Developer intro

If you wanted to make a change to OpenGrok code base (small fix, new feature, anything) here's couple of steps to get you started.

Before we begin

Firstly, create Github account (of course !).

OpenGrok can be built and tested on any system with Java. You will need:

1. At least JDK8
2. Universal ctags (built from source)
3. Git (for cloning the repository), Subversion (needed for successful build)
4. Any modern IDE (e.g. IDEA, Eclipse)
5. other SCMs

Also, it would not hurt if you created an issue (https://github.com/oracle/OpenGrok/issues) and mentioned that you will be working on it.

Setting up your code repository

You'll want to create a fork of the oracle/OpenGrok repo. On Github this is as simple as clicking 'Fork' on the main project page. Getting the source code of your fork is easy, just use the instructions on the front page of the project and select the right method for you for getting the source (https://help.github.com/articles/which-remote-url-should-i-use).

Here's an example on getting the source from command line (assuming your fork is called 'foo/OpenGrok' (where 'foo' is your username on Github)

git clone [email protected]:foo/OpenGrok.git opengrok-git-mine

You'll want to setup remotes (mainly the path to the upstream repo) using the steps on https://help.github.com/articles/fork-a-repo For OpenGrok it would be:

cd opengrok-git-mine
git remote add upstream [email protected]:oracle/OpenGrok.git

Then create a branch (again, the help document on https://help.github.com/articles/fork-a-repo contains detailed steps) and switch to it, e.g.:

git branch myfix
git checkout myfix

Building

To build the code from command line, just run ./mvnw compile. It will download all necessary pre-requisities. When using IDE you can open or import the project.

Deploy the web app to the web server

There are multiple possibilities.

Deploy .war directly

Run command ./mvnw package -DskipTests which will create a release as in the releases page. This release can be found in the distribution/target directory with file name opengrok-{version}.tar.gz. Unzip the file by tar xvf opengrok-{version}.tar.gz. The .war file is located in opengrok-{version}/lib/source.war. You can copy this file to the web applications directory of your application server (e.g. webapps for Tomcat). Or you can use different means specific to application servers (e.g. Tomcat Manager).

Use Tomcat Maven plugin

Modify ~/.m2/settings.xml to:

<settings>
  <servers>
    <server>
      <id>OpenGrok</id>
      <username>admin1</username> <!-- you can use different name -->
      <password>admin1</password> <!-- you can use different password -->
    </server>
  </servers>
</settings>

Add to tomcat_users.xml located in conf directory of your Tomcat server:

<user password="admin1" roles="manager-script,admin" username="admin1"/>

Note: Do not add manager-gui role because it won't work.

Type following commands to the console:

./mvnw install
cd opengrok-web
./mvnw tomcat7:redeploy

or use IDE - in IDEA simply click on the Maven projects tab on the upper right side of the window. Then invoke the redeploy target of the Maven Tomcat plugin. It is advisable to toggle the 'Skip tests' button in order to speed up the redeploy.

Setup sources and index them

Now setup the sources to be indexed under e.g. /var/opengrok/src and create data directory for storing indexes under e.g. /var/opengrok/data. Make sure both directories have correct permissions so that the user running the process can read and write to them.

In IDEA, go to the Run menu and select Edit Configurations and create a configuration based on Application so it looks e.g. like this:

The arguments are nicely editable by expanding the field:

You can then run the indexer from the Run item in the Run menu. Of course, there can be multiple indexer runs preconfigured.

Or, you can run the main method org.opengrok.indexer.index.Indexer e.g. like this from command line (once the Maven package phase is done):

java -Djava.util.logging.config.file=logging.properties \
    -cp 'distribution/target/dist/*' org.opengrok.indexer.index.Indexer \
    -W /var/opengrok/etc/configuration.xml \
    -s /var/opengrok/src \
    -d /var/opengrok/data \
    -c /usr/local/bin/ctags \
    -H -S -P \
    -U http://localhost:8080/source

This is assuming the ctags binary of your ctags installation resides in /usr/local/bin/ctags.

If you now refresh the web page mentioned above it will reflect the reindex and you can do searches etc.

Debugging

See Debugging wiki for more information on debugging.

Test

To run tests type ./mvnw test command. For specific tests you can use -Dtest option, e.g. ./mvnw test -Dtest=IndexerTest -DfailIfNoTests=false.

Also, OpenGrok repository is setup so that pushes will trigger Travis builds so it is not necessary to run tests on your workstation - just commit and push to Github (first it is necessary to enable Travis for your fork on the Travis web).

See Testing wiki for more info on testing.

Publish changes

Once done with your changes, save them, git commit and push them to your repository (or you can do the Git dance directly from your IDE). From there it is possible to create new pull request to the upstream master branch using the standard Github process (https://help.github.com/articles/creating-a-pull-request - again Github help describes this in detail).

Make sure to sign the Oracle Contributor Agreement (http://www.oracle.com/technetwork/goto/oca) and ideally post the contributor number to the pull request.