Bug 464116 - DAnalysis.getModels returns different result after SiriusRepairProcess
Summary: DAnalysis.getModels returns different result after SiriusRepairProcess
Status: NEW
Alias: None
Product: Sirius
Classification: Modeling
Component: Core (show other bugs)
Version: 2.0.0   Edit
Hardware: PC Windows 7
: P3 normal (vote)
Target Milestone: ---   Edit
Assignee: Project inbox CLA
QA Contact:
URL:
Whiteboard:
Keywords: triaged
Depends on:
Blocks:
 
Reported: 2015-04-08 03:49 EDT by Laurent Fasani CLA
Modified: 2015-04-22 03:29 EDT (History)
3 users (show)

See Also:


Attachments

Note You need to log in before you can comment on or make changes to this bug.
Description Laurent Fasani CLA 2015-04-08 03:49:22 EDT
First, DAnalysisModelsUpdateTests.testDAnalysisModelsPreservationOnRepairMigrate must be updated to be relevant. The test needs to reopen the session before testing the result of DAnalysis.getModels
After this test fix, the test is KO but pertinent.
Comment 1 Eclipse Genie CLA 2015-04-08 04:09:48 EDT
New Gerrit change created: https://git.eclipse.org/r/45436
Comment 2 Esteban DUGUEPEROUX CLA 2015-04-14 12:11:04 EDT
Could you provide an example with a scenario to reproduce?
Comment 3 Laurent Fasani CLA 2015-04-21 08:39:41 EDT
In reply to comment 2:
After the first commit, the test is correctly written but is KO. 
The aim of this ticket is to understand why it is KO and provide a fix.
Hence, the scenario is to have this test OK.
Comment 5 Pierre-Charles David CLA 2015-04-22 03:29:18 EDT
Seen with Laurent. What happened is this: the changes performed in the context of bug 456351 made DAnalysisModelsUpdateTests.testDAnalysisModelsPreservationOnRepairMigrate fail, but after analysis it failed because it was actually incorrectly written and passed earlier for wrong reasons (what it tested does not correspond to what actually happens in a running system).

So Laurent's patch, merged as 62a157b3bfb9c196b41affe22d88667406807aa5 rewrites the test to actually check what it should always have checked. The rewrite reveals that the corresponding behavior was actually broken (and has been for a long time probably), so the test is currently disabled.

The tested behavior is somewhat independent of the work done for bug 456351, hence this new separate ticket, whose goals is to fix the underlying issue revealed by the rewritten test (and re-enable the test once this is done).