Summary: | Problems encountered while deleting resources | ||
---|---|---|---|
Product: | [Eclipse Project] JDT | Reporter: | Dani Megert <daniel_megert> |
Component: | UI | Assignee: | JDT-UI-Inbox <jdt-ui-inbox> |
Status: | VERIFIED FIXED | QA Contact: | |
Severity: | normal | ||
Priority: | P3 | CC: | frederic_fusier, philippe_mulet |
Version: | 3.4 | Keywords: | test |
Target Milestone: | 3.4.1 | ||
Hardware: | PC | ||
OS: | Windows XP | ||
Whiteboard: |
Description
Dani Megert
2008-07-10 03:25:24 EDT
I cannot see one of the changes of the JavaModel done in R3_4_maintenance which could explain this failure: - bug 235778: Potential race condition while resolving classpath - bug 222665: CCE while computing non Java resources - 2 other bugs around binding keys ...but of course, I may be wrong. So, the best thing to do until Jerome comes back from vacations is to make the test safe and retry if the resource is not deleted. Note that all JDT/Core tests files/directories deletions are done using utils method retrying until the resources are really deleted. So, I'm not sure there's really something to do on our side for this kind of issue, but Jerome will confirm or infirm this assumption... Good. Just wanted to be sure as this happened in 3.4.x stream. We've added the panic code on our side for 3.4.1. Please close once Jérome confirmed this. Correct. One cannot assume that a delete or a move will always succeed (many reasons can explain the failure to delete/move, the most common being another thread holding a lock on the resource, e.g. the indexer). Retrying is the best to do in a test suite. Moving to JDT/UI to dispose. Got fixed a while ago for 3.5 and 3.4.1. Verified in code of M20080813-0800. |