I have been using SsanF to translate a string with two parameters separated by a comma to generate two float values.
The problem is that I use DOUBLES and they are not the same between 32bit and 64 bit, one is 10 bytes in 32 bit land while the other is only 8 which is just a double.
Otherwise, It turns into a cluster of spaghetti code using other methods to split the string and use StrToFloat etc.
The shortest I have come up with is using a local Extended and then from there define the actual values which are DOUBLEs
Is there some sort of hidden flag char that is not documented to indicate a DOUBLE or even a single float type?
Currently I only see "%f" for that.
SscanF('123.5,234.8','%f,%f',[@X,@Y]);
Here X,Y are doubles and it works ok in 64 bit land because there is no Extended because it gets resplaced by a Double
But to compile that in 32 bit mode crashes with a memory flood.