CMAKE_BUILD_TYPE的Debug、Release、RelWithDebInfo和MinSizeRel分别是什么？

架构老林

2026-5-19

CMAKE_BUILD_TYPE: Explaining Debug, Release, RelWithDebInfo, and MinSizeRel

Great question! Let's break down each CMAKE_BUILD_TYPE value and clarify when this variable comes into play:

First, a quick context: CMAKE_BUILD_TYPE only matters for single-config generators (like Makefile Generators and Ninja). These tools lock in one build configuration the moment you run CMake to generate your build tree. Multi-config generators (like Visual Studio) don't use this variable—you pick your desired configuration directly when building instead.

Now, let's dive into the four core build types:

Debug:
Built for active development and debugging. This type enables full debug symbols, turns off all compiler optimizations, and keeps code unobfuscated. You can step through line-by-line with debuggers (like gdb or lldb), inspect variables in real time, and get clear error traces. The tradeoff? The resulting executable runs slowly and takes up more disk space.
Release:
Tuned for production deployment. Debug symbols are stripped entirely, and the compiler applies maximum performance optimizations (e.g., -O3 for GCC/Clang). It will inline functions, remove unused code, and prioritize speed wherever possible. You get a fast, compact executable—but debugging it is nearly impossible since all debug context is gone.
RelWithDebInfo:
A practical middle ground between Debug and Release. It enables most performance optimizations (typically -O2) while retaining debug symbols. This is perfect when you need a performant build but might have to debug occasional issues in a production-like environment. You get near-Release speed with the ability to pull meaningful debug traces if something breaks.
MinSizeRel:
Optimized specifically for minimal executable size. The compiler uses optimizations focused on shrinking binary size (e.g., -Os for GCC/Clang) rather than raw speed. Unused code, data, and debug symbols are stripped aggressively. This is ideal for environments where storage is limited—like embedded systems or apps where download size is a critical factor.