This talk takes you through the life cycle of a bug in Eclipse from the filing of the bug to the state of commit of the solution of that bug. As all are aware, all the issues are tracked via bugzilla, and the story starts with the creation of the bug

New -> Analyze and move to Assigned. The owner of the bug analyzes, finds the solution, writes the test cases and commits into Gerrit. Gerrit runs all the tests and if found successful will flag so. Then the owner adds the reviewers, Reviewers review the code and suggests any changes inline and owner acts upon them. This repeats till there are no more review comments. Owner commits code using gerrit itself.

This talk will take you through a demo of this process.

Have you ever wondered how the results automatically appear in the website? This talk will take you through this journey from the code to the results –the multi-stage, multi-tech release engineering process internals.

There two different streams producing three different regular builds ( not taking into account the special P and Y) with the final bits running four different platforms making it a gruesome 32 hour journey optimized by parallelization. This process has four pipelined stages, with each stage driven by separate owners – one of them being Maven-Tycho configuration and the rest being a combination of shell Scripts, Ant, Junit in sync with multiple Hudson servers (HIPP).

To give a peek into the stages : The first stage has private as well as publicly visible parts, using two stage cloning with bare bones and normal git repos owned by Scripts. Second Stage, owned by Maven-Tycho, is responsible for three maven-tycho builds. This output feeds into the next phase, owned by shell-scripts, which does a few things resulting in a transient pseudo-private download zone which again feeds into the last stage owned by Hudson-Ant-Junit Config and is responsible for actual firing of the tests in multiple platforms. Finally you have the publicly viewable html based downloads and results.