float: preserve NaN sign when propagating NaNs

[HuzaifaAbdulRehman]

Closes #1188.

Soft-float addition classified NaNs using sign-stripped representations and then reused those representations to construct the result. This discarded the sign bit when propagating an input NaN.

Use the original operand representation when setting the quiet bit, preserving the input sign and payload while quieting signaling NaNs. This also matches the corresponding compiler-rt behavior.

The regression tests check signaling and quiet NaNs in both operand positions for f16, f32, f64, and f128. They fail against main with sign-stripped results and pass with this change.

Validation:

• cargo +nightly test -p builtins-test --test addsub --no-default-features -j 1
• cargo +nightly test --workspace --no-default-features --exclude musl-math-sys -j 1
• cargo +nightly fmt --all -- --check
• git diff --check

The unexcluded workspace command requires the musl source fixture, which is not present in this shallow checkout.

AI assistance: I used OpenAI Codex to assist with investigation, implementation, and validation. I reviewed the changes and test results.

[HuzaifaAbdulRehman]

Hi @tgross35, this has been open for a couple weeks with all 44 checks passing and no conflicts with main. Happy to make any changes if you'd like a different approach, just let me know. Thanks for taking a look when you get a chance.

[tgross35]

Yeah sorry, I have seen this but have been a bit backed up for reviews, it may be another week or two since we've come across some higher priority things.

As a quick note, it looks like your PR description may be machine-generated; please rewrite it, all communication with humans needs to be handwritten. Also the title should be more specific, this is only touches addition (similarly, it shouldn't close the linked issue).