clang

Compare original and translation side by side

🇺🇸

Original

English

🇨🇳

Translation

Chinese

Clang

Purpose

用途

Guide agents through Clang-specific features: superior diagnostics, sanitizer integration, optimization remarks, static analysis, and LLVM tooling. Covers divergences from GCC and Apple/FreeBSD specifics.

指导使用者了解Clang专属特性：更出色的诊断功能、Sanitizer集成、优化备注、静态分析以及LLVM工具链。内容涵盖与GCC的差异，以及Apple/FreeBSD平台的专属细节。

Triggers

触发场景

"I want better compiler diagnostics/errors"
"How do I use clang-tidy / clang-format?"
"How do I see what the compiler optimised or didn't?"
"I'm on macOS / FreeBSD using clang"
"clang-cl for MSVC-compatible builds" — see
```
skills/compilers/msvc-cl
```
Sanitizer queries — see
```
skills/runtimes/sanitizers
```

"我想要更友好的编译器诊断/错误提示"
"如何使用clang-tidy / clang-format？"
"如何查看编译器优化了哪些内容，或是未优化哪些内容？"
"我在macOS / FreeBSD上使用clang"
"用于兼容MSVC构建的clang-cl" — 请查看
```
skills/compilers/msvc-cl
```
Sanitizer相关查询 — 请查看
```
skills/runtimes/sanitizers
```

Workflow

操作流程

1. Build mode flags (identical to GCC)

1. 构建模式选项（与GCC一致）

Clang accepts most GCC flags. Key differences:

Feature	GCC	Clang
Min size	`-Os`	`-Os` or `-Oz` (more aggressive)
Optimise only hot	—	`-fprofile-instr-use` (LLVM PGO)
Thin LTO	`-flto`	`-flto=thin` (faster)
Static analyser	`-fanalyzer`	`clang --analyze` or `clang-tidy`

Clang支持大多数GCC编译选项。核心差异如下：

特性	GCC	Clang
最小体积	`-Os`	`-Os` 或 `-Oz` （优化更激进）
仅优化热点代码	—	`-fprofile-instr-use` （LLVM PGO）
Thin LTO	`-flto`	`-flto=thin` （速度更快）
静态分析器	`-fanalyzer`	`clang --analyze` 或 `clang-tidy`

2. Clang-specific diagnostic flags

2. Clang专属诊断选项

bash

undefined

bash

undefined

Show fix-it hints inline

内联显示修复提示

clang -Wall -Wextra --show-fixits src.c

Limit error count

限制错误数量

clang -ferror-limit=5 src.c

Verbose template errors (disable elision)

详细模板错误（禁用省略）

clang -fno-elide-type src.cpp

Show tree diff for template mismatch

显示模板不匹配的树形差异

clang -fdiagnostics-show-template-tree src.cpp


Clang's diagnostics include exact range highlighting and fix-it suggestions that GCC lacks.

clang -fdiagnostics-show-template-tree src.cpp


Clang的诊断功能包含精确的范围高亮和GCC所没有的修复建议。

3. Optimization remarks

3. 优化备注

Optimization remarks let you see what Clang did or refused to do:

bash

undefined

优化备注可让你查看Clang执行了哪些优化，或是拒绝了哪些优化：

bash

undefined

Inliner decisions

内联器决策

clang -O2 -Rpass=inline src.c

Missed vectorisation

未完成向量化的情况

clang -O2 -Rpass-missed=loop-vectorize src.c

Why a loop was not vectorized

循环未被向量化的原因

clang -O2 -Rpass-analysis=loop-vectorize src.c

Save all remarks to YAML for post-processing

将所有备注保存为YAML以便后续处理

clang -O2 -fsave-optimization-record src.c

Produces src.opt.yaml

生成src.opt.yaml文件


Interpret remarks:

