Find file History



The usage of Repairnator is described in the dedicated documentation.

Content of the directory

This directory contains the different subfolders:

  • repairnator-*: are the different Repairnator modules
  • docker-images: contains the content of the different docker images produced for Repairnator
  • scripts: contains all the script to run Repairnator

How does Repairnator work?


Repairnator is decomposed in 6 different modules:

  • repairnator-core: contains shared elements for all other elements, it can be also use to generate Google Spreadsheets Credentials
  • repairnator-scanner: scan automatically Travis Build and produce metrics and list of build ids (only for Bears)
  • repairnator-dockerpool: this part can be used to create a pool of docker containers to launch taking as input a list of build ids
  • repairnator-pipeline: it is the engine of Repairnator. Given a build id this part, compile, test and repair it, gathering data on it, and push on Github and on MongoDB
  • repairnator-realtime: inspects in realtime the builds from Travis (typically every minute)
  • repairnator-checkbranches: this part is used to validate efficiently a large set of data produced by Repairnator

Usage of each parts are detailed in their own Readme file.


The following process is generally used.

First, repairnator-pipeline is built using shared element from repairnator-core and a docker image is created.

If the user wants to follow and repair in realtime the builds procuced by Travis, then she should use repairnator-realtime. The component is build using elements from repairnator-core and repairnator-dockerpool: the docker image previously created will be used automatically for each failing build.

Then if the user wants to repair scanned builds on a timerange, she should use repairnator-scanner and then repairnator-dockerpool. First she built and launch repairnator-scanner, using a list of Github project as input, and a list of build ids will be produced. Then she can build and launch repairnator-dockerpool against this list of build ids: the previously created docker image will be then used.

Those process can be automatically triggered using the scripts given in scripts directory.