astral-tokio-tar-0.5.1/.cargo_vcs_info.json0000644000000001360000000000100142340ustar { "git": { "sha1": "86d43db84e48cce481d6e906b5f17f331f621aee" }, "path_in_vcs": "" }astral-tokio-tar-0.5.1/.github/workflows/ci.yml000064400000000000000000000013321046102023000175360ustar 00000000000000name: CI on: push: branches: [main] pull_request: workflow_dispatch: env: CARGO_TERM_COLOR: always jobs: ci: runs-on: ${{ matrix.os }} strategy: matrix: os: [ubuntu-latest, windows-latest, macos-latest] name: ${{ matrix.os }} steps: - uses: actions/checkout@v2 - uses: Swatinem/rust-cache@v2 - run: cargo fmt --all -- --check - uses: taiki-e/install-action@v2 with: tool: cargo-hack - run: cargo hack check --all --ignore-private --each-feature --no-dev-deps - run: cargo check --all --all-targets --all-features - run: cargo test - run: cargo clippy --workspace --all-targets --all-features --locked -- -D warnings astral-tokio-tar-0.5.1/.github/workflows/publish.yml000064400000000000000000000005601046102023000206130ustar 00000000000000name: Publish to crates.io on: push: tags: - 'v*' env: CARGO_TERM_COLOR: always jobs: publish: runs-on: ubuntu-latest environment: name: crates-io steps: - uses: actions/checkout@v4 - name: Publish to crates.io run: cargo publish env: CARGO_REGISTRY_TOKEN: ${{ secrets.CARGO_REGISTRY_TOKEN }} astral-tokio-tar-0.5.1/.gitignore000064400000000000000000000000101046102023000150030ustar 00000000000000/target astral-tokio-tar-0.5.1/CHANGELOG.md000064400000000000000000000017171046102023000146430ustar 00000000000000# Changelog ## 0.5.1 * Add test to reproduce issue in `impl Stream for Entries` causing filename truncation by @charliermarsh in https://github.com/astral-sh/tokio-tar/pull/41 * Avoid truncation during pending reads by @charliermarsh in https://github.com/astral-sh/tokio-tar/pull/40 ## 0.5.0 * Setting `preserve_permissions` to `false` will avoid setting _any_ permissions on extracted files. In [`alexcrichton/tar-rs`](https://github.com/alexcrichton/tar-rs), setting `preserve_permissions` to `false` will still set read, write, and execute permissions on extracted files, but will avoid setting extended permissions (e.g., `setuid`, `setgid`, and `sticky` bits). * Avoid creating directories outside the unpack target (see: [`alexcrichton/tar-rs#259`](https://github.com/alexcrichton/tar-rs/pull/259)). * Added `unpack_in_raw` which memoizes the set of validated paths (and assumes a pre-canonicalized) unpack target to avoid redundant filesystem operations. astral-tokio-tar-0.5.1/Cargo.lock0000644000000321040000000000100122070ustar # This file is automatically @generated by Cargo. # It is not intended for manual editing. version = 4 [[package]] name = "addr2line" version = "0.24.2" source = "registry+https://github.com/rust-lang/crates.io-index" checksum = "dfbe277e56a376000877090da837660b4427aad530e3028d44e0bffe4f89a1c1" dependencies = [ "gimli", ] [[package]] name = "adler2" version = "2.0.0" source = "registry+https://github.com/rust-lang/crates.io-index" checksum = "512761e0bb2578dd7380c6baaa0f4ce03e84f95e960231d1dec8bf4d7d6e2627" [[package]] name = "astral-tokio-tar" version = "0.5.1" dependencies = [ "filetime", "futures-core", "libc", "portable-atomic", "rustc-hash", "tempfile", "tokio", "tokio-stream", "xattr", ] [[package]] name = "autocfg" version = "1.4.0" source = "registry+https://github.com/rust-lang/crates.io-index" checksum = "ace50bade8e6234aa140d9a2f552bbee1db4d353f69b8217bc503490fc1a9f26" [[package]] name = "backtrace" version = "0.3.74" source = "registry+https://github.com/rust-lang/crates.io-index" checksum = "8d82cb332cdfaed17ae235a638438ac4d4839913cc2af585c3c6746e8f8bee1a" dependencies = [ "addr2line", "cfg-if", "libc", "miniz_oxide", "object", "rustc-demangle", "windows-targets", ] [[package]] name = "bitflags" version = "2.8.0" source = "registry+https://github.com/rust-lang/crates.io-index" checksum = "8f68f53c83ab957f72c32642f3868eec03eb974d1fb82e453128456482613d36" [[package]] name = "bytes" version = "1.10.0" source = "registry+https://github.com/rust-lang/crates.io-index" checksum = "f61dac84819c6588b558454b194026eb1f09c293b9036ae9b159e74e73ab6cf9" [[package]] name = "cfg-if" version = "1.0.0" source = "registry+https://github.com/rust-lang/crates.io-index" checksum = "baf1de4339761588bc0619e3cbc0120ee582ebb74b53b4efbf79117bd2da40fd" [[package]] name = "errno" version = "0.3.10" source = "registry+https://github.com/rust-lang/crates.io-index" checksum = "33d852cb9b869c2a9b3df2f71a3074817f01e1844f839a144f5fcef059a4eb5d" dependencies = [ "libc", "windows-sys 0.59.0", ] [[package]] name = "fastrand" version = "2.3.0" source = "registry+https://github.com/rust-lang/crates.io-index" checksum = "37909eebbb50d72f9059c3b6d82c0463f2ff062c9e95845c43a6c9c0355411be" [[package]] name = "filetime" version = "0.2.25" source = "registry+https://github.com/rust-lang/crates.io-index" checksum = "35c0522e981e68cbfa8c3f978441a5f34b30b96e146b33cd3359176b50fe8586" dependencies = [ "cfg-if", "libc", "libredox", "windows-sys 0.59.0", ] [[package]] name = "futures-core" version = "0.3.31" source = "registry+https://github.com/rust-lang/crates.io-index" checksum = "05f29059c0c2090612e8d742178b0580d2dc940c837851ad723096f87af6663e" [[package]] name = "getrandom" version = "0.3.1" source = "registry+https://github.com/rust-lang/crates.io-index" checksum = "43a49c392881ce6d5c3b8cb70f98717b7c07aabbdff06687b9030dbfbe2725f8" dependencies = [ "cfg-if", "libc", "wasi 0.13.3+wasi-0.2.2", "windows-targets", ] [[package]] name = "gimli" version = "0.31.1" source = "registry+https://github.com/rust-lang/crates.io-index" checksum = "07e28edb80900c19c28f1072f2e8aeca7fa06b23cd4169cefe1af5aa3260783f" [[package]] name = "libc" version = "0.2.169" source = "registry+https://github.com/rust-lang/crates.io-index" checksum = "b5aba8db14291edd000dfcc4d620c7ebfb122c613afb886ca8803fa4e128a20a" [[package]] name = "libredox" version = "0.1.3" source = "registry+https://github.com/rust-lang/crates.io-index" checksum = "c0ff37bd590ca25063e35af745c343cb7a0271906fb7b37e4813e8f79f00268d" dependencies = [ "bitflags", "libc", "redox_syscall", ] [[package]] name = "linux-raw-sys" version = "0.4.15" source = "registry+https://github.com/rust-lang/crates.io-index" checksum = "d26c52dbd32dccf2d10cac7725f8eae5296885fb5703b261f7d0a0739ec807ab" [[package]] name = "lock_api" version = "0.4.12" source = "registry+https://github.com/rust-lang/crates.io-index" checksum = "07af8b9cdd281b7915f413fa73f29ebd5d55d0d3f0155584dade1ff18cea1b17" dependencies = [ "autocfg", "scopeguard", ] [[package]] name = "memchr" version = "2.7.4" source = "registry+https://github.com/rust-lang/crates.io-index" checksum = "78ca9ab1a0babb1e7d5695e3530886289c18cf2f87ec19a575a0abdce112e3a3" [[package]] name = "miniz_oxide" version = "0.8.3" source = "registry+https://github.com/rust-lang/crates.io-index" checksum = "b8402cab7aefae129c6977bb0ff1b8fd9a04eb5b51efc50a70bea51cda0c7924" dependencies = [ "adler2", ] [[package]] name = "mio" version = "1.0.3" source = "registry+https://github.com/rust-lang/crates.io-index" checksum = "2886843bf800fba2e3377cff24abf6379b4c4d5c6681eaf9ea5b0d15090450bd" dependencies = [ "libc", "wasi 0.11.0+wasi-snapshot-preview1", "windows-sys 0.52.0", ] [[package]] name = "object" version = "0.36.7" source = "registry+https://github.com/rust-lang/crates.io-index" checksum = "62948e14d923ea95ea2c7c86c71013138b66525b86bdc08d2dcc262bdb497b87" dependencies = [ "memchr", ] [[package]] name = "once_cell" version = "1.20.2" source = "registry+https://github.com/rust-lang/crates.io-index" checksum = "1261fe7e33c73b354eab43b1273a57c8f967d0391e80353e51f764ac02cf6775" [[package]] name = "parking_lot" version = "0.12.3" source = "registry+https://github.com/rust-lang/crates.io-index" checksum = "f1bf18183cf54e8d6059647fc3063646a1801cf30896933ec2311622cc4b9a27" dependencies = [ "lock_api", "parking_lot_core", ] [[package]] name = "parking_lot_core" version = "0.9.10" source = "registry+https://github.com/rust-lang/crates.io-index" checksum = "1e401f977ab385c9e4e3ab30627d6f26d00e2c73eef317493c4ec6d468726cf8" dependencies = [ "cfg-if", "libc", "redox_syscall", "smallvec", "windows-targets", ] [[package]] name = "pin-project-lite" version = "0.2.16" source = "registry+https://github.com/rust-lang/crates.io-index" checksum = "3b3cff922bd51709b605d9ead9aa71031d81447142d828eb4a6eba76fe619f9b" [[package]] name = "portable-atomic" version = "1.10.0" source = "registry+https://github.com/rust-lang/crates.io-index" checksum = "280dc24453071f1b63954171985a0b0d30058d287960968b9b2aca264c8d4ee6" [[package]] name = "proc-macro2" version = "1.0.93" source = "registry+https://github.com/rust-lang/crates.io-index" checksum = "60946a68e5f9d28b0dc1c21bb8a97ee7d018a8b322fa57838ba31cc878e22d99" dependencies = [ "unicode-ident", ] [[package]] name = "quote" version = "1.0.38" source = "registry+https://github.com/rust-lang/crates.io-index" checksum = "0e4dccaaaf89514f546c693ddc140f729f958c247918a13380cccc6078391acc" dependencies = [ "proc-macro2", ] [[package]] name = "redox_syscall" version = "0.5.8" source = "registry+https://github.com/rust-lang/crates.io-index" checksum = "03a862b389f93e68874fbf580b9de08dd02facb9a788ebadaf4a3fd33cf58834" dependencies = [ "bitflags", ] [[package]] name = "rustc-demangle" version = "0.1.24" source = "registry+https://github.com/rust-lang/crates.io-index" checksum = "719b953e2095829ee67db738b3bfa9fa368c94900df327b3f07fe6e794d2fe1f" [[package]] name = "rustc-hash" version = "2.1.0" source = "registry+https://github.com/rust-lang/crates.io-index" checksum = "c7fb8039b3032c191086b10f11f319a6e99e1e82889c5cc6046f515c9db1d497" [[package]] name = "rustix" version = "0.38.44" source = "registry+https://github.com/rust-lang/crates.io-index" checksum = "fdb5bc1ae2baa591800df16c9ca78619bf65c0488b41b96ccec5d11220d8c154" dependencies = [ "bitflags", "errno", "libc", "linux-raw-sys", "windows-sys 0.59.0", ] [[package]] name = "scopeguard" version = "1.2.0" source = "registry+https://github.com/rust-lang/crates.io-index" checksum = "94143f37725109f92c262ed2cf5e59bce7498c01bcc1502d7b9afe439a4e9f49" [[package]] name = "signal-hook-registry" version = "1.4.2" source = "registry+https://github.com/rust-lang/crates.io-index" checksum = "a9e9e0b4211b72e7b8b6e85c807d36c212bdb33ea8587f7569562a84df5465b1" dependencies = [ "libc", ] [[package]] name = "smallvec" version = "1.13.2" source = "registry+https://github.com/rust-lang/crates.io-index" checksum = "3c5e1a9a646d36c3599cd173a41282daf47c44583ad367b8e6837255952e5c67" [[package]] name = "socket2" version = "0.5.8" source = "registry+https://github.com/rust-lang/crates.io-index" checksum = "c970269d99b64e60ec3bd6ad27270092a5394c4e309314b18ae3fe575695fbe8" dependencies = [ "libc", "windows-sys 0.52.0", ] [[package]] name = "syn" version = "2.0.98" source = "registry+https://github.com/rust-lang/crates.io-index" checksum = "36147f1a48ae0ec2b5b3bc5b537d267457555a10dc06f3dbc8cb11ba3006d3b1" dependencies = [ "proc-macro2", "quote", "unicode-ident", ] [[package]] name = "tempfile" version = "3.16.0" source = "registry+https://github.com/rust-lang/crates.io-index" checksum = "38c246215d7d24f48ae091a2902398798e05d978b24315d6efbc00ede9a8bb91" dependencies = [ "cfg-if", "fastrand", "getrandom", "once_cell", "rustix", "windows-sys 0.59.0", ] [[package]] name = "tokio" version = "1.43.0" source = "registry+https://github.com/rust-lang/crates.io-index" checksum = "3d61fa4ffa3de412bfea335c6ecff681de2b609ba3c77ef3e00e521813a9ed9e" dependencies = [ "backtrace", "bytes", "libc", "mio", "parking_lot", "pin-project-lite", "signal-hook-registry", "socket2", "tokio-macros", "windows-sys 0.52.0", ] [[package]] name = "tokio-macros" version = "2.5.0" source = "registry+https://github.com/rust-lang/crates.io-index" checksum = "6e06d43f1345a3bcd39f6a56dbb7dcab2ba47e68e8ac134855e7e2bdbaf8cab8" dependencies = [ "proc-macro2", "quote", "syn", ] [[package]] name = "tokio-stream" version = "0.1.17" source = "registry+https://github.com/rust-lang/crates.io-index" checksum = "eca58d7bba4a75707817a2c44174253f9236b2d5fbd055602e9d5c07c139a047" dependencies = [ "futures-core", "pin-project-lite", "tokio", ] [[package]] name = "unicode-ident" version = "1.0.16" source = "registry+https://github.com/rust-lang/crates.io-index" checksum = "a210d160f08b701c8721ba1c726c11662f877ea6b7094007e1ca9a1041945034" [[package]] name = "wasi" version = "0.11.0+wasi-snapshot-preview1" source = "registry+https://github.com/rust-lang/crates.io-index" checksum = "9c8d87e72b64a3b4db28d11ce29237c246188f4f51057d65a7eab63b7987e423" [[package]] name = "wasi" version = "0.13.3+wasi-0.2.2" source = "registry+https://github.com/rust-lang/crates.io-index" checksum = "26816d2e1a4a36a2940b96c5296ce403917633dff8f3440e9b236ed6f6bacad2" dependencies = [ "wit-bindgen-rt", ] [[package]] name = "windows-sys" version = "0.52.0" source = "registry+https://github.com/rust-lang/crates.io-index" checksum = "282be5f36a8ce781fad8c8ae18fa3f9beff57ec1b52cb3de0789201425d9a33d" dependencies = [ "windows-targets", ] [[package]] name = "windows-sys" version = "0.59.0" source = "registry+https://github.com/rust-lang/crates.io-index" checksum = "1e38bc4d79ed67fd075bcc251a1c39b32a1776bbe92e5bef1f0bf1f8c531853b" dependencies = [ "windows-targets", ] [[package]] name = "windows-targets" version = "0.52.6" source = "registry+https://github.com/rust-lang/crates.io-index" checksum = "9b724f72796e036ab90c1021d4780d4d3d648aca59e491e6b98e725b84e99973" dependencies = [ "windows_aarch64_gnullvm", "windows_aarch64_msvc", "windows_i686_gnu", "windows_i686_gnullvm", "windows_i686_msvc", "windows_x86_64_gnu", "windows_x86_64_gnullvm", "windows_x86_64_msvc", ] [[package]] name = "windows_aarch64_gnullvm" version = "0.52.6" source = "registry+https://github.com/rust-lang/crates.io-index" checksum = "32a4622180e7a0ec044bb555404c800bc9fd9ec262ec147edd5989ccd0c02cd3" [[package]] name = "windows_aarch64_msvc" version = "0.52.6" source = "registry+https://github.com/rust-lang/crates.io-index" checksum = "09ec2a7bb152e2252b53fa7803150007879548bc709c039df7627cabbd05d469" [[package]] name = "windows_i686_gnu" version = "0.52.6" source = "registry+https://github.com/rust-lang/crates.io-index" checksum = "8e9b5ad5ab802e97eb8e295ac6720e509ee4c243f69d781394014ebfe8bbfa0b" [[package]] name = "windows_i686_gnullvm" version = "0.52.6" source = "registry+https://github.com/rust-lang/crates.io-index" checksum = "0eee52d38c090b3caa76c563b86c3a4bd71ef1a819287c19d586d7334ae8ed66" [[package]] name = "windows_i686_msvc" version = "0.52.6" source = "registry+https://github.com/rust-lang/crates.io-index" checksum = "240948bc05c5e7c6dabba28bf89d89ffce3e303022809e73deaefe4f6ec56c66" [[package]] name = "windows_x86_64_gnu" version = "0.52.6" source = "registry+https://github.com/rust-lang/crates.io-index" checksum = "147a5c80aabfbf0c7d901cb5895d1de30ef2907eb21fbbab29ca94c5b08b1a78" [[package]] name = "windows_x86_64_gnullvm" version = "0.52.6" source = "registry+https://github.com/rust-lang/crates.io-index" checksum = "24d5b23dc417412679681396f2b49f3de8c1473deb516bd34410872eff51ed0d" [[package]] name = "windows_x86_64_msvc" version = "0.52.6" source = "registry+https://github.com/rust-lang/crates.io-index" checksum = "589f6da84c646204747d1270a2a5661ea66ed1cced2631d546fdfb155959f9ec" [[package]] name = "wit-bindgen-rt" version = "0.33.0" source = "registry+https://github.com/rust-lang/crates.io-index" checksum = "3268f3d866458b787f390cf61f4bbb563b922d091359f9608842999eaee3943c" dependencies = [ "bitflags", ] [[package]] name = "xattr" version = "1.4.0" source = "registry+https://github.com/rust-lang/crates.io-index" checksum = "e105d177a3871454f754b33bb0ee637ecaaac997446375fd3e5d43a2ed00c909" dependencies = [ "libc", "linux-raw-sys", "rustix", ] astral-tokio-tar-0.5.1/Cargo.toml0000644000000045420000000000100122370ustar # THIS FILE IS AUTOMATICALLY GENERATED BY CARGO # # When uploading crates to the registry Cargo will automatically # "normalize" Cargo.toml files for maximal compatibility # with all versions of Cargo and also rewrite `path` dependencies # to registry (e.g., crates.io) dependencies. # # If you are reading this file be aware that the original Cargo.toml # will likely look very different (and much more reasonable). # See Cargo.toml.orig for the original contents. [package] edition = "2021" name = "astral-tokio-tar" version = "0.5.1" authors = [ "Alex Crichton ", "dignifiedquire ", "Artem Vorotnikov ", "Aiden McClelland ", ] build = false exclude = ["tests/archives/*"] autolib = false autobins = false autoexamples = false autotests = false autobenches = false description = """ A Rust implementation of an async TAR file reader and writer. This library does not currently handle compression, but it is abstract over all I/O readers and writers. Additionally, great lengths are taken to ensure that the entire contents are never required to be entirely resident in memory all at once. """ homepage = "https://github.com/astral-sh/tokio-tar" documentation = "https://docs.rs/tokio-tar" readme = "README.md" keywords = [ "tar", "tarfile", "encoding", ] license = "MIT OR Apache-2.0" repository = "https://github.com/astral-sh/tokio-tar" [lib] name = "tokio_tar" path = "src/lib.rs" [[example]] name = "extract_file" path = "examples/extract_file.rs" [[example]] name = "list" path = "examples/list.rs" [[example]] name = "raw_list" path = "examples/raw_list.rs" [[example]] name = "write" path = "examples/write.rs" [[test]] name = "all" path = "tests/all.rs" [[test]] name = "entry" path = "tests/entry.rs" [dependencies.filetime] version = "0.2" [dependencies.futures-core] version = "0.3" [dependencies.portable-atomic] version = "1" [dependencies.rustc-hash] version = "2.1.0" [dependencies.tokio] version = "1" features = [ "fs", "io-util", "rt", ] [dependencies.tokio-stream] version = "0.1" [dev-dependencies.tempfile] version = "3" [dev-dependencies.tokio] version = "1" features = ["full"] [features] default = ["xattr"] [target."cfg(unix)".dependencies.libc] version = "0.2" [target."cfg(unix)".dependencies.xattr] version = "1.0" optional = true astral-tokio-tar-0.5.1/Cargo.toml.orig000064400000000000000000000023761046102023000157230ustar 00000000000000[package] name = "astral-tokio-tar" version = "0.5.1" authors = [ "Alex Crichton ", "dignifiedquire ", "Artem Vorotnikov ", "Aiden McClelland ", ] homepage = "https://github.com/astral-sh/tokio-tar" repository = "https://github.com/astral-sh/tokio-tar" documentation = "https://docs.rs/tokio-tar" license = "MIT OR Apache-2.0" keywords = ["tar", "tarfile", "encoding"] readme = "README.md" edition = "2021" exclude = ["tests/archives/*"] description = """ A Rust implementation of an async TAR file reader and writer. This library does not currently handle compression, but it is abstract over all I/O readers and writers. Additionally, great lengths are taken to ensure that the entire contents are never required to be entirely resident in memory all at once. """ [dependencies] filetime = "0.2" futures-core = "0.3" portable-atomic = "1" rustc-hash = "2.1.0" tokio = { version = "1", features = ["fs", "io-util", "rt"] } tokio-stream = "0.1" [dev-dependencies] tempfile = "3" tokio = { version = "1", features = ["full"] } [target."cfg(unix)".dependencies] xattr = { version = "1.0", optional = true } libc = "0.2" [features] default = ["xattr"] [lib] name = "tokio_tar" astral-tokio-tar-0.5.1/LICENSE-APACHE000064400000000000000000000251371046102023000147600ustar 00000000000000 Apache License Version 2.0, January 2004 http://www.apache.org/licenses/ TERMS AND CONDITIONS FOR USE, REPRODUCTION, AND DISTRIBUTION 1. Definitions. "License" shall mean the terms and conditions for use, reproduction, and distribution as defined by Sections 1 through 9 of this document. "Licensor" shall mean the copyright owner or entity authorized by the copyright owner that is granting the License. "Legal Entity" shall mean the union of the acting entity and all other entities that control, are controlled by, or are under common control with that entity. For the purposes of this definition, "control" means (i) the power, direct or