Community
Participate
Working Groups
We have a benchmark that did "new Shell()" that was slower on GTK than Windows. Here is how we fixed it: Shell creation on a single monitor system takes longer then it needs to. It calls gdk_screen_get_monitor_at_window() when there is only ever one monitor possible. We optimized this case.
Fixed in HEAD > 20080627
We need to rerun the benchmarks.
Is it possible to get a patch for 3.4?
It's not worth it. Unless you are creating tons of shell, you won't see a difference.
I was asked to bring this in for Notes 8.5. I will take your recommendation back to them.. thanks.
Looks like we do create quite a lot. Some data was sent to you via email. Please evaluate and let me know if we would benefit from the patch.
I have no data, other than some benchmark code that was entered in another bug report that I cannot find right now (@#$%@#%). Bogdan, we entered this bug report when we fixed one of the issues in the benchmark code that I am talking about. Can you find the original bug report? I searched on "performance", "slow" etc. and failed.
Ok, Bogdan please find the bug report anyways but I have found the data. NOTE: It is not common (but also not wrong in the open source sense) for clients of Eclipse or SWT to decide that they wish to communicate data about a bug report via email. This can happen for any number of reasons, but like I said, it's not very common.
Yes, I understand. I would have liked the data and all communications to be posted here as well. It is easier to track.