Gradle Summit 2015

gradle-logo-horizontal (1)

Gradle Summit 2015 has officially concluded.

Here are some of the highlights of the most exciting things coming to Gradle :

– Build receipts.  A cloud-based, and eventually hosted solution will be released to track all sorts of recorded information about local, and remote CI builds.  Soon to be offered in beta, the idea is that this will help to diagnose common build issues, and capture build patterns within an organization.  For example, if I find my build completes successfully locally, but for some reason fails when I deploy to CI;  I may be able to diagnose and infer through the Gradle build receipt system that the last branch built successfully built in CI was built against a different version of procbuf installed on CI, or something to that effect.  Some of this information can be found in a crude fashion through a Jenkin’s job build console history, but even this doesn’t have the power to record local build events.  Another hope is that this will help save developer time by helping to record, and link to forums; documenting solutions to pain-staking one-off problems.

– New rule-based model configuration, and managed models will soon deprecate an older, flakier unmanaged project extension based model for developing Gradle plugins.  This sort of managed model paradigm will allow Gradle to “know” much more about the inputs/outputs of a Gradle plugin, and it’s set of tasks; as well as giving developers a much more elegant declarative style to developing plugins.  This kind of knowledge will help the Gradle team drive more towards task execution build performance, and build result cacheability.

– Continuous mode, released in 2.4 is cool new feature that allows for a build task, or set of tasks to run continuously in automatic response to a change of task inputs;  Giving the developer faster, background executed feedback.

The main focus for the Gradle team going forward is to continue to improve build performance, and fast feedback.  Gradle 2.4 introduced absolutely massive improvements to overall build time, but the bar has only begun to be set higher.  Distributed cache, distributed builds, and better parallelization of execution are all areas of current improvement, and enhancement.