You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
This repository was archived by the owner on May 20, 2020. It is now read-only.
Now that WebAssembly/spec#1104 is merged and the new syntax is appearing in spec tests, I am hoping to add support for nan:canonical and nan:arithmetic to assert_return in wast. Here's what I think is necessary:
deprecate WastDirective::AssertReturn[.+] with #[deprecated(since = "5.1.0", note = "Please use AssertReturn instead")]; I think it makes sense to leave these around until all of the spec tests are converted over and users of wast update their spec tests
modify Lexer::number to lex nan:arithmetic to a new variant FloatVal::NanArithmetic and nan:canonical to a new variant FloatVal::NanCanonical (I was thinking of just lexing to FloatVal::Nan and setting val but I'm not sure this works for different sizes)
in float!, implement the variants FloatVal::NanArithmetic and FloatVal::NanCanonical; note that this means that Float32 and Float64 will be able to be parsed from nan:canonical and nan:arithmetic everywhere (not just in assert_return statements) which I don't think is technically correct but might actually be useful
bump to 5.1.0; I think this should be a minor version bump because no existing functionality is removed
The text was updated successfully, but these errors were encountered:
abrown
added a commit
to abrown/wat
that referenced
this issue
Jan 9, 2020
Now that WebAssembly/spec#1104 is merged and the new syntax is appearing in spec tests, I am hoping to add support for
nan:canonical
andnan:arithmetic
toassert_return
inwast
. Here's what I think is necessary:WastDirective::AssertReturn[.+]
with#[deprecated(since = "5.1.0", note = "Please use AssertReturn instead")]
; I think it makes sense to leave these around until all of the spec tests are converted over and users ofwast
update their spec testsLexer::number
to lexnan:arithmetic
to a new variantFloatVal::NanArithmetic
andnan:canonical
to a new variantFloatVal::NanCanonical
(I was thinking of just lexing toFloatVal::Nan
and settingval
but I'm not sure this works for different sizes)float!
, implement the variantsFloatVal::NanArithmetic
andFloatVal::NanCanonical
; note that this means thatFloat32
andFloat64
will be able to be parsed fromnan:canonical
andnan:arithmetic
everywhere (not just inassert_return
statements) which I don't think is technically correct but might actually be useful5.1.0
; I think this should be a minor version bump because no existing functionality is removedThe text was updated successfully, but these errors were encountered: