TICALCS.net

Full Version: Bug
You're currently viewing a stripped down version of our content. View the full version with proper formatting.
Wow... I just finished fixing a huge bug in Mathematica... Let me post about it tomorrow...
???

here we are...the day after tomorrow...
The bug that I discovered two days ago was that one of the arbitrary precision addition/subtraction functions was not allocating enough memory in the case where there was not a final carry in the completion of an addition operation. For example (I'm using base 10 here), the cases that I tested exercised carrying, such as 2+8 = 10. When there was no carry, such as 2+7 = 9, the system allocated exactly 2 times less memory than needed. This resulted in the corruption of some memory, and it also crashed the calculator when the Mathematica program terminated. I spent approximately 2 hours tracing this bug to the function where it originated. I do not have access to the source right now, so I am not sure for which function it was. I suspect it was in the case where an arbitrary precision number and an unsigned short were being added.

I guess, it is not really that interesting...
Reference URL's