Then the precision would be 0, just like with Float.

Apply the same change to Complex, which had a different validation

Avoid memory leak when TypedData_Wrap_Struct fail.
First, prepare TypedData_Wrap_Struct with NULL.
Next, allocate Real.
Finally, bind VALUE and Real.

Use exp(x) = exp(x.round(k)) * exp(x - x.round(k)).
exp(x.round(k)) is fast because calculation of x**n is fast.
exp(x-x.round(k) is fast because x**n/n! converges fast.

Now that assert_fixed_point_precision is removed and there is only assert_relative_precision,
"relative" part is not important anymore. Change it to `assert_converge_in_precision` which express the assertion more accurate.

* Add support for tangent function

* Apply review comments for tests

* Make tan depend only on sin

* Use Ziv's loop method suggested in review

* Test failing test

* Use first implementation based on sin and cos

* Remove unused code

Use BigDecimal(float, 0) instead of BigDecimal(float).
Add a ruby-implemented polyfill of BigDecimal#_decimal_shift.
This is needed to let math.rb depend on bigdecimal.rb utility methods.

…error (#415)

* Round result of sqrt and other BigMath methods

BigMath methods no longer returns extra digits which may be inaccurate

* Remove workaround for removing extra digits from sqrt result in log calculation

…ulo (#429)

Bumps [step-security/harden-runner](https://github.com/step-security/harden-runner) from 2.13.0 to 2.13.1.
- [Release notes](https://github.com/step-security/harden-runner/releases)
- [Commits](step-security/harden-runner@ec9f2d5...f4a75cf)

---
updated-dependencies:
- dependency-name: step-security/harden-runner
  dependency-version: 2.13.1
  dependency-type: direct:production
  update-type: version-update:semver-patch
...

Signed-off-by: dependabot[bot] <support@github.com>

…rity/harden-runner-2.13.1

Bump step-security/harden-runner from 2.13.0 to 2.13.1

When y.exponent is several thousand or more, x**y was slow because exponentiation by squaring requires several thousands of multiplications.
Use exp and log in such case.
Needed to calaculate (1+1/n).power(n, prec)

…439)

* Update changelog for v3.3.0

* Bump version to 3.3.0