Yes, the plain ceil function is fine, but the XL_ceil, and ceil_qtrs functions are (nearly certainly, without actually testing) broken for negative numbers.

I'll have to play a bit later and see if I can fix those functions in a similar way...

Update: Indeed there were issues with that code. I have /msg'ed the QandAEditors, and asked them to replace the bad code, but until they do, I've placed the new code on my scratchpad.

-Scott

Update: Erroneous code has since been replaced with working code.