Enable/Disable Mutation Analysis Plugin per SonarQube project

[princefisher]

Hello Gerald, (@gmuecke )
I have been work with your great plugin in my locally hosted SonarQube instance for a few weeks and found it quite useful.

I have now asked my admins to consider installing this plugin on our company's central SonarQube cluster and they came back with a few concerns. Since I am no expert and only briefly looked through the plugin's source code, I was wondering if you could help me out with the following concerns/questions.

Note:

• Our central SonarQube is enterprise edition and running the latest LTS version and hosts thousands of projects.
• My local SonarQube is community edition and running in Docker with only a few projects for testing.

1. The plugin seems to be enabled by default for every project
2. The plugin produces "vast" amounts of info-type logging regardless of server/project settings
3. Performance while parsing the XML files (especially when they are from larger code bases)

I have already tested some of these concern and noticed the following:

1. The mutation analysis seems to run regardless of whether the Java/Kotlin sensors are active on every project in a given SonarQube. (At least there are logs for every project on my local/docker hosted SonarQube [community edition], latest LTS)
2. The amount of logs in default setting... devcon5io-mutation-analysis-plugin-generated-logs-per-scan.log
   • I attached above log file snippet as an example... if I analyze the same project twice with two different keys, where one project is enabled with mutation analysis and the other is not, then I still see all logs for each project.
3. This one is probably not up to this plugin, since XML parsing is usually straight-forward. I guess we will have to experiment with incremental analysis and consider not enabling all mutations for all projects to reduce the XML file size and thus speed up the parsing on the SonarQube server.

[princefisher]

Hi @gmuecke ,
any chance you could take a look at above questions?

[gmuecke]

Hi @princefisher
sorry for the very late response on my part, I had other priorities the last two years.
Anyway, regarding your questions:

1. you could enable/disable the sensor per project. I don't know how SQ handles this for other projects, as all share the basic concepts (global / project settings).
2. I wouldn't consider this amount as "vast" :) The logs mentioned are on the server-side (as far as I can tell), not sure if I can control the behavior of Sonarqube
3. I haven't done any performance optimizations especially not for large input files.

Hope this helps,
Gerald