Hi Hank,
>
>that's why I use CVP, which let's me boil it down and find the long-running statements. Without it, however, I agree it's much easier to take the time to output time logs.

I have run CVP only testwise but not on very long files and I had assumed the times taken would grow the same as in the vfp coverage compiler - reasoning by analogy is not really QED <g>. But I have not really verified that set coverage to will be able to write out files > 2 Gig, which it did not in vfp6 but should happen if the coverage runs in one part for an H - which would mean at least one component to be free of that ridicolous 2GB barrier even exisiting on LLF.

I remember creating a hook im my logging fwk to toggle coverages depending on rules, so that they could be analyzed in the vfp coverage profiler in reasonable time, and some routines to SQL-massage huge log files to create estimates on times running with and witout called code, then loading a .txt piped analysis of those parts into a memo field of the log table - those were fine days early this century <bg>.

regards

thomas