Returning an exponent after adding a negative value that sums to 0.0000000

[tmbobbins]

Appears to happen at 7 decimal places on each side (at 6 it works fine), as long as it sums to 0

[rtheunissen]

I believe this is expected behaviour, as toString determines whether to use scientific notation based on the exponent, even for zero. The precision is preserved in the output.

You should use toFixed(), or if that is not quite what you need, we may need to look at an adjustment of string representation.

Just to be clear, was the expectation for this to return "0"?

[tmbobbins]

For the first 6 decimal places toString() will return 0.000000 (6) as soon as it hits the 7th decimal place it returns an exponent, hence I would have expected it to return 0.0000000 (7), this doesn't appear to happen if the value doesn't land on 0.

Is there any way at current to make toString not return exponents if the decimal lands on a large precision number that equals 0? Or are you saying that toFixed will return the decimal representation rather than an exponent?

As an example the above image returns all of the decimal places rather than an exponent for any other value than 0

[rtheunissen]

Looks like a special case to me. This issue is along the same lines as #22 so I think we should reconsider how string formatting is done. Maybe:

• __toString / toString / (string): No exponent, no trailing zeroes.
• toFixed: No exponent, trailing zeroes up to the number of places given.
• toScientific or toExponentString, same as toString but always using the exponent?

[tmbobbins]

aye the list looks good, consistency is the main take away.

[rtheunissen]

My only concern is that the default toString could end up creating a very long string when precision is high, but in those cases users should just use toExponentString instead, being aware of the precision they are working with.

[rtheunissen]

Let's roll with toString(), toFixed(int places) and toExponentString.

toString() will return a fixed representation of as many decimal places that the instance is aware of.

I think that should work.

[rtheunissen]

See #22 (comment)

[rybakit]

@rtheunissen I've stumbled upon the same problem and saw that it had already been solved in 2.0. Is there any plan to release version 2.0 in the near future?