indirect, to cause the direction or management of such entity, whether by contract or otherwise, or (ii) ownership of fifty percent (50%) or more of the outstanding shares, or (iii) beneficial ownership of such entity. "You" (or "Your") shall mean an individual or Legal Entity exercising permissions granted by this License. "Source" form shall mean the preferred form for making modifications, including but not limited to software source code, documentation source, and configuration files. "Object" form shall mean any form resulting from mechanical transformation or translation of a Source form, including but not limited to compiled object code, generated documentation, and conversions to other media types. "Work" shall mean the work of authorship, whether in Source or Object form, made available under the License, as indicated by a copyright notice that is included in or attached to the work (an example is provided in the Appendix below). "Derivative Works" shall mean any work, whether in Source or Object form, that is based on (or derived from) the Work and for which the editorial revisions, annotations, elaborations, or other modifications represent, as a whole, an original work of authorship. For the purposes of this License, Derivative Works shall not include works that remain separable from, or merely link (or bind by name) to the interfaces of, the Work and Derivative Works thereof. "Contribution" shall mean any work of authorship, including the original version of the Work and any modifications or additions to that Work or Derivative Works thereof, that is intentionally submitted to Licensor for inclusion in the Work by the copyright owner or by an individual or Legal Entity authorized to submit on behalf of the copyright owner. For the purposes of this definition, "submitted" means any form of electronic, verbal, or written communication sent to the Licensor or its representatives, including but not limited to communication on electronic mailing lists, source code control systems, and issue tracking systems that are managed by, or on behalf of, the Licensor for the purpose of discussing and improving the Work, but excluding communication that is conspicuously marked or otherwise designated in writing by the copyright owner as "Not a Contribution." "Contributor" shall mean Licensor and any individual or Legal Entity on behalf of whom a Contribution has been received by Licensor and subsequently incorporated within the Work. 2. Grant of Copyright License. Subject to the terms and conditions of this License, each Contributor hereby grants to You a perpetual, worldwide, non-exclusive, no-charge, royalty-free, irrevocable copyright license to reproduce, prepare Derivative Works of, publicly display, publicly perform, sublicense, and distribute the Work and such Derivative Works in Source or Object form. 3. Grant of Patent License. Subject to the terms and conditions of this License, each Contributor hereby grants to You a perpetual, worldwide, non-exclusive, no-charge, royalty-free, irrevocable (except as stated in this section) patent license to make, have made, use, offer to sell, sell, import, and otherwise transfer the Work, where such license applies only to those patent claims licensable by such Contributor that are necessarily infringed by their Contribution(s) alone or by combination of their Contribution(s) with the Work to which such Contribution(s) was submitted. If You institute patent litigation against any entity (including a cross-claim or counterclaim in a lawsuit) alleging that the Work or a Contribution incorporated within the Work constitutes direct or contributory patent infringement, then any patent licenses granted to You under this License for that Work shall terminate as of the date such litigation is filed. 4. Redistribution. You may reproduce and distribute copies of the Work or Derivative Works thereof in any medium, with or without modifications, and in Source or Object form, provided that You meet the following conditions: (a) You must give any other recipients of the Work or Derivative Works a copy of this License; and (b) You must cause any modified files to carry prominent notices stating that You changed the files; and (c) You must retain, in the Source form of any Derivative Works that You distribute, all copyright, patent, trademark, and attribution notices from the Source form of the Work, excluding those notices that do not pertain to any part of the Derivative Works; and (d) If the Work includes a "NOTICE" text file as part of its distribution, then any Derivative Works that You distribute must include a readable copy of the attribution notices contained within such NOTICE file, excluding those notices that do not pertain to any part of the Derivative Works, in at least one of the following places: within a NOTICE text file distributed as part of the Derivative Works; within the Source form or documentation, if provided along with the Derivative Works; or, within a display generated by the Derivative Works, if and wherever such third-party notices normally appear. The contents of the NOTICE file are for informational purposes only and do not modify the License. You may add Your own attribution notices within Derivative Works that You distribute, alongside or as an addendum to the NOTICE text from the Work, provided that such additional attribution notices cannot be construed as modifying the License. You may add Your own copyright statement to Your modifications and may provide additional or different license terms and conditions for use, reproduction, or distribution of Your modifications, or for any such Derivative Works as a whole, provided Your use, reproduction, and distribution of the Work otherwise complies with the conditions stated in this License. 5. Submission of Contributions. Unless You explicitly state otherwise, any Contribution intentionally submitted for inclusion in the Work by You to the Licensor shall be under the terms and conditions of this License, without any additional terms or conditions. Notwithstanding the above, nothing herein shall supersede or modify the terms of any separate license agreement you may have executed with Licensor regarding such Contributions. 6. Trademarks. This License does not grant permission to use the trade names, trademarks, service marks, or product names of the Licensor, except as required for reasonable and customary use in describing the origin of the Work and reproducing the content of the NOTICE file. 7. Disclaimer of Warranty. Unless required by applicable law or agreed to in writing, Licensor provides the Work (and each Contributor provides its Contributions) on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied, including, without limitation, any warranties or conditions of TITLE, NON-INFRINGEMENT, MERCHANTABILITY, or FITNESS FOR A PARTICULAR PURPOSE. You are solely responsible for determining the appropriateness of using or redistributing the Work and assume any risks associated with Your exercise of permissions under this License. 8. Limitation of Liability. In no event and under no legal theory, whether in tort (including negligence), contract, or otherwise, unless required by applicable law (such as deliberate and grossly negligent acts) or agreed to in writing, shall any Contributor be liable to You for damages, including any direct, indirect, special, incidental, or consequential damages of any character arising as a result of this License or out of the use or inability to use the Work (including but not limited to damages for loss of goodwill, work stoppage, computer failure or malfunction, or any and all other commercial damages or losses), even if such Contributor has been advised of the possibility of such damages. 9. Accepting Warranty or Additional Liability. While redistributing the Work or Derivative Works thereof, You may choose to offer, and charge a fee for, acceptance of support, warranty, indemnity, or other liability obligations and/or rights consistent with this License. However, in accepting such obligations, You may act only on Your own behalf and on Your sole responsibility, not on behalf of any other Contributor, and only if You agree to indemnify, defend, and hold each Contributor harmless for any liability incurred by, or claims asserted against, such Contributor by reason of your accepting any such warranty or additional liability. END OF TERMS AND CONDITIONS APPENDIX: How to apply the Apache License to your work. To apply the Apache License to your work, attach the following boilerplate notice, with the fields enclosed by brackets "[]" replaced with your own identifying information. (Don't include the brackets!) The text should be enclosed in the appropriate comment syntax for the file format. We also recommend that a file or class name and description of purpose be included on the same "printed page" as the copyright notice for easier identification within third-party archives. Copyright [yyyy] [name of copyright owner] Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with the License. You may obtain a copy of the License at http://www.apache.org/licenses/LICENSE-2.0 Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the specific language governing permissions and limitations under the License. astral-tokio-tar-0.5.1/LICENSE-MIT000064400000000000000000000020411046102023000144550ustar 00000000000000Copyright (c) 2014 Alex Crichton Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the "Software"), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions: The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software. THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. astral-tokio-tar-0.5.1/README.md000064400000000000000000000032761046102023000143130ustar 00000000000000# `astral-tokio-tar` A `tokio`-based tar archive reader and writer. ## Provenance This crate is a fork of [`edera-dev/tokio-tar`](https://github.com/edera-dev/tokio-tar), which was a fork of [`vorot93/tokio-tar`](https://github.com/vorot93/tokio-tar), which was a fork of [`dignifiedquire/async-tar`](https://github.com/dignifiedquire/async-tar), which is based on [`alexcrichton/tar-rs`](https://github.com/alexcrichton/tar-rs). As compared to the async tar crates, this crate includes a variety of performance improvements and missing patches from [`alexcrichton/tar-rs`](https://github.com/alexcrichton/tar-rs). As compared to [`alexcrichton/tar-rs`](https://github.com/alexcrichton/tar-rs), this crate features the following modifications: - Setting `preserve_permissions` to `false` will avoid setting _any_ permissions on extracted files. In [`alexcrichton/tar-rs`](https://github.com/alexcrichton/tar-rs), setting `preserve_permissions` to `false` will still set read, write, and execute permissions on extracted files, but will avoid setting extended permissions (e.g., `setuid`, `setgid`, and `sticky` bits). See the [changelog](CHANGELOG.md) for a more detailed list of changes. ## License This project is licensed under either of * Apache License, Version 2.0, ([LICENSE-APACHE](LICENSE-APACHE) or http://www.apache.org/licenses/LICENSE-2.0) * MIT license ([LICENSE-MIT](LICENSE-MIT) or http://opensource.org/licenses/MIT) at your option. ## Contribution Unless you explicitly state otherwise, any contribution intentionally submitted for inclusion in this project by you, as defined in the Apache-2.0 license, shall be dual licensed as above, without any additional terms or conditions. astral-tokio-tar-0.5.1/examples/extract_file.rs000064400000000000000000000015701046102023000176640ustar 00000000000000//! An example of extracting a file in an archive. //! //! Takes a tarball on standard input, looks for an entry with a listed file //! name as the first argument provided, and then prints the contents of that //! file to stdout. extern crate tokio_tar as async_tar; use std::{env::args_os, path::Path}; use tokio::io::{copy, stdin, stdout}; use tokio_stream::*; use async_tar::Archive; fn main() { tokio::runtime::Runtime::new().unwrap().block_on(async { let first_arg = args_os().nth(1).unwrap(); let filename = Path::new(&first_arg); let mut ar = Archive::new(stdin()); let mut entries = ar.entries().unwrap(); while let Some(file) = entries.next().await { let mut f = file.unwrap(); if f.path().unwrap() == filename { copy(&mut f, &mut stdout()).await.unwrap(); } } }); } astral-tokio-tar-0.5.1/examples/list.rs000064400000000000000000000011061046102023000161610ustar 00000000000000//! An example of listing the file names of entries in an archive. //! //! Takes a tarball on stdin and prints out all of the entries inside. extern crate tokio_tar as async_tar; use tokio::io::stdin; use tokio_stream::*; use async_tar::Archive; fn main() { tokio::runtime::Runtime::new().unwrap().block_on(async { let mut ar = Archive::new(stdin()); let mut entries = ar.entries().unwrap(); while let Some(file) = entries.next().await { let f = file.unwrap(); println!("{}", f.path().unwrap().display()); } }); } astral-tokio-tar-0.5.1/examples/raw_list.rs000064400000000000000000000040271046102023000170370ustar 00000000000000//! An example of listing raw entries in an archive. //! //! Takes a tarball on stdin and prints out all of the entries inside. extern crate tokio_tar as async_tar; use tokio::io::stdin; use tokio_stream::*; use async_tar::Archive; fn main() { tokio::runtime::Runtime::new().unwrap().block_on(async { let mut ar = Archive::new(stdin()); let mut i = 0; let mut entries = ar.entries_raw().unwrap(); while let Some(file) = entries.next().await { println!("-------------------------- Entry {}", i); let mut f = file.unwrap(); println!("path: {}", f.path().unwrap().display()); println!("size: {}", f.header().size().unwrap()); println!("entry size: {}", f.header().entry_size().unwrap()); println!("link name: {:?}", f.link_name().unwrap()); println!("file type: {:#x}", f.header().entry_type().as_byte()); println!("mode: {:#o}", f.header().mode().unwrap()); println!("uid: {}", f.header().uid().unwrap()); println!("gid: {}", f.header().gid().unwrap()); println!("mtime: {}", f.header().mtime().unwrap()); println!("username: {:?}", f.header().username().unwrap()); println!("groupname: {:?}", f.header().groupname().unwrap()); if f.header().as_ustar().is_some() { println!("kind: UStar"); } else if f.header().as_gnu().is_some() { println!("kind: GNU"); } else { println!("kind: normal"); } if let Ok(Some(extensions)) = f.pax_extensions().await { println!("pax extensions:"); for e in extensions { let e = e.unwrap(); println!( "\t{:?} = {:?}", String::from_utf8_lossy(e.key_bytes()), String::from_utf8_lossy(e.value_bytes()) ); } } i += 1; } }); } astral-tokio-tar-0.5.1/examples/write.rs000064400000000000000000000006661046102023000163520ustar 00000000000000extern crate tokio_tar as async_tar; use async_tar::Builder; use tokio::fs::File; fn main() { tokio::runtime::Runtime::new().unwrap().block_on(async { let file = File::create("foo.tar").await.unwrap(); let mut a = Builder::new(file); a.append_path("README.md").await.unwrap(); a.append_file("lib.rs", &mut File::open("src/lib.rs").await.unwrap()) .await .unwrap(); }); } astral-tokio-tar-0.5.1/src/archive.rs000064400000000000000000000634621046102023000156150ustar 00000000000000use portable_atomic::{AtomicU64, Ordering}; use rustc_hash::FxHashSet; use std::{ cmp, collections::VecDeque, path::Path, pin::Pin, sync::Arc, task::{Context, Poll}, }; use tokio::{ fs, io::{self, AsyncRead as Read, AsyncReadExt}, sync::Mutex, }; use tokio_stream::*; use crate::header::BLOCK_SIZE; use crate::{ entry::{EntryFields, EntryIo}, error::TarError, other, Entry, GnuExtSparseHeader, GnuSparseHeader, Header, }; /// A top-level representation of an archive file. /// /// This archive can have an entry added to it and it can be iterated over. #[derive(Debug)] pub struct Archive { inner: Arc>, } impl Clone for Archive { fn clone(&self) -> Self { Archive { inner: self.inner.clone(), } } } #[derive(Debug)] pub struct ArchiveInner { pos: AtomicU64, unpack_xattrs: bool, preserve_permissions: bool, preserve_mtime: bool, overwrite: bool, ignore_zeros: bool, obj: Mutex, } /// Configure the archive. pub struct ArchiveBuilder { obj: R, unpack_xattrs: bool, preserve_permissions: bool, preserve_mtime: bool, overwrite: bool, ignore_zeros: bool, } impl ArchiveBuilder { /// Create a new builder. pub fn new(obj: R) -> Self { ArchiveBuilder { unpack_xattrs: false, preserve_permissions: false, preserve_mtime: true, overwrite: true, ignore_zeros: false, obj, } } /// Indicate whether extended file attributes (xattrs on Unix) are preserved /// when unpacking this archive. /// /// This flag is disabled by default and is currently only implemented on /// Unix using xattr support. This may eventually be implemented for /// Windows, however, if other archive implementations are found which do /// this as well. pub fn set_unpack_xattrs(mut self, unpack_xattrs: bool) -> Self { self.unpack_xattrs = unpack_xattrs; self } /// Indicate whether the permissions on files and directories are preserved /// when unpacking this entry. /// /// This flag is disabled by default and is currently only implemented on /// Unix. pub fn set_preserve_permissions(mut self, preserve: bool) -> Self { self.preserve_permissions = preserve; self } /// Indicate whether files and symlinks should be overwritten on extraction. pub fn set_overwrite(mut self, overwrite: bool) -> Self { self.overwrite = overwrite; self } /// Indicate whether access time information is preserved when unpacking /// this entry. /// /// This flag is enabled by default. pub fn set_preserve_mtime(mut self, preserve: bool) -> Self { self.preserve_mtime = preserve; self } /// Ignore zeroed headers, which would otherwise indicate to the archive that it has no more /// entries. /// /// This can be used in case multiple tar archives have been concatenated together. pub fn set_ignore_zeros(mut self, ignore_zeros: bool) -> Self { self.ignore_zeros = ignore_zeros; self } /// Construct the archive, ready to accept inputs. pub fn build(self) -> Archive { let Self { unpack_xattrs, preserve_permissions, preserve_mtime, overwrite, ignore_zeros, obj, } = self; Archive { inner: Arc::new(ArchiveInner { unpack_xattrs, preserve_permissions, preserve_mtime, overwrite, ignore_zeros, obj: Mutex::new(obj), pos: 0.into(), }), } } } impl Archive { /// Create a new archive with the underlying object as the reader. pub fn new(obj: R) -> Archive { Archive { inner: Arc::new(ArchiveInner { unpack_xattrs: false, preserve_permissions: false, preserve_mtime: true, overwrite: true, ignore_zeros: false, obj: Mutex::new(obj), pos: 0.into(), }), } } /// Unwrap this archive, returning the underlying object. pub fn into_inner(self) -> Result { let Self { inner } = self; match Arc::try_unwrap(inner) { Ok(inner) => Ok(inner.obj.into_inner()), Err(inner) => Err(Self { inner }), } } /// Construct an stream over the entries in this archive. /// /// Note that care must be taken to consider each entry within an archive in /// sequence. If entries are processed out of sequence (from what the /// stream returns), then the contents read for each entry may be /// corrupted. pub fn entries(&mut self) -> io::Result> { if self.inner.pos.load(Ordering::SeqCst) != 0 { return Err(other( "cannot call entries unless archive is at \ position 0", )); } Ok(Entries { archive: self.clone(), pending: None, current: (0, None, 0, None), gnu_longlink: (false, None), gnu_longname: (false, None), pax_extensions: (false, None), }) } /// Construct an stream over the raw entries in this archive. /// /// Note that care must be taken to consider each entry within an archive in /// sequence. If entries are processed out of sequence (from what the /// stream returns), then the contents read for each entry may be /// corrupted. pub fn entries_raw(&mut self) -> io::Result> { if self.inner.pos.load(Ordering::SeqCst) != 0 { return Err(other( "cannot call entries_raw unless archive is at \ position 0", )); } Ok(RawEntries { archive: self.clone(), current: (0, None, 0), }) } /// Unpacks the contents tarball into the specified `dst`. /// /// This function will iterate over the entire contents of this tarball, /// extracting each file in turn to the location specified by the entry's /// path name. /// /// This operation is relatively sensitive in that it will not write files /// outside of the path specified by `dst`. Files in the archive which have /// a '..' in their path are skipped during the unpacking process. /// /// # Examples /// /// ```no_run /// # fn main() -> Result<(), Box> { tokio::runtime::Runtime::new().unwrap().block_on(async { /// # /// use tokio::fs::File; /// use tokio_tar::Archive; /// /// let mut ar = Archive::new(File::open("foo.tar").await?); /// ar.unpack("foo").await?; /// # /// # Ok(()) }) } /// ``` pub async fn unpack>(&mut self, dst: P) -> io::Result<()> { let mut entries = self.entries()?; let mut pinned = Pin::new(&mut entries); let dst = dst.as_ref(); if fs::symlink_metadata(dst).await.is_err() { fs::create_dir_all(&dst) .await .map_err(|e| TarError::new(format!