Bug 604517

Summary: RFE: filtering of matrix report - display only lines where results differ or regressions
Product: [Community] Beaker Reporter: Marian Csontos <mcsontos>
Component: web UIAssignee: beaker-dev-list
Status: NEW --- QA Contact:
Severity: medium Docs Contact:
Priority: low    
Version: 0.5CC: bpeck, mcsontos, tools-bugs
Target Milestone: ---Keywords: FutureFeature, Triaged
Target Release: ---   
Hardware: All   
OS: Linux   
Whiteboard: UX
Fixed In Version: Doc Type: Enhancement
Doc Text:
Story Points: ---
Clone Of: Environment:
Last Closed: Type: ---
Regression: --- Mount Type: ---
Documentation: --- CRM:
Verified Versions: Category: ---
oVirt Team: --- RHEL 7.3 requirements from Atomic Host:
Cloudforms Team: ---

Description Marian Csontos 2010-06-16 06:37:03 UTC
Description of problem:
I have a job with about 100 tasks (this applies to all TierN tests).
Matrix is great for reviewing regressions but 90%+ of the results are the same.

See: https://beaker.engineering.redhat.com/matrix/?whiteboard_filter=&job_ids=3229

I would like to be able to see only lines where:
- the results are not same
- the result is worse than base result

I run as two recipesets in single job.
It might be tricky, how to specify what's the base and what its update...

Comment 1 Marian Csontos 2010-06-16 06:39:46 UTC
See also Bug 596410 requesting filtering of columns.

This would add filter for rows as well...

Comment 2 Raymond Mancy 2010-06-16 07:09:15 UTC
(In reply to comment #0)
> Description of problem:
> I have a job with about 100 tasks (this applies to all TierN tests).
> Matrix is great for reviewing regressions but 90%+ of the results are the same.
> 
> See:
> https://beaker.engineering.redhat.com/matrix/?whiteboard_filter=&job_ids=3229
> 
> I would like to be able to see only lines where:
> - the results are not same
> - the result is worse than base result
> 
> I run as two recipesets in single job.
> It might be tricky, how to specify what's the base and what its update...    

You wouldn't really design it on the recipeset level, that would only work for people who have set up their Jobs to work in such a way, and I assume most don't.

Perhaps a 1:N relationship could be created on jobs, and then a special kind of matrix report could be run on these, which would basically do a complement of the base against all the other 'non base' jobs?

Comment 3 Marian Csontos 2010-06-16 08:30:58 UTC
> You wouldn't really design it on the recipeset level, that would only work for
> people who have set up their Jobs to work in such a way, and I assume most
> don't.

You are right of course. It was just convenient having it in one job as I need to rerun it... :-(

More info:

Display/Hide should be done with javascript - so one does not have to reload results. Rationale: Bug 604541 I just added...