Closed
Milestone
expired on May 19, 2016
8.0.1
The 8.0.1 release will be a new, super-major release.
- [https://mail.haskell.org/pipermail/ghc-devs/2016-January/010966.html "[ANNOUNCE] Glasgow Haskell Compiler 8.0.1, release candidate 1"]
- [https://mail.haskell.org/pipermail/ghc-devs/2016-February/011309.html "[ANNOUNCE] Glasgow Haskell Compiler 8.0.1, release candidate 2"]
- [https://mail.haskell.org/pipermail/ghc-devs/2016-April/011862.html "[ANNOUNCE] GHC 8.0.1 release candidate 3 available"]
- [https://mail.haskell.org/pipermail/ghc-devs/2016-April/011955.html "[ANNOUNCE] GHC 8.0.1 release candidate 4 available"]
- [https://mail.haskell.org/pipermail/ghc-devs/2016-May/012098.html "[ANNOUNCE] GHC 8.0.1 is available!"]
See also [[Status/GHC-8.0.1]] for more details.
Migration Guide: [[Migration/8.0]]
Unstarted Issues (open and unassigned)
261
- [bug] ModOrigin: hidden module redefined
- -fth-dec-file uses qualified names in binding positions
- -fth-dec-file uses qualified names from hidden modules
- UnicodeSyntax documentation lists wrong symbols
- Several profiling tests give different results optimised vs. unoptimised
- cross building integer-gmp is running target program on build host
- Touching a file that uses TH triggers TH recompilation flood
- malloc and mallocArray ignore Storable alignment requirements
- Debug.Trace.trace is too strict
- dep_orphs in Dependencies redundantly records type family orphans
- Simplifer non-determinism leading to 8 fold difference in run time performance
- Improve flag description in the user guide
- prefetch primops are not currently useful
- LLVM vs NCG: floating point numbers close to zero have different sign
- Lack of type information in GHC error messages when the liberage coverage condition is unsatisfied
- GHC not recognizing INPUT(-llibrary) in linker scripts
- ghc-pkg complains about missing haddock interface files
- GHC unnecessarily sign/zero-extends C call arguments
- Inferring Safe mode with GeneralizedNewtypeDeriving is wrong
- I/O manager causes unnecessary syscalls in send/recv loops
- long compilation time for module with large data type and partial record selectors
- Arithmetic overflow from (minBound :: Int) `quot` (-1)
- ghc -M doesn't handle addDependentFile or #included files
- Building an empty module with profiling requires profiling libraries for integer-gmp
- -ffull-laziness does more harm than good
- one-shot compilation + TH doesn't see instances that is seen in batch mode
- type nats solver is too weak!
- reify module list in TH
- Filesystem related tests failed on solaris (SmartOS)
- unexpected behavior with encodeFloat on large inputs
- bad alignment in code gen yields substantial perf issue
- Interface hashes include time stamp of dependent files (UsageFile mtime)
- Assertion failure when using multithreading in debug mode.
- Bug in hsc2hs --cross-safe
- Error on pattern matching of an existential whose context includes a type function
- RebindableSyntax and Arrow
- Superclass methods are left unspecialized
- GHCI core dumps when used with VTY
- building GHC overwrites the installed package database if GHC_PACKAGE_PATH is set
- StablePtrs should be organized by generation for efficient minor collections
- Location in -fdefer-type-errors
- Handling ImplicitParams in Instance Declaration
- Cross compilation support for LLVM backend
- Hard ghc api crash when calling runStmt on code which has not been compiled
- Using -with-rtsopts=-N should fail unless -threaded is also specified
- Room for GHC runtime improvement >~5%, inlining related
- Generated C code under -prof -fprof-auto -fprof-cafs very slow to compile
- GHC API reports CPP errors in confusing ways
- GHC compile times are seriously non-linear in program size
- DefaultSignatures conflict with default implementations
- CAPI doesn't work with ghci
- Panic: mkNoTick: Breakpoint loading modules with -O2 via API
- Identical alts/bad divInt# code
- rule not firing
- When building GHC: Failed to load interface for `GHC.Fingerprint'
- Recompilation check fails for TH unless functions are inlined
- Callstack depends on way (prof, profasm, profthreaded
- INLINEing top-level patterns causes ghc to emit 'arity missing' traces
- Stack trace truncated too much with indirect recursion
- GHC's -fprof-auto does not work with LINE pragmas
- Cannot recover (good) inlining behaviour from 7.0.2 in 7.4.1
- reject reading rationals with exponent notation
- Profiling with -p not written if killed with SIGTERM
- GHCi runtime linker cannot link with duplicate common symbols
- Type operators are not accepted as variables in contexts
- Top level splice in Template Haskell has over-ambitious lexical scope?
