From Source to Bytecode: Understanding the Kotlin Compilation Process Presented by Theophilus Kibet Android Engineer

What is a compiler? Kotlin Compiler Overview Resources What's Up today? Questions

A program which translates a program in a source language to a program in target language. . What is a compiler? P. 04 Bemefits OF Coroutines Source Code Compiler Target/machine code

01 P. 04 Bemefits OF Coroutines .kt files Kotlin Compiler Kotlin Compiler JVM Bytecode JavaScript LLVM Bitcode Wasm

02 P. 04 Bemefits OF Coroutines .kt Kotlin Compiler frontend backend target code

03 P. 04 Bemefits OF Coroutines .kt Kotlin Compiler backend target code lexer parser semantic analyzer tokens syntax tree syntax tree + semantic info

04 P. 04 Bemefits OF Coroutines .kt Kotlin Compiler frontend target code intermediate code generator & optimizer machine code generator & optimizer syntax tree + semantic info intermidiate representation

Consumes plain text of a program and groups together individual characters to form complete tokens. Token - has a value and a token type. Lexer P. 04 fun getSpeaker(id: String) { }

fun KtModifierKeywordToken.FUN getSpeaker KtToken.IDENTIFIER ( KtSingleValueToken.LPAR id KtToken.IDENTIFIER : KtSingleValueToken.COLON and so on..... Tokens

Consumes tokens and groups them together into complete statements and expressions, guided by the grammar. It outputs a syntax tree. Parser P. 04

A set of rules defining the language structure. ‘fun’ identifier ‘(’ parameters ‘)’ ‘{’ statements ‘}’ NB:this is just a sample rule, Kotlin has many rules for functions. Grammar

-> grammatical structure of the program. Abstract syntax tree(AST) contains only information which are needed for processing the source input. Concrete syntax tree(CST) contains all the information derived from the source input, this includes things like whitespaces and comments. NB:Kotlin Compiler produces a CST, referred as Program Structure Interface(PSI). Syntax tree

Element.FUN Token.fun Element.IDENTIFIER Element.VALUE_PARAMETER_ LIST Token.LBRACE Token.RBRACE Token.LPAR Token.RPAR Element.VALUE_PARAMETER Element.IDENTIFIER Token.COLON Element.TYPE_REFERENCE fun getSpeaker(id: String) { }

fun getSpeaker(id: String) { } Use PSIViewer plugin to view

P. 04 Bemefits OF Coroutines .kt PSI2FIR backend target code lexer parser semantic analyzer tokens PSI syntax tree + semantic info Transforms PSI to FIR.

-> Enhanced PSI designed for code resolution. -> Some optimization is done on the FIR. e.g replacing if statements with when and for loops with while Frontend Intermediate Representation Use FIR Tree plugin to view

Resolves Semantic info. imports. types Checks errors and report. Semantic Analyzer

P. 04 Bemefits OF Coroutines .kt frontend target code intermediate code generator & optimizer machine code generator & optimizer FIR intermidiate representation FIR2IR Transforms FIR to IR.

-> Enhanced FIR designed for code generation. -> More optimization is done on the IR, these are mostly architecture independent. Intermediate Representation

Constant folding. converting an expression by combining multiple constants into a single constant. Strength Reduction. converting a special case of an expensive operation into a less expensive. e.g 3^3 becomes 3*3 Optimizations on the IR

Compiler Explorer

-> Hard to navigate and run the project. -> No resources resources on getting dumps for every stage, you have to write a compiler plugin. NB: You can get help on Kotlin Slack - #compiler Challenges

18 Resources Kotlin-Compiler-Crash-Course by Amanda What Everyone Must Know About The NEW Kotlin K2 Compiler Exploring Kotlin Performance with Romain Guy Kotlin-script-parser-test Kotlin Repository Compiler Explorer Kotlin Explorer

No content

Thank @_kibetheophilus Theophilus Kibet You Contact kibettheophilus