Mike,

I see your point clearly, but I do wonder at how many production applications this has already made its way into?!?!?!?

Would it have hurt to put a little more info in the article, like which buffers, under what conditions, are freed up? surely 3-4 lines would have covered it.

By the way, the first time I saw your picture here I thought that the immortal Bard himself was now into VFP. (now THAT would be a real coup!).

Cheers and thanks,

JimN

>>It comes with no guarantee other than there is no gaurantee! I wonder if it ensure buffers are written out before purging them (since it says no gaurantee).
>
>If we guaranteed the reliability of the function, then we'd have to support it. If we support it, we can't axe it in the future because then it might break existing code. If we support, it also means we have to formally test it (which we may or may not have done, I don't know). So, apparently no one thought this would be useful enough to customers to put the work in necessary to support it. IIRC, it was originally put in there for testing purposes (like a lot of undocumented functions).
>
>All of this does not mean it does not work reliably, it just means we don't know whether it's reliable or not, and we're currently not going to go to the trouble (read: use finite resources that we could use to ship 7.0) to find out.
>
>>Wouldn't someone have to be real hard up to use this?? Would anyone *dare* to even consider this for a production system?
>
>See above. I would never recommend relying on using undocumented calls of any type.