| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
* Rename GHC.Core.Opt.WorkWrap.Utils.wantToUnboxArg to canUnboxArg
and similarly wantToUnboxResult to canUnboxResult.
* Add GHC.Core.Opt.DmdAnal.wantToUnboxArg as a wrapper for
the (new) GHC.Core.Opt.WorkWrap.Utils.canUnboxArg,
avoiding some yukky duplication.
I decided it was clearer to give it a new data type for its
return type, because I nedeed the FD_RecBox case which was not
otherwise readiliy expressible.
* Add dcpc_args to WorkWrap.Utils.DataConPatContext for the payload
* Get rid of the Unlift constructor of UnboxingDecision, eliminate
two panics, and two arguments to canUnboxArg (new name). Much
nicer now.
|
|
|
|
|
|
|
|
| |
This fix, in the definition of profitableFloat,
is just for consistency. `floatConsts` should
do what it says!
I don't think it'll affect anything much, though.
|
|
|
|
|
|
|
|
|
| |
This function is called in inner loops in the compiler, and it's
overloaded and higher order. Best just to inline it.
This popped up when I was looking at something else. I think
perhaps GHC is delicately balanced on the cusp of inlining this
automatically.
|
|
|
|
| |
I don't know why this hasn't bitten us before, but it was plain wrong.
|
| |
|
| |
|
|
|
|
|
|
|
| |
* Rename `docs` to `doc`
* Place pdf documentation in `doc/` instead of `doc/pdfs/`
Fixes #21164.
|
|
|
|
|
|
| |
It appears that Centos 7 has a more strict C++ compiler than most
distributions since std::runtime_error is defined in <stdexcept> rather
than <exception>. In T11829 we mistakenly imported the latter.
|
| |
|
| |
|
|
|
|
| |
Darwin expects a leading underscore.
|
| |
|
|
|
|
|
|
|
| |
It appears that GCC 6t (at least on i386) fails to give
init_array/fini_array sections the correct SHT_INIT_ARRAY/SHT_FINI_ARRAY
section types, instead marking them as SHT_PROGBITS. This caused T20494
to fail on Debian.
|
| |
|
| |
|
| |
|
| |
|
|
|
|
|
|
|
| |
RtlDeleteFunctionTable expects a pointer to the .pdata section
yet we passed it the .xdata section.
Happily, this fixes #21354.
|
| |
|
|
|
|
|
|
|
| |
Previously we would run constructors and destructors in arbitrary order
despite explicit priorities.
Fixes #21847.
|
|
|
|
|
|
|
|
| |
It turns out that PE objects may have multiple `.ctors`/`.dtors`
sections but the RTS linker had assumed that there was only one. Fix
this.
Fixes #21618.
|
|
|
|
|
|
|
|
|
|
|
| |
Unfortunately on Windows we have no RPATH-like facility, making dynamic
linking extremely fragile. Since we cannot assume that the user will
add their GHC installation to `$PATH` (and therefore their DLL
search path) we cannot assume that the loader will be able to locate our
`libc++.dll`. To avoid this, we instead statically link against `libc++.a` on
Windows.
Fixes #21435.
|
|
|
|
|
|
| |
Gnu ld allows `-l` to be passed an absolute file path,
signalled by a `:` prefix. Implement this in the GHC's
loader search logic.
|
|
|
|
|
|
| |
This implements #20494 for the PEi386 linker.
Happily, this also appears to fix `T9405`, resolving #21361.
|
|
|
|
|
|
|
|
|
|
|
| |
Previously we would free oc->info after running initializers. However,
we can't do this is we want to also run finalizers.
Moreover, freeing oc->info so early was wrong for another reason:
we will need it in order to unregister the exception tables (see the
call to `RtlDeleteFunctionTable`).
In service of #20494.
|
|
|
|
| |
fini is short for "finalizer", which does not contain a "t".
|
| |
|
|
|
|
| |
Addresses #20494.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
At some point we used the clang distribution from msys2's `MINGW64`
environment for our Windows toolchain. This defaulted to using libgcc
and libstdc++ for its runtime library. However, we found for a variety
of reasons that compiler-rt, libunwind, and libc++ were more reliable,
consequently we explicitly overrode the CXXFLAGS to use these.
