Bug #7993
closedSome attributes are not shown
100%
Description
The sync:race rule may output several unsafes per module. Strange, but attributes "Linux kernel", "LKVOG strategy" and "Verification object" are shown only for the first unsafe. The other unsafes have "-" in the cells. And other attributes are visualized completely.
An example of job is attached.
Files
Updated by Vladimir Gratinskiy about 7 years ago
Please restart this job with "fullweight" option selected.
Updated by Pavel Andrianov about 7 years ago
- File Job-5304a14c5c-0.zip Job-5304a14c5c-0.zip added
It turned out that in the "fullweight" launch all attributes are shown correctly. Just in case I attached a job. Is it possible to show attributes in "middleweight" launch too?
Updated by Vladimir Gratinskiy about 7 years ago
The attached job is medium-weight too (650 KB).
Pavel Andrianov wrote:
It turned out that in the "fullweight" launch all attributes are shown correctly. Just in case I attached a job. Is it possible to show attributes in "middleweight" launch too?
No, it is impossible as reports doesn't have needed attributes. Maybe wrong parent identifier set for unsafe report during its uploading.
Updated by Pavel Andrianov about 7 years ago
- File Job-5304a14c5c-0(1).zip Job-5304a14c5c-0(1).zip added
Real "full-weight" job is attached.
Updated by Evgeny Novikov about 7 years ago
- Category changed from Tasks generation to Bridge
Core does not distinguish a verification results weight - it always upload the same reports. I guess that you didn't test medium- and lightweight verification results when there are several unsafes or safes with the same attributes like verification object and rule specification.
Updated by Vladimir Gratinskiy about 7 years ago
I've found my comment in code where unsafe report is uploaded: # Each verification report must have only one unsafe child. In other cases unsafe reports will be without attributes
In this job all unsafes belongs to one verification report.
Updated by Vladimir Gratinskiy about 7 years ago
Vladimir Gratinskiy wrote:
I've found my comment in code where unsafe report is uploaded:
# Each verification report must have only one unsafe child. In other cases unsafe reports will be without attributes
In this job all unsafes belongs to one verification report.
The reason is after unsafe report is uploaded, the system checks that parent (verification report) has archive (with report files). If it does then verification report parent is set to Core. If doesn't then unsafe's parent is set to Core.
So in first case after the second unsafe report is uploaded there are no components (with their attributes) between Core and Verification report and the second and other unsafes are uploaded without attributes of these components.
Updated by Evgeny Novikov about 7 years ago
Was it fixed to support several child unsafes and perhaps unknown after all?
Updated by Vladimir Gratinskiy about 7 years ago
- Due date set to 02/25/2017
- Status changed from New to Resolved
- % Done changed from 0 to 100
Evgeny Novikov wrote:
Was it fixed to support several child unsafes and perhaps unknown after all?
Fixed in fix_7993.
Updated by Pavel Andrianov about 7 years ago
- Status changed from Resolved to Open
I ran the same task, middle-weight configuration on the fixed version (branch fix_7993). Now, if I clicked on the unsafes, I get an empty web page. Full-weight launch is still working, I get a page with list of unsafes.
Updated by Vladimir Gratinskiy about 7 years ago
Please decide the job in branch "test_7993" then send me file "klever/bridge/REPORTLOG.txt".
Updated by Pavel Andrianov about 7 years ago
- File REPORTLOG.txt REPORTLOG.txt added
Updated by Vladimir Gratinskiy about 7 years ago
- Assignee changed from Vladimir Gratinskiy to Evgeny Novikov
There are no "verification finish" reports at all. But it is required for medium- and light- weight jobs.
Updated by Evgeny Novikov about 7 years ago
- Status changed from Open to Closed
- Assignee changed from Evgeny Novikov to Vladimir Gratinskiy
I examined verification finish reports are uploaded and the suggested fix does help. So I merged the branch to master in 82cf4c8.
Perhaps Pavel has some incorrect version of Klever or a browser cache hits so that one can't expect something good. If this isn't the case, please, report this.