Performance with large amount of events

[fff7d1bc]

Hi,

I am trying to watch a mountpoint with a large amount of files, a simple call of du -sh /mnt/somemountpoint goes over hundreds of thousands of files. Since the tree is in cache it exits in less than 3s however the get_events() takes age to get over all of those events generated simply by du -sh call. This is a generator and if my loop over poll()` I go with

        while poll.poll():
            events = list(cli.get_events())
            print(len(events)

I see that it usually have only 11 events there, rarely 12, never more, and there's some delay between next iteration. I am not sure where it comes from and how could I modify the code to clear it faster. Any hints would be great. Thanks!