Bugzilla – Full Text Bug Listing
| Summary: | [planner] Projector should get fragments dependencies | ||||||||
|---|---|---|---|---|---|---|---|---|---|
| Product: | [Eclipse Project] Equinox | Reporter: | Pascal Rapicault <pascal> | ||||||
| Component: | p2 | Assignee: | Pascal Rapicault <pascal> | ||||||
| Status: | RESOLVED FIXED | QA Contact: | |||||||
| Severity: | normal | ||||||||
| Priority: | P3 | CC: | adreghiciu, leberre | ||||||
| Version: | 3.4 | Flags: | leberre:
review+
|
||||||
| Target Milestone: | 3.6 RC1 | ||||||||
| Hardware: | PC | ||||||||
| OS: | Windows XP | ||||||||
| Whiteboard: | |||||||||
| Bug Depends on: | 222158 | ||||||||
| Bug Blocks: | |||||||||
| Attachments: |
|
||||||||
|
Description
Pascal Rapicault
A similar problem happens for the patches and their concept of lifecycle. Created attachment 163355 [details] Proposed patch for fragments & unit tests Attached patch solves the described problem. This is done by not using requirements looping as a trigger for determining potential hosts but host requirements directly. I also changed IUFragment implementation to not automatically add host requirements as required capabilities, so solving also bug #222158. As proven by added unit tests, the solution also solve bug #307271, so host is not required to be present case when the fragment does not get attached. Forgot to mention that if the fragment has other sorts of requirements those will still be processed as till now. For the existing situation, where people expressed host requirements also as requirements situation will remain unchanged. Created attachment 167574 [details]
New patch
Here is a new patch. At this point I have not done the changes wrt removals of the host requirements from the regular requirements of the fragment.
Daniel, could you please review this patch.
+1 I have released additional changes for the persistence aspect of this work. . . Fixed |