("failed to create `{}`", dst.display()), e))?; } // Canonicalizing the dst directory will prepend the path with '\\?\' // on windows which will allow windows APIs to treat the path as an // extended-length path with a 32,767 character limit. Otherwise all // unpacked paths over 260 characters will fail on creation with a // NotFound exception. let dst = fs::canonicalize(dst).await?; // Memoize filesystem calls to canonicalize paths. let mut targets = FxHashSet::default(); // Delay any directory entries until the end (they will be created if needed by // descendants), to ensure that directory permissions do not interfere with descendant // extraction. let mut directories = Vec::new(); while let Some(entry) = pinned.next().await { let mut file = entry.map_err(|e| TarError::new("failed to iterate over archive", e))?; if file.header().entry_type() == crate::EntryType::Directory { directories.push(file); } else { file.unpack_in_raw(&dst, &mut targets).await?; } } // Apply the directories. // // Note: the order of application is important to permissions. That is, we must traverse // the filesystem graph in topological ordering or else we risk not being able to create // child directories within those of more restrictive permissions. See [0] for details. // // [0]: directories.sort_by(|a, b| b.path_bytes().cmp(&a.path_bytes())); for mut dir in directories { dir.unpack_in_raw(&dst, &mut targets).await?; } Ok(()) } } /// Stream of `Entry`s. pub struct Entries { archive: Archive, current: (u64, Option
, usize, Option), /// The [`Entry`] that is currently being processed. pending: Option>>, /// GNU long name extension. /// /// The first element is a flag indicating whether the long name entry has been fully read. /// The second element is the buffer containing the long name, or `None` if the long name entry /// has not been encountered yet. gnu_longname: (bool, Option>), /// GNU long link extension. /// /// The first element is a flag indicating whether the long link entry has been fully read. /// The second element is the buffer containing the long link, or `None` if the long link entry /// has not been encountered yet. gnu_longlink: (bool, Option>), /// PAX extensions. /// /// The first element is a flag indicating whether the extension entry has been fully read. /// The second element is the buffer containing the extension, or `None` if the extension entry /// has not been encountered yet. pax_extensions: (bool, Option>), } macro_rules! ready_opt_err { ($val:expr) => { match futures_core::ready!($val) { Some(Ok(val)) => val, Some(Err(err)) => return Poll::Ready(Some(Err(err))), None => return Poll::Ready(None), } }; } macro_rules! ready_err { ($val:expr) => { match futures_core::ready!($val) { Ok(val) => val, Err(err) => return Poll::Ready(Some(Err(err))), } }; } impl Stream for Entries { type Item = io::Result>>; fn poll_next(mut self: Pin<&mut Self>, cx: &mut Context<'_>) -> Poll> { loop { let archive = self.archive.clone(); let entry = if let Some(entry) = self.pending.take() { entry } else { let (next, current_header, current_header_pos, _) = &mut self.current; ready_opt_err!(poll_next_raw( archive, next, current_header, current_header_pos, cx )) }; let is_recognized_header = entry.header().as_gnu().is_some() || entry.header().as_ustar().is_some(); if is_recognized_header && entry.header().entry_type().is_gnu_longname() { if self.gnu_longname.0 { return Poll::Ready(Some(Err(other( "two long name entries describing \ the same member", )))); } let mut ef = EntryFields::from(entry); let cursor = self.gnu_longname.1.get_or_insert_with(|| { let cap = cmp::min(ef.size, 128 * 1024); Vec::with_capacity(cap as usize) }); if let Poll::Ready(result) = Pin::new(&mut ef).poll_read_all(cx, cursor) { if let Err(err) = result { return Poll::Ready(Some(Err(err))); } } else { self.pending = Some(ef.into_entry()); return Poll::Pending; } self.gnu_longname.0 = true; continue; } if is_recognized_header && entry.header().entry_type().is_gnu_longlink() { if self.gnu_longlink.0 { return Poll::Ready(Some(Err(other( "two long name entries describing \ the same member", )))); } let mut ef = EntryFields::from(entry); let cursor = self.gnu_longlink.1.get_or_insert_with(|| { let cap = cmp::min(ef.size, 128 * 1024); Vec::with_capacity(cap as usize) }); if let Poll::Ready(result) = Pin::new(&mut ef).poll_read_all(cx, cursor) { if let Err(err) = result { return Poll::Ready(Some(Err(err))); } } else { self.pending = Some(ef.into_entry()); return Poll::Pending; } self.gnu_longlink.0 = true; continue; } if is_recognized_header && entry.header().entry_type().is_pax_local_extensions() { if self.pax_extensions.0 { return Poll::Ready(Some(Err(other( "two pax extensions entries describing \ the same member", )))); } let mut ef = EntryFields::from(entry); let cursor = self.pax_extensions.1.get_or_insert_with(|| { let cap = cmp::min(ef.size, 128 * 1024); Vec::with_capacity(cap as usize) }); if let Poll::Ready(result) = Pin::new(&mut ef).poll_read_all(cx, cursor) { if let Err(err) = result { return Poll::Ready(Some(Err(err))); } } else { self.pending = Some(ef.into_entry()); return Poll::Pending; } self.pax_extensions.0 = true; continue; } let mut fields = EntryFields::from(entry); if self.gnu_longname.0 { fields.long_pathname = self.gnu_longname.1.take(); self.gnu_longname.0 = false; } if self.gnu_longlink.0 { fields.long_linkname = self.gnu_longlink.1.take(); self.gnu_longlink.0 = false; } if self.pax_extensions.0 { fields.pax_extensions = self.pax_extensions.1.take(); self.pax_extensions.0 = false; } let archive = self.archive.clone(); let (next, _, current_pos, current_ext) = &mut self.current; ready_err!(poll_parse_sparse_header( archive, next, current_ext, current_pos, &mut fields, cx )); return Poll::Ready(Some(Ok(fields.into_entry()))); } } } /// Stream of raw `Entry`s. pub struct RawEntries { archive: Archive, current: (u64, Option
, usize), } impl Stream for RawEntries { type Item = io::Result>>; fn poll_next(mut self: Pin<&mut Self>, cx: &mut Context<'_>) -> Poll> { let archive = self.archive.clone(); let (next, current_header, current_header_pos) = &mut self.current; poll_next_raw(archive, next, current_header, current_header_pos, cx) } } fn poll_next_raw( mut archive: Archive, next: &mut u64, current_header: &mut Option
, current_header_pos: &mut usize, cx: &mut Context<'_>, ) -> Poll>>>> { let mut header_pos = *next; loop { // Seek to the start of the next header in the archive if current_header.is_none() { let delta = *next - archive.inner.pos.load(Ordering::SeqCst); match futures_core::ready!(poll_skip(&mut archive, cx, delta)) { Ok(_) => {} Err(err) => return Poll::Ready(Some(Err(err))), } *current_header = Some(Header::new_old()); *current_header_pos = 0; } let header = current_header.as_mut().unwrap(); // EOF is an indicator that we are at the end of the archive. match futures_core::ready!(poll_try_read_all( &mut archive, cx, header.as_mut_bytes(), current_header_pos, )) { Ok(true) => {} Ok(false) => return Poll::Ready(None), Err(err) => return Poll::Ready(Some(Err(err))), } // If a header is not all zeros, we have another valid header. // Otherwise, check if we are ignoring zeros and continue, or break as if this is the // end of the archive. if !header.as_bytes().iter().all(|i| *i == 0) { *next += BLOCK_SIZE; break; } if !archive.inner.ignore_zeros { return Poll::Ready(None); } *next += BLOCK_SIZE; header_pos = *next; } let header = current_header.as_mut().unwrap(); // Make sure the checksum is ok let sum = header.as_bytes()[..148] .iter() .chain(&header.as_bytes()[156..]) .fold(0, |a, b| a + (*b as u32)) + 8 * 32; let cksum = header.cksum()?; if sum != cksum { return Poll::Ready(Some(Err(other("archive header checksum mismatch")))); } let file_pos = *next; let size = header.entry_size()?; let mut data = VecDeque::with_capacity(1); data.push_back(EntryIo::Data(archive.clone().take(size))); let header = current_header.take().unwrap(); let ret = EntryFields { size, header_pos, file_pos, data, header, long_pathname: None, long_linkname: None, pax_extensions: None, unpack_xattrs: archive.inner.unpack_xattrs, preserve_permissions: archive.inner.preserve_permissions, preserve_mtime: archive.inner.preserve_mtime, overwrite: archive.inner.overwrite, read_state: None, }; // Store where the next entry is, rounding up by 512 bytes (the size of // a header); let size = size .checked_add(BLOCK_SIZE - 1) .ok_or_else(|| other("size overflow"))?; *next = next .checked_add(size & !