- `remark: foo inlined into bar` — inlining happened; good for hot paths
- `remark: loop not vectorized: loop control flow is not understood` — restructure the loop
- `remark: not vectorized: cannot prove it is safe to reorder...` — add `__restrict__` or `#pragma clang loop vectorize(assume_safety)`


解读优化备注：

- `remark: foo inlined into bar` — 已执行内联；适合热点路径
- `remark: loop not vectorized: loop control flow is not understood` — 需要重构循环
- `remark: not vectorized: cannot prove it is safe to reorder...` — 添加`__restrict__`或`#pragma clang loop vectorize(assume_safety)`

4. Static analysis

4. 静态分析

bash

undefined

bash

undefined

Built-in analyser (CSA)

内置分析器（CSA）

clang --analyze -Xanalyzer -analyzer-output=text src.c

clang-tidy (separate tool, richer checks)

clang-tidy（独立工具，检查功能更丰富）

clang-tidy src.c -- -std=c++17 -I/usr/include

Enable specific check families

启用特定检查家族

clang-tidy -checks='clang-analyzer-,modernize-,bugprone-*' src.cpp --

Apply fixits automatically

自动应用修复建议

clang-tidy -fix src.cpp --


Common `clang-tidy` check families:

- `bugprone-*`: real bugs (use-after-move, dangling, etc.)
- `clang-analyzer-*`: CSA checks (memory, null deref)
- `modernize-*`: C++11/14/17 modernisation
- `performance-*`: unnecessary copies, move candidates
- `readability-*`: naming, complexity

clang-tidy -fix src.cpp --


常见的`clang-tidy`检查家族：

- `bugprone-*`: 真实bug检查（比如use-after-move、悬空指针等）
- `clang-analyzer-*`: CSA检查（内存问题、空指针解引用等）
- `modernize-*`: C++11/14/17语法现代化检查
- `performance-*`: 不必要的拷贝、可移动候选检查
- `readability-*`: 命名规范、代码复杂度检查

5. LTO with lld

5. 借助lld实现LTO

bash

undefined

bash

undefined

Full LTO

完整LTO

clang -O2 -flto -fuse-ld=lld src.c -o prog

Thin LTO (faster link, nearly same quality)

Thin LTO（链接速度更快，质量几乎一致）

clang -O2 -flto=thin -fuse-ld=lld src.c -o prog

Check lld is available

检查lld是否可用

clang -fuse-ld=lld -Wl,--version 2>&1 | head -1


For large projects, ThinLTO is preferred: link times 5-10x faster than full LTO with comparable code quality.

clang -fuse-ld=lld -Wl,--version 2>&1 | head -1


对于大型项目，推荐使用ThinLTO：链接速度比完整LTO快5-10倍，同时代码质量相当。

6. PGO (LLVM instrumentation)

6. PGO（LLVM插桩式）

bash

undefined

bash

undefined

Step 1: instrument

步骤1：生成插桩程序

clang -O2 -fprofile-instr-generate prog.c -o prog_inst

Step 2: run with representative input

步骤2：使用代表性输入运行程序

./prog_inst < workload.input

Generates default.profraw

生成default.profraw文件

Step 3: merge profiles

步骤3：合并性能数据

llvm-profdata merge -output=prog.profdata default.profraw

Step 4: use profile

步骤4：使用性能数据编译

clang -O2 -fprofile-instr-use=prog.profdata prog.c -o prog


AutoFDO (sampling-based, less intrusive): collect with `perf`, convert with `create_llvm_prof`, use with `-fprofile-sample-use`. See `skills/profilers/linux-perf`.

clang -O2 -fprofile-instr-use=prog.profdata prog.c -o prog


AutoFDO（基于采样，侵入性更低）：使用`perf`采集数据，通过`create_llvm_prof`转换，配合`-fprofile-sample-use`使用。详情请查看`skills/profilers/linux-perf`。

7. GCC compatibility

7. GCC兼容性

Clang is intentionally GCC-compatible for driver flags. Key differences:

