fix: bad logic, add tests for every module

[nberlette]

• tests(cbrt): add tests in cbrt.test.ts
• tests(log2): add tests in log2.test.ts
• tests(log10): add tests in log10.test.ts
• tests(trunc): add tests in trunc.test.ts
• tests(ceil): add tests in ceil.test.ts
• tests(asin): add tests in asin.test.ts
• tests(random): add tests in random.test.ts
• tests(log): add tests in log.test.ts
• tests(float32): add tests in float32/guards.test.ts
• tests(float32): add tests in float32/constants.test.ts
• tests(float32): add tests in float32/index.test.ts
• tests(float32): add tests in float32/encode.test.ts
• tests(float32): add tests in float32/decode.test.ts
• tests(float32): add tests in float32/round.test.ts
• tests(clamp): add tests in clamp.test.ts
• tests(clz32): add tests in clz32.test.ts
• tests(abs): add tests in abs.test.ts
• tests(constants): add tests in constants/log2e.test.ts
• tests(constants): add tests in constants/max_safe_integer.test.ts
• tests(constants): add tests in constants/ln2.test.ts
• tests(constants): add tests in constants/log10e.test.ts
• tests(constants): add tests in constants/positive_zero.test.ts
• tests(constants): add tests in constants/epsilon.test.ts
• tests(constants): add tests in constants/negative_zero.test.ts
• tests(constants): add tests in constants/index.test.ts
• tests(constants): add tests in constants/nan.test.ts
• tests(constants): add tests in constants/sqrt1_2.test.ts
• tests(constants): add tests in constants/ln10.test.ts
• tests(constants): add tests in constants/positive_infinity.test.ts
• tests(constants): add tests in constants/min_safe_integer.test.ts
• tests(constants): add tests in constants/e.test.ts
• tests(constants): add tests in constants/max_value.test.ts
• tests(constants): add tests in constants/negative_infinity.test.ts
• tests(constants): add tests in constants/pi.test.ts
• tests(constants): add tests in constants/sqrt2.test.ts
• tests(constants): add tests in constants/infinity.test.ts
• tests(constants): add tests in constants/min_value.test.ts
• tests(expm1): add tests in expm1.test.ts
• tests(sin): add tests in sin.test.ts
• tests(constants): add tests in constants.test.ts
• tests(atan): add tests in atan.test.ts
• tests(max): add tests in max.test.ts
• tests(asinh): add tests in asinh.test.ts
• tests(sinh): add tests in sinh.test.ts
• tests(index): add tests in index.test.ts
• tests(acos): add tests in acos.test.ts
• tests(atan2): add tests in atan2.test.ts
• tests(guards): add tests in guards/positive_zero.test.ts
• tests(guards): add tests in guards/safe_integer.test.ts
• tests(guards): add tests in guards/negative_zero.test.ts
• tests(guards): add tests in guards/index.test.ts
• tests(guards): add tests in guards/nan.test.ts
• tests(guards): add tests in guards/integer.test.ts
• tests(guards): add tests in guards/positive_infinity.test.ts
• tests(guards): add tests in guards/negative_infinity.test.ts
• tests(guards): add tests in guards/infinity.test.ts
• tests(guards): add tests in guards/finite.test.ts
• tests(tan): add tests in tan.test.ts
• tests(ieee754): add tests in ieee754.test.ts
• tests(floor): add tests in floor.test.ts
• tests(tanh): add tests in tanh.test.ts
• tests(imul): add tests in imul.test.ts
• tests(float16): add tests in float16/guards.test.ts
• tests(float16): add tests in float16/constants.test.ts
• tests(float16): add tests in float16/index.test.ts
• tests(float16): add tests in float16/encode.test.ts
• tests(float16): add tests in float16/decode.test.ts
• tests(float16): add tests in float16/round.test.ts
• tests(types): add tests in types/precision.test.ts
• tests(types): add tests in types/safe_integer.test.ts
• tests(types): add tests in types/index.test.ts
• tests(types): add tests in types/integer.test.ts
• tests(types): add tests in types/finite.test.ts
• tests(types): add tests in types/float.test.ts
• tests(cos): add tests in cos.test.ts
• tests(sqrt): add tests in sqrt.test.ts
• tests(atanh): add tests in atanh.test.ts
• tests(min): add tests in min.test.ts
• tests(pow): add tests in pow.test.ts
• tests(trigonometry): add tests in trigonometry.test.ts
• tests(sign): add tests in sign.test.ts
• tests(exp): add tests in exp.test.ts
• tests(f16round): add tests in f16round.test.ts
• tests(log1p): add tests in log1p.test.ts
• tests(round): add tests in round.test.ts
• tests(hypot): add tests in hypot.test.ts
• tests(cosh): add tests in cosh.test.ts
• tests(fround): add tests in fround.test.ts
• tests(acosh): add tests in acosh.test.ts
• tests(internal): create shared test helpers module
• chore(internal): add ArrayFrom to primordials
• refactor(random): clean up MT19937 class, use ArrayFrom primordial
• fix(float16,float32): use type-only imports for FloatFormat
• fix(ieee754): ensure floating point rounding logic is sound
• chore: update license
• docs(readme): improve doctests, use node:assert
• config: rewrite CI workflow for npm trusted publishing
• fix(constants): ensure negative-zero module is exported
• tests: fix doctest error in src/index.ts
• config: update deno.json tasks and settings
• fix(abs): ensure abs behaves like Math.abs
• fix(asinh): ensure asinh behaves like Math.asinh
• fix(atan): ensure atan behaves like Math.atan
• fix(atan2): ensure atan2 behaves like Math.atan2
• fix(cbrt): ensure cbrt behaves like Math.cbrt
• fix(ceil): ensure ceil behaves like Math.ceil
• fix(clz32): ensure clz32 behaves like Math.clz32
• fix(floor): ensure floor behaves like Math.floor
• fix(hypot): ensure hypot behaves like Math.hypot
• fix(min): ensure min behaves like Math.min
• fix(round): ensure round behaves like Math.round
• fix(sinh): ensure sinh behaves like Math.sinh
• fix(sqrt): ensure sqrt behaves like Math.sqrt
• fix(trunc): ensure trunc behaves like Math.trunc
• refactor(float16,float32)!: remove constants re-exports
• ci: add build script

