Difference between revisions of "Gyrex/Contributor Guide"
(added content from other pages)
|(6 intermediate revisions by the same user not shown)|
|Line 1:||Line 1:|
= Committer Resources =
= Committer Resources =
= Workspace Setup =
= Workspace Setup =
Revision as of 15:26, 28 November 2012
|Mailing List • Forums • IRC|
|Browse Source • Project Set File|
A guide to contributing and committing to the Gyrex project.
- 1 Committer Resources
- 2 Workspace Setup
- 3 Builds
- 3.1 Really Quick HowTo
- 3.2 Prerequisites
- 3.3 Building on Windows
- 3.4 Build Environment Setup
- 3.5 Cloning Gyrex Source Tree
- 3.6 Running The Build
- 3.7 Build Process Insights
This article lists useful tips and articles for developers interested in working with/at the Gyrex code base.
Policies & Guidelines
The Eclipse Gyrex Project follows the Eclipse Project API guidelines. Please look at API Central and read the resources carefully. In case of any questions please don't hesitate to ask on email@example.com mailing list.
Ramp Down Plan
Typically the last week of a Milestone is for testing, and fixing only regressions and P1 or blocking defects. For milestones, the component lead (or delegate) is enough to review and approve a bug.
For any M6 build, we plan to be API complete, so there will not be any more breaking API changes or unsolicited API change requests accepted.
After RC1 is produced, the time for general functional improvements is long past. The following describes the types of bugs that would be appropriate:
- A regression
- A P1 or P2 bug, one that is blocking or critical, and some cases of major severities.
- Documentation and PII files are exceptions to the normal PMC required review, since there is little chance of that breaking anything, though it is still expected to be complete by M6, and remaining work to be only documentation fixes (that is, no refactoring of plugins, build changes, etc, without PMC review and approval).
- In addition to a bug meeting the above priority/severity conditions, there should be a simple, safe, well understood fix that is well isolated from effecting other components, that doesn't affect API or adopters, that has been well reviewed and well tested.
- As each Release Candidate passes, the criteria for weighing the benefit-to-risk ratio criteria gets higher and higher, and as such requires a larger number of project members to review.
- After RC1: Another committer from the project must review and vote + using Bugzilla (see below).
- After RC2: Another committer from the project and a project lead must review and vote +1 after reviewing the bug for appropriateness and risk.
Bugzilla Usage for RC Fixes
Committers can signify their approval for submitting fixes as required after RC1 by simply indicating +1 in a comment.
Bugzilla has a flag feature that is to be used for obtaining PMC approval. Enter the email address of the PMC member you are requesting approval from and submit. The PMC member is notified via email and will change the flag from '?' to '+' or '-' as appropriate.
Backporting fixes for service releases
- Bugfixes should be backported to the maintenance branch only if they match the criteria above.
- A review is required if bug fixes should be backported.
- A patch must be attached to every bug which is backported.
- If a bug has been backported for a service release, a tag of the form "srX_Y_Z" must be added to the status whiteboard field in bugzilla, where X_Y_Z is the service release number without dots (e.g. "sr0_10_1" for Gyrex 0.10.1).
- To request backporting for a service release, a tag of the form "srX_Y_Z?" can be used.
Here's how to set up your workspace for developing Gyrex.
Download and install any Eclipse distribution suitable for developing Eclipse plug-ins. We recommend trying the latest Eclipse milestone for access to the latest features in JDT and PDE.
- Eclipse Plug-in Development Environment
- Eclipse CVS Client
- Eclipse Git Client
- Download and install your preferred Java 6 JDK
- Window -> Preferences -> Java -> Installed JREs -> Add... -> Standard VM -> Next
- Set JRE home to the Java 6 JDK folder -> Finish
- Activate check-box to set Java 6 JDK as the default JDK in Eclipse
Please use the "Import Existing Projects" wizard to import all Gyrex projects from any cloned Git repository. We also provide a collection of Working Sets for your convenience. Working Sets can be imported using the AnyEdit Tools plug-in available via Eclipse Marketplace. You can find the Working Sets file in the RelEng project (
Setup Target Platform
We maintain a shared target platform. It's located in the "org.eclipse.gyrex.releng" project in the
- Open file
gyrex-dev.targetin Target Definition editor
- Wait till the target platform is resolved
- Click on link "Set as Target Platform" (located in the upper right corner of the Target Definition editor)
Import Code Settings
Optimize Project Browsing
- Switch to the Java perspective
- Open the Package Explorer view menu (little triangle icon in the upper right corner) and select "Top Level Elements -> Working Sets"
- Select all working sets ("Select All")
- We recommend sorting the working sets (check "Sort working sets")
- We also recommend hierarchical package browsing (open the Package Explorer view menu and this time select "Package Presentation -> Hierarchical")
Gyrex builds are produced on the Eclipse.org infrastructure. It's also possible to build locally. We use Maven and Tycho as our build technology which is also the standard build technology recommended by CBI.
