This Bugzilla instance is a read-only archive of historic NetBeans bug reports. To report a bug in NetBeans please follow the project's instructions for reporting issues.

Bug 40472 - Sluggish response of editor when typing and navigating
Summary: Sluggish response of editor when typing and navigating
Status: VERIFIED INVALID
Alias: None
Product: editor
Classification: Unclassified
Component: -- Other -- (show other bugs)
Version: 3.x
Hardware: PC Linux
: P2 blocker (vote)
Assignee: Antonin Nebuzelsky
URL:
Keywords: PERFORMANCE
Depends on:
Blocks:
 
Reported: 2004-02-25 14:28 UTC by Antonin Nebuzelsky
Modified: 2011-11-24 15:24 UTC (History)
3 users (show)

See Also:
Issue Type: DEFECT
Exception Reporter:


Attachments

Note You need to log in before you can comment on or make changes to this bug.
Description Antonin Nebuzelsky 2004-02-25 14:28:18 UTC
Editor is sluggish and responds with
interruptions. This is noticable especially on my
machine when I reuse a userdir with many
filesystems for a long time and large files open.

This should be investigated further, culprits
found (drawing engine of something else?) and
eliminated.
Comment 1 Jan Lahoda 2004-02-25 14:53:45 UTC
Hi,
   I have a question/comment: I did not understand what you do with
the userdir:
a) on the particular userdir, after the start of the IDE, the
responsiveness is OK, and gets worse over the time, or
b) on the particular userdir, the responsiveness is bad from the IDE
start and never gets better?

If the case is a), I would guess it is a leak (some listeners are not
unregistered correcty, etc.). In this case I would suggest to get  a
few memory dumps (for example by the hprof or insane) when the
situation happens. The leak may be visible in these dumps.

If the case is b), I think that answers to these questions may be
interesting:
What is the number of filesystems?
What are types of the filesystems (local/jar/CVS/???).
How the memory behaves?
Is it nearly full all the time, or is there plenty of free memory?
How many editors are open and how long are the documents?
What are the document types?
Comment 2 Antonin Nebuzelsky 2004-02-25 14:56:05 UTC
b) is correct.

I will have a look into it and comment to this issue if I find something.
Comment 3 Miloslav Metelka 2004-02-26 14:36:00 UTC
Tondo, I do not know how to evaluate this issue. You write that the
interruptions happen on your machine but there is "All" in the "OS"
field. Does it mean that you notice it on all platforms? I understand
that you have entered the issue in the way that you fill in some more
information later. But you've made it P2 issue so we need to evaluate
it and resolve it ASAP but we cannot because we miss the critical
information such as the thread dumps from the interruptions etc. You
well know that such thing can happen because of many reasons possibly
totally unrelated to the editor (such as any sort of background
scanning). So please give us more info or downgrade the issue to P4.
Thanks.
Comment 4 Antonin Nebuzelsky 2004-02-26 16:44:57 UTC
I am trying to find more info. Currently looking into it in a profiler.
Comment 5 Antonin Nebuzelsky 2004-03-01 16:14:57 UTC
Hmm, unfortunatelly profiler does not show anything obviously bad
(like filesystem/datasystem operations, touching clipboard frequently,
parsing, or something like that). All hotspots are from editor module,
which is IMHO expected.

Running two instances of a trunk build (first with my complex userdir
and second with a simple new userdir with just one FS mounted) using
-verbosegc does not reveal any differences in garbage collection
cycles. So, this is not the cause as well.

Still looking into it...
Comment 6 Antonin Nebuzelsky 2004-03-03 15:21:18 UTC
Hmm, very very long ago I turned on the option "Options -> Editing ->
Editor Settings -> Expert -> Text Antialiasing"...
No more comments needed I hope. ;-)
Comment 7 pfelenda 2004-03-03 16:13:27 UTC
Verified in dev 200403021900.
Comment 8 Miloslav Metelka 2004-03-04 12:15:13 UTC
I can imagine how long time it takes to discover such a subtle thing ...
Anyway I'm glad it's now resolved.