>Does anyone have a clue why very infrequently I might be getting the following error while running a VFP.exe?
>
>STFUFF.exe caused an exception 10H in module VFP500.DLL at 0137:100657b7.Registers:
>EAX=00000001 CS=0137 EIP=100657b7 EFLGS=00010246
>EBX=0063a5f4 SS=013f ESP=0063a0ec EBP=0063a654
>ECX=00000000 DS=013f ESI=01945fb8 FS=1347
>EDX=0063a0c4 ES=013f EDI=00000001 GS=0000Bytes at CS:EIP:
>dc 4d bc e8 3c 38 fa ff 6a 00 8d 4d fc 51 8d 55 Stack dump:
>01946154 01945fb8 00000000 00000000 00000003 00000006 00000000 00000000
>00000006 019c23b0 10024ab4 00000000 102ff52c 1002dd95 31202020 00003933

Exception 0x10 is a floating point error. The only thing I've seen where VFP caused this was actually a problem with some of HP's printer drivers; there are a couple of references in Microsoft's Knowledgebase that recommend using an API call the alters the state of the floating point registers prior to calling print routines where a suspect HP driver is in use. My own solution has been to use print drivers from Microsoft's operating system distributions rather than the latest, greatest (and admittedly, more feature-laden) HP-supplied drivers.

If there's an HP printer in use, and the problem occurs printing reports or shortly afterwards, I'd start by switching print drivers. I tend to use an older printer's drivers that are known to be stable - in my case, the most frequent change is to use the HP IIIsi drivers in place of the HP 5si drivers.

There are other things that can cause the exception from an operating system's point of view, like not properly saving and restoring the state of floating point and MMX registers (they're actually the same registers, just used a bit differently), or a bad jump table vector.

Can you find some set of actions that consistently reproduce the problem?