[4.2.7] FastFill behaves oddly

[cheese]

On execution of what is essentially the Optimize Daily script (with one additional zone for cache files right behind Zone 3),
the unaltered script part for Zone 4 produces during the FastFill run huge gaps (see attached image Note: Zone 5 on the screen equals Zone 4 of the plain vanilla script. This is a boot disk with 45 GB)
Zone 3 and 4 both ran FastFill  without problems.
Consecutive runs did overcome the issue, but were still hampered by large gaps though fewer
-C

[jeroen]

This is not a bug. Fastfill is designed to be fast, not designed to be perfect. If all files above a gap are bigger than the gap then the gap will be skipped. For exact details on what is happening see the debug logfile.

tip: you can zoom in on the diskmap, and see small files that are not visible in the normal resolution, lurking in gaps.

[cheese]

This is not a bug. Fastfill is designed to be fast ...

You might be right, though I like to believe that I understand how fastfill is supposed to work and posted because:
- the gaps are huge and largely empty (no small file pieces on zooming in, just the larger ones visible here)
- the gap pattern was created into a perfectly empty area that I created before in order to move/defrag the pagefile there (unsucessfully as it ended up on the top but that is not the point here)
- this zone does not contain large files at all (they should nicely fill in)
- there were no other apps running on the computer (no file deletions, creation at the same time)
- an immediate re-run of the same script did fill this zone perfectly (suggesting that it could have been accomplished)
It was an interesting experience to watch MyDefrag filling files in funny patterns into a perfectly empty gap.

Over night I had a chance to put MyDefrag through it's paces on a few more disks (all very rarely defragmented before, if at all).
I observed one similar case though not as pronounced, also perfectly fine after a second run.
My conclusion is, that under certain circumstances (maybe especially when a lot of optimization work is required) MyDefrag can get its volume map wrong. Unfortunately I can not think of any other way of helping to pinpoint this behavior than switching on debug logging on any future attempts.
You might like to file my report under 'anecdotal' until I come up with some more tangible data.

In any case, I am very grateful for your work on MyDefrag. Thanks.

-C

[jeroen]

Thanks for your insights, I appreciate the time you put into writing down your observations. Only with a debug logfile could I explain in detail what you have observed. I understand that what you have seen is not what you wanted or expected to see, and perhaps fastfill could be improved in some manner, but that does not mean a bug.