ghc-8.10.4-bp153.1.35

- always build against llvm9, also on SLE15 (suse_version == 1500)
  otherwise the resulting build is not working (llvm is 7.x there) (bsc#1181571)

- update to 8.10.4
  * Fix a crash caused by inappropriate garbage of heap-allocated data
    reachable from foreign exports

- update to 8.10.3
  * More robust support for architectures with weak memory ordering
    guarantees (e.g. modern ARM hardware).
  * GHC can now split dynamic objects to accomodate macOS' RPATH size
    limitation when building large projects (#14444)
  * Several correctness bugs in the new low-latency garbage collector
  * Many, many other bug-fixes))

- disable %{ix86} build

- force multiversion for subpackages

- add --enable-unregistered to configure phase on unregistered
  architectures

- update to 8.10.2
  * Fixes a code generation bug resulting in incorrect code generation
    for foreign calls with complex arguments
  * Fixes a bug causing object merging to fail when the lld linker is in use
  * Introduces a workaround for a Linux kernel bug in the implementation
    of the timerfd mechanism
  * Fixes a few specialiser regressions as well introduces a variety
    of miscellaneous specialiser improvements
  * Fixes a potential loss of sharing due to left operator sections
  * Fix bootstrapping of GHC with the LLVM backend on x86-64
  * few important correctness fixes for the low-latency garbage collector.

- cleanup specfile from Leap 42 and older remrants
- drop fix-unregisterised-v8.4-8.6.patch
- drop fix-build-using-unregisterized-v8.4.patch

- Update ghc to version 8.10.1. Important changes are:
  * The UnliftedNewtypes extension, allowing newtypes to be wrap
    types of kind other than Type.
  * The StandaloneKindSignatures extension, allowing explicit
    signatures on type constructors.
  * A new, low-latency garbage collector for the oldest generation.
  Please refer to the compiler documentation for all details:
  https://downloads.haskell.org/~ghc/8.10.1/docs/html/users_guide/8.10.1-notes.html

- update to 8.8.3
  * a compiler panic due to over-zealous eta reduction
  * the ability to bootstrap with earlier 8.8 releases has been restored
  * the directory submodule has been updated
  * the process submodule has been updated

- Increase memory limit for qemu user-space builds

- Fix typo in last change

- use ghc-bootstrap-helpers on all archs
- lower build env requirements on riscv64
  * make sure we can build on smaller systems
  * adjust parallelism so we don't hog big workers for almost a week
  * many thanks to Stefan Bruens for pointing out the issue and
    helping with the solution

- Update _constraints for aarch64

- drop 0001-Fix-check-whether-GCC-supports-__atomic_-builtins.patch
  * fixed upstream
- drop D5212.patch
  * fixed upstream
- drop fix-build-using-unregisterized-v8.4.patch
  * bootstrap compiler is 8.6.5 now
- refresh fix-unregisterised-v8.4-8.6.patch
- clean up spec file

- update to 8.8.2
  * A bug (Issue #17088) in the compacting garbage collector resulting in segmentation
  faults under specific circumstances. Note that this may affect user programs even
  if they did not explicitly request the compacting GC (using the -c RTS flag)
  since GHC may fallback to compacting collection during times of high memory pressure.
  * A code generator bug (Issue #17334) resulting in GHC panics has been fixed.
  * A typechecker bug (Issue #17112) resulting in GHC panicking on some programs
  has been fixed.
  * A bug (Issue #16707) in the implementation of software transactional memory
  potentially resulting in hanging transactions has been fixed.
  * A bug in the implementation of compact normal forms resulting in segmentation
  faults in some uses (Issue #17044) has been fixed.

- add fix-unregisterised-v8.4-8.6.patch
  * fix build on s390x and riscv64

- refresh buildpath-abi-stability.patch
- refresh ghc-8.0.2-Cabal-dynlibdir.patch
- refresh ghc-pie.patch
- enable riscv64

- use prebuilt alex and happy on Tumbleweed to avoid buildcycle
- Don't require llvm in ghc-compiler for other architectures than arm

- update to 8.8.1 -new stable release
- refresh Disable-unboxed-arrays.patch
- drop allow-riscv-and-riscv64-CPU.patch - merged in upstream

