Hide Forgot
Description of problem ====================== Colors in result.html don't reflect nature of the result properly: RESULT || COLOR ||-------.------- || bar | box =================||=======|======= passed || green | green fixed || red | green failed || ??? | red informational || red | grey needs_action || red | grey needs_inspection || red | grey notapplicable || ??? | grey error || ??? | red Legend: * 'bar' - the colored proportion bar under "Compliance and Scoring" section * 'box' - individual result boxes you can see when you click "Show all result details" * 'color: ???' - the result is not represented or misrepresented (e.g. XYZ counted as "passed") in the bar See attached screenshots and reports for illustration. Version-Release number of selected component ============================================ preupgrade-assistant-2.1.10-6.el6 How reproducible ================ Always Steps to Reproduce ================== 1. Run upgrade path with various results For example, test /CoreOS/preupgrade-assistant/preupg/result-files runs a fake upgrade path that contains one of each valid results 2. Open result.html 3. Observe result severity color coding Actual results ============== Colors and gauges don't match nature of result * (A) in bar, 'informational' and 'fixed' are red * (B) in bar, something is misinterpreted as "passed" (there's only one 'passed' in the test but the bar counts 2) * (C) 'needs_inspection', 'needs_action' and 'informational' boxes are all grey * (D) there's also another similar bar called "Severity of failed rules", which apparently only contains 'failed' statistics, however it's coded green here. * (E) 'error' is missing from the bar * could be more... Expected results ================ Colors should match nature of result, the report should be intuitive. For example: * (A) 'informational' and 'fixed' should have good/neutral color * (B) only 'passed' should show as 'passed' here * (C) at leasr 'needs_action' and 'needs_inspection' should not be grey * (D) hard to say what to expect and whether this bar could make sense given we have only one "failed" severity type. * (E) 'error' should not be missing from the bar Additional info =============== We're using custom results inherited from original XCCDF results. My first guess is that the definition we give is not complete so these custom results inherit some "default" values inappropriate fot their new meaning, which further messes up the stats and color coding.
Created attachment 1211900 [details] realistic use case: bar
Created attachment 1211901 [details] realistic use case: boxes
Created attachment 1211902 [details] realistic use case: full report
Created attachment 1211903 [details] fake upgrade path: bar
Created attachment 1211904 [details] fake upgrade path: full report
Note about attachments: * "realistic use case" is simple preupg run on a fresh ppc64 install. * "fake upgrade path" is a mock upgrade path forged to generate each result type exactly once (a modified version from test mentioned above).
Closing this bugzilla as the component is under a maintenance mode in which the maintainers are going to fix only critical bugs. If you consider the bugzilla critical, feel free to reopen the bug with an explanation.