Commit message (Collapse) | Author | Age | Files | Lines | |
---|---|---|---|---|---|
* | fix typo | Carter Tazio Schonwald | 2020-01-05 | 1 | -1/+1 |
| | |||||
* | adding some flags and debug info, plus smarter mktrans/sym attempt | Carter Tazio Schonwald | 2020-01-05 | 5 | -16/+26 |
| | | | | need to add nthco at some point | ||||
* | improving debugging for now :) | Carter Tazio Schonwald | 2020-01-05 | 6 | -22/+21 |
| | |||||
* | easier to debug with stack traces not requires a 3 layer boot strap :) | Carter Tazio Schonwald | 2020-01-05 | 2 | -3/+19 |
| | |||||
* | fix imports | Carter Tazio Schonwald | 2020-01-04 | 2 | -0/+2 |
| | |||||
* | nuclear debugging | Carter Tazio Schonwald | 2020-01-04 | 12 | -34/+37 |
| | |||||
* | drop unused import | Carter Tazio Schonwald | 2020-01-04 | 1 | -1/+0 |
| | |||||
* | current state of play | Carter Tazio Schonwald | 2020-01-04 | 4 | -5/+10 |
| | |||||
* | push the code into the UnivCo axiom framework and drop all the "smart ↵ | Carter Tazio Schonwald | 2020-01-04 | 16 | -160/+57 |
| | | | | | | | compression" of the helper constructors from the coercion/type checking machinery, because my patch is focused on addressing issues in the simplifier rather than in the type checker | ||||
* | more stuff with the things | Carter Tazio Schonwald | 2020-01-04 | 7 | -11/+32 |
| | |||||
* | things seem to now all fit! (may need to add some extra wiring) | Carter Tazio Schonwald | 2020-01-04 | 21 | -85/+275 |
| | |||||
* | fix typo in smarter coercion erasure | Carter Tazio Schonwald | 2020-01-04 | 1 | -2/+2 |
| | |||||
* | if the coercion is a Refl or MRefl, lets just keep it. | Carter Tazio Schonwald | 2020-01-04 | 1 | -3/+9 |
| | |||||
* | halfway kicked along the coercion proofs, still probably need to handle the ↵ | Carter Tazio Schonwald | 2020-01-04 | 1 | -1/+30 |
| | | | | | | other cases of the Type syntax | ||||
* | step one of many for threading this stuff through the coercions, not ↵ | Carter Tazio Schonwald | 2020-01-04 | 2 | -7/+30 |
| | | | | finished by any means | ||||
* | i should have seen a coverage warning about this one, how'd i miss it? :) | Carter Tazio Schonwald | 2020-01-04 | 1 | -0/+1 |
| | |||||
* | more dumb typos, i really need to learn the ghci fast type checking dev work ↵ | Carter Tazio Schonwald | 2020-01-04 | 4 | -5/+12 |
| | | | | | | flow for ghc also added a note about an idea i'm probably gonna do tomorrow | ||||
* | more late to catching type errors and typos | Carter Tazio Schonwald | 2020-01-04 | 9 | -20/+26 |
| | |||||
* | fixing typo | Carter Tazio Schonwald | 2020-01-04 | 1 | -1/+1 |
| | |||||
* | setup the correct coercion erasure spot | Carter Tazio Schonwald | 2020-01-04 | 1 | -1/+2 |
| | |||||
* | fixed the approach to erasure proof to have the from / to targets, still ↵ | Carter Tazio Schonwald | 2020-01-04 | 15 | -29/+54 |
| | | | | need to update the erasure code in core | ||||
* | fixed some typoes and missing cases in backpack/interface file conections | Carter Tazio Schonwald | 2020-01-04 | 5 | -1/+5 |
| | |||||
* | fixed more type errors that were boring | Carter Tazio Schonwald | 2020-01-04 | 1 | -6/+6 |
| | |||||
* | massaging ErasedCoercion info through the compiler, | Carter Tazio Schonwald | 2020-01-04 | 8 | -8/+21 |
| | | | | | | | | | | | also getting things type checking the choices i've made atm in compiler/types/**.hs aren't ideal, but seem like an ok starting point. the implication of the design does mean core lint will only work when every dep has been built with core lint flags. In some respects, this is a similar stance to how profiling builds work | ||||
* | fixing gaps in supporting coercion erasure | Carter Tazio Schonwald | 2020-01-04 | 6 | -5/+16 |
| | |||||
* | experiment towards reducing normal simplifier blowups | Carter Tazio Schonwald | 2020-01-04 | 5 | -2/+63 |
| | | | | idea: erase are coercion terms unless O0 or core linting are enabled | ||||
* | Simplify mrStr | Gabor Greif | 2020-01-03 | 3 | -10/+3 |
| | |||||
* | Tweak Cmm dumps to avoid generating sections for empty groups | Ömer Sinan Ağacan | 2019-12-31 | 1 | -15/+20 |
| | | | | | | | | | | When dumping Cmm groups check if the group is empty, to avoid generating empty sections in dump files like ==================== Output Cmm ==================== [] Also fixes a few bad indentation in the code around changes. | ||||
* | Fix some sloppy indentation | Kevin Buhr | 2019-12-31 | 1 | -3/+3 |
| | |||||
* | Add additional Note explaining the -Iw flag | Kevin Buhr | 2019-12-31 | 1 | -2/+49 |
