Skip to content

Latest commit

 

History

History

testdb

Folders and files

NameName
Last commit message
Last commit date

parent directory

..
README.md
README.md
 
 
comparebuildrepos.rb
comparebuildrepos.rb
 
 
comparepip.py
comparepip.py
 
 
fetch.rb
fetch.rb
 
 
findflake.rb
findflake.rb
 
 
grep.rb
grep.rb
 
 
jq.rb
jq.rb
 
 
packageoutput.rb
packageoutput.rb
 
 

README.md

Nuclearsandwich's experimental test report database scripts.

Some scripts for generating and querying a local flat-file database of Jenkins test reports and build logs.

Database "schema"

The root of the test data base defaults to $HOME/osrf/testdb but can be configured with the TESTDB_DIR environment variable. Within the root directory are domain directories with names matching the Jenkins domain names such as build.osrfoundation.org and ci.ros2.org. Within each domain directory are job directories with names matching the Jenkins job name such as ign_gui_ci-win and nightly_linux-rhel_release. within each job directory are build directories with names matching the Jenkins build number such as 1948 or 810. Within each build directory are files related to that build, most commonly testreport.json.gz which contains a compresed json document with the Jenkins test report results. Build directories may also contain a build.txt file containing the plaintext console log for the job or other files in the future.

Example layout:

TESTDB_DIR/
  build.osrfoundation.org/
  ...
  ci.ros2.org/
    nightly_linux_debug/
      1848/...
      1849/...
      ...
      1948/
        testreport.json.gz
        build.txt
    nightly_linux_release/

Scripts

These scripts are written in Ruby because they weren't ever really intended to leave my machine. If this project matures and these scripts start gaining wider adoption and Ruby is a barrier to contribution I'll rewrite them. Probably in awk.

Any relatively recent Ruby interpreter (>=2.0.0) and standard library should be sufficient to run the scripts. If a performant JSON parser is installed (Simdjson or Oj, in Ruby) they will be used instead of the stdlib parser. However processing time is not egregious without the performant JSON parsers so they are not required.