stupid behaviour in Trac #16029. And to my joy I discovered the let/app invariant rendered it unnecessary. (Maybe the special case pre-dated the let/app invariant.) Result: less special-case handling in the compiler, and better perf for the compiled code. * In WwLib.mkWWstr_one, treat seqDmd like U(AAA). It was not being so treated before, which again led to stupid code. * Update and improve Notes There are .stderr test wibbles because we get slightly different strictness signatures for an argumment of unlifted type: <L,U> rather than <S,U> for Int# <S,U> rather than <S(S),U(U)> for Int
-
- Nov 19, 2018
-
-
Sebastian Graf authored
Summary: Currently, `CoreToStg` annotates `StgRhsClosure`s with their set of non-global free variables. This free variable information is only needed in the final code generation step (i.e. `StgCmm.codeGen`), which leads to transformations such as `StgCse` and `StgUnarise` having to maintain this information. This is tiresome and unnecessary, so this patch introduces a trees-to-grow-like approach that only introduces the free variable set into the syntax tree in the code gen pass, along with a free variable analysis on STG terms to generate that information. Fixes #15754. Reviewers: simonpj, osa1, bgamari, simonmar Reviewed By: osa1 Subscribers: rwbarton, carter GHC Trac Issues: #15754 Differential Revision: https://phabricator.haskell.org/D5324
-
- Nov 07, 2018
-
-
davide authored
This patch makes the following improvement: - Automatically records test metrics (per test environment) so that the programmer need not supply nor update expected values in *.T files. - On expected metric changes, the programmer need only indicate the direction of change in the git commit message. - Provides a simple python tool "perf_notes.py" to compare metrics over time. Issues: - Using just the previous commit allows performance to drift with each commit. - Currently we allow drift as we have a preference for minimizing false positives. - Some possible alternatives include: - Use metrics from a fixed commit per test: the last commit that allowed a change in performance (else the oldest metric) - Or use some sort of aggregate since the last commit that allowed a change in performance (else all available metrics) - These alternatives may result in a performance issue (with the test driver) having to heavily search git commits/notes. - Run locally, performance tests will trivially pass unless the tests were run locally on the previous commit. This is often not the case e.g. after pulling recent changes. Previously, *.T files contain statements such as: ``` stats_num_field('peak_megabytes_allocated', (2, 1)) compiler_stats_num_field('bytes allocated', [(wordsize(64), 165890392, 10)]) ``` This required the programmer to give the expected values and a tolerance deviation (percentage). With this patch, the above statements are replaced with: ``` collect_stats('peak_megabytes_allocated', 5) collect_compiler_stats('bytes allocated', 10) ``` So that programmer must only enter which metrics to test and a tolerance deviation. No expected value is required. CircleCI will then run the tests per test environment and record the metrics to a git note for that commit and push them to the git.haskell.org ghc repo. Metrics will be compared to the previous commit. If they are different by the tolerance deviation from the *.T file, then the corresponding test will fail. By adding to the git commit message e.g. ``` # Metric (In|De)crease <metric(s)> <options>: <tests> Metric Increase ['bytes allocated', 'peak_megabytes_allocated'] \ (test_env='linux_x86', way='default'): Test012, Test345 Metric Decrease 'bytes allocated': Test678 Metric Increase: Test711 ``` This will allow the noted changes (letting the test pass). Note that by omitting metrics or options, the change will apply to all possible metrics/options (i.e. in the above, an increase for all metrics in all test environments is allowed for Test711) phabricator will use the message in the description Reviewers: bgamari, hvr Reviewed By: bgamari Subscribers: rwbarton, carter GHC Trac Issues: #12758 Differential Revision: https://phabricator.haskell.org/D5059
-
- Oct 15, 2018
-
-
Enabling -Werror=compat in the testsuite allows us to easily see the impact that a new warning has on code. It also means that in the period between adding the warning and making the actual breaking change, all new test cases that are being added to the testsuite will be forwards-compatible. This is good because it will make the actual breaking change contain less irrelevant testsuite updates. Things that -Wcompat warns about are things that are going to break in the future, so we can be proactive and keep our testsuite forwards-compatible. This patch consists of two main changes: * Add `TEST_HC_OPTS += -Werror=compat` to the testsuite configuration. * Fix all broken test cases. Test Plan: Validate Reviewers: hvr, goldfire, bgamari, simonpj, RyanGlScott Reviewed By: goldfire, RyanGlScott Subscribers: rwbarton, carter GHC Trac Issues: #15278 Differential Revision: https://phabricator.haskell.org/D5200
-
Unfortunately the implementation has confused the ability to make dynamic libraries with dynamic way. This constraint is only true for systems that require `-fPIC` for shared libraries. Since the implementation has this implicit assumption, mark the tests as requiring dynway. Test Plan: ./validate Reviewers: bgamari Reviewed By: bgamari Subscribers: simonpj, rwbarton, carter Differential Revision: https://phabricator.haskell.org/D5174
-
- Sep 26, 2018
-
-
Simon Peyton Jones authored
In this patch commit 869f69fd Author: Simon Peyton Jones <simonpj@microsoft.com> Date: Wed Dec 11 18:19:34 2013 +0000 Guarding against silly shifts we deal with silly shifts like (Sll 1 9223372036854775807). But I only dealt with primops that Int# and Word#. Alas, the same problem affects shifts of Integer, as Trac #15673 showed. Fortunately, the problem is easy to fix.
