What Does a Compiler Do?

Roughly speaking, the compiler has two tasks: converting our Objective-C code into low-level code, and analyzing our code to make sure we didn’t make any obvious mistakes.

These days, Xcode ships with clang as the compiler. Wherever we write compiler, you can read it as clang. clang is the tool that takes Objective-C code, analyzes it, and transforms it into a more low-level representation that resembles assembly code: LLVM Intermediate Representation. LLVM IR is low level, and operating system independent. LLVM takes instructions and compiles them into native bytecode for the target platform. This can either be done just-in-time or at the same time as compilation.

The nice thing about having those LLVM instructions is that you can generate and run them on any platform that is supported by LLVM. For example, if you write your iOS app, it automatically runs on two very different architectures (Intel and ARM), and it’s LLVM that takes care of translating the IR code into native bytecode for those platforms.

LLVM benefits from having a three-tier architecture, which means it supports a lot of input languages (e.g. C, Objective-C, and C++, but also Haskell) in the first tier, then a shared optimizer in the second tier (which optimizes the LLVM IR), and different targets in the third tier (e.g. Intel, ARM, and PowerPC). If you want to add a language, you can focus on the first tier, and if you want to add another compilation target, you don’t have to worry too much about the input languages.