| | |||||
* | Add "-Iw" RTS flag for minimum wait between idle GCs (#11134) | Kevin Buhr | 2019-12-31 | 4 | -20/+63 |
| | |||||
* | Testsuite: update some Haddock tests | Vladislav Zavialov | 2019-12-31 | 12 | -8/+60 |
| | | | | | | | | | | | | Fixed tests: * haddockA039: added to all.T * haddockE004: replaced with T17561 (marked as expect_broken) New tests: * haddockA040: deriving clause for a data instance * haddockA041: haddock and CPP #include | ||||
* | Module hierarchy (#13009): Stg | Sylvain Henry | 2019-12-31 | 46 | -228/+226 |
| | |||||
* | testsuite: Enlarge acceptance window for T1969 | Ben Gamari | 2019-12-30 | 1 | -1/+1 |
| | | | | | | | | As noted in #17624, it's quite unstable, especially, for some reason, on i386 and armv7 (something about 32-bit platforms perhaps?). Metric Increase: T1969 | ||||
* | perf_notes: Add --zero-y argument | Ben Gamari | 2019-12-30 | 1 | -12/+25 |
| | | | | | This makes it easier to see the true magnitude of fluctuations. Also do some house-keeping in the argument parsing department. | ||||
* | TcIface: Fix inverted logic in typechecking of source ticks | Ben Gamari | 2019-12-30 | 1 | -1/+1 |
| | | | | | | | | | | | | | Previously we would throw away source ticks when the debug level was non-zero. This is precisely the opposite of what was intended. Fixes #17616. Metric Decrease: T13056 T9020 T9961 T12425 | ||||
* | hadrian: Track hash of Cabal Setup builder arguments | Ben Gamari | 2019-12-30 | 1 | -1/+6 |
| | | | | Lest we fail to rebuild when they change. Fixes #17611. | ||||
* | testsuite: Disable derefnull when built with LLVM | Ben Gamari | 2019-12-30 | 1 | -4/+4 |
| | | | | | | LLVM does not guarantee any particular semantics when dereferencing null pointers. Consequently, this test actually passes when built with the LLVM backend. | ||||
* | rts: Fix --debug-numa mode under Docker | Ben Gamari | 2019-12-30 | 2 | -0/+3 |
| | | | | | | | | | As noted in #17606, Docker disallows the get_mempolicy syscall by default. This caused numerous tests to fail under CI in the `debug_numa` way. Avoid this by disabling the NUMA probing logic when --debug-numa is in use, instead setting n_numa_nodes in RtsFlags.c. Fixes #17606. | ||||
* | rts: Error on invalid --numa flags | Ben Gamari | 2019-12-30 | 1 | -1/+6 |
| | | | | | Previously things like `+RTS --numa-debug` would enable NUMA support, despite being an invalid flag. | ||||
* | llvmGen: Drop old fix for #11649 | Ben Gamari | 2019-12-30 | 1 | -36/+1 |
| | | | | | This was a hack which is no longer necessary now since we introduce a dedicated entry block for each procedure. | ||||
* | llvmGen: Ensure that entry labels don't have predecessors | Ben Gamari | 2019-12-30 | 1 | -7/+14 |
| | | | | | | | | | | The LLVM IR forbids the entry label of a procedure from having any predecessors. In the case of a simple looping function the LLVM code generator broke this invariant, as noted in #17589. Fix this by moving the function prologue to its own basic block, as suggested by @kavon in #11649. Fixes #11649 and #17589. | ||||
* | rts: Ensure that nonmoving gc isn't used with profiling | Ben Gamari | 2019-12-30 | 2 | -2/+7 |
| | |||||
* | driver: Include debug level in the recompilation check hash | Ben Gamari | 2019-12-30 | 5 | -1/+19 |
| | | | | Fixes #17586. | ||||
* | use shell variable CcLlvmBackend for test | Gabor Greif | 2019-12-27 | 1 | -1/+1 |
| | | | Previously we used `AC_DEFINE`d variable `CC_LLVM_BACKEND` which has an empty shell expansion. | ||||
* | testsuite: Mark cgrun057 as fragile on ARM | Ben Gamari | 2019-12-27 | 1 | -0/+1 |
| | | | | | As reported in #17554. Only marking on ARM for now although there is evidence to suggest that the issue may occur on other platforms as well. | ||||
* | while at it rename XCode to the official Xcode | Gabor Greif | 2019-12-27 | 4 | -17/+17 |
| | |||||
* | suppress popup dialog about missing Xcode at configure | Gabor Greif | 2019-12-27 | 1 | -1/+1 |
| | | | | tested with `bash` and `zsh`. | ||||
* | testsuite: Skip T17499 when built against integer-simple | Ben Gamari | 2019-12-27 | 1 | -1/+3 |
| | | | | Since it routinely times out in CI. | ||||
* | Replace panic/notHandled with noExtCon in DsMeta | Ryan Scott | 2019-12-26 | 1 | -35/+42 |
| | | | | | | | There are many spots in `DsMeta` where `panic` or `notHandled` is used after pattern-matching on a TTG extension constructor. This is overkill, however, as using `noExtCon` would work just as well. This patch switches out these panics for `noExtCon`. |