| Summary: | can't find model while the associated file is being moved | ||
|---|---|---|---|
| Product: | [WebTools] WTP Source Editing | Reporter: | Hirotaka Matsumoto <jljlmatu> |
| Component: | wst.sse | Assignee: | David Williams <david_williams> |
| Status: | CLOSED DUPLICATE | QA Contact: | |
| Severity: | critical | ||
| Priority: | P1 | CC: | sisikawa |
| Version: | 0.7 | ||
| Target Milestone: | --- | ||
| Hardware: | PC | ||
| OS: | Windows XP | ||
| Whiteboard: | |||
|
Description
Hirotaka Matsumoto
I'll help investigate this one, but not sure if we can fix on our end. I think we (you) need to move more to the Eclipse-base-established-way that if a resource is moved, that you basically "abandon" what you have, and let the "new stuff" perculate up from the bottom. I belive, for example, that in this case 'setInput' is called again on an editor (I'd have to check to know which method, exactly). Does this sound feasible? Or am I missing the point? To abandon the old stuff, we need to know which model is the one we need to get rid of. In the above scenaio, the workbench tells us that "/dynamicWeb/WebContent/a.html is moved to /dynamicWeb/WebContent/b.html" so that we need to abondon the model of "/dynamicWeb/WebContent/a.html". Because our editor supports the multiple document editing ( i.e. frame/jsp include ), we need to pick up the model in the models our editor keeps for the editing. To do this, we get the id by calling IModelManager#calculateId(IFile) and we will look for the target model. However, the id in the model and the id which is returned from this api is different ( I think this is a root problem ), we can't find out the model which we need to release. I would guess fixing this bug (https://bugs.eclipse.org/bugs/show_bug.cgi?id=107316) solves this problem as well. . Agreed, probably a dup of 107316, but will leave open and up priority until confirmed. I believe fixing 106324 has fixed this problem as well. As always, if I've missed something, don't heesitate to re-open. *** This bug has been marked as a duplicate of 106324 *** no, fixing bug106324 seems not to fix this. I still believe this is the same as 107316. . Agreed. Closing as 'dup' which doesn't mean this is invalid description/use case, I'd just liked to track one place. *** This bug has been marked as a duplicate of 107316 *** . |