11 c. crash

[P Padil]

During some calcs I was converting large values of arbitrary precision integers to a floating point by mulitplying by 1.0, which at some point it overflows the range of double so the result was _. Tried using extended precision floats to see if that still gave a _, but it resulted in a "floating point exception” which crashed jconsole.

Example:
1.0*!10x^6 => _
1.0fq*!10x^6 => crash

Regards
Pete

[P Padil]

Always forget:
Engine: j9.7.0-beta14/j64arm/darwin
Build: commercial/2026-03-25T22:06:46/clang-17-0-0/SLEEF=1
Library: 9.7.11
Platform: Darwin 64
Installer: j9.7 install

[Devon McCormick]

You appear to be trying to convert numbers which cannot be represented in floating point.  According to Gemini, the range for the IEEE 64-bit floating point specification is  about 1.7977E308.  The value of !10x^6 is about 1e5565709 so is out of range.

Devon McCormick

To unsubscribe from this group and stop receiving emails from it, send an email to forum+un...@jsoftware.com.

[P Padil]

I know is out of range for 64bit but even then the answer is _ (signifying overflow - not crashing).

[P Padil]

Just to be clear, the example is just for demonstrating the issue, not what I was trying to do which is using the log10 of some values...

[Henry Rich]

I'll get to this, but it won't be till May.

Henry Rich

[Henry Rich]

If you add this to the GitHub issues list you can be sure I won't forget about it. 

Henry Rich

[P Padil]

Ok will do.