Dan Ingalls wrote:

...

I'm trying to create a typed FloatArray. There's not a lot in the system for converting floats to bytes, so I hacked up something where I use basicAt: to extract the first and second dwords, and then strip the dwords down a byte at a time (a bit more complex then VW where changeClassToThatOf: #[] works handily). But now that I've got it ripped down into 8 bytes, I can't figure out how to get the float back again. I put the 4 byte chunks back together (via bitShift: and +), but can't use basicAt:put: on an instance of a float. Feature, bug,...? Is there a better way to do this anyway?

Travis Griggs

I bet you forgot to use Float new: 2 instead of just Float new (which defaults to new: 1). [This should probably be fixed]

Try the following example:

     | r |
    r := Float new: 2.
    r basicAt: 1 put: (Float pi basicAt: 1).
    r basicAt: 2 put: (Float pi basicAt: 2).
    r

Yup. That was it. Actually I used Float basicNew which made it's size 0. Thanks much!

Travis Griggs