Skip to content

Latest commit

 

History

History
106 lines (91 loc) · 6.16 KB

CONTRIBUTING.md

File metadata and controls

106 lines (91 loc) · 6.16 KB
Raw

Contributing to Skyve

Skyve is free and open source and we always are excited to receive contributions from our community. There are many ways to contribute, from writing tutorials, providing translations into new languages, improving the documentation or writing code to fix bugs or add new features to the platform itself.

Issues and Bugs

If you find a bug in the source code or a mistake in the documentation, you can help us by creating a GitHub Issue.

If you have a question on using or getting started with a Skyve project, please see our Getting Started page and watch the videos, follow the tutorial, use the skyve tag on StackOverflow or join our Slack channel.

Feature Requests

If there is something you feel is missing from Skyve, we want to hear your ideas. We are constantly trying to improve the platform to make it as versatile as possible, and many of the features in Skyve exist because our users saw a need.

Create a new GitHub Issue which describes the feature, how it would be useful to you, and how it should work. The core team will discuss with you in the issue, and if you would like to contribute you can then start working on a Pull Request.

Pull Requests

If you have contributed to a feature or addressed an issue, please create a Pull Request. Don't forget to link the Pull Request to the related issue.

If you are creating a branch to address an issue, please name it in the form of issue-##-name-of-issue.

If you have been assigned a Trello card for this feature/issue, you can link the relevant PR to the card.

Skyve Development Setup

This Skyve repository is intended for developers with existing experience with Java projects and development environments. If you are trying to start a new Skyve project, please see our Getting Started page instead.

Coding Guidelines

  • skyve-core - enough to power metadata reading and interpretation and domain generation
  • skyve-ext - Bring in other dependencies, caching, reporting, addins etc.
  • skyve-web - The web container and faces code.
  • Remember, every move you make is amplified when you make it in Skyve
  • Keep dependencies to a minimum
    • Reduce attack surface
    • Reduce technical debt
    • Reduce dependency complexity (especially transitive dependencies)
  • Keep performance in front of mind
    • Balance of memory and processor when choosing data structures and algorithms
    • Reduce temp object creation
      • Keeps GC looking sawtooth and predictable
    • Normalise data structures in memory unless performance impact
    • Minimise database access
      • Use joins over multiple queries to reduce network access (to a point)
      • Minimise Hibernate domain loading
    • Look for better than linear log(N) and avoid powers of N.
    • Increase flexibility of deployment
      • Faster starts
      • Raise usage obtained before scale out is required
  • Keep thread-safety in front of mind
    • Order of preference when in a free thread model situation
      • Use stateless objects
      • Create stateful objects per thread assignment
      • Use concurrency control mechanisms
      • Use ThreadLocal
    • Static loading and initialisation guaranteed thread-safe by JVM
    • ConcurrentHashMap - High throughput with locked buckets
      • MetaData
      • Thumbnails
      • SingletonCachedBizlets cache
      • Accesses
      • Session-scope repository delegates
    • ConcurrentReadWriteLock - Many readers, 1 blocking writer
      • LockableDynamicRepository
      • WidgetId fragments
      • Audit Logging
    • CopyOnWriteArray(List/Set) - for small collections (listeners and such)
      • GeoIPCountryCodes - Set
      • Repository delegates - List
    • ThreadLocal
      • Persistence
      • Request/Response
      • CustomerResourceServlet
      • RDBMSAuditInterceptor
    • Careful with singleton lazy instantiation - probably easier to lean on static initialisation
      • SkyveContextListener is thread safe and can set up singleton states
      • For truly lazy instantiation, use a static inner class for thead-safety
        • FileLockRepo
    • Usually avoid synchronized keyword or use with double test with volatile keyword
      • InterceptorMetaDataImpl
    • Atomics
      • View Renderers for resource references
      • Skyve’s Multi-threaded test runners
      • Session count
      • HealthServlet response
  • Ensure startup time is low - aim for 200 millis for serverless cold starts.
  • Use json configuration as a single configuration point if possible.
  • Do not hard code javaee/spring resources with annotations - Use web.xml as a second point of container configuration to remain configurable and flexible.
  • Aim for null-safety over Optional.
  • Remain agnostic of DI - no DI in Skyve (except faces)
  • Usually 1 exit from a function called result
  • Always use braces with if statement (unless trace test)
  • Watch boxing and unboxing - be explicit
    • Source of NPEs or temp object creation
  • Functional coding
    • Use to simplify (don’t overuse) - stream processing.
    • Introduces more run overhead and start delays
      • invokedynamic jvm instruction
      • Can be lots of work at the call site / capture point
      • Class engineering at first run.
      • Usually the same as or better than anonymous inner classes
    • Good for parallelism and scale-out.
  • Call getters once
  • Don't use String.format() unless you are using the formatting features - 50 times slower than StringBuilder.