Bad File Finder behavior with large list

tullik

Situation: C:\ and D:\ (a partitioned drive) contain 97786 files and 7979 directories (per win explorer select all & properties). Told FF to find *.* on both drives skipping no directories. Expected FF to pick up the ~97000 files. Deselected size and three date columns on the fly to speed the search.

Result: FF finished in 7 minutes :) and probably listed all the files (last upper right count noticed was over 94000).

However: :(

(1) the lower left count froze at 6499 and remained at that count when FF finished; select all selected the entire list, but the count showed 6499 selected.

(2) could not close any jv16pt window (three were present: main, FF tool, and FF results list); all could be selected. Told FF tool to look in a small directory for *.*. Result: large results list reset to blank grids (expected), but FF found /looked for nothing ... tried several variations of this with same result (found /looked for nothing). Still, none of the three windows would close. Had to kill jv16pt with task manager.

Added note: this was still beta4 -- did the test at the same time RC1 was put out.

redseujac

jv16 answered in another thread:

"The jv16 PT 2005 is quite slow in handling of very long lists, but the upcoming jv16 PT 2006 can handle long lists much better. Perhaps you should try the latest PT 2006 beta?"

So, there seems to be supplementary work to do by jv16 :

jv16

Hmmm... Could you try the same with Windows' Search feature? If it doesn't hang on such list maybe PT shouldn't either :

redseujac

Sorry, we encountered an error while displaying this content. If you're a user, please try again later. If you're an administrator, take a look in your Flarum log files for more information.

tullik

Sorry, we encountered an error while displaying this content. If you're a user, please try again later. If you're an administrator, take a look in your Flarum log files for more information.

tullik

Tried the same test as the topic head, but using RC1

The upper right count (URC) and lower left count (LLC) continued to change for quite a long time; LLC always lagged behind URC.

Time elapsed: about 1 hour, 5 minutes: the URC vanished -- it was about 110000, while LLC was 85567.

Time elapsed: about 1 hour, 25 minutes: LLC was 87318; stayed at this number.

Time elapsed: about 1 hour 40 minutes: FF doing little or nothing per Task mgr. Clicked on a line -- CPU usage went up, nothing else happened -- chose select all, cpu usage went up some more -- all three PT windows became unresponsive --CPU usage about 40% of each CPU -- let it loop overnight.

Time elapsed: about 9 hours, 45 minutes: CPU usage still the same; the three PT windows were just frames filled with white, still unresponsive; killed jv16pt with Task mgr.