- add fix-build-using-unregisterized-v8.4.patch
  * fixes unregisterised builds when bootstrap compiler is GHC 8.4
  * fixes build on s390x

- Enable build on riscv64
- Add memory constraints for riscv64, restrict to 2 jobs
- allow-riscv-and-riscv64-CPU.patch: recognize riscv64 as architecture
- Apply Disable-unboxed-arrays.patch also on s390x

- update to 8.6.5
  * A build system issue where use of GCC with -flto broke configure was fixed

- add D4089.patch
  * fix GHCi issue on ppc64
  * see Haskell Trac #11262

- add 0001-Use-__atomic-intrinsics-for-atomicread-write.patch
  * add missing memory barriers on atomic read/write ops
  * fixes upstream ticket #14244
- adjust comment for patch 32
  * the situation with upstream #12537 is improved but not
    solved. Patch 32, however, is correct. There are even more
    memory barriers missing elsewhere in the compiler.

- add 0001-PPC-Implement-Atomic-operations.patch
  * fixes parallel programs on all PowerPC platforms
  * uptream ticket #12537
- refresh buildpath-abi-stability.patch

- make compactible with new upstream macros
- added patches for more stable abi:
  * buildpath-abi-stability.patch
  * reproducible-tmp-names.patch

- enable for global PIE support:
  - the compiler is built without PIE support due to the bootstrap
    libraries being non-PIE using BuildIgnore: gcc-PIE
  - the compiler itself builds -fPIC objects all the time,
    which should make all ghc compiled binaries be PIE.
- ghc-pie.patch: enable default -fPIC on Linux/x86_64

- comment for patch 100
- refresh ghc-no-madv-free.patch with upstream patch
  * upstream fix for Haskell Trac #12865
  * Haskell Trac #12495 is a duplicate

- update to 8.0.2
- drop patches merged by upstream:
  * D2495.patch
- refreshed patches:
  * 0001-PPC-CodeGen-fix-lwa-instruction-generation.patch
  * 0001-StgCmmPrim-Add-missing-write-barrier.patch
  * ghc-no-madv-free.patch
- added ghc-8.0.2-Cabal-dynlibdir.patch by Jens Petersen simplifing
    haskell packaging
  + Interface files produced by GHC should now be deterministic ( boo#998972 )
  + A bug has been fixed that caused standalone derived Ix instances
    to fail for GADTs with exactly one constructor
  + Compatibility fixes recent Linux distributions.

- drop D2844.patch
  * the patch causes more issues with parallel builds
- fixes boo#1020909

- add D2844.patch
  * fix issues with parallel builds

- make compatible with new ghc-rpm-macros

- update to 8.0.1
- drop patches fixed upstream:
  * drop atomic-cast.patch
  * drop cabal-show-detail-direct.patch
  * drop 0001-link-command-line-libs-to-temp-so.patch
  * drop 0001-implement-native-code-generator-for-ppc64.patch
  * drop ghc.git-b29f20.patch
  * drop u_terminfo_0402.patch
  * drop u_Cabal_update.patch
  * drop u_haskeline_update.patch
  * drop 0001-Fix-misspelled-WORDS_BIGENDIAN-macro.patch
  * drop D2214.patch
  * drop D2225.patch
- refresh 0001-StgCmmPrim-Add-missing-write-barrier.patch
- refresh 0001-PPC-CodeGen-fix-lwa-instruction-generation.patch
- add ghc-no-madv-free.patch to fix "unable to decommit memory:
  Invalid argument" errors GHC reports due to a misdetected
  MADV_FREE capability. The configure script sees the symbol define
  by glibc, but unfortunately this doesn't mean that the running
  kernel actually supports the feature, too. The upstream ticket
  https://ghc.haskell.org/trac/ghc/ticket/12495 has more details.
  Thanks to @psimons
- GHC produces debug information on x86 and x86_64
- aarch64 has LLVM backend (requires llvm-3.7)
- native code generator for powerpc64[le] (openSUSE contribution!)
- disable html doc on SLE, we don't have python-sphinx
  * Thanks @mimi_vx

- add D2495.patch
  * missing memory barrier on PowerPC, ARM
- add 0001-StgCmmPrim-Add-missing-write-barrier.patch
  * another issing memory barrier on PowerPC, ARM
- add 0001-PPC-CodeGen-fix-lwa-instruction-generation.patch
  * fix bug in code generator
  * fixes build of ghc-zeromq4-haskell on powerpc64[le]