Really Quick HowTo
# create temp directory mkdir -p /tmp/gyrex cd /tmp/gyrex # clone source into local directory git clone git://git.eclipse.org/gitroot/gyrex/gyrex-server.git . # run Maven mvn clean verify -f releng/aggregator/pom.xml
The build is modular with build-order dependencies between the modules. We do not use a single "aggregator" project that aggregates the various Git repositories. But the build process expects that repositories are cloned into the same parent directory in order to allow relative path reference in build scripts/descriptor files.
Oracle Java 7 or higher
- Oracle Java 7 JDK installed
Note: Open JDK might work too but hasn't been tested yet. Feedback is welcome!
- Environment variable JAVA_HOME set to the JDK directory
- Please ensure that the java executable is available in the PATH!
- Download from http://maven.apache.org/download.html or use your Linux distribution provided version (whether it works depends on the distribution).
Note: At the time of writing, Maven 3.0.4 has been used and tested.
- Please ensure that the mvn executable is available in the PATH!
- Please ensure that a recent version of the git executable is on the PATH!
- Depending on your operating system the setup might vary. A good setup guide is available at Set Up Git (GitHub).
Building on Windows
In general, the build should also work on Windows. However, please be aware that there might be some limitations which make it a difficult to produce good results. For example, the path length is limited. Thus, the code tree should not be cloned into an exciting deep hierarchy of folders.
Additionally, builds on Windows can't produce the correct permissions in produced archives on Linux. Therefore, good quality builds for multiple platforms should not be produced from a Windows system.
Build Environment Setup
The following instructions are required to prepare your build environment. Unless otherwise stated, they apply to Linux, Mac OS X, and Windows platforms.
To avoid getting hit by bug 365727 and to make sure your local environment configuration does not interfere with CBI build, make sure build user account does not have a Maven settings.xml
your_user@desktop:~$ ls -l ~/.m2/settings.xml ls: cannot access /home/your_user/.m2/settings.xml: No such file or directory
On Windows XP this is located in:
C:\Documents and Settings\your_user\.m2\settings.xml File should not exist, if it does delete it.
On Windows 7 this is located in:
C:\Users\your_user\.m2\settings.xml File should not exist, if it does delete it.
Maven Memory Consumption
It may be necessary to grant Maven more memory. In this case, 2GB of RAM should be enough.
If you're using Windows:
Cloning Gyrex Source Tree
All Gyrex Git repositories must be cloned into the same parent directory.
Running The Build
From the aggregator directory (directory
mvn clean verify
If you build multiple streams on the same system, you'll want to add
'-Dmaven.repo.local=/some/directory/somewhere' to the end of each
'mvn' command to avoid collisions (using a different local repo for each stream). Most casual developers won't be affected.
Alternatively, you can add the additional argument to the
MAVEN_OPTS environment variable instead.
On Windows use
Building directly from Eclipse
Instead of using Maven and Tycho, it is also possible to simply use the PDE Product Export wizard that's available in Eclipse. It requires to have the target platform configured and all Gyrex bundles imported in your workspace. The export can be started via File -> Export... -> Product configuration menu.
Note, however, that this is not recommended for producing official, production build results (to be published on Eclipse.org). But it's a convenient way of testing a product setup.
Build Process Insights
The root for all build related stuff is the Gyrex RelEng (Release Engineering) tree. It contains:
git clone git://git.eclipse.org/gitroot/gyrex/gyrex-server.git cd gyrex-server/releng
The Target Definition is used for dependency management. All dependencies are listed as p2 repositories hosted at Eclipse.
Product Configuration and Features
The Product Configuration generates the various downloadable artifacts (eg. Gyrex Server ZIP). It references Eclipse Features which define collection of bundles that are meant to be build/deployed together. There are two types of features - regular Gyrex features and dependency-only features. Gyrex features collect Gyrex bundles. Dependency-only feature collect all dependencies. This makes it easy to customize Gyrex based products with different dependencies by using custom features.
Headless Build Process
The build process is entirely based on CBI. CBI is the common build infrastructure at Eclipse.org which is based on Maven and Tycho and has some additional Maven plug-ins for signing of bundles/binaries, etc.
We use Hudson for CI. The build process on Hudson is as follows:
- Checkout all Git repos into same directory
- Invoke Maven using
- Build target platform from target definition
- Build modules referenced in POM
- Assemble Gyrex Server Product
- Produce final p2 Repo (including signing of artifacts when executed on Eclipse.org infrastructure with Maven profile)
- Publish to download.eclipse.org (when executed on Eclipse.org infrastructure with Maven profile)
Hudson Build Jobs
The integration job is used for producing I, S and R builds. Care must be taken when producing S and R builds. The results are automatically published to download.eclipse.org (when running on Eclipse.org infrastructure).
The integration job is used for producing M and R builds of the maintenance branch.