Scientific notation of bases other than decimal

[Pol Welter]

Try 0x5e2.

Experienced result: 1506.

Expected result: Not sure. But not 1506.

Could be

• 1280 = 0x500
• 500 = 0x5 * 100
• NaN, because of ambiguity.

This seems to come from the parser deep down in floatio.c. I don't feel like digging over there...

[Pol Welter]

Wow, I must have seriously brainfarted when writing this. What was I thinking?

[Tey']

Looks like we all missed that one :)

For the records, the math engine support exponents for other bases (something like 0x5(+0d2) in your case), but not the parser. It also supports complement (0xsF5).

[Pol Welter]

What exactly do you mean by complement?

I was aware of the scientific thing, but this is new to me.

[Tey']

It computes the two's complement and adds a minus sign (not sure why though). With the decimal number 5, the syntax is 0xsF5 for hexa (=> -0xB), 0bs10101 for binary (=> -0b1011) and 0os75 for octal (=> -0o3). Adding support for that in the evaluator is pretty easy, but I'm not sure there will be any uses of it.

[Pol Welter]

I agree that this is not very useful. For computing the two's complement one should specify the bit length explicitly. For that we already have the unmask function.