Bugzilla – Full Text Bug Listing
| Summary: | [tests] Port Equinox tests to the draft r4.3 implementation | ||||||||
|---|---|---|---|---|---|---|---|---|---|
| Product: | [Eclipse Project] Equinox | Reporter: | Lazar Kirchev <l.kirchev> | ||||||
| Component: | Framework | Assignee: | equinox.framework-inbox <equinox.framework-inbox> | ||||||
| Status: | RESOLVED WONTFIX | QA Contact: | |||||||
| Severity: | normal | ||||||||
| Priority: | P3 | CC: | tjwatson | ||||||
| Version: | 3.5 | ||||||||
| Target Milestone: | 3.7 | ||||||||
| Hardware: | PC | ||||||||
| OS: | Windows Vista | ||||||||
| Whiteboard: | |||||||||
| Attachments: |
|
||||||||
|
Description
Lazar Kirchev
Created attachment 172725 [details]
Tests ported to r4.3 implementation
The patch contains port of several classes, which did not compiled with r4.3 due to changes in classes of the framework. I made the changes needed for the tests to compile and pass.
One of the tests - SecurityAdminUnitTest - has 5 failing test cases when run against r4.3. These are the tests for postponed conditions. They fail because of a change in the SecurityAdmin.
Previously the SecurityAdmin had a reference to a supported security manager, which was EquinoxSecurityManager, which is removed in r4.3 draft implementation. Now its method getSupportedSecurityManager() returns the system security manager. The checks for postponed conditions fail, because it is checked if there is a supported security manager (a condition is postponed if a supported security manager is used). Before, the EquinoxSecurityManager from the SecurityAdmin was used, but now there is no supported security manager (the system security managed is not set). If the test sets the EquinoxSecurityManager as a system security manager many other checks fail. Is this an expected and desired effect of the change in the SecurityAdmin?
There are also other 5 test cases in other tests, which are failing with r4.3 and will have a look at them.
Thanks Lazar. I had looked at the security ones a bit. I think I released some changes to allow for easier testing. I will have to look at that. Created attachment 173742 [details]
Test fixes
These are fixes for the other 5 failing tests I have mentioned.
Actually the tests in the OSGI_R4_3_RI branch are taken not from the latest version of the tests (in Head), so they miss some changes which are present in the Head.
These 5 tests were failing because they missed these changes. In the patch I attach I have included these latest changes. But probably instead of applying the patch the failing tests may be merged with the code in their newer versions.
(In reply to comment #3) > Created an attachment (id=173742) [details] > Test fixes > > These are fixes for the other 5 failing tests I have mentioned. > > Actually the tests in the OSGI_R4_3_RI branch are taken not from the latest > version of the tests (in Head), so they miss some changes which are present in > the Head. > > These 5 tests were failing because they missed these changes. In the patch I > attach I have included these latest changes. But probably instead of applying > the patch the failing tests may be merged with the code in their newer > versions. The composite bundles tests remained. Where can I find documentation about the composite bundles? The R4.3 draft is not available yet on the OSGi Alliance page. I want to read it before having a look on the tests. *** Bug 312784 has been marked as a duplicate of this bug. *** There is no need to do this port anymore. The R4.3 specification is not going to have composites. We are not doing implementation of the R4.3 framework specification in HEAD. Eventually we will remove the composite implementation in HEAD. |