Community
Participate
Working Groups
In the event some plugin adds a ServerEditorSection to many or all servers' editors, and that ServerEditorSection makes incorrect assumptions (all servers have runtimes, all servers have launch configs, etc), and causes a runtime exception of some sort, the entire first page of the editor refuses to show up. This is very confusing to the user. It also blows away the stack. Is there some other solution we can provide here? Perhaps a try/catch in ServerEditorPart.insertSections(etc) ? Link: https://issues.jboss.org/browse/JBIDE-8643
Rob, I don't understand the description on this bug. Sure we can add a try/catch, but can you please explain more clearly the problem? 1) What is the scenario, some adopter doing something wrong? 2) You say the first page doesn't show, is there something that shows instead? Does it show grey with an exception inside? 3) With a try/catch block what behaviour do you expect? For the first page to show but that section not to show, I assume?
1) What is the scenario, some adopter doing something wrong? Some adopter, specifically JRebel, has added a section for all servers. If you have JRebel installed, then you will see this bug. Their section's createSection(etc) method throws a runtime exception, and this propagates up the stack. 2) You say the first page doesn't show, is there something that shows instead? Does it show grey with an exception inside? In our specific example, the server we tested it with had two pages, the default "Overview" page which every server should have, and an additional page. When we experienced this bug, all we saw was our second page. There was no first page at all, nor a blank page with an error on it. The only indication something went wrong was a stack-trace in the Error log view. 3) With a try/catch block what behaviour do you expect? For the first page to show but that section not to show, I assume? I would expect that the first page (since it contains critical information) does show, and that the offending section did not show, or that it showed in whatever state it was in when the runtime exception was thrown.
Created attachment 200693 [details] Adds a try/catch around the creation of the section; logs error
Created attachment 201342 [details] v1.0 Rob, thanks for submitting the patch. The logic in the patch looks good. Updating the patch based on the original one to clean up the trace code.
Hey Elson: Great addition! Will this also make it into 3.3.1?
Yes, all changes on 3.2.5 will be ported to 3.3.1 and 3.4.
Code released to 32M, 33M and HEAD
New Gerrit change created: https://git.eclipse.org/r/109064