(BLOCK_SIZE - 1)) .ok_or_else(|| other("size overflow"))?; Poll::Ready(Some(Ok(ret.into_entry()))) } fn poll_parse_sparse_header( mut archive: Archive, next: &mut u64, current_ext: &mut Option, current_ext_pos: &mut usize, entry: &mut EntryFields>, cx: &mut Context<'_>, ) -> Poll> { if !entry.header.entry_type().is_gnu_sparse() { return Poll::Ready(Ok(())); } let gnu = match entry.header.as_gnu() { Some(gnu) => gnu, None => return Poll::Ready(Err(other("sparse entry type listed but not GNU header"))), }; // Sparse files are represented internally as a list of blocks that are // read. Blocks are either a bunch of 0's or they're data from the // underlying archive. // // Blocks of a sparse file are described by the `GnuSparseHeader` // structure, some of which are contained in `GnuHeader` but some of // which may also be contained after the first header in further // headers. // // We read off all the blocks here and use the `add_block` function to // incrementally add them to the list of I/O block (in `entry.data`). // The `add_block` function also validates that each chunk comes after // the previous, we don't overrun the end of the file, and each block is // aligned to a 512-byte boundary in the archive itself. // // At the end we verify that the sparse file size (`Header::size`) is // the same as the current offset (described by the list of blocks) as // well as the amount of data read equals the size of the entry // (`Header::entry_size`). entry.data.truncate(0); let mut cur = 0; let mut remaining = entry.size; { let data = &mut entry.data; let reader = archive.clone(); let size = entry.size; let mut add_block = |block: &GnuSparseHeader| -> io::Result<_> { if block.is_empty() { return Ok(()); } let off = block.offset()?; let len = block.length()?; if len != 0 && (size - remaining) % BLOCK_SIZE != 0 { return Err(other( "previous block in sparse file was not \ aligned to 512-byte boundary", )); } else if off < cur { return Err(other( "out of order or overlapping sparse \ blocks", )); } else if cur < off { let block = io::repeat(0).take(off - cur); data.push_back(EntryIo::Pad(block)); } cur = off .checked_add(len) .ok_or_else(|| other("more bytes listed in sparse file than u64 can hold"))?; remaining = remaining.checked_sub(len).ok_or_else(|| { other( "sparse file consumed more data than the header \ listed", ) })?; data.push_back(EntryIo::Data(reader.clone().take(len))); Ok(()) }; for block in gnu.sparse.iter() { add_block(block)? } if gnu.is_extended() { let started_header = current_ext.is_some(); if !started_header { let mut ext = GnuExtSparseHeader::new(); ext.isextended[0] = 1; *current_ext = Some(ext); *current_ext_pos = 0; } let ext = current_ext.as_mut().unwrap(); while ext.is_extended() { match futures_core::ready!(poll_try_read_all( &mut archive, cx, ext.as_mut_bytes(), current_ext_pos, )) { Ok(true) => {} Ok(false) => return Poll::Ready(Err(other("failed to read extension"))), Err(err) => return Poll::Ready(Err(err)), } *next += BLOCK_SIZE; for block in ext.sparse.iter() { add_block(block)?; } } } } if cur != gnu.real_size()? { return Poll::Ready(Err(other( "mismatch in sparse file chunks and \ size in header", ))); } entry.size = cur; if remaining > 0 { return Poll::Ready(Err(other( "mismatch in sparse file chunks and \ entry size in header", ))); } Poll::Ready(Ok(())) } impl Read for Archive { fn poll_read( self: Pin<&mut Self>, cx: &mut Context<'_>, into: &mut io::ReadBuf<'_>, ) -> Poll> { let mut r = if let Ok(v) = self.inner.obj.try_lock() { v } else { return Poll::Pending; }; let res = futures_core::ready!(Pin::new(&mut *r).poll_read(cx, into)); match res { Ok(()) => { self.inner .pos .fetch_add(into.filled().len() as u64, Ordering::SeqCst); Poll::Ready(Ok(())) } Err(err) => Poll::Ready(Err(err)), } } } /// Try to fill the buffer from the reader. /// /// If the reader reaches its end before filling the buffer at all, returns `false`. /// Otherwise returns `true`. fn poll_try_read_all( mut source: R, cx: &mut Context<'_>, buf: &mut [u8], pos: &mut usize, ) -> Poll> { while *pos < buf.len() { let mut read_buf = io::ReadBuf::new(&mut buf[*pos..]); match futures_core::ready!(Pin::new(&mut source).poll_read(cx, &mut read_buf)) { Ok(()) if read_buf.filled().is_empty() => { if *pos == 0 { return Poll::Ready(Ok(false)); } return Poll::Ready(Err(other("failed to read entire block"))); } Ok(()) => *pos += read_buf.filled().len(), Err(err) => return Poll::Ready(Err(err)), } } *pos = 0; Poll::Ready(Ok(true)) } /// Skip n bytes on the given source. fn poll_skip( mut source: R, cx: &mut Context<'_>, mut amt: u64, ) -> Poll> { let mut buf = [0u8; 4096 * 8]; while amt > 0 { let n = cmp::min(amt, buf.len() as u64); let mut read_buf = io::ReadBuf::new(&mut buf[..n as usize]); match futures_core::ready!(Pin::new(&mut source).poll_read(cx, &mut read_buf)) { Ok(()) if read_buf.filled().is_empty() => { return Poll::Ready(Err(other("unexpected EOF during skip"))); } Ok(()) => { amt -= read_buf.filled().len() as u64; } Err(err) => return Poll::Ready(Err(err)), } } Poll::Ready(Ok(())) } astral-tokio-tar-0.5.1/src/builder.rs000064400000000000000000000602641046102023000156170ustar 00000000000000use crate::{ header::{path2bytes, HeaderMode}, other, EntryType, Header, }; use std::{fs::Metadata, path::Path, str}; use tokio::{ fs, io::{self, AsyncRead as Read, AsyncReadExt, AsyncWrite as Write, AsyncWriteExt}, }; /// A structure for building archives /// /// This structure has methods for building up an archive from scratch into any /// arbitrary writer. pub struct Builder { mode: HeaderMode, follow: bool, finished: bool, obj: Option, cancellation: Option>, } const TERMINATION: &[u8; 1024] = &[0; 1024]; impl Builder { /// Create a new archive builder with the underlying object as the /// destination of all data written. The builder will use /// `HeaderMode::Complete` by default. /// /// On drop, would write [`TERMINATION`] into the end of the archive, /// use `skip_termination` method to disable this. pub fn new(obj: W) -> Builder { let (tx, rx) = tokio::sync::oneshot::channel::(); tokio::spawn(async move { if let Ok(mut w) = rx.await { let _ = w.write_all(TERMINATION).await; } }); Builder { mode: HeaderMode::Complete, follow: true, finished: false, obj: Some(obj), cancellation: Some(tx), } } } impl Builder { /// Create a new archive builder with the underlying object as the /// destination of all data written. The builder will use /// `HeaderMode::Complete` by default. /// /// The [`TERMINATION`] symbol would not be written to the archive in the end. pub fn new_non_terminated(obj: W) -> Builder { Builder { mode: HeaderMode::Complete, follow: true, finished: false, obj: Some(obj), cancellation: None, } } /// Changes the HeaderMode that will be used when reading fs Metadata for /// methods that implicitly read metadata for an input Path. Notably, this /// does _not_ apply to `append(Header)`. pub fn mode(&mut self, mode: HeaderMode) { self.mode = mode; } /// Follow symlinks, archiving the contents of the file they point to rather /// than adding a symlink to the archive. Defaults to true. pub fn follow_symlinks(&mut self, follow: bool) { self.follow = follow; } /// Skip writing final termination bytes into the archive. pub fn skip_termination(&mut self) { drop(self.cancellation.take()); } /// Gets shared reference to the underlying object. pub fn get_ref(&self) -> &W { self.obj.as_ref().unwrap() } /// Gets mutable reference to the underlying object. /// /// Note that care must be taken while writing to the underlying /// object. But, e.g. `get_mut().flush()` is claimed to be safe and /// useful in the situations when one needs to be ensured that /// tar entry was flushed to the disk. pub fn get_mut(&mut self) -> &mut W { self.obj.as_mut().unwrap() } /// Unwrap this archive, returning the underlying object. /// /// This function will finish writing the archive if the `finish` function /// hasn't yet been called, returning any I/O error which happens during /// that operation. pub async fn into_inner(mut self) -> io::Result { if !self.finished { self.finish().await?; } Ok(self.obj.take().unwrap()) } /// Adds a new entry to this archive. /// /// This function will append the header specified, followed by contents of /// the stream specified by `data`. To produce a valid archive the `size` /// field of `header` must be the same as the length of the stream that's /// being written. Additionally the checksum for the header should have been /// set via the `set_cksum` method. /// /// Note that this will not attempt to seek the archive to a valid position, /// so if the archive is in the middle of a read or some other similar /// operation then this may corrupt the archive. /// /// Also note that after all entries have been written to an archive the /// `finish` function needs to be called to finish writing the archive. /// /// # Errors /// /// This function will return an error for any intermittent I/O error which /// occurs when either reading or writing. /// /// # Examples /// /// ``` /// # fn main() -> Result<(), Box> { tokio::runtime::Runtime::new().unwrap().block_on(async { /// # /// use tokio_tar::{Builder, Header}; /// /// let mut header = Header::new_gnu(); /// header.set_path("foo")?; /// header.set_size(4); /// header.set_cksum(); /// /// let mut data: &[u8] = &[1, 2, 3, 4]; /// /// let mut ar = Builder::new(Vec::new()); /// ar.append(&header, data).await?; /// let data = ar.into_inner().await?; /// # /// # Ok(()) }) } /// ``` pub async fn append( &mut self, header: &Header, mut data: R, ) -> io::Result<()> { append(self.get_mut(), header, &mut data).await?; Ok(()) } /// Adds a new entry to this archive with the specified path. /// /// This function will set the specified path in the given header, which may /// require appending a GNU long-name extension entry to the archive first. /// The checksum for the header will be automatically updated via the /// `set_cksum` method after setting the path. No other metadata in the /// header will be modified. /// /// Then it will append the header, followed by contents of the stream /// specified by `data`. To produce a valid archive the `size` field of /// `header` must be the same as the length of the stream that's being /// written. /// /// Note that this will not attempt to seek the archive to a valid position, /// so if the archive is in the middle of a read or some other similar /// operation then this may corrupt the archive. /// /// Also note that after all entries have been written to an archive the /// `finish` function needs to be called to finish writing the archive. /// /// # Errors /// /// This function will return an error for any intermittent I/O error which /// occurs when either