- INLINABLE fails to specialize in presence of simple wrapper
- hsc2hs forces wordsize (i.e. -m32 or -m64) to be the choice of GHC instead of allowing a different (or no/default choice)
- Offer a compiler warning for failable pattern matches
- -faggressive-primops change caused a failure in perf/compiler/parsing001
- The -L flag should not exist
- simple program fails with -shared on mac
- Slow 64-bit primops on 32 bit system
- unreg compiler: warning: conflicting types for built-in function ‘memcpy’
- Inlined functions aren't fully specialised
- Compiling with -O makes some expressions too lazy and causes space leaks
- Runtime error when allocating lots of memory
- Add blockST for nested ST scopes
- Better inlining test in CoreUnfold
- increase error message detail for module lookups failure due to hi references
- Another SpecConstr infelicity
- GHC.ConsoleHandler does not call back application when Close button is pressed
- literate markdown not handled correctly by unlit
- Finding the right loop breaker
- Windows: Dynamic linking doesn't work out-of-the-box
- Re-linking avoidance is too aggressive
- Accept expressions in left-hand side of quasiquotations
- LLVM compiles Updates.cmm badly
- Optimisations give bad core for foldl' (flip seq) ()
- Poor -fspec-constr-count=n warning messages
- do not consider associativity for unary minus for fixity resolution
- reject unary minus in infix left hand side function bindings that resolve differently as expressions
- GHC API messes up signal handlers
- CPP+QuasiQuotes confuses compilation errors' line numbers
- ffi005 fails on OS X
- Support for ABI versioning of C libraries
- ghc-pkg should check for existence of extra-libraries
- Strange display behaviour in GHCi
- Handle multiline input in GHCi history
- "Fix" pervasive-but-unnecessary signedness in GHC.Prim
- section parse errors, e.g. ( let x=1 in x + )
- Avoid Haddock-links to the Prelude
- Improve inlining for local functions
- SpecConstr for join points
- Parsing of lambdas is not consistent with Haskell'98 report.
- Comparisons against minBound/maxBound not optimised for (Int|Word)(8|16|32)
- Bad error reporting when calling a function in a module which depends on a DLL on Windows
- Layout and pragmas
- ghc -M should emit dependencies on CPP headers
- Bizzarely bloated binaries
- Allocation where none should happen
- Generated ghc man page missing xrefs
- CInt FFI exports do not use C int in _stub.h header file
- package.conf.d should be under /var, not /usr
- Returning a known constructor: GHC generates terrible code for cmonad
- Over-eager GC when blocked on a signal in the non-threaded runtime
- Avoid reconstructing dictionaries in recursive instance methods
- GHC's GC default heap growth strategy is not as good as other runtimes
- Int / Word / IntN / WordN are unequally optimized
- ghc FFI doesn't support thiscall
- divInt# floated into a position which leads to low arity
- LDFLAGS ignored by build system
- Make a way to tell GHC that a pragma name should be "recognised"
- Document -pgmL (Use cmd as the literate pre-processor)
- Avoid unnecessary evaluation when unpacking constructors
- Improve SpecConstr for join points
- aborting an STM transaction should throw an exception
- Optimizer misses unboxing opportunity
- MutableByteArray# is slower than Addr#
- num009 fails on OS X 10.5?
