Bugzilla – Full Text Bug Listing
| Summary: | if target platform has an existing configuration two workbenches will launch | ||
|---|---|---|---|
| Product: | [Eclipse Project] PDE | Reporter: | Mik Kersten <mik.kersten> |
| Component: | UI | Assignee: | PDE-UI-Inbox <pde-ui-inbox> |
| Status: | RESOLVED WORKSFORME | QA Contact: | |
| Severity: | normal | ||
| Priority: | P3 | CC: | daniel_megert, david_williams, ekuleshov |
| Version: | 3.3 | ||
| Target Milestone: | 3.3 M7 | ||
| Hardware: | PC | ||
| OS: | Windows XP | ||
| Whiteboard: | |||
|
Description
Mik Kersten
Honestly, I didn't believe this until I saw it with my own eyes (with recent I-build) ... weeiiirrrd stuff! I am having problems reproducing this. In the steps listed, for "Run another workbench (e.g. your normal development workbench)", was the development workbench I20070205-1300 or was it a previous build? If it was a previous build, which build? I talked to David Williams and found when he saw the problem he was using M4 with the target platform pointed to I20070207-1800. I have been unsuccessful running M4 and I20070206-0010 with the target platform being I20070206-0010. I am downloading I20070207-1800 to try to reproduce on that build. In my steps I was always using I20070205-1300, no other Eclipse versions. When I said "run another workbench" I should have mentioned via "-data". let me know if you can't reproduce, and I can try again with I20070207-1800. As David points out the effect of this is *really* weird and causes a way longer runtime workbench startup time with no clear work-around, so if the bug is reproducible I think it would be very good to have this fixed for M5. Mik, thanks for your quick response. I would imagine the problem didn't magically get fixed in the past couple days, so don't worry too much about trying the latest version (though if you have time, you might try it on the outside chance we get lucky). First, do you mind attaching the config.ini from the runtime workbench. It will be in <workspace>/.metadata/.plugins/org.eclipse.pde.core/<workbench name>. We did some config.ini work that reads the config.ini file of the target platform. I was hoping to see if that might have caused the problem. Second, I want to confirm how many java process are launched. This should be in the Debug perspective. Want to double check we aren't kicking off two different java processes. Third, I just want to make sure I have the steps right. 1. Download the zip file (I20060205) 2. Unzip it into a directory 3. start the installation, and choose an arbitrary workspace (wasn't sure if it HAD to be a subfolder) 4. Close the running instance. This should have created the configuration directory. 5. Run the installation again, with a -data to specify a workspace other than the one used in step 3. 6. Create a runtime workbench with unique workspace 7. Run. Then two instances lauch. Does that sounds right? I left out resetting the target platform since it should automatically be set right when self hosting. I moved to eclipse-SDK-I20070208-1200 today and I can no longer reproduce this problem. I'm not sure if this is a good or a bad thing! I'm certain that I was not using M4 because that is not possible for me (due to working self-hosted). There is on thing I was was doing differently on step 3 when I was able to reproduce this: * I would create an eclipse/workspace folder * I would then run eclipse/eclipse.exe, point the dialog that came up at that (empty) folder, and click the option to always use that folder * I would then do exactly what you say in steps 5, 6, 7 Unfortunate I no longer have the config.ini corresponding to the failure. Sorry that I can't provide more information. I'm confused now as to why this is no longer happening. But if other people aren't reporting it perhaps there was something peculiar about my configuration. The only interesting things I can think of is that I do have an extension directory (subclipse, findbugs and target management installed). The only change between when when I was seeing this behavior and today is that I removed RSE (http://www.eclipse.org/dsdp/tm/) from my install. from the timing of the bug report, I suspect it had something to do with the new Equinox launching story, which has stabilized since. |