-
- Sep 13, 2018
-
-
Summary: Fix test for #15578 By allowing 0-arity values to be inlined, we end up changing boringness annotations, and this gets reflected in the Core output for this particular test. Add Notes for #15578 Test Plan: ./validate Reviewers: simonpj, bgamari Reviewed By: simonpj Subscribers: rwbarton, carter GHC Trac Issues: #15578 Differential Revision: https://phabricator.haskell.org/D5137
-
- Sep 12, 2018
-
-
Simon Peyton Jones authored
This patch takes up the missed opportunity described in Trac #15631, by convering a case into a let slightly more agressively. See Simplify.hs Note [Case-to-let for strictly-used binders] There is no measurable perf impact for good or ill. But the code is simpler and easier to explain.
-
- Sep 05, 2018
-
-
Simon Peyton Jones authored
Trac #15445 showed that, as a result of CSE, a function with an automatically generated specialisation RULE could be inlined before the RULE had a chance to fire. This patch attaches a NOINLINE[2] activation to the Id, during CSE, to stop this happening. See Note [Delay inlining after CSE] ---- Historical note --- This patch is simpler and more direct than an earlier version: commit 2110738b Author: Simon Peyton Jones <simonpj@microsoft.com> Date: Mon Jul 30 13:43:56 2018 +0100 Don't inline functions with RULES too early We had to revert this patch because it made GHC itself slower. Why? It delayed inlining of /all/ functions with RULES, and that was very bad in TcFlatten.flatten_ty_con_app * It delayed inlining of liftM * That delayed the unravelling of the recursion in some dictionary bindings. * That delayed some eta expansion, leaving flatten_ty_con_app = \x y. let <stuff> in \z. blah * That allowed the float-out pass to put sguff between the \y and \z. * And that permanently stopped eta expasion of the function, even once <stuff> was simplified. -- End of historical note ---
-
- Aug 27, 2018
-
-
Summary: The current pattern-match coverage checker implements the formalism presented in the //GADTs Meet Their Match// paper in a fairly faithful matter. However, it was discovered recently that there is a class of unreachable patterns that //GADTs Meet Their Match// does not handle: unreachable code due to strict argument types, as demonstrated in #15305. This patch therefore goes off-script a little and implements an extension to the formalism presented in the paper to handle this case. Essentially, when determining if each constructor can be matched on, GHC checks if its associated term and type constraints are satisfiable. This patch introduces a new form of constraint, `NonVoid(ty)`, and checks if each constructor's strict argument types satisfy `NonVoid`. If any of them do not, then that constructor is deemed uninhabitable, and thus cannot be matched on. For the full story of how this works, see `Note [Extensions to GADTs Meet Their Match]`. Along the way, I did a little bit of much-needed refactoring. In particular, several functions in `Check` were passing a triple of `(ValAbs, ComplexEq, Bag EvVar)` around to represent a constructor and its constraints. Now that we're adding yet another form of constraint to the mix, I thought it appropriate to turn this into a proper data type, which I call `InhabitationCandidate`. Test Plan: make test TEST=T15305 Reviewers: simonpj, bgamari Reviewed By: simonpj Subscribers: rwbarton, carter GHC Trac Issues: #15305 Differential Revision: https://phabricator.haskell.org/D5087
-
- Aug 21, 2018
-
-
Simon Peyton Jones authored
We were failing to keep correct strictness info when eta-expanding join points; Trac #15517. The situation was something like \q v eta -> let j x = error "blah -- STR Lx bottoming! in case y of A -> j x eta B -> blah C -> j x eta So we spot j as a join point and eta-expand it. But we must also adjust the stricness info, else it vlaimes to bottom after one arg is applied but now it has become two. I fixed this in two places: - In CoreOpt.joinPointBinding_maybe, adjust strictness info - In SimplUtils.tryEtaExpandRhs, return consistent values for arity and bottom-ness
-
- Aug 18, 2018
-
-
Summary: Write tests with fewer lines. See comments of nomeata in https://phabricator.haskell.org/D5038. Test Plan: make test TESTS='T9441a T9441b T9441c' Reviewers: nomeata, dfeuer, bgamari Reviewed By: nomeata Subscribers: rwbarton, carter GHC Trac Issues: #9441 Differential Revision: https://phabricator.haskell.org/D5076
-
- Aug 12, 2018
-
-
Summary: In order to disambiguate names from different modules, qualify all names that don't originate in the current module. Also update docs for QueryQualifyName Test Plan: validate Reviewers: simonpj, bgamari Reviewed By: simonpj Subscribers: rwbarton, thomie, carter, tdammers GHC Trac Issues: #15269 Differential Revision: https://phabricator.haskell.org/D4852
-
Summary: Add testcase for #9441 Test Plan: make test TESTS="T9441a T9441b T9441c" Reviewers: dfeuer, simonpj, thomie, austin, bgamari Reviewed By: dfeuer Subscribers: rwbarton, carter GHC Trac Issues: #9441 Differential Revision: https://phabricator.haskell.org/D5038
-
- Aug 11, 2018
-
-
Krzysztof Gogolewski authored
Summary: - remove clean_cmd - framework_failures was undefined - times_file was not used - if_verbose_dump was called only when verbose >= 1; remove the check - simplify normalise_whitespace Test Plan: validate Reviewers: bgamari, thomie Reviewed By: thomie Subscribers: rwbarton, carter Differential Revision: https://phabricator.haskell.org/D5061
-
