使用 Homebrew, apk, dnf, MacPorts, Nix, pacman 安装 bloaty

项目链接

• 主页
• 上游文档
• 软件包管理器页面

命令和别名

• bloaty

项目历史

The official repository was created in 2016 and describes Bloaty as a size profiler for binaries. Its README frames the core question directly: finding what makes a binary big by showing size profiles across file size and virtual-memory size.

Bloaty's implementation is unusually deep for a size-reporting CLI. The README and internals documentation describe custom ELF, DWARF, and Mach-O parsing, experimental PE/COFF and WebAssembly support, disassembly-assisted attribution, hierarchical profiles, size diffs, demangling, separate debug files, and custom data sources.

The project reached Bloaty McBloatface v1.0 in August 2018 and v1.1 in May 2020. Its README also notes that it is not an official Google product, a common marker for Google-hosted open source tools that are useful outside their original internal context.

采用历史

Bloaty's adoption followed the path of a specialist but widely useful developer tool: first-party docs and releases on GitHub, then packaging across major developer distributions. The input metadata records packages in Alpine apk, Homebrew, Fedora dnf, MacPorts, Nix, and Arch/pacman.

Its cross-platform binary-format support helped adoption outside one operating-system niche. Engineers working on native applications, toolchains, mobile code, embedded binaries, and Wasm can all use the same CLI vocabulary to ask where bytes went.

使用方式

The basic workflow is `bloaty FILE`, optionally with `-d` data sources such as sections, symbols, segments, or compile units. Bloaty reports both file size and VM size because debug data, zero-initialized data, mapped segments, and metadata can affect disk and runtime footprints differently.

A second major workflow is size diffing: users pass a current binary and a base binary separated by `--` to see exactly which sections, symbols, compile units, or other sources grew or shrank. That makes Bloaty especially attractive for release engineering and CI checks where size regressions need blameable output.

为什么软件包爱好者会关心

Bloaty is package-nerd catnip because it explains package bloat with evidence instead of vibes. It bridges low-level object-file knowledge and everyday package maintenance by making debug sections, symbol tables, compile units, `.text`, `.rodata`, Mach-O load commands, and ELF program headers visible in a table humans can act on.

The tool is also a useful counterweight to vague binary-size debates. Package maintainers can compare stripped and unstripped builds, bundled dependencies, LTO changes, debug-info packaging, static-linking choices, and architecture-specific output using one reproducible CLI.

Compared with traditional tools such as `size`, `nm`, `objdump`, `readelf`, or linker map files, Bloaty packages several views into one profiler and preserves totals so the reported rows account for the whole file. That makes it unusually practical for maintainers who need to decide whether a size increase belongs in code, debug metadata, linked libraries, compiler output, or packaging policy.

时间线

• 2016: Official Google-hosted Bloaty repository created.
• 2018: Bloaty McBloatface v1.0 released.
• 2020: Bloaty McBloatface v1.1 released.
• Package era: Bloaty appears across apk, Homebrew, dnf, MacPorts, Nix, and pacman according to package metadata.

Related projects

• `size`, `nm`, `objdump`, `readelf`, linker map files, and platform object-file tools are the traditional neighbors Bloaty complements.
• ELF, DWARF, Mach-O, PE/COFF, and WebAssembly are the binary/debug formats named by Bloaty's official docs.
• Bloaty bundles or can use dependencies including protobuf, RE2, Capstone, and pkg-config according to its README.

来源

• GitHub repository metadata via gh api repos/google/bloaty
• https://github.com/google/bloaty#readme
• https://github.com/google/bloaty/blob/main/doc/how-bloaty-works.md
• https://github.com/google/bloaty/blob/main/doc/using.md
• https://github.com/google/bloaty/releases
• source_facts.package-manager