- Compilation of large source files requires a lot of RAM
- Needless reboxing of values when returning from a tight loop
- inlining defeats seq
- unhelpful error message for a misplaced DEPRECATED pragma
- STM slightly conservative on write-only transactions
- Precedence and associativity rules ignored when mixing infix type and data constructors in a single expression
- Make distclean (still) doesn't
- Broken link testing
- Derived Read instances for recursive datatypes with infix constructors are too inefficient
- ghc runs preprocessor too much
- ghc panic with mutually recursive modules and template haskell
- GHCi on x86_64, cannot link to static data in shared libs
- Control.Exception.assert should perhaps take an implicit call stack
- Test suite: Support non-utf8 locale
- Make quot/rem/div/mod with known divisors fast
- Template Haskell for cross compilers (port from GHCJS)
- Improve `mkInteger` interface
- ghc does not expose branchless max/min operations as primops
- Improve parser error reporting in `ghc-pkg`
- cross compiling for x86_64 solaris2
- explore ways to possibly use more tag bits in x86_64 pointers
- Implement unloading of shared libraries
- make better/more robust loopbreaker choices
- Reading ./.ghci files raises security issues
- Defer other kinds of errors until runtime, not just type errors
- Remove in-tree gmp
- Improve join point inlining
- Track -dynamic/-fPIC to avoid obscure linker errors
- GHC should use the standard binary package
- Refactor Template Haskell syntax conversions
- Improve CPR analysis
- Optimisation: Nested CPR
- Make it easy to find documentation for GHC and installed packages
- Optimisation: eliminate unnecessary heap check in recursive function
- Generalise the ! and UNPACK mechanism for data types, to unpack function arguments
- Avoidance of unaligned loads is overly conservative
- Explore when to apply static argument transformation
- Warning Suppression
- Program location for thread error messages
- Access to module renaming with reifyModule, in TemplateHaskell
- Applicative Comprehensions
- Implement deprecation-warnings for class-methods to non-method transitions
- Expose ghc-bin code as a library
- Allow expressions in patterns
- Allow `State# s` argument/result types in `ccall` FFI imports
- Branchless arithmetic operations
- add anyToAddr# :: (#a#)-> Addr# primop (inverse of addrToAny#)
- Show parenthesised output of expressions in ghci
- Pattern synonym used in an expression context could have different constraints to pattern used in a pattern context
- Annotation reification with types in TH
- equip GHC with an accurate internal model of floating point
- add idris style EDSL support for deep embedding lambdas
- prefetch# isn't as general as it should be (currently the general version isn't type safe)
- data families and TH names do not mix well (e.g. cannot use TH deriving)
- UNPACK polymorphic fields
- SafeHaskell implying other options
- Allow compatible type synonyms to be the return type of a GADT data constructor.
- runghc (runhaskell) should be able to reload code on editing
- GHC does not generate great code for bit-level rotation
- Need for extra warning pragma for accidental pattern matching in do blocks
- Allow the evaluation of declaration splices in GHCi
- Allow CAFs kept reachable by FFI to be forcibly made unreachable for GC
- Specialise INLINE functions
- Give more detailed information about PINNED data in a heap profile
- Use a class to control FFI marshalling
- GHCi commands case insensitive
- Allow type signature in export list
- arrow analogs of lambda case and multi-way if
- Add the ability to statically define a `FunPtr` to a haskell function
- Allow defining kinds alone, without a datatype
- Add compilation stage plugins
- Holes with other constraints
- Allow both INLINE and INLINABLE for the same function
- FFI and CAPI needs {-# INCLUDE #-} back?
