Downloaded 11 times
Metaprogramming with javascript
- 1.
- 2. Jakarta What is Metaprogramming isthe writing of computer programs that write or manipulate other program (or themselves) as their data, or that do part of the work at compile time that otherwise be done at runtime “ writing code that writes code
- 3. Jakarta The Purpose To havesuper power! • More elegant code • Less boilerplate code • Hide the how, focus on what • Essential in framework building
- 4. Jakarta Background • My firstencounter: Prototype.js, jQuery • Experimenting with my own parsing (zparse) and angular like framework foo.js (2007) • Then created chain.js batik etc. (2008-2010) • Now maintaining Merapi and its plugins (for YesBoss / Kata.ai) I love Metaprogramming
- 5. Jakarta Uses in RealWorld • Almost every framework uses Meta Programming in some way • Domain Specific Languages • Add functionalities without littering main logic (logging, measurement, etc.)
- 6. Jakarta Two Flavours Dynamic LanguageFeatures Program can inspect itself using built-in features of the language. Code Generation Translate meta features into generated code in target language. examples: jquery spring (java) rails (ruby) merapi ;) examples: QT (C++) CoffeeScript React Babel / Typescript
- 7.
- 8.
- 9.
- 10. Jakarta Dynamic Language Features •Reflection • Getter Setter • Proxy • Decorator • Generator • DOM Manipulation • Eval
- 11. Jakarta Reflection • for (leti in object); • .prototype, getPrototypeOf() • getOwnPropertyNames(), defineProperties() • typeof etc…
- 12.
- 13.
- 14. Jakarta Getter Setter • {setprop() {}, get prop() {}} • Object.defineProperty() • Object.defineProperties()
- 15.
- 16.
- 17. Jakarta Proxy • it’s atrap! • replace object access by function invocation • create truly dynamic object • warning: slow performance • handlers: • get(), set(), has(), ownKeys(), …
- 18.
- 19.
- 20. Jakarta Generator • Function thatstops on “yield” • Returns an Iterator
- 21.
- 22.
- 23. Jakarta Decorator • @Decorator • Annotatingmethod / property • manipulate / give context
- 24.
- 25. Jakarta Performance • Code Generation(compile/transpile): • No impact performance (however depends on implementation) • Runtime (processed during initialization only) • Could lead to slower initialization (still ok in most cases) • Runtime (heavy use) • Be careful!
- 26. Jakarta Conclusion • It islike doing black magic: • know what you are doing • can have side effects if not used properly • Performance Trade-Off • Unclear Program Flow • Devs not familiar with the core can get confused
- 27.