v1.5.53

New features

• Buddy key regeneration
• AWS CodeDeploy: added 'Wait until deployment has finished' option
• Heroku: you can now deploy files processed in build actions, for example you can build assets with Gulp or Grunt, and push the source code to Heroku

Improvements

• Better error support in AWS CodeDeploy & Elastic Beanstalk when authentication data does not allow to add an S3 bucket (deployment to EB requires permissions to S3 buckets). Click here for the full list of AWS policies
• Fixed error support when too little RAM is assigned to build actions. At the moment, all build actions are executed in Docker containers with 2GB of RAM. In a couple of weeks you will be able to buy custom plans with more RAM and vCPU's
• It is now possible to define trigger conditions with API and YAML
• We've changed the way changes are fetched to the pipeline's filesystem, which solved lots of problems with .git metadata
• Short hash has been added to the execution view (until now you had to open an extra link to copy the revision of the execution)

Bugfixes

• Fixed bug with synchronizing private repositories in case authorization data used for checkout contained a question mark ?
• Fixed bug with the 'Wait for apply' action not working correctly for free accounts (caused by the global queue)
• Fixed bug with pipelines not marking actions as disabled when cloning
• Fixed bug with browsing paths in the 'Deploy to AWS S3' action (paths different than root were not being listed properly). Shout out to Roland for pointing that out!

Updates

• Docker upgraded to v17.09