- Shorter qualified import statements
- Support pin-changing on ByteArray#s
- Template Haskell: support for Haddock comments
- Warnings about impossible MPTCs would be nice
- Less noisy version of -fwarn-name-shadowing
- Licensing requirements and copyright notices
- need a version of hs_init that returns an error code for command-line errors
- Support static linker semantics for archives and weak symbols
- Misfeature of Cmm optimiser: no way to extract a branch of expression into a separate statement
- Allow unicode sub/superscript symbols in operators
- Pragma to SPECIALISE on value arguments
- Warning about module abbreviation clashes
- Warning about variables with leading underscore that are used anyway
- Remove indirections caused by sum types, such as Maybe
- Make event tracing conditional on an RTS flag only
- Consider usage files in the GHCi recompilation check
- Loop strength reduction for array indexing
- Implement TDNR
- Loop optimization: identical counters
- Allow specifying .hi files of imports on command line in batch mode
- Interactive "do" notation in GHCi
- Relax restrictions on type family instance overlap
- Template Haskell lets you reify supposedly-abstract data types
- New primops for indexing: index*OffAddrUsing# etc
- Two sided sections
- Please consider adding support for local type synonyms
- RTS GC Statistics from -S should be logged via the eventlog system
- Make ghci's -l option consistent with GNU ld's -l option
- allow existential wrapper newtypes
- lift restrictions on records with existential fields, especially in the presence of class constraints
- allow to set ghc search path globally (a'la CPATH)
- Improve granularity of UndecidableInstances
- Some mechanism for eliminating "absurd" patterns
- Find import declaration importing a certain function
- Show type of most recent expression in GHCi
- GHC API is not thread safe
- Allow for multiple linker instances
- How to start an emacs editor within ghci asynchronously with :edit filename.hs :set editor emacs & don't go
- split rts headers into public and private
- Add dynCompileCoreExpr :: GhcMonad m => Bool -> Expr CoreBind -> m Dynamic to ghc-api
- warn about language extensions that are not used
- Report out of date interface files robustly
- Revise the rules for -XExtendedDefaultRules
- Treat -X flags consistently in GHCi
- Avoid excessive specialisation in SpecConstr
- Add an option to read file names from a file instead of the command line
- Warning for missing export lists
- provide -mwindows option like gcc
- Warn about usage of `OPTIONS_GHC -XLanguageExtension`
- :browse limitations (browsing virtual namespaces, listing namespaces)
- Improve Template Haskell error recovery
- Warn if functions are exported whose types cannot be written
- explicitly importing deprecated symbols should elicit the deprecation warning
- Allow unconstrained existential contexts in newtypes
- Add a total order on type constructors
- [Debugger] Watch on accesses of "variables"
- Automatic heap profile intervals
Ongoing Issues (open and assigned)
106
- -fth-dec-file outputs invalid case clauses
- Linker script patch in rts/Linker.c doesn't work for (non-C or non-en..) locales
- Clean up GHC.RTS.Flags
- map/coerce rule does not fire until the coercion is known
- More lazy orphan module loading
- Foldable doesn't have any laws
- Comment in GHC.Base about GHC.Prim does not appear to be correct
- ghc --print-(gcc|ld)-linker-flags broken
- Add warning for invalid digits in integer literals
- Massive blowup of code size on trivial program
- Global big object heap allocator lock causes contention
- Strange slowness when using async library with FFI callbacks
- GHC.Base.{breakpoint, breakpointCond} do nothing
- defer StackOverflow exceptions (rather than dropping them) when exceptions are masked
- One-shot mode is buggy w.r.t. hs-boot files
- need types to express constant argument for primop correctness
- Demand analyser is unpacking too deeply
- dynamicToo001 fails on Windows
- Hackage docs for base library contain broken links
- Opportunity to improve CSE
- GHC doesn't optimize (strict) composition with id
- Bug with PolyKinds, type synonyms & GADTs
- Regression in optimisation time of functions with many patterns (6.12 to 7.4)?