- fix boo#994268

- add D2225.patch
  * backport of upstream patch accepted for ghc 8.0.1
  * fix SMP primitives on all powerpc archs

- add D2214.patch
  * backport of upstream patch slated for ghc 8.0.2
  * fix issue in code PowerPC code generator (32 bit and 64 bit)
- add 0001-Fix-misspelled-WORDS_BIGENDIAN-macro.patch
  * backport of upstream fix slated for ghc 8.0.2
  * fix bytestring library on big-endian archs
  * fix wrong results in ghc-pureMD5 and others on big-endian

- update haskeline to 0.7.2.3
- refresh u_haskeline_update.patch
  * Fix hsc2hs-related warning on ghc-8
  * Fix the behavior of ctrl-W in the emacs bindings
  * Point to github instead of trac

- update Cabal to 1.22.8.0
- refresh u_Cabal_update.patch
  * Distribution.Simple.Setup: remove job cap
  * Check all object file suffixes for recompilation
  * Move source files under 'src/'.

- limit parallel jobs to two on aarch64
  * aarch64 runs out of memory with too many parallel compiles

- add missing C library dependency to terminfo devel package

- update haskeline to 0.7.2.2
- add u_haskeline_update.patch
  * canonicalize AMP instances to make the code more future proof
  * Generalize constraints for InputT instances
  * Bump upper bounds on base and transformers
  * Make Haskeline `-Wtabs` clean

- update Cabal to 1.22.7.0
- refresh u_Cabal_update.patch
  * Backport #3012 to the 1.22 branch
  * Add foldl' import
  * The Cabal part for fully gcc-like response files

- update tarball to 7.10.3b
  * some documentation files were missing in the original
  7.10.3 tarball
- add 0001-link-command-line-libs-to-temp-so.patch
  * fix panic in GHCi when linking against C library specified
  as command line argument (-l<lib>)
  * currently under upstream review
  * fix Haskell trac #10458

- update Cabal to 1.22.6.0
- add u_Cabal_update.patch --  update of Cabal
  * Relax upper bound to allow upcoming binary-0.8

- update terminfo to 0.4.0.2
- drop remove-cur_term-usage.patch - merged by upstream
- add u_terminfo_0402.patch -- update of terminfo

- update Cabal to 1.22.5.0
  * Don't recompile C sources unless needed
  * Support Haddock response files.
  * Add frameworks when linking a dynamic library.

- use upstream variant of remove-cur_term-usage.patch

- update to 7.10.3
- add cabal-show-detail-direct.patch -- Fix testsuites with large output.
    Debian patch. Thanks Joachim for suggesting it in Haskell #10870!
- add ghc.git-b29f20.patch
  * fixes upstream #7830
  * fix ghc 7.10.2 failing to bootstrap ghc head on ppc
  and perhaps ppc64 with some optimizations enabled
  * backport of my patch for upcoming ghc 8.0.1

- atomic-cast.patch: fix signature of atomic bultins
- ghc-7.8-arm-use-ld-gold.patch: remove obsolete patch

- Remove ncurses6.patch - it's not so easy ;)
- Add remove-cur_term-usage.patch to remove the need to access the
  cur_term global which isn't available in the openSUSE ncurses6
  build.

- Add choice for libffi in openSUSE:Leap:42.1

- remove llvm from requires
- add ncurses6.patch to fix build with ncurses6, all thanks goes to
  Tomas Cech sleep_walker@opensuse.org

- update to 7.10.2
  * type checker fixes
  * fixes for Aarch64 support
  * fix deadlock in runtime system when scheduling garbage collection
- 7.10.1 highlights
  * implement Applicative Monad Proposal
  * implement Burning Bridges Proposal
  * support for partial type signatures
  * reimplement integer-gmp
  * support plugins in type checker (experimental!)
- drop llvm-powerpc64-datalayout.patch
  * this patch was incomplete all along and now we have our native
  code generator
- drop ghc-cabal-unversion-docdir.patch
  * ghc-rpm-macros is following ghc's doc layout so no need to patch
- drop D349.patch
  * fixed upstream
- drop integer-gmp.patch
  * we do not support SLE11 anymore
- drop ghc-7.8.2-cgen-constify.patch
  * fixed upstream
- drop D560.patch
  * fixed upstream
- drop ghc-glibc-2.20_BSD_SOURCE.patch
  * fixed upstream
- drop ghc-arm64.patch
  * fixed upstream
- drop ghc-config.mk.in-Enable-SMP-and-GHCi-support-for-Aarch64.patch
  * fixed upstream
- refresh 0001-implement-native-code-generator-for-ppc64.patch
- disable ghc-7.8-arm-use-ld-gold.patch
  * not sure we need this, must be rebased if we do

