- Feb 29, 2020
-
-
-
-
In #14335 we want to be able to use both the internal interpreter (for the plugins) and the external interpreter (for TH and GHCi) at the same time. This patch performs some preliminary refactoring: the `hsc_interp` field of HscEnv replaces `hsc_iserv` and is now used to indicate which interpreter (internal, external) to use to execute TH and GHCi. Opt_ExternalInterpreter flag and iserv options in DynFlags are now queried only when we set the session DynFlags. It should help making GHC multi-target in the future by selecting an interpreter according to the selected target.
-
- Added a few comments in StgPAP - Added a few comments and assertions in scavenge_small_bitmap and walk_large_bitmap - Did tiny refactor in GHC.Data.Bitmap: added some comments, deleted dead code, used PlatformWordSize type.
-
-
loadInterface replaces the `mi_decls`, `mi_insts`, `mi_fam_insts`, `mi_rules`, `mi_anns` fields of ModIface with `undefined` before inserting the interface into the EPS. However, we still want to give loadInterface plugins access to these fields. Consequently, we want to pass the unmodified `ModIface` the plugin.
-
When GHC is parsing a file generated by a tool, e.g. by the C preprocessor, the tool may insert #line pragmas to adjust the locations reported to the user. As the result, the locations recorded in RealSrcLoc are not monotonic. Elements that appear later in the StringBuffer are not guaranteed to have a higher line/column number. In fact, there are no guarantees whatsoever, as #line pragmas can arbitrarily modify locations. This lack of guarantees makes ideas such as #17544 infeasible. This patch adds an additional bit of information to every SrcLoc: newtype BufPos = BufPos { bufPos :: Int } A BufPos represents the location in the StringBuffer, unaffected by any pragmas. Updates haddock submodule. Metric Increase: haddock.Cabal haddock.base haddock.compiler MultiLayerModules Naperian parsing001 T12150 -
Add arm-unknown-linux-gnueabi, which is used by Debian's ARM EABI port (armel), as an LLVM target.
-
This fixes #17792 See Note [VTA for out-of-scope functions] in TcExpr
-
- Feb 28, 2020
-
-
This was making -Werror builds fail on Windows (at least with Hadrian).
-
Previously we were tracing the object we were asked to mark, even if it lives in a compact region. However, there is no need to do this; we need only to mark the region itself as live. I have seen a segfault due to this due to the concurrent mark seeing a an object in the process of being compacted by the mutator.
-
Previously, we used relative paths in submodules. When cloning from GitHub, they had to be manually tweaked. Since a76b233d we use absolute paths, so this workaround can be removed.
-
- Feb 27, 2020
-
-
-
Consider ```hs data T = A | B | C f :: T -> Int f A = 1 f x = case x of A -> 2 B -> 3 C -> 4 ``` Clearly, the RHS returning 2 is redundant. But we don't currently see that, because our approximation to the covered set of the inner case expression just picks up the positive information from surrounding pattern matches. It lacks the context sensivity that `x` can't be `A` anymore! Therefore, we adopt the conceptually and practically superior approach of reusing the covered set of a particular GRHS from an outer pattern match. In this case, we begin checking the `case` expression with the covered set of `f`s second clause, which encodes the information that `x` can't be `A` anymore. After this MR, we will successfully warn about the RHS returning 2 being redundant. Perhaps surprisingly, this was a great simplification to the code of both the coverage checker and the desugarer. Found a redundant case alternative in `unix` submodule, so we have to bump it with a fix. Metric Decrease: T12227
-
- Feb 26, 2020
-
-
Update haddock submodule
-
-
Fixes name of C compiler.
-
-
-
This exposes the flag of the same name supported by the testsuite driver.
-
This allows us to work-around distribution-specific breakage easily.
-
I've ruined two trees already by failing to pass --flavour to hadrian. Let's factor this out so it can be reused during troubleshooting.
-
Otherwise we are sensitive to libc's buffering strategy. Similar to the issue fixed in 543dfaab.
-
Darwin sh doesn't support \e.
-
-
Needs to `write` bytes, not str.
-
This should fix #17861.
-
* FailablePattern can no longer be created since ab51bee4 Therefore, Opt_WarnMissingMonadFailInstances has no effect anymore. * XWrap is no longer used, it was moved to an extension field
-
-
Move the gmp configuration to its own binding.
-
-
This reverts commit 8924224e and fixes #17787.
-
- Feb 25, 2020
-
-
-
* Remove outdated Note [HsForAllTy tyvar binders] and [Context quantification]. Since the wildcard refactor 1e041b73, HsForAllTy no longer has an flag controlling explicity. The field `hsq_implicit` is gone too. The current situation is covered by Note [HsType binders] which is already linked from LHsQTyVars. * Small refactor in CoreLint, extracting common code to a function * Remove "not so sure about WpFun" in TcEvidence, per Richard's comment !852 (comment 223226) * Use mkIfThenElse in Foreign/Call, as it does exactly what we need.
-
-
- Feb 24, 2020
-
-
-
-
-
Before this patch, GHC relied on Ord SrcSpan to identify source elements, by using SrcSpan as Map keys: blackList :: Map SrcSpan () -- compiler/GHC/HsToCore/Coverage.hs instanceMap :: Map SrcSpan Name -- compiler/GHC/HsToCore/Docs.hs Firstly, this design is not valid in presence of UnhelpfulSpan, as it distinguishes between UnhelpfulSpan "X" and UnhelpfulSpan "Y", but those strings are messages for the user, unfit to serve as identifiers for source elements. Secondly, this design made it hard to extend SrcSpan with additional data. Recall that the definition of SrcSpan is: data SrcSpan = RealSrcSpan !RealSrcSpan | UnhelpfulSpan !FastString Say we want to extend the RealSrcSpan constructor with additional information: data SrcSpan = RealSrcSpan !RealSrcSpan !AdditionalInformation | UnhelpfulSpan !FastString getAdditionalInformation :: SrcSpan -> AdditionalInformation getAdditionalInformation (RealSrcSpan _ a) = a Now, in order for Map SrcSpan to keep working correctly, we must *ignore* additional information when comparing SrcSpan values: instance Ord SrcSpan where compare (RealSrcSpan r1 _) (RealSrcSpan r2 _) = compare r1 r2 ... However, this would violate an important law: a == b therefore f a == f b Ignoring AdditionalInformation in comparisons would mean that with f=getAdditionalInformation, the law above does not hold. A more robust design is to avoid Ord SrcSpan altogether, which is what this patch implements. The mappings are changed to use RealSrcSpan instead: blackList :: Set RealSrcSpan -- compiler/GHC/HsToCore/Coverage.hs instanceMap :: Map RealSrcSpan Name -- compiler/GHC/HsToCore/Docs.hs All SrcSpan comparisons are now done with explicit comparison strategies: SrcLoc.leftmost_smallest SrcLoc.leftmost_largest SrcLoc.rightmost_smallest These strategies are not subject to the law mentioned above and can easily discard both the string stored in UnhelpfulSpan and AdditionalInformation. Updates haddock submodule.
-
- Feb 23, 2020
-
-
