[wasm] eqz, copysign, trunc, and nearest #797
Conversation
|
Hi @commonlisp, I'm your friendly neighborhood Microsoft Pull Request Bot (You can call me MSBOT). Thanks for your contribution! The agreement was validated by Microsoft and real humans are currently evaluating your PR. TTYL, MSBOT; |
| @@ -743,6 +742,11 @@ MACRO_BACKEND_ONLY( SlotArrayCheck, Empty, OpCanCSE) | |||
| MACRO_BACKEND_ONLY( FrameDisplayCheck, Empty, OpCanCSE) | |||
| MACRO_EXTEND( BeginBodyScope, Empty, OpSideEffect) | |||
|
|
|||
| MACRO_BACKEND_ONLY(Copysign_A, Empty, OpInlinableBuiltIn | OpTempNumberSources | OpCanCSE | OpProducesNumber) | |||
| MACRO_BACKEND_ONLY(Trunc_A, Empty, OpInlinableBuiltIn | OpTempNumberSources | OpCanCSE | OpProducesNumber) | |||
| MACRO_BACKEND_ONLY(Nearest_A, Empty, OpInlinableBuiltIn | OpTempNumberSources | OpCanCSE | OpProducesNumber) | |||
There was a problem hiding this comment.
I don't believe we need the OpInlinableBuiltIn flag
commonlisp
force-pushed
the
wasm_operators
branch
4 times, most recently
from
c16577e
to
d6efe54
Compare
- Implemented interpreter support for eqz, copysign, trunc, and nearest - Implemented codegen for eqz and copysign - Updated SExprParser generated files - Added tests and trunc and nearest fp ops - F64 constant space and fix - Moved ops from AsmjsMath to WasmMath - Corrected compile-flags in tests and propagating opcodes through - Punting on NaN sign canonicalization - Adding dblextended prefix since overrunning extended prefix
commonlisp
force-pushed
the
wasm_operators
branch
from
d6efe54
to
6c72715
Compare
|
@dotnet-bot please test this |
|
Hi @commonlisp, I'm your friendly neighborhood Microsoft Pull Request Bot (You can call me MSBOT). Thanks for your contribution! The agreement was validated by Microsoft and real humans are currently evaluating your PR. TTYL, MSBOT; |
| @@ -27,6 +27,7 @@ | |||
| $(IntDir); | |||
| %(AdditionalIncludeDirectories); | |||
| </AdditionalIncludeDirectories> | |||
| <PreprocessToFile Condition="'$(Configuration)|$(Platform)'=='Debug|Win32'">false</PreprocessToFile> | |||
There was a problem hiding this comment.
Yeah, good catch. This was left over when I was hunting down extended prefix exhaustion.
|
lgtm |
Merge pull request chakra-core#797 from commonlisp:wasm_operators This PR covers interpreter & codegen for eqz and copysign and interpreter support for trunc and nearest with codegen forthcoming.
This PR covers interpreter & codegen for eqz and copysign and interpreter support for trunc and nearest with codegen forthcoming.