reading or writing. /// /// # Examples /// /// ``` /// # fn main() -> Result<(), Box> { tokio::runtime::Runtime::new().unwrap().block_on(async { /// # /// use tokio_tar::{Builder, Header}; /// /// let mut header = Header::new_gnu(); /// header.set_size(4); /// header.set_cksum(); /// /// let mut data: &[u8] = &[1, 2, 3, 4]; /// /// let mut ar = Builder::new(Vec::new()); /// ar.append_data(&mut header, "really/long/path/to/foo", data).await?; /// let data = ar.into_inner().await?; /// # /// # Ok(()) }) } /// ``` pub async fn append_data, R: Read + Unpin>( &mut self, header: &mut Header, path: P, data: R, ) -> io::Result<()> { prepare_header_path(self.get_mut(), header, path.as_ref()).await?; header.set_cksum(); self.append(header, data).await?; Ok(()) } /// Adds a file on the local filesystem to this archive. /// /// This function will open the file specified by `path` and insert the file /// into the archive with the appropriate metadata set, returning any I/O /// error which occurs while writing. The path name for the file inside of /// this archive will be the same as `path`, and it is required that the /// path is a relative path. /// /// Note that this will not attempt to seek the archive to a valid position, /// so if the archive is in the middle of a read or some other similar /// operation then this may corrupt the archive. /// /// Also note that after all files have been written to an archive the /// `finish` function needs to be called to finish writing the archive. /// /// # Examples /// /// ```no_run /// # fn main() -> Result<(), Box> { tokio::runtime::Runtime::new().unwrap().block_on(async { /// # /// use tokio_tar::Builder; /// /// let mut ar = Builder::new(Vec::new()); /// /// ar.append_path("foo/bar.txt").await?; /// # /// # Ok(()) }) } /// ``` pub async fn append_path>(&mut self, path: P) -> io::Result<()> { let mode = self.mode; let follow = self.follow; append_path_with_name(self.get_mut(), path.as_ref(), None, mode, follow).await?; Ok(()) } /// Adds a file on the local filesystem to this archive under another name. /// /// This function will open the file specified by `path` and insert the file /// into the archive as `name` with appropriate metadata set, returning any /// I/O error which occurs while writing. The path name for the file inside /// of this archive will be `name` is required to be a relative path. /// /// Note that this will not attempt to seek the archive to a valid position, /// so if the archive is in the middle of a read or some other similar /// operation then this may corrupt the archive. /// /// Also note that after all files have been written to an archive the /// `finish` function needs to be called to finish writing the archive. /// /// # Examples /// /// ```no_run /// # fn main() -> Result<(), Box> { tokio::runtime::Runtime::new().unwrap().block_on(async { /// # /// use tokio_tar::Builder; /// /// let mut ar = Builder::new(Vec::new()); /// /// // Insert the local file "foo/bar.txt" in the archive but with the name /// // "bar/foo.txt". /// ar.append_path_with_name("foo/bar.txt", "bar/foo.txt").await?; /// # /// # Ok(()) }) } /// ``` pub async fn append_path_with_name, N: AsRef>( &mut self, path: P, name: N, ) -> io::Result<()> { let mode = self.mode; let follow = self.follow; append_path_with_name( self.get_mut(), path.as_ref(), Some(name.as_ref()), mode, follow, ) .await?; Ok(()) } /// Adds a file to this archive with the given path as the name of the file /// in the archive. /// /// This will use the metadata of `file` to populate a `Header`, and it will /// then append the file to the archive with the name `path`. /// /// Note that this will not attempt to seek the archive to a valid position, /// so if the archive is in the middle of a read or some other similar /// operation then this may corrupt the archive. /// /// Also note that after all files have been written to an archive the /// `finish` function needs to be called to finish writing the archive. /// /// # Examples /// /// ```no_run /// # fn main() -> Result<(), Box> { tokio::runtime::Runtime::new().unwrap().block_on(async { /// # /// use tokio::fs::File; /// use tokio_tar::Builder; /// /// let mut ar = Builder::new(Vec::new()); /// /// // Open the file at one location, but insert it into the archive with a /// // different name. /// let mut f = File::open("foo/bar/baz.txt").await?; /// ar.append_file("bar/baz.txt", &mut f).await?; /// # /// # Ok(()) }) } /// ``` pub async fn append_file>( &mut self, path: P, file: &mut fs::File, ) -> io::Result<()> { let mode = self.mode; append_file(self.get_mut(), path.as_ref(), file, mode).await?; Ok(()) } /// Adds a directory to this archive with the given path as the name of the /// directory in the archive. /// /// This will use `stat` to populate a `Header`, and it will then append the /// directory to the archive with the name `path`. /// /// Note that this will not attempt to seek the archive to a valid position, /// so if the archive is in the middle of a read or some other similar /// operation then this may corrupt the archive. /// /// Also note that after all files have been written to an archive the /// `finish` function needs to be called to finish writing the archive. /// /// # Examples /// /// ``` /// # fn main() -> Result<(), Box> { tokio::runtime::Runtime::new().unwrap().block_on(async { /// # /// use tokio::fs; /// use tokio_tar::Builder; /// /// let mut ar = Builder::new(Vec::new()); /// /// // Use the directory at one location, but insert it into the archive /// // with a different name. /// ar.append_dir("bardir", ".").await?; /// # /// # Ok(()) }) } /// ``` pub async fn append_dir(&mut self, path: P, src_path: Q) -> io::Result<()> where P: AsRef, Q: AsRef, { let mode = self.mode; append_dir(self.get_mut(), path.as_ref(), src_path.as_ref(), mode).await?; Ok(()) } /// Adds a directory and all of its contents (recursively) to this archive /// with the given path as the name of the directory in the archive. /// /// Note that this will not attempt to seek the archive to a valid position, /// so if the archive is in the middle of a read or some other similar /// operation then this may corrupt the archive. /// /// Also note that after all files have been written to an archive the /// `finish` function needs to be called to finish writing the archive. /// /// # Examples /// /// ``` /// # fn main() -> Result<(), Box> { tokio::runtime::Runtime::new().unwrap().block_on(async { /// # /// use tokio::fs; /// use tokio_tar::Builder; /// /// let mut ar = Builder::new(Vec::new()); /// /// // Use the directory at one location, but insert it into the archive /// // with a different name. /// ar.append_dir_all("bardir", ".").await?; /// # /// # Ok(()) }) } /// ``` pub async fn append_dir_all(&mut self, path: P, src_path: Q) -> io::Result<()> where P: AsRef, Q: AsRef, { let mode = self.mode; let follow = self.follow; append_dir_all( self.get_mut(), path.as_ref(), src_path.as_ref(), mode, follow, ) .await?; Ok(()) } /// Finish writing this archive, emitting the termination sections. /// /// This function should only be called when the archive has been written /// entirely and if an I/O error happens the underlying object still needs /// to be acquired. /// /// In most situations the `into_inner` method should be preferred. pub async fn finish(&mut self) -> io::Result<()> { if self.finished { return Ok(()); } self.finished = true; self.get_mut().write_all(&[0; 1024]).await?; Ok(()) } } async fn append( mut dst: &mut Dst, header: &Header, mut data: &mut Data, ) -> io::Result<()> { dst.write_all(header.as_bytes()).await?; let len = io::copy(&mut data, &mut dst).await?; // Pad with zeros if necessary. let buf = [0; 512]; let remaining = 512 - (len % 512); if remaining < 512 { dst.write_all(&buf[..remaining as usize]).await?; } Ok(()) } async fn append_path_with_name( dst: &mut Dst, path: &Path, name: Option<&Path>, mode: HeaderMode, follow: bool, ) -> io::Result<()> { let stat = if follow { fs::metadata(path).await.map_err(|err| { io::Error::new( err.kind(), format!("{} when getting metadata for {}", err, path.display()), ) })? } else { fs::symlink_metadata(path).await.map_err(|err| { io::Error::new( err.kind(), format!("{} when getting metadata for {}", err, path.display()), ) })? }; let ar_name = name.unwrap_or(path); if stat.is_file() { append_fs( dst, ar_name, &stat, &mut fs::File::open(path).await?, mode, None, ) .await?; Ok(()) } else if stat.is_dir() { append_fs(dst, ar_name, &stat, &mut io::empty(), mode, None).await?; Ok(()) } else if stat.file_type().is_symlink() { let link_name = fs::read_link(path).await?; append_fs( dst, ar_name, &stat, &mut io::empty(), mode, Some(&link_name), ) .await?; Ok(()) } else { #[cfg(unix)] { append_special(dst, path, &stat, mode).await } #[cfg(not(unix))] { Err(other(&format!("{} has unknown file type", path.display()))) } } } #[cfg(unix)] async fn append_special( dst: &mut Dst, path: &Path, stat: &Metadata, mode: HeaderMode, ) -> io::Result<()> { use ::std::os::unix::fs::{FileTypeExt, MetadataExt}; let file_type = stat.file_type(); let entry_type; if file_type.is_socket() { // sockets can't be archived return Err(other(&format!( "{}: socket can not be archived", path.display() ))); } else if file_type.is_fifo() { entry_type = EntryType::Fifo; } else if file_type.is_char_device() { entry_type = EntryType::Char; } else if file_type.is_block_device() { entry_type = EntryType::Block; } else { return Err(other(&format!