Second argument to parseFloat in JavaScript?

Question

In this font-size resizing tutorial:

Quick and easy font resizing

the author uses parseFloat with a second argument, which I read here:

parseFloat() w/two args

Is supposed to specify the base of the supplied number-as-string, so that you can feed it '0x10' and have it recognized as HEX by putting 16 as the second argument.

The thing is, no browser I've tested seems to do this.

Are these guys getting confused with Java?

Answer 1

No, they're getting confused with parseInt(), which can take a radix parameter. parseFloat(), on the other hand, only accepts decimals. It might just be for consistency, as you should always pass a radix parameter to parseInt() because it can treat numbers like 010 as octal, giving 8 rather than the correct 10.

Here's the reference for parseFloat(), versus parseInt().

Answer 2

Here's a quickie version of parseFloat that does take a radix. It does NOT support scientific notation. Undefined behavior when given strings with digits outside the radix. Also behaves badly when given too many digits after the decimal point.

function parseFloatWithRadix(s, r) {
  r = (r||10)|0;
  const [b,a] = ((s||'0') + '.').split('.');
  const l1 = parseInt('1'+(a||''), r).toString(r).length;
  return parseInt(b, r) + 
    parseInt(a||'0', r) / parseInt('1' + Array(l1).join('0'), r);
}

parseFloatWithRadix('10.8', 16) gives 16.5

gist: https://gist.github.com/Hafthor/0a60f918d50113600d7c67252e68a02d