I fixed a rare bug that was causing about 1 out of every 1000 WUs to crash. Not a big deal, unless you were one of the unlucky ones to get such a WU.
The app also checkpoints more often now and gives finer resolution with the progress meter.

Bad news: I got another computation error - After only 13 hours this time ;)

http://stat.la.asu.edu/NumberFields/result.php?resultid=49570

http://stat.la.asu.edu/NumberFields/workunit.php?wuid=31816

It's a different problem this time though:

segmentation fault: bug in PARI or calling program

Al.

Bad news: I got another computation error - After only 13 hours this time ;)

http://stat.la.asu.edu/NumberFields/result.php?resultid=49570

http://stat.la.asu.edu/NumberFields/workunit.php?wuid=31816

It's a different problem this time though:

segmentation fault: bug in PARI or calling program

Al.

Yeah, the other bug I fixed had an error message like:
*** impossible assignment I-->I
*** Error in the PARI system. End of program.

The seg fault error message is going to be harder to track down, because it doesn't happen to me when I run stand alone. It's like there is some interaction between the boinc client and my app which is causing the fault. But I now have a private "mirror image" of the project on my home network which I use for testing. So I can attach a debugger to the actual boinc client and watch what happens...