The largest negative integer value we misinterpret

[anonymous]

The following code illustrates the problem

ConsoleWrite("- 9223372036854775808 = (int)" & Int (-9223372036854775808) & "= (hex)" & Hex (-9223372036854775808) &CRLF)
ConsoleWrite("- 9223372036854775807-1 = (int)" & Int (-9223372036854775807 - 1) & "= (hex)" & Hex (-9223372036854775807 - 1) &CRLF)

at -9223372036854775808 I get as a result -9223372036854775807
at (-9223372036854775807 - 1) I get the correct result -9223372036854775808

tested with the last stable and beta version

[jchd18]

This is rather the largest negative int64 value, but yes there seems to be an issue with parsing the literal value -9223372036854775808.

[BrewManNH]

That's not an Int64 value, it's too large. When converting it to a 64 bit value, -9223372036854775808 and -9223372036854775807 yield the same result.
​http://www.binaryconvert.com/result_double.html?decimal=045057050050051051055050048051054056053052055055053056048056

[anonymous]

Not Int with Double confused !!
​http://calc.penjee.com/?s=HJJJJJJJJJJJJJJJ

[jchd18]

-9223372036854775808 and -9223372036854775807 yield the same result.

Only when using a site which limits integral values to Int32 or under the double tab.

And no they are not the same value and both are valid int64 values.

[czardas]

This appears to be a bug connected with string conversion. See the following topic : ​https://www.autoitscript.com/forum/topic/176690-number-puzzle/ It seems to occur with various functions which convert the integer to a string - including ConsoleWrite.

ConsoleWrite(-9223372036854775808 & @LF) ; -9223372036854775807

[J-Paul Mesnage]

same as #3772