ParseJSON - Differences for Double Values

[Martin_T]

I see different results after parsing a JSON file (using ParseJSON) with following double values:

Original:

{"pgSz":{"w":597.60000000000002274,"h":842.39999999999997726}}

Parsed:

{"pgSz":{"w":597.5999755859375,"h":842.4000244140625}}

Can anyone tell if this is a bug or if it is automatically rounded due to the length of the decimal point? Are the double values too long after the decimal point?

[DerkJ]

Double is by definition not precise so this could happen. Perhaps an engineer could let you know if this could be a bug or not.

But note that with doubles you won’t get the exact values you expect mostly.

[Jeff_Tullin]

In fact for paper sizes, you may as well be storing these as integer sizes, that’s how they are defined.
A1 paper is 594 x 841

[anon20074439]

I’d say it was a bug, the original double seems to be being changed into something shorter.

597.60000000000002274
0100000010000010101011001100110011001100110011001100110011001101

597.5999755859375
0100000010000010101011001100110011000000000000000000000000000000

[Christian_Schmitz]

Do they crop it to 32-bit float instead of 64-bit double?

[anon20074439]

It looks like its being converted to a single.

[Christian_Schmitz]

Works JSONMBS class in with MBS Xojo Util Plugin:

and the ToString outputs it like this:

{
	"pgSz":	{
		"w":	597.60000000000002274,
		"h":	842.39999999999997726
	}
}

As you see we preserve the digits.

[Martin_T]

Sounds logical. I‘ll create a feedback case.