https://bz.apache.org/bugzilla/show_bug.cgi?id=60687

--- Comment #15 from Felix Schumacher <[email protected]> ---
(In reply to Vladimir Sitnikov from comment #14)
> I think SummaryReport uses unsafe approach of updating ObjectTableModel in
> the sampling threads.

Could you be a bit more specific here? I think the only problem we have is
visibility, as the swing thread locks on lock and the other threads on lock and
rows. So I think of it as a kind of eventual visibility.

> 
> Does it make sense to decouple stats collection from stat visualization?

That was my first thought, but it would introduce another thread. We could drop
the row locks, though.

> 
> If I get it right, we should avoid using swing.* as a stat
> collection/aggregation facility.
> Otherwise we have to use Swing.invokeLater on the hot path.

Right. Get swing out of the path and be happy.

-- 
You are receiving this mail because:
You are the assignee for the bug.

Reply via email to