| Summary: | [linked source] deleting a linked resource from a read-only project leaves project in an undefined state | ||||||
|---|---|---|---|---|---|---|---|
| Product: | [Eclipse Project] Platform | Reporter: | Helmut J. Haigermoser <helmut.haigermoser> | ||||
| Component: | Resources | Assignee: | Szymon Brandys <Szymon.Brandys> | ||||
| Status: | RESOLVED FIXED | QA Contact: | |||||
| Severity: | normal | ||||||
| Priority: | P3 | CC: | mober.at+eclipse, sptaszkiewicz, Szymon.Brandys | ||||
| Version: | 3.7 | ||||||
| Target Milestone: | 3.8 M4 | ||||||
| Hardware: | All | ||||||
| OS: | All | ||||||
| Whiteboard: | |||||||
| Bug Depends on: | |||||||
| Bug Blocks: | 369493 | ||||||
| Attachments: |
|
||||||
|
Description
Helmut J. Haigermoser
Created attachment 199484 [details]
Pic showing the error dialog following an attempt to remove the linked resource from a readonly .project file
CQ:WIND00164690 This might be related to bug 322821 Hi All :) Can we get a comment from the platform team, this bugs seems like an easy fix: check the write-flag of .project before actually removing linked resources from the project. However, what looks easy from the outside usually is a lot different internally... Helmut I can't speak for the Undo/Abort dialog - that comes from the refactoring system. I'm not sure what exactly Undo tries to do when the operation fails. I can see why the inconsistency occurs - Resource#deleteResource removes the resource from the workspace tree before attempting to write the description. Thus when writing the description fails the resource is already gone. These steps could probably be reversed to avoid the immediate problem - delete links/filters before deleting the resource from the tree. We should probably also perform a Workspace#validateSave here to give a team hook a chance to checkout the resource. Szymon B, can you take this, or perhaps it is a good one for Gosia or Szymon P. (In reply to comment #4) > Szymon B, can you take this, or perhaps it is a good one for Gosia or Szymon P. Szymon, Gosia, what's your thoughts on this? Thanks a lot! :) Helmut (In reply to comment #5) > (In reply to comment #4) > > Szymon B, can you take this, or perhaps it is a good one for Gosia or Szymon P. > Szymon, Gosia, what's your thoughts on this? > Thanks a lot! :) > Helmut I haven't looked at it closer yet. I'll let you know soon. (In reply to comment #6) > (In reply to comment #5) > > (In reply to comment #4) > > > Szymon B, can you take this, or perhaps it is a good one for Gosia or Szymon P. > > Szymon, Gosia, what's your thoughts on this? > > Thanks a lot! :) > > Helmut > I haven't looked at it closer yet. I'll let you know soon. Thanks Szymon, please let me know if you need any more input to reproduce or fix this! :) Helmut This has been reported by a customer of ours. Is the fix safe enough for backporting into 3.7.2 ? Or is the time window for 3.7.2 fixes already closed ? Thanks, Martin (In reply to comment #9) > This has been reported by a customer of ours. > > Is the fix safe enough for backporting into 3.7.2 ? > Or is the time window for 3.7.2 fixes already closed ? > > Thanks, > Martin It's pretty late. We are in RC3 for 3.7.2, but still can release important fixes. The fix is safe, so I'm raising a bug to backport it to 3.7.2. |