However, since then we have switched to use the `CLANG64` packaging,
which default to these already. Consequently we can drop these
arguments, silencing some redundant argument warnings from clang.
Fixes #21669.
|
| |
|
| |
|
|
|
|
|
|
|
| |
Here we reorganize `GHC.Cmm` to eliminate the orphan `Outputable` and
`OutputableP` instances for the Cmm AST. This makes it significantly
easier to use the Cmm pretty-printers in tracing output without
incurring module import cycles.
|
|
|
|
| |
Fixes #21824
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Before the behaviour of `dopt` and `logHasDumpFlag` (and the underlying
function `log_dopt`) were different as the latter did not take the
verbosity level into account. This led to problems during the
refactoring as we cannot simply replace calls to `dopt` with calls to
`logHasDumpFlag`.
In addition to that a subtle bug in the GHC module was fixed:
`setSessionDynFlags` did not update the logger and as a consequence the
verbosity value of the logger was not set appropriately.
Fixes #21861
|
|
|
|
| |
Since we have backported the fix to 9.4.1.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This is a naive approach to fixing the unsoundness noticed in #21708.
Specifically, we remove the lowering of `keepAlive#` via CorePrep and
instead turn it into an out-of-line primop.
This is simple, inefficient (since the continuation must now be heap
allocated), but good enough for 9.4.1. We will revisit this
(particiularly via #16098) in a future release.
Metric Increase:
T4978
T7257
T9203
|
|
|
|
|
|
|
| |
As noted in #21868, several primops accepting continuations producing
RuntimeRep-polymorphic results aren't nearly as polymorphic as their
types suggest. Document this limitation and adapt the `UnliftedWeakPtr`
test to avoid breaking this limitation in `keepAlive#`.
|
| |
|
|
|
|
|
| |
Ensures type information for record-dot-syntax
is included in HieASTs. See #21797
|
|
|
|
|
|
|
|
|
| |
This changes the bytecode return convention for unlifted
algebraic datatypes to be the same as for lifted
types, i.e. ENTER/PUSH_ALTS instead of
RETURN_UNLIFTED/PUSH_ALTS_UNLIFTED
Fixes #20849
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Previously the implementation of bitmap_first_unset assumed that
`__builtin_clz` would accept `uint8_t` however it apparently rather
extends its argument to `unsigned int`.
To fix this we simply revert to a naive implementation since handling
the various corner cases with `clz` is quite tricky. This should be
fine given that AdjustorPool isn't particularly hot. Ideally we would
have a single, optimised bitmap implementation in the RTS but I'll leave
this for future work.
Fixes #21838.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
to previous ModuleGraphs, in particular the lazy `mg_non_boot` field.
This manifests in `extendMG`.
Solution: Delete `mg_non_boot` as it is only used for `mgLookupModule`, which
is only called in two places in the compiler, and should only be called at most
once for every home unit:
GHC.Driver.Make:
mainModuleSrcPath :: Maybe String
mainModuleSrcPath = do
ms <- mgLookupModule mod_graph (mainModIs hue)
ml_hs_file (ms_location ms)
GHCI.UI:
listModuleLine modl line = do
graph <- GHC.getModuleGraph
let this = GHC.mgLookupModule graph modl
Instead `mgLookupModule` can be a linear function that looks through the entire
list of `ModuleGraphNodes`
Fixes #21816
|
|
|
|
|
|
|
|
|
|
| |
The documentation for key-value settings was a bit out of date.
This patch updates it to account for `cabal.configure.opts` and
`hsc2hs.run.opts`.
The user-settings document was also re-arranged, to make the key-value
settings more prominent (as it doesn't involve changing the Hadrian
source code, and thus doesn't require any recompilation of Hadrian).
|
| |
|
|
|
|
|
| |
This is useful for hie files to reconstruct the evidence that default methods
depend on.
|
| |
|
| |
|
| |
|