1548908 - Clean up NaN handling in Value.h

Status: RESOLVED FIXED

(Core :: JavaScript Engine, task, P2)

Description

[Iain Ireland [:iain]]

While reviewing Bug 1401624, it became clear that our handling of NaNs could use some tidying. Putting these patches in a separate bug because they are only tangentially related to object-biased NaN-boxing.

Comment 1

[Iain Ireland [:iain]]

Attachment: Bug 1548908: Add InfinityBits to FloatingPoint.h r=jwalden

We can't generate a constexpr uint64_t containing the bits for positive/negative infinity, because of a (very sensible) static_assert in SpecificNaNBits. This patch adds support to FloatingPoint.h for infinity, and uses it to make JS::InfinityValue constexpr (to match JS::NaNValue).

Comment 2

[Iain Ireland [:iain]]

Attachment: Bug 1548908: Remove IsCanonical and simplify StructuredCloneReader r=tcampbell

IsCanonical was only used in StructuredCloneReader::CheckDouble. Two of CheckDouble's three uses already canonicalized the double before checking it, and the third should have already done so.

Comment 3

[Iain Ireland [:iain]]

Attachment: Bug 1548908: Remove "well-known numbers" from JSRuntime r=tcampbell

Back when we first freed the lizard in 1998, it made sense to have copies of NaN/+Inf/-Inf living on the runtime, because Values didn't store doubles inline. That hasn't been true for a long time. This patch gets rid of those.

Depends on D29867

Comment 4

[Iain Ireland [:iain]]

Attachment: Bug 1548908: Clean up JS::CanonicalizedDoubleValue r=tcampbell

1. CanonicalizedDoubleValue should reuse the logic in CanonicalizeNaN.
2. Places that call DoubleValue(CanonicalizeNaN(d)) should just use CanonicalizedDoubleValue(d).

Comment 5

[BugBot [:suhaib / :marco/ :calixte]]

There are some r+ patches which didn't land and no activity in this bug for 2 weeks.
:iain, could you have a look please?
For more information, please visit auto_nag documentation.

Comment 6

[Iain Ireland [:iain]]

These patches (and the related patches in Bug 1401624) were waiting until 68.

Comment 7

[Iain Ireland [:iain]]

Attachment: Bug 1548908: Verify during JS_Init that the hardware NaN is sane

Depends on D33503

Comment 8

[Pulsebot]

Pushed by [e-mail]:
https://hg.mozilla.org/integration/autoland/rev/cfde75f084e0
Clean up JS::CanonicalizedDoubleValue r=tcampbell
https://hg.mozilla.org/integration/autoland/rev/f42ce0774d43
Add InfinityBits to FloatingPoint.h r=jwalden
https://hg.mozilla.org/integration/autoland/rev/2953a85f6b5f
Remove "well-known numbers" from JSRuntime r=tcampbell
https://hg.mozilla.org/integration/autoland/rev/411b72386829
Remove IsCanonical and simplify StructuredCloneReader r=tcampbell
https://hg.mozilla.org/integration/autoland/rev/b7340dfe4325
Verify during JS_Init that the hardware NaN is sane r=jwalden

Comment 9

[Andrei Ciure[:aciure]]

https://hg.mozilla.org/mozilla-central/rev/cfde75f084e0
https://hg.mozilla.org/mozilla-central/rev/f42ce0774d43
https://hg.mozilla.org/mozilla-central/rev/2953a85f6b5f
https://hg.mozilla.org/mozilla-central/rev/411b72386829
https://hg.mozilla.org/mozilla-central/rev/b7340dfe4325