- Share top-level code for strings
- LLVM incorrectly hoisting loads
- ghc-7 assumes incoherent instances without requiring language `IncoherentInstances`
- Compiling DynFlags is jolly slow
- hSetNewlineMode and hSetEncoding can be performed on closed and semi-closed handles
- Add flag to configure that skips overwriting of symlinks on install
- Inlining the single method of a class can shadow rules
- Inlining depends on datatype size, even with INLINE pragmas
- Panic using mixing list with parallel arrays incorrectly
- isInstance does not work for compound types
- Register allocators can't handle non-uniform register sets
- The `Read` instance of `Rational` does not support decimal notation
- Show instance for integer-simple is not lazy enough
- Liberate case not happening
- GHC retains unnecessary binding
- Adding a type signature changes heap allocation into stack allocation without changing the actual type
- Performance regression 7.0 -> 7.2 (still in 7.4)
- Cannot tell from an exception handler whether the exception was asynchronous
- core lint error with arrow notation and GADTs
- Getting stdout and stderr as a single handle from createProcess does not work on Windows
- Sharing across functions causing space leak
- Polymorphic instances aren't automatically specialised
- check_overlap panic (7.1 regression)
- Missing type checks for arrow command combinators
- Improve consistency checking for family instances
- stub header files don't work with the MS C compiler
- Non-standard compile plus Template Haskell produces spurious "unknown symbol" linker error
- (^^) is not correct for Double and Float
- The dreaded SkolemOccurs problem
- Primitive constant unfolding
- Strict constructor fields inspected in loop
- GHC Bindist is Broken on FreeBSD/amd64
- SpecConstr should exploit cases where there is exactly one call pattern
- The Ord instance for unboxed arrays is very inefficient
- Permission denied error with runProcess/openFile
- floor(0/0) should not be defined
- Missed optimisation with dictionaries and loops
- -fhpc inteferes/prevents rewrite rules from firing
- cpuTimePrecision is wrong
- hpc mix files for Main modules overwrite each other
- Stack check for AP_STACK
- Missed opportunity for let-no-esape
- arrow notation: incorrect scope of existential dictionaries
- Smooth out the differences between `compiler/utils/Pretty.hs` and `libraries/pretty`
- The list modules need a bit of post-BBP shaking
- audit ghc floating point support for IEEE (non)compliance
- Using GADT's to maintain invariant in GHC libraries
- Add rules involving `coerce` to the libraries
- clean up dependency and usages handling in interface files
- Add dummy undefined symbols to indicate ways
- LLVM: Improve alias analysis / performance
- Soft heap limit flag
- Fix LLVM backend for PowerPC
- Make a proper options parser for the RTS
- LLVM: Stack alignment on OSX
- New codegen: allocate large objects using allocateLocal()
- Avoid generating C trigraphs
- implement waitForProcess using signals
- Make the External Package Table contain ModDetails not ModIface
- Useful optimisation for set-cost-centre
- State a law for foldMap
- Improve enumFromX support for OverloadedLists
- symbols should/might be type level lists of chars
- Get rid of HEAP_ALLOCED on Windows (and other non-Linux platforms)
- Resource limits for Haskell
- Add SIMD support to x86/x86_64 NCG
- Stride scheduling for Haskell threads with priorities
- generalizing overloaded list syntax to Sized Lists, HLists, HRecords, etc
- Generic1 deriving: Can we replace Rec1 f with f :.: Par1?
- mkWeakMVar is non-compositional
- Add tryWriteTBQueue to Control.Concurrent.STM.TBQueue
- Allow declaration splices inside declaration brackets
- parBufferWHNF could be less subtle
- "guarded instances": instance selection can add extra parameters to the class
- CPR optimisation for sum types if only one constructor is used
- Polymorphic Data.Dynamic
- GHC.Conc modifyTVar primitive
- System.Posix.Signals should provide a way to set the SA_NOCLDWAIT flag
- :info printing instances often isn't wanted
- Enumeration of values for `Sys.Info.os`, `Sys.Info.arch`
- Word type to Double or Float conversions are slower than Int conversions
- ghc --cleanup
- Offer control over branch prediction
Completed Issues (closed)
928
- SMP primitives broken on power(pc)
- Documentation will not build on platforms where GNU make is not called make
- PowerPC: Unsupported relocation against x0
- Split ghc-boot so we have better dependency hygiene
- Undefined stg_sel_17_upd_info symbols on OS X
- Configure script doesn't check libdw version
- Filtering of cost-center profiler output no longer works
- Merge "Skip TEST=TcCoercibleFail when compiler_debugged"
- Implement `-f(no-)version-macros` flag for controlling version macro generation
- GHC 8 superclass chain constraint regression
- Error in optCoercion
- 'Strict' extension is incompatible with 'deriving' mechanism
- TYPE 'UnboxedTupleRep is a lie
- Merge some TypeInType fixes
- Constant folding on 'mod/Word' - incorrect result
- Regression using NamedFieldPuns with qualified field names
- Allow plugins to define built-in rules
- LLVM code generator produces mal-formed LLVM blocks
- assertPprPanic, called at compiler/types/TyCoRep.hs:1932
- Core lint error in result of Specialise for TEST=T3220 WAY=optasm
- Core lint error in simplifier when compiling Rules1 with -O -dcore-lint
- Heterogeneous type equality evidence ignored
- Cannot declare hs-boot declaration if there is already a value in scope
- document TypeInType
- Possible type-checker regression in GHC 8.0 when compiling `microlens`
- Panic (ASSERT failed) in compiler/types/TyCoRep.hs:1939
- Optimize cmpTypeX
- GHC 8.0 can't be bootstrapped with GHC 8.0
- BangPatterns-related behavior regressions on GHC 8.0
- Using Cabal 1.22 against GHC 8.0 results in unhelpful errors
- catch _|_ breaks at -O1
- GHC falls into a hole if given incorrect kind signature
- stg_ap_pp_fast doesn't pass the argument in the arity=1 case
- Levity polymorphism checks are inadequate
- Use of typechecker plugin erroneously triggers "unbound implicit parameter" error
- Terrible failure of type inference in visible type application
- GHC.Prim does not export Constraint
- Cannot export operator newtype
- Make unrecognised `-W` flags a warning rather than an error
- GHC 8.0-rc1's linker does not work in OSX
- Binary distributions seem to lack haddock docs
- Panic with -XStrict: StgCmmEnv: variable not found
- GHC 8.1.20160111 fails to bootstrap itself.
- -XTypeInType uses up all memory when used in data family instance
- Incorrect failure of type-level skolem escape check
- No match in record selector ctev_dest
- Ill-kinded instance head involving -XTypeInType can invoke GHC panic
- Type mismatch in local definitions in Haskell 98 code
- deriving Ix with custom ifThenElse causes "Bad call to tagToEnum#"
- Solver hits iteration limit in code without recursive constraints
- `-Woverlapping-patterns` induced memory-blowup
- Redundant superclass warnings being included in -Wall destroys the "3 Release Policy"
- Possible type-checker regression in GHC 8.0
- Regression when deriving Generic1 on poly-kinded data family
- GHC panic when calling typeOf on a promoted data constructor
- panic! TEST=tc198: lookupVers2 GHC.Stack.Types CallStack
- Dysfunctional `__GLASGOW_HASKELL_TH` macro
- haddock and Cabal regression
- Type checker regression introduced by visible type-application
- Programs compiled with GHC master segfault when run with +RTS -h
- GHC HEAD uses up all memory while compiling `genprimcode`
- DfltProb1(optasm): panic CoreToStg.myCollectArgs
- T6031: *** Core Lint errors : in result of Common sub-expression ***
- Undeclared `CCS_MAIN` in unregisterised build
- GHC hangs/takes an exponential amount of time with simple program
- Extend ghc environment file features
- Regression typechecking type synonym which includes `Any`.
- GHCi on Windows segfaults
- Program doesn't preserve semantics after pattern synonym inlining.
- Please add initial platform support for sparc64
- GHCi doesn't qualify types anymore
- Turning on optimisations produces SEGFAULT or Impossible case alternative
