aboutsummaryrefslogtreecommitdiff
path: root/src/parseTools.js
AgeCommit message (Collapse)Author
2014-05-06emit >>0 in HEAP8, for consistency with HEAP16|32Alon Zakai
2014-04-22ignore readnone in paramsAlon Zakai
2014-02-25Introduce the asmjs-unknown-emscripten target triple.Dan Gohman
2014-02-11escape curly braces in regexps for more portability (unescaped curly braces ↵Sami Vaarala
are technically a SyntaxError in Ecmascript E5/E5.1)
2014-01-28rely on js opts to do safe heap for asm, to avoid duplicating logic for ↵Alon Zakai
proper asm ffi coercions
2014-01-18handle #if !X in preprocessorAlon Zakai
2014-01-17handle unsigned reads in safe heapAlon Zakai
2014-01-17remove current memory initialization chunking via .concat(..)Alon Zakai
2014-01-16enable a form of safe heap in asm, using js optimizer pass to ensure full ↵Alon Zakai
coverage and support for fastcomp
2014-01-15fix Math_floor coercion in unrecommended codegen modes (no roundings ↵Alon Zakai
correction or imprecise i64)
2014-01-13Handle negative zero values correctly.Bruce Mitchener
This handles the situation when they're part of a global variable whether it just be a float/double or embedded within a struct. Fixes #1898.
2014-01-07Merge branch 'incoming' into llvm-3.41.8.6Alon Zakai
Conflicts: tests/test_benchmark.py tools/shared.py
2014-01-07properly overflow gep arguments; fixes #1975Alon Zakai
2014-01-06preprocess recursively into #include's; fixes #1984Alon Zakai
2014-01-06ignore new readonly attribute in llvm 3.4Alon Zakai
2014-01-06ignore new readnone attribute in llvm 3.4Alon Zakai
2014-01-02support #include x.js in our preprocessed js sourcesAlon Zakai
2013-12-03fix memset return value in makeSetValuesAlon Zakai
2013-11-26getting some initial set of x86 intrinsics mapped to JS SIMD ops; update ↵Heidi Pan
existing SIMD code to use updated JS SIMD API (SIMD.type.op instead of SIMD.op & SIMD.opu32; int32x4 instead of uint32x4); bug fix: right type of withX() for int32x4 initialization; bug fix: splat(0) instead of zero() for int32x4 initialization
2013-11-26parse hex floats in vector literalsAlon Zakai
2013-11-14ensure a >>> coercion on unsigned comparisons, even if the input is an i29 ↵Alon Zakai
which would be coerced using an &; fixes asm1.test_sqlite on llvm 3.3
2013-11-14handle odd types like i24 in makeSignOp, fixes asm1.test_fuzz on llvm 3.3Alon Zakai
2013-11-11Merge branch 'incoming' into f32Alon Zakai
2013-11-11coerce to double before converting a float to an i64Alon Zakai
2013-11-09support inline asm comments; fixes #1766Alon Zakai
2013-11-09fix ensureDot logicAlon Zakai
2013-11-08emit proper large float constants when js opts are to be runAlon Zakai
2013-11-08Merge branch 'incoming' into f32Alon Zakai
Conflicts: src/parseTools.js src/preamble.js
2013-11-04add missing parens for select mathopAlon Zakai
2013-11-01do not emit unnecessary coercions when calling a math builtin with a floatAlon Zakai
2013-11-01Merge branch 'incoming' into f32Alon Zakai
Conflicts: src/parseTools.js
2013-11-01add vector support for and, or and xorAlon Zakai
2013-10-30handle vector zeroinitializersAlon Zakai
2013-10-30remove finalizeParam and properly use finalizeLLVMParameter everywhereAlon Zakai
2013-10-30use proper base type when finalizing vectorsAlon Zakai
2013-10-30fix negative zero handlingAlon Zakai
2013-10-30simplify parseParamTokens now that it uses parseLLVMSegmentAlon Zakai
2013-10-30make parseParamTokens use parseLLVMSegmentAlon Zakai
2013-10-30handle more return types in getReturnTypeAlon Zakai
2013-10-30optimize isIllegalTypeAlon Zakai
2013-10-29add test for precise float32 support, allow 3 modes of precise float32 ↵Alon Zakai
support, and rename option to PRECISE_F32 to be consistent with other precision options
2013-10-29fix uitofp of i64 to float32Alon Zakai
2013-10-29ensure a dot in large float32 constantsAlon Zakai
2013-10-29clean up simd code, fix uint32x4 heap offsets, and add testingAlon Zakai
2013-10-29handle vector types in calcAllocatedSizeAlon Zakai
2013-10-29use asm ffi coercing for floats in load abort callsAlon Zakai
2013-10-25send and receive only doubles through ffis, not floatsAlon Zakai
2013-10-25fix splitI64 use of float instead of doubleAlon Zakai
2013-10-25do not mangle hex float constants in function callsAlon Zakai
2013-10-25simplify auto-return adding code, and make it generic so it can handle ↵Alon Zakai
floats as well