>See http://fox.wikis.com/wc.dll?wiki~C0000005ExError for more info.
>
>I talked to Robert Green (MS guy) who verified that it is indeed a bug in VFP.
>
>It dosn't "fix the bug", but running the same code under '98 dosn't give me as much of a headache as running the same thing under nt.
>

It's not 'a' bug, it's a whole class of errors, most of which can be avoided in VFP5 and VFP6 SP3 - there were things that were not handled properly in the initial release of VFP6, some of which were documented in detail. The article on the wiki refers primarily to errors addressed in VFP6 in SP3 - there are VFP5 problems which can cause the same type of error, but the resolution is different than what was done to fix errors in the initial VFP6 release. I can trivially cause C5 errors by passing invalid values to an API function, so it is not a single cause or single cure scenario.

NT is less tolerant of memory mismanagement in general, but is far less likely to have an app crash and take the whole OS with it. The actual trap is done by the CPU's memory mangement subsystem; Windows grabs the ISR for the hardware error, determines the cause and shuts down the app that overstepped it bounds if necessary, messily in most cases, having a tendency to leave DLLs in an inconsistent state.