| Summary: | Baseline is disposed | ||
|---|---|---|---|
| Product: | [Eclipse Project] PDE | Reporter: | Benjamin Muskalla <b.muskalla> |
| Component: | API Tools | Assignee: | PDE API Tools Inbox <pde-apitools-inbox> |
| Status: | RESOLVED WORKSFORME | QA Contact: | |
| Severity: | normal | ||
| Priority: | P3 | CC: | darin.eclipse, Michael_Rennie, Olivier_Thomann |
| Version: | 3.5 | ||
| Target Milestone: | 3.6 M7 | ||
| Hardware: | All | ||
| OS: | All | ||
| Whiteboard: | |||
| Bug Depends on: | |||
| Bug Blocks: | 285123 | ||
|
Description
Benjamin Muskalla
Validating / updating our tests for bug 275020 and bug 282537 I had a chance to debug this exception (partly). I found that the baseline manager was responding to a resource changed event and tossing out the workspace baseline in the middle of a build (in the example I debugged). I think to properly fix this we need to do a few things: 1. we need to move away from the "toss it all because there was a change" to something more selective, i.e. only dispose the API component that actually has changes. 2. we need to make our API containers recoverable - meaning they can be re-opened / re-initialized once close / disposed. Currently if a container is closed / disposed we just die when we try to access it (the cause of this bug). (In reply to comment #1) Since the changes to amalgamate all of our delta processing and the updates to builder to not dispose baselines as well as the move to use PDE state instead of handling it all ourselves, I have not seen this exception. It also no longer ever appears in the testsuite either. Marking as worksforme |