[
Date Prev][
Date Next][
Thread Prev][
Thread Next][
Date Index][
Thread Index]
[
List Home]
Re: [ptp-dev] Abstract element viewer selection bug
|
I understand the part where you say that if the mouse y-coordinate is
returning you the wrong number then clearly you cannot correctly locate
that node in the window - certainly that's not your fault. What I don't
understand is who's problem it is. Is it a problem with the Linux GTK2
SWT? Do you have any idea WHY there's a problem? You say the widget
cannot support a large size - do you mean a large pixel size?
That canvas is essentially a massively sized widget and you're only
showing a portaion of it that's on the screen at a time. Are you saying
if we try and great a JContainer and set it's size at something very
large (say 1000x20,000 pixels) it will not work under Linux GTK2 SWT?
-- Nathan
Correspondence
---------------------------------------------------------------------
Nathan DeBardeleben, Ph.D.
Los Alamos National Laboratory
Parallel Tools Team
High Performance Computing Environments
phone: 505-667-3428
email: ndebard@xxxxxxxx
---------------------------------------------------------------------
Clement Chu wrote:
Hi Nathan,
I know this error before. If you created a large number of
elements, this error may occur. Let me explain what it is. One of
the problems is that the widget cannot support a large size. 50000
elements = (50000 * (16+4)) pixel where 16 and 4 are icon's pixel and
spacing between each icon respectively. Another one is that the
y-coordinate of the mouse cannot locate at accurately location of the
widget. So when you clicked the node and another node will be
selected. My code is depending on the mouse coordinate on the widget
to find out the element. If the y-coordinate is not correct, I will
find out the wrong element. Finally, this error is depending on what
platform you are using. Different platform SWT will have different
result. I tried on my laptop which is running under windows. I
haven't the problem you mentioned, but the scrollbar only support to
node 27845 which is depended on the size of your view. Sorry, it is
quite hard to explain. Do you understand what I am talking about so far.
Regards,
Clement
Nathan DeBardeleben wrote:
More information.
Attached are 2 more screenshots. The first, selectbug2.gif, shows me
selecting a region where the region isn't off the screen. You can
see that as I select a rectangular region it's going 5 or so nodes
north of where I'm selecting and grabbing those nodes. I've hovered
over the node number to give you an estimate of where this happens.
The second screenshot, selectbug3, is hard to explain. What I'm
doing here is just selecting a SINGLE node. The mouse isn't showing
up in the screenshots :(. However, you know that the hover tooltip
is displayed near/on top of the mouse. So you have an idea of where
my house is. Anyway, in this screenshot as I hover over this node
and select it, notice that the node about 7 rows north gets selected
instead.
There's some sort of math translation that's wrong here. It's
incorrectly dealing with machines that have this high number of
nodes. Again, you can fix it temporarily by resizing the machine
window. So it makes me think there's some relationship to the
rows/columns (number of elements displayed) code.
Hope this helps others see what I'm seeing.
-- Nathan
Correspondence
---------------------------------------------------------------------
Nathan DeBardeleben, Ph.D.
Los Alamos National Laboratory
Parallel Tools Team
High Performance Computing Environments
phone: 505-667-3428
email: ndebard@xxxxxxxx
---------------------------------------------------------------------
Nathan DeBardeleben wrote:
I've attached a screenshot here that shows the bug. Basically on a
Linux box I set the simulator preferences to 1 machine and 50,000
nodes. I then was scrolling around the machine view and saw that
when I scrolled towards the bottom and hovered over a node it didn't
display the tool tip. I then tried to select a region and it
totally screwed up.
Take a look at the screenshot.
Notice that I'm halfway down the list fof machines, trying to select
a rectangular region, but it's zooming the bottom of the rectangle
off the screen and is ALSO selecting the nodes above my region, off
the screen. Something's wrong with the logic when the number of
nodes is very large.
Clement, can you take a look at this since you wrote the code? We
kind of need this fixed because we're going to want to demonstrate
that we can simulate on a massive number of machines.
Thanks! Hopefully it's clear what error I'm seeing by looking at
the screenshot.
------------------------------------------------------------------------
------------------------------------------------------------------------
_______________________________________________
ptp-dev mailing list
ptp-dev@xxxxxxxxxxx
https://dev.eclipse.org/mailman/listinfo/ptp-dev
------------------------------------------------------------------------
------------------------------------------------------------------------
------------------------------------------------------------------------
_______________________________________________
ptp-dev mailing list
ptp-dev@xxxxxxxxxxx
https://dev.eclipse.org/mailman/listinfo/ptp-dev
_______________________________________________
ptp-dev mailing list
ptp-dev@xxxxxxxxxxx
https://dev.eclipse.org/mailman/listinfo/ptp-dev