- add ghc-config.mk.in-Enable-SMP-and-GHCi-support-for-Aarch64.patch
  * enable SMP and GHCi on Aarch64
- add ghc-7.8-arm-use-ld-gold.patch
  * Aarch64 needs gold linker

- refresh ghc-arm64.patch
- unconditionally apply ghc-arm64.patch

- add backport of powerpc64le native code generator to
  0001-implement-native-code-generator-for-ppc64.patch
- refresh D349.patch

- on s390/s390x use "make -j2" to stabilize abi hashes
  (see also redhat bug 1212374)

- disable llvm on arm64

- re-diffed ghc-arm64.patch

- add ghc-arm64.patch for aarch64
- add aarch64, s390, s390x to ExclusiveArch

- remove unneeded conditions

- chnage constraints to 12G for disk and 4G for ram

- add ghc-glibc-2.20_BSD_SOURCE.patch
  * fixes build on unregisterised architectures (openSUSE: ppc64le)

- drop ghc-cabal-unversion-docdir.patch
  * ghc-rpm-macros follows upstream ghc again

+ add ghc-glibc-2.20_BSD_SOURCE.patch from fedora
- updated to 7.8.4
  + removed D177.patch (in upstream)
  + removed D173.patch (in upstream)
  + removed ghc.git-e18525f.patch (in upstream)
  + refresh 0001-implement-native-code-generator-for-ppc64.patch
  * A critical bug in the LLVM backend which would cause the compiler to generate incorrect code has been fixed (issue #9439).
  * Several bugs in the code generator have been fixed for unregisterised platforms, such as 64bit PowerPC (issue #8819 and #8849).
  * A bug that could cause GHC's constructor specialization pass (enabled by default at -O2, or via -fspec-constr) to loop forever and consume large amounts of memory has been fixed (issue #8960).
  * A bug that would cause GHC to fail when attempting to determine GCC's version information in non-english locales has been fixed (issue #8825).
  * A minor bug that allowed GHC to seemingly import (but not use) private data constructors has been fixed (issue #9006).
  * A bug in the register allocator which would cause GHC to crash during compilation has been fixed (issue #9303).
  * A bug that caused the compiler to panic on some input C-- code has been fixed (issue #9329).
  * A few various minor deadlocks in the runtime system when using forkProcess have been fixed.
  * A bug which made blocked STM transactions non-interruptible has been fixed (issue #9379).
  * A bug in the compiler which broke pattern synonym imports across modules in Haddock has been fixed (issue #9417).
  * A minor bug in the code generator in which the popCnt16# did not zero-extend its result has been fixed (issue #9435).
  * A bug which caused the compiler to panic on pattern synonyms inside a class declaration has been fixed (issue #9705).
  * A bug in the typechecker revolving around un-saturated type family applications has been fixed (issue #9433).
  * Several bugs have been fixed causing problems with building GHC on ARM (issues #8951, #9620, #9336, and #9552).
  * A bug in the typechecker that could cause an infinite loop when using superclasses in a cycle has been fixed (issue #9415).
  * A bug causing corruption in signal handling with the single-threaded runtime system has been fixed (issue #9817).
  * A bug that could cause compiled programs to crash due to use of overlapping type families has been fixed (issue #9371).
  * A bug in the inliner that caused certain expressions within unboxed tuples to not be properly evaluated has been fixed (issue #9390).
  * A bug that caused the compiler to not always properly detect LLVM tools (particularly on Windows) has been fixed (issue #7143).
  * A bug that prevented GHC from deriving Generic1 instances for data families has been fixed (#9563).
  * A bug that caused type inference to infer the incorrect type in the presence of certain type families and constraints has been fixed (issue #9316).

- add constraints file

- remove dblatex dependency