From 042ac8d14d3c80e51fd032e706256da726adf9d2 Mon Sep 17 00:00:00 2001 From: Dan Gohman Date: Thu, 25 Jun 2015 12:10:23 -0700 Subject: Correct behavior of int to float conversions. IEEE 754 specifies that these operations round rather than overflow. --- AstSemantics.md | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/AstSemantics.md b/AstSemantics.md index f3e8155..a4c7a7d 100644 --- a/AstSemantics.md +++ b/AstSemantics.md @@ -458,8 +458,8 @@ the opcode, and the operand. Reinterpretations always succeed. -Conversions from integer to floating point always succeed, though they may -overflow to infinity or negative infinity as specified by IEEE-754. +Conversions from integer to floating point always succeed, and use +round-to-nearest ties-to-even rounding. Truncation from floating point to integer where IEEE-754 would specify an invalid operation exception (e.g. when the floating point value is NaN or -- cgit v1.2.3