-
1. Re: [forge-dev] Seems like the performance of validation is not actually due to Forge...
gastaldi Jun 19, 2014 2:18 AM (in response to lincolnthree)Maybe the select/unselect algorithm should be improved? Or could this be caused by a listener registered in widgetSelected?
Em 19/06/2014, às 01:37, "Lincoln Baxter, III" <lincolnbaxter@gmail.com> escreveu:
Check out these results. Most of the time is actually spent in Eclipse!! What is going on here? I profiled this for a long time, selecting all checkboxes, then unselecting, so it's not just a fluke.
<Screen Shot 2014-06-19 at 12.30.03 AM.png>
--
Lincoln Baxter, III
"Simpler is better."
-
att1.html.zip 678 bytes
-
-
2. Re: [forge-dev] Seems like the performance of validation is not actually due to Forge...
maxandersen Jun 19, 2014 11:09 AM (in response to lincolnthree)On 19 Jun 2014, at 6:37, Lincoln Baxter, III wrote:
Check out these results. Most of the time is actually spent in
Eclipse!!
What is going on here? I profiled this for a long time, selecting all
checkboxes, then unselecting, so it's not just a fluke.
I think what you are seeing here is that Window.open() is actually
looping and pushing through
SWT.events for the dialog. Thus it will also count when "nothing"
changes.
That is my guess.
I would look inside the tree of widgetSelected() what is going on there.
Another explanation would be that you somehow filtered things out in
this view and that doesn't show up since
the %'s on the right does not add up.
/max
_______________________________________________
forge-dev mailing list
https://lists.jboss.org/mailman/listinfo/forge-dev