Clang does not support all GCC-specific attributes; check with
```
__has_attribute(foo)
```
```
-Weverything
```
enables all Clang warnings (no GCC equivalent); too noisy for production, useful for one-off audits
Some GCC intrinsics need
```
#include <x86intrin.h>
```
on Clang too
```
__int128
```
is supported;
```
__float128
```
requires
```
-lquadmath
```
on some targets

Clang在驱动程序选项上刻意保持与GCC兼容。核心差异如下：

Clang不支持所有GCC专属属性；可通过
```
__has_attribute(foo)
```
检查
```
-Weverything
```
启用所有Clang警告（GCC无对应选项）；对于生产环境过于嘈杂，适合一次性审计
部分GCC内建函数在Clang上也需要
```
#include <x86intrin.h>
```
支持
```
__int128
```
；
```
__float128
```
在部分目标平台上需要
```
-lquadmath
```

8. macOS specifics

8. macOS平台专属细节

On macOS,

clang

is the system compiler (Apple LLVM). Key points:

```
ld64
```
is the default linker;
```
lld
```
requires explicit
```
-fuse-ld=lld
```
and Homebrew LLVM
Use
```
-mmacosx-version-min=X.Y
```
to set deployment target
Sanitizers on macOS use
```
DYLD_INSERT_LIBRARIES
```
; do not strip the binary
```
xcrun clang
```
resolves to the Xcode toolchain clang

For flag reference, see references/flags.md. For clang-tidy config examples, see references/clang-tidy.md.

在macOS上，

clang

是系统默认编译器（Apple LLVM）。核心要点：

默认链接器是
```
ld64
```
；使用
```
lld
```
需要显式添加
```
-fuse-ld=lld
```
，且需通过Homebrew安装LLVM
使用
```
-mmacosx-version-min=X.Y
```
设置部署目标版本
macOS上的Sanitizers使用
```
DYLD_INSERT_LIBRARIES
```
；请勿剥离二进制文件
```
xcrun clang
```
会解析为Xcode工具链中的clang

编译选项参考请查看references/flags.md。 clang-tidy配置示例请查看references/clang-tidy.md。

clang

Original

Translation

Clang

Clang

Purpose

用途

Triggers

触发场景

Workflow

操作流程

1. Build mode flags (identical to GCC)

1. 构建模式选项（与GCC一致）

2. Clang-specific diagnostic flags

2. Clang专属诊断选项

Show fix-it hints inline

内联显示修复提示

Limit error count

限制错误数量

Verbose template errors (disable elision)

详细模板错误（禁用省略）

Show tree diff for template mismatch

显示模板不匹配的树形差异

3. Optimization remarks

3. 优化备注

Inliner decisions

内联器决策

Missed vectorisation

未完成向量化的情况

Why a loop was not vectorized

循环未被向量化的原因

Save all remarks to YAML for post-processing

将所有备注保存为YAML以便后续处理

Produces src.opt.yaml

生成src.opt.yaml文件

4. Static analysis

4. 静态分析

Built-in analyser (CSA)

内置分析器（CSA）

clang-tidy (separate tool, richer checks)

clang-tidy（独立工具，检查功能更丰富）

Enable specific check families

启用特定检查家族

Apply fixits automatically

自动应用修复建议

5. LTO with lld

5. 借助lld实现LTO

Full LTO

完整LTO

Thin LTO (faster link, nearly same quality)

Thin LTO（链接速度更快，质量几乎一致）

Check lld is available

检查lld是否可用

6. PGO (LLVM instrumentation)

6. PGO（LLVM插桩式）

Step 1: instrument

步骤1：生成插桩程序

Step 2: run with representative input

步骤2：使用代表性输入运行程序

Generates default.profraw

生成default.profraw文件

Step 3: merge profiles

步骤3：合并性能数据

Step 4: use profile

步骤4：使用性能数据编译

7. GCC compatibility

7. GCC兼容性

8. macOS specifics

8. macOS平台专属细节

Related skills

相关技能