("{} has unknown file type", path.display()))); } let mut header = Header::new_gnu(); header.set_metadata_in_mode(stat, mode); prepare_header_path(dst, &mut header, path).await?; header.set_entry_type(entry_type); let dev_id = stat.rdev(); let dev_major = ((dev_id >> 32) & 0xffff_f000) | ((dev_id >> 8) & 0x0000_0fff); let dev_minor = ((dev_id >> 12) & 0xffff_ff00) | ((dev_id) & 0x0000_00ff); header.set_device_major(dev_major as u32)?; header.set_device_minor(dev_minor as u32)?; header.set_cksum(); dst.write_all(header.as_bytes()).await?; Ok(()) } async fn append_file( dst: &mut Dst, path: &Path, file: &mut fs::File, mode: HeaderMode, ) -> io::Result<()> { let stat = file.metadata().await?; append_fs(dst, path, &stat, file, mode, None).await?; Ok(()) } async fn append_dir( dst: &mut Dst, path: &Path, src_path: &Path, mode: HeaderMode, ) -> io::Result<()> { let stat = fs::metadata(src_path).await?; append_fs(dst, path, &stat, &mut io::empty(), mode, None).await?; Ok(()) } fn prepare_header(size: u64, entry_type: EntryType) -> Header { let mut header = Header::new_gnu(); let name = b"././@LongLink"; header.as_gnu_mut().unwrap().name[..name.len()].clone_from_slice(&name[..]); header.set_mode(0o644); header.set_uid(0); header.set_gid(0); header.set_mtime(0); // + 1 to be compliant with GNU tar header.set_size(size + 1); header.set_entry_type(entry_type); header.set_cksum(); header } async fn prepare_header_path( dst: &mut Dst, header: &mut Header, path: &Path, ) -> io::Result<()> { // Try to encode the path directly in the header, but if it ends up not // working (probably because it's too long) then try to use the GNU-specific // long name extension by emitting an entry which indicates that it's the // filename. if let Err(e) = header.set_path(path) { let data = path2bytes(path)?; let max = header.as_old().name.len(); // Since e isn't specific enough to let us know the path is indeed too // long, verify it first before using the extension. if data.len() < max { return Err(e); } let header2 = prepare_header(data.len() as u64, EntryType::GNULongName); // null-terminated string let mut data2 = data.chain(io::repeat(0).take(1)); append(dst, &header2, &mut data2).await?; // Truncate the path to store in the header we're about to emit to // ensure we've got something at least mentioned. Note that we use // `str`-encoding to be compatible with Windows, but in general the // entry in the header itself shouldn't matter too much since extraction // doesn't look at it. let truncated = match str::from_utf8(&data[..max]) { Ok(s) => s, Err(e) => str::from_utf8(&data[..e.valid_up_to()]).unwrap(), }; header.set_truncated_path_for_gnu_header(truncated)?; } Ok(()) } async fn prepare_header_link( dst: &mut Dst, header: &mut Header, link_name: &Path, ) -> io::Result<()> { // Same as previous function but for linkname if let Err(e) = header.set_link_name(link_name) { let data = path2bytes(link_name)?; if data.len() < header.as_old().linkname.len() { return Err(e); } let header2 = prepare_header(data.len() as u64, EntryType::GNULongLink); let mut data2 = data.chain(io::repeat(0).take(1)); append(dst, &header2, &mut data2).await?; } Ok(()) } async fn append_fs( dst: &mut Dst, path: &Path, meta: &Metadata, read: &mut R, mode: HeaderMode, link_name: Option<&Path>, ) -> io::Result<()> { let mut header = Header::new_gnu(); prepare_header_path(dst, &mut header, path).await?; header.set_metadata_in_mode(meta, mode); if let Some(link_name) = link_name { prepare_header_link(dst, &mut header, link_name).await?; } header.set_cksum(); append(dst, &header, read).await?; Ok(()) } async fn append_dir_all( dst: &mut Dst, path: &Path, src_path: &Path, mode: HeaderMode, follow: bool, ) -> io::Result<()> { let mut stack = vec![(src_path.to_path_buf(), true, false)]; while let Some((src, is_dir, is_symlink)) = stack.pop() { let dest = path.join(src.strip_prefix(src_path).unwrap()); // In case of a symlink pointing to a directory, is_dir is false, but src.is_dir() will return true if is_dir || (is_symlink && follow && src.is_dir()) { let mut entries = fs::read_dir(&src).await?; while let Some(entry) = entries.next_entry().await.transpose() { let entry = entry?; let file_type = entry.file_type().await?; stack.push((entry.path(), file_type.is_dir(), file_type.is_symlink())); } if dest != Path::new("") { append_dir(dst, &dest, &src, mode).await?; } } else if !follow && is_symlink { let stat = fs::symlink_metadata(&src).await?; let link_name = fs::read_link(&src).await?; append_fs(dst, &dest, &stat, &mut io::empty(), mode, Some(&link_name)).await?; } else { #[cfg(unix)] { let stat = fs::metadata(&src).await?; if !stat.is_file() { append_special(dst, &dest, &stat, mode).await?; continue; } } append_file(dst, &dest, &mut fs::File::open(src).await?, mode).await?; } } Ok(()) } impl Drop for Builder { fn drop(&mut self) { // TODO: proper async cancellation if !self.finished { if let Some(cancellation) = self.cancellation.take() { cancellation.send(self.obj.take().unwrap()).ok(); } } } } astral-tokio-tar-0.5.1/src/entry.rs000064400000000000000000001120131046102023000153200ustar 00000000000000use crate::{ error::TarError, header::bytes2path, other, pax::pax_extensions, Archive, Header, PaxExtensions, }; use filetime::{self, FileTime}; use rustc_hash::FxHashSet; use std::{ borrow::Cow, cmp, collections::VecDeque, convert::TryFrom, fmt, io::{Error, ErrorKind, SeekFrom}, marker, path::{Component, Path, PathBuf}, pin::Pin, task::{Context, Poll}, }; use tokio::{ fs, fs::{remove_file, OpenOptions}, io::{self, AsyncRead as Read, AsyncReadExt, AsyncSeekExt, AsyncWriteExt}, }; /// A read-only view into an entry of an archive. /// /// This structure is a window into a portion of a borrowed archive which can /// be inspected. It acts as a file handle by implementing the Reader trait. An /// entry cannot be rewritten once inserted into an archive. pub struct Entry { fields: EntryFields, _ignored: marker::PhantomData>, } impl fmt::Debug for Entry { fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result { f.debug_struct("Entry") .field("fields", &self.fields) .finish() } } // private implementation detail of `Entry`, but concrete (no type parameters) // and also all-public to be constructed from other modules. pub struct EntryFields { pub long_pathname: Option>, pub long_linkname: Option>, pub pax_extensions: Option>, pub header: Header, pub size: u64, pub header_pos: u64, pub file_pos: u64, pub data: VecDeque>, pub unpack_xattrs: bool, pub preserve_permissions: bool, pub preserve_mtime: bool, pub overwrite: bool, pub(crate) read_state: Option>, } impl fmt::Debug for EntryFields { fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result { f.debug_struct("EntryFields") .field("long_pathname", &self.long_pathname) .field("long_linkname", &self.long_linkname) .field("pax_extensions", &self.pax_extensions) .field("header", &self.header) .field("size", &self.size) .field("header_pos", &self.header_pos) .field("file_pos", &self.file_pos) .field("data", &self.data) .field("unpack_xattrs", &self.unpack_xattrs) .field("preserve_permissions", &self.preserve_permissions) .field("preserve_mtime", &self.preserve_mtime) .field("read_state", &self.read_state) .finish() } } pub enum EntryIo { Pad(io::Take), Data(io::Take), } impl fmt::Debug for EntryIo { fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result { match self { EntryIo::Pad(t) => write!(f, "EntryIo::Pad({})", t.limit()), EntryIo::Data(t) => write!(f, "EntryIo::Data({})", t.limit()), } } } /// When unpacking items the unpacked thing is returned to allow custom /// additional handling by users. Today the File is returned, in future /// the enum may be extended with kinds for links, directories etc. #[derive(Debug)] #[non_exhaustive] pub enum Unpacked { /// A file was unpacked. File(fs::File), /// A directory, hardlink, symlink, or other node was unpacked. Other, } impl Entry { /// Returns the path name for this entry. /// /// This method may fail if the pathname is not valid Unicode and this is /// called on a Windows platform. /// /// Note that this function will convert any `\` characters to directory /// separators, and it will not always return the same value as /// `self.header().path()` as some archive formats have support for longer /// path names described in separate entries. /// /// It is recommended to use this method instead of inspecting the `header` /// directly to ensure that various archive formats are handled correctly. pub fn path(&self) -> io::Result> { self.fields.path() } /// Returns the raw bytes listed for this entry. /// /// Note that this function will convert any `\` characters to directory /// separators, and it will not always return the same value as /// `self.header().path_bytes()` as some archive formats have support for /// longer path names described in separate entries. pub fn path_bytes(&self) -> Cow<[u8]> { self.fields.path_bytes() } /// Returns the link name for this entry, if any is found. /// /// This method may fail if the pathname is not valid Unicode and this is /// called on a Windows platform. `Ok(None)` being returned, however, /// indicates that the link name was not present. /// /// Note that this function will convert any `\` characters to directory /// separators, and it will not always return the same value as /// `self.header().link_name()` as some archive formats have support for /// longer path names described in separate entries. /// /// It is recommended to use this method instead of inspecting the `header` /// directly to ensure that various archive formats are handled correctly. pub fn link_name(&self) -> io::Result>> { self.fields.link_name() } /// Returns the link name for this entry, in bytes, if listed. /// /// Note that this will not always return the same value as /// `self.header().link_name_bytes()` as some archive formats have support for /// longer path names described in separate entries. pub fn link_name_bytes(&self) -> Option> { self.fields.link_name_bytes() } /// Returns an iterator over the pax extensions contained in this entry. /// /// Pax extensions are a form of archive where extra metadata is stored in /// key/value pairs in entries before the entry they're intended to /// describe. For example this can be used to describe long file name or /// other metadata like atime/ctime/mtime in more precision. /// /// The returned iterator will yield key/value pairs for each extension. /// /// `None` will be returned if this entry does not indicate that it itself /// contains extensions, or if there were no previous extensions describing /// it. /// /// Note that global pax extensions are intended to be applied to all /// archive entries. