- Feb 06, 2021
-
-
Fixes #19307
-
Function arguments passed to the interpreter are extended to whole words. However, foreign function interface expects correctly typed argument pointers. Accordingly, we have to adjust argument pointers in case of a big-endian architecture. In contrast to function arguments where subwords are passed in the low bytes of a word, the return value is expected to reside in the high bytes of a word.
-
Just marking the `SimplTopEnv` parameter as one-shot was not enough to eta-expand `simplExpr`. Fixes #19302.
-
-
As found by @phadej in ghc/ghc!4740 (diffs, comment 327510) Also fix FastMutInt which allocating the size in bits instead of bytes.
-
Since commit be5d74ca the payload of a closure of Int<N> or Word<N> is not extended anymore to the machines word size. Instead, only the first N bits of a payload are written. This patch ensures that only those bits are read/written independent of the machines endianness.
-
Also updates the note with the case of multi-argument lambdas. Seems slightly beneficial based on the Cabal test: -O0: -1MB allocations (out of 50GB) -O : -1MB allocations (out of ~200GB)
-
-
This is avoids pushing the entire list to the stack before we can begin computing the result.
-
-
-
- Feb 02, 2021
-
-
These instances are useful so that a `GenClosure` form `ghc-heap` can be used as a key in a `Map`. Therefore the order itself is not important but just the fact that there is one.
-
Previously tc_eq_type would allocate a number of closures due to the two boolean "mode" flags, despite the fact that these were always statically known. To avoid this we force tc_eq_type to inline into its call sites, allowing the simplifier to eliminate both some runtime branches and the closure allocations.
-
- Feb 01, 2021
-
-
This commit introduces a DecoratedSDoc type which replaces the old ErrDoc, and hopefully better reflects the intent.
-
Updates Haddock submodule
-
This commit boldly removes the ErrDoc and the MsgDoc from the codebase. The former was introduced with the only purpose of classifying errors according to their importance, but a similar result can be obtained just by having a simple [SDoc], and placing bullets after each of them. On top of that I have taken the perhaps controversial decision to also banish MsgDoc, as it was merely a type alias over an SDoc and as such it wasn't offering any extra type safety. Granted, it was perhaps making type signatures slightly more "focused", but at the expense of cognitive burden: if it's really just an SDoc, let's call it with its proper name.
-
After commit 55ef3bdc, running `./configure` now generates a `driver/ghci/ghci-wrapper.cabal` file from `driver/ghci/ghci-wrapper.cabal.in`, which pollutes the `git` tree: ``` $ git status On branch master Your branch is up to date with 'origin/master'. Untracked files: (use "git add <file>..." to include in what will be committed) driver/ghci/ghci-wrapper.cabal nothing added to commit but untracked files present (use "git add" to track) ``` Since `driver/ghci/ghci-wrapper.cabal` is autogenerated, the sensible thing to do is to add it to `.gitignore`. While I was in town, I also added the standard `*.in` file disclaimer to `driver/ghci/ghci-wrapper.cabal.in`. [ci skip]
-
An accidental use of `tcSymbol` instead of `tcNat` in the `TypeLitNat` case of `mkTypeLitFromString` meant that it was possible to unsafely equate `Nat` with `Symbol`. A consequence of this is that you could write `unsafeCoerce`, as observed in #19288. This is fixed easily enough, thankfully. Fixes #19288.
-
- Jan 30, 2021
-
-
Krzysztof Gogolewski authored
- Add missing :since: for NondecreasingIndentation and OverlappingInstances - Remove duplicated descriptions for Safe Haskell flags and UndecidableInstances. Instead, the sections contain a link. - compare-flags: Also check for options supported by ghci. This uncovered two more that are not documented. The flag -smp was removed. - Formatting fixes - Remove the warning about -XNoImplicitPrelude - it was written in 1996, the extension is no longer dangerous. - Fix misspelled :reverse: flags Fixes #18958.
-
-
See T19264 for a tricky corner case when explicitly importing GHC.Num.BigNat and another module. With -dynamic-too, the FinderCache contains paths for non-dynamic interfaces so they must be loaded first, which is usually the case, except for some interfaces loaded in the backend (e.g. in CorePrep). So we must run the backend for the non-dynamic way first for -dynamic-too to work as it is but I broke this invariant in c85f4928 by mistakenly making the backend run for the dynamic way first.
-
The motivation is given in Note [tcFamTyPats: zonking the result kind]. Fixes #19250 -- the fix is easy.
-
Otherwise we end up with terminating \r characters on Windows.
-
This wraps the existing GHCi wrapper script (driver/ghci/ghci.c) in a cabal file and adds the package to Hadrian.
-
-
Following the example of `git`, as noted in #19030. Fixes #19030.
-
-fstg-lift-lams-rec-* and -fstg-lift-lams-non-rec-* were setting the same field. Fix manual: -fstg-lift-lams-non-rec-args is disabled by -fstg-lift-lams-non-rec-args-any, there's no -fno-stg-lift-*.
-
Provoked by #19074, this patch makes GHC.Core.PatSyn.PatSyn immutable, by recording only the *Name* of the matcher and builder rather than (as currently) the *Id*. See Note [Keep Ids out of PatSyn] in GHC.Core.PatSyn. Updates haddock submodule.
-
- Jan 29, 2021
-
-
It was checking the old path compiler/prelude/*, outdated with the new module hierarchy. I added a sanity check to avoid this in the future.
-
SDoc string literals created for example with `text "xyz"` are converted into `PtrString` (`Addr#` + size in bytes) with a rewrite rule to avoid allocating a String. Before this patch, the size in bytes was still computed at runtime. For every literal, we obtained the following pseudo STG: x :: Addr# x = "xzy"# s :: PtrString s = \u [] case ffi:strlen [x realWorld#] of (# _, sz #) -> PtrString [x sz] But since GHC 9.0, we can use `cstringLength#` instead to get: x :: Addr# x = "xzy"# s :: PtrString s = PtrString! [x 3#] Literals become statically known constructor applications. Allocations seem to decrease a little in perf tests (between -0.1% and -0.7% on CI). -
-
-
Related to a future change in Data.List, https://downloads.haskell.org/ghc/8.10.3/docs/html/users_guide/using-warnings.html?highlight=wcompat#ghc-flag--Wcompat-unqualified-imports Companion pull&merge requests: - https://github.com/judah/haskeline/pull/153 - https://github.com/haskell/containers/pull/762 - https://gitlab.haskell.org/ghc/packages/hpc/-/merge_requests/9 After these the actual change in Data.List should be easy to do.
-
StgLam is used exclusively in the work of CoreToStg, but there's nothing in the type of StgExpr that indicates this, so we're forced throughout the Stg.* codebase to handle cases like: case expr of ... StgLam lam -> panic "Unexpected StgLam" ... This patch removes the StgLam constructor from the base StgExpr so these cases no longer need to be handled. Instead, we use a new intermediate type in CoreToStg, PreStgRhs, to represent the RHS expression of a binding.
-
When -XStrict is enabled the rules for irrefutability are slightly modified. Specifically, the pattern in a program like do ~(Just hi) <- expr cannot be considered irrefutable. The ~ here merely disables the bang that -XStrict would usually apply, rendering the program equivalent to the following without -XStrict do Just hi <- expr To achieve make this pattern irrefutable with -XStrict the user would rather need to write do ~(~(Just hi)) <- expr Failing to account for this resulted in #19027. To fix this isIrrefutableHsPat takes care to check for two the irrefutability of the inner pattern when it encounters a LazyPat and -XStrict is enabled.
-
- Jan 28, 2021
-
-
Metric Decrease: T12545 T14683 T16577 T5321Fun T5642 -
-
to get this commit: commit 0952d94a2e30a3e7cddbede811b15fa70f7b9462 (HEAD) Author: Joachim Breitner <mail@joachim-breitner.de> Date: Tue Jan 19 11:39:38 2021 +0100 Make haddock more robust to changes to the `Language` data type With the introduction of GHC2021, the `Languages` data type in GHC will grow. In preparation of that (and to avoid changing haddock with each new language), this change makes the code handle extensions to that data type gracefully. (cherry picked from commit c341dd7c9c3fc5ebc83a2d577c5a726f3eb152a5) This can go in as preparation for !4853 -
-
-
