fix(compress): toFixed, toPrecision, toExponential and toString of Number

[magic-akari]

Description:

The code snippet from StackOverflow is for handling toPrecision instead of toFixed

BREAKING CHANGE:

Related issue (if exists):

• fix 'compress' option breaks number literals ending with '0' when .toFixed() is called #6957

[magic-akari]

The num_to_precision cannot handle JS toPrecision. It will return unexpected results.

assert.sameValue((10).toPrecision(1), "1e+1");
assert.sameValue((11).toPrecision(1), "1e+1");
assert.sameValue((17).toPrecision(1), "2e+1");
assert.sameValue((19).toPrecision(1), "2e+1");
assert.sameValue((20).toPrecision(1), "2e+1");

compress result:

assert.sameValue("10", "1e+1");
assert.sameValue("11", "1e+1");
assert.sameValue("17", "2e+1");
assert.sameValue("19", "2e+1");
assert.sameValue("20", "2e+1");

[kdy1]

Do you mean it's not possible at all?

[magic-akari]

Do you mean it's not possible at all?

No, we can calculate the result step by step according to the ECMA262 specification.

[magic-akari]

There are two test case files from test262.
I don't know if there is a license issue. Is it OK to include them?

[kdy1]

I think it's fine

[kdy1]

Thank you!

---

swc-bump:

• swc_ecma_minifier

[kdy1]

Thank you!

---

swc-bump:

• swc_ecma_transforms_base