oopsla12-reproduce

Below is an explanation about results referred to in the OOPSLA 2012 paper submission.

How to Duplicate the Paper's Results

1. Follow the installation instruction page to get Prime ready on your machine.
2. Download one of the cached analysis results from the "cached" repository in this project.
   • Alternatively, create a cached file by yourself: download the Python script attached to this page, and run it with the query import%20full.api.package (e.g. import%20org.openqa.selenium). This will search Github for all files that contain that text and download them. Then use Prime to create a cached file, as explained in the running instructions page.
3. Create a new Eclipse Application run configuration, activating the product org.eclipse.sdk.ide with the arguments -os ${target.os} -ws ${target.ws} -arch ${target.arch} -nl "en_US" -consoleLog and the VM arguments -Xmx2500m -XX:MaxPermSize=256m -Xss4m.
4. Run the created Eclipse Application. The top menu should contain the Prime menu item:
   
5. Select the code you want to use as a query, then select "Code Search on Prime" from the menu (or just click the pickaxe icon on the menu):
   
   • The code segment selected should be a complete class called $Search.
   • Remember to use fully-qualified type names in the query, as shown in the screenshot.
   • The syntax is 100% Java; to simulate unknown sequences, simply call some made-up unknown method on the object you're interested in.
   • The object chosen as the query is automatically selected from the code.
6. In the opened dialog, select the appropriate .cached file(s) and feed in the regex to use for specifying API types, as demonstrated in this screenshot:
   
7. Wait until the results view opens. It should look something like:
   
8. Double-clicking any of the entries in the tree will generate the associate .dot (Graphviz) file in the output folder, and will also immediately attempt to open it with the default system editor. .dot file viewers can be downloaded from the Graphviz site.