Max 8 Message box truncates float input (wonkish)

Chris Rolfe's icon

I recently switched to using 64-bit and Max 8. I expected message boxes to use 64-bit numerics but instead they seem to truncate input prior to binary storage.

From the FAQ:

"The bit resolution of Max messages is dependent on the version of software that you are using. Versions before Max 6.1, and the 32-bit versions of Max 6.1+, use 32-bit numeric data. The 64-bit version of Max 6.1+ uses 64-bit numeric data for all message contents."

Nuh-uh.

Message objects appear to limit input to 6 decimal places as text, i.e.,, typing 261.xxxxxxxxxxxx is truncated to 261.xxxxxx". To my mind, truncating the text string before parsing (whether as a 32-bit or 64-bit value) is incorrect, speaking numerically. Max 7 in 32-bit mode did it correctly, albeit in 32 bits.

Loadmess (newobject) works correctly in Max 8, but outputs 32-bit rather than 64-bit.

Manually entering the value into a number box and storing it as a snapshot works correctly (64-bit binary float).

Here's a patcher for Max * that demonstrates the problem with message:

FloatTest.maxpat
Max Patch

Edit: Hmm. Not sure how to post a Max patcher properly in this forum (with snapshots).

Chris Rolfe's icon

Here's the patcher as a zip file…

FloatTest.maxpat.zip
application/zip 2.11 KB
Zip