Community
Participate
Working Groups
The package explorer is listening explicity for the F5 key event. With the new key binding architecture, it is not recommended that individual views listen for key events. Instead, they should register handlers (actions) with one of the global commands. This allows the user to modify the key binding, if they wish. It also guarantees that your key binding will always be respected (i.e., not clobbered by someone defining a key binding on a global command that "hides" yours). The new key configuration provided in I20031202 binds F5 to "Refresh". This seems to trigger a refresh in the package explorer properly. If this is not the case, please register an appropriate handler for the package explorer view. In either case, please stop attaching a key listener to the view.
This bug is related to Bug 47093. It is also part of the work involved in Bug 43851.
Classes in JDT/UI that reference SWT.F5 CallHierarchyViewPart.java JavaBrowsingPart.java TypeHierarchyViewPart.java PackageExplorerActionGroup.java HierarchyInformationControl.java Douglas, the reason why we did this was that the debugger uses F5 in the "Old Eclipse Keybinding" and we had lots of requests F5 does a refresh. If we change this to use the keybinding service what will happen if users use the "old eclipse" key binding. Can you please comment on this.
Ping ...
Sorry for not replying earlier, but I've been waiting to come up with a complete answer. I'll give you what I have.... I have tested the scenario you described (both debug and jdt binding "F5" as a keyboard shortcut). It works for me. However, in speaking with Chris, it is not clear that this is necessarily true in all situations. I haven't had a chance yet to test more thoroughly, so I'm loathe to recommend anything at this point.
I'm going to have to go back on this. Or, at least, fixing this isn't strictly necessary. In the ideal world, both debug and package explorer should define a context, and in their individual contexts they should bind "F5". "F5" shouldn't be bound in the global context. However, we've found a way around the problem, which seems to work (i.e., unhandled commands don't eat keys). I'd recommend closing this bug for now. Either "INVALID" or "LATER".
dirk, do you want to suggest a retargable 'refresh' action? Close it if you don't.
Markus, for your information.
