Community
Participate
Working Groups
I consider this to be a serious new bug introduced since F3. We were running GM1 today, and several times we ended up with a situation where the editor was telling us that some class was undefined. It suggested that we import it. Problem is, it was already imported. The only workaround we found was to completely exit eclipse, and restart, at which point it was quite happy. We saw this at least three times today. We could not successfully run the code. We're running the linux/gtk version.
Would you have simple steps to reproduce this problem?
Instead of exiting, you may want to close/reopen the offending projects from the navigator. It should have the same effect. We had such defects at some point due to some concurrency situations, but thought we had fixed them (all?). There must still be a way for the JavaModel to become inconsistent at some point, but without steps to reproduce, there is little we can do in the short term. Also, no one else has reported such defects lately, downgrading it to P2 until further information is provided. Anyway, it is getting late for addressing before 2.0
Unfortunately, we have no simple steps to reproduce. Today, as we're working, we'll try to pay attention to what we were doing just before the problem occurred. I am suspicious that some change between F3 and GM1 has made this problem worse for us. For our medium-sized project (30kloc), we had never before seen this problem over a month's heavy use of F2 and F3, yet it occurred 3 times in our first day on GM1. One feature that we used heavily yesterday, but only infrequently in the past, is the Refactoring/Pull Up..., but whether this is significant is only speculation.
Closing. Quite a number of fixes occurred in this area since then. Please reopen if steps can be determined.