[Copilot]

Pull request overview

This pull request adds comprehensive test coverage for every module in the math library (100+ new test files), fixes critical bugs related to signed zero handling and floating-point edge cases, refactors the Mersenne Twister implementation, and introduces build tooling with an npm publish workflow.

Changes:

• Adds test files for all math functions, constants, guards, types, and float16/32 modules
• Fixes bugs in 15+ math functions (abs, asinh, atan, atan2, cbrt, ceil, clz32, floor, hypot, min, round, sinh, sqrt, trunc) to properly handle signed zero, infinity, and NaN edge cases
• Refactors MT19937 class to use module-level constants and ArrayFrom primordial
• Changes float16/float32 imports to type-only imports for FloatFormat
• Adds build script for npm publishing and rewrites CI workflow

Reviewed changes

Copilot reviewed 118 out of 118 changed files in this pull request and generated 6 comments.

Show a summary per file

File	Description
src/internal/_test_utils.ts	New test helper utilities (assertSameValue, assertClose)
src/internal/ieee754.ts	Improved floating-point rounding logic with overflow handling
src/internal/primordials.ts	Added ArrayFrom primordial binding
src/abs.ts	Fixed to return +0 for both +0 and -0 inputs
src/asinh.ts, src/cbrt.ts, src/sinh.ts	Added signed zero preservation
src/atan.ts	Fixed Taylor series calculation errors
src/atan2.ts	Comprehensive rewrite for correct signed zero handling
src/ceil.ts, src/floor.ts	Rewrote to eliminate circular dependencies
src/clz32.ts	Replaced logarithm-based implementation with bit manipulation
src/hypot.ts	Fixed NaN/infinity handling, removed max() dependency
src/min.ts	Added signed zero preference handling
src/round.ts, src/trunc.ts	Rewrote for correct signed zero behavior
src/sqrt.ts	Added explicit infinity handling
src/random.ts	Refactored MT19937 to use constants at module level
src/float16/constants.ts, src/float32/constants.ts	Changed to type-only imports
src/float16/index.ts, src/float32/index.ts	Removed constants re-exports (breaking change)
src/constants/index.ts	Fixed missing negative_zero export
All test files (90+)	Comprehensive test coverage added
scripts/build.ts	New build script for dnt-based npm publishing
.github/workflows/ci.yml	Rewritten CI with JSR/NPM/GPR publishing support
LICENSE	Updated copyright year and formatting
README.md	Updated documentation examples to use node:assert
deno.json	Added tasks, lock file, nodeModulesDir config

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[Copilot]

In the atan2 function, lines 49-50 and 56 have coverage ignore comments. The logic at line 50 appears unreachable - if x is negative infinity and y is finite, the condition at line 51 would be true (since !isFinite(y) is false), so the code would fall through to line 55. This suggests the logic flow may need review.

[Copilot]

The comment at line 23 states "deno-lint-ignore-file no-explicit-any" but this is a build script that will be run with Deno's linter. Consider whether ignoring all explicit any usages in this file is appropriate, or if specific locations should be marked instead for better type safety.

[Copilot]

The LICENSE file header changed from "The MIT License (MIT)" to "# The MIT License (MIT)" (adding a markdown heading). This may cause issues with automated license detection tools that expect the standard MIT License format. Consider keeping the original format without the markdown heading symbol.

Suggested change

	# The MIT License (MIT)
	The MIT License (MIT)

[Copilot]

Using BigInt(value < 0 ? 1 : 0) is more verbose than the original value < 0 ? 1n : 0n. The original pattern using bigint literals is clearer and more idiomatic for bigint operations.

[Copilot]

The condition in line 110 checks if mant_bits === mant_scale, which would mean the mantissa has overflowed. However, this should trigger rounding up the exponent and resetting mantissa to 0. Consider whether this logic correctly handles the edge case where incrementing the exponent causes it to overflow into infinity territory - the check at line 116 should catch this, but verify the flow is correct.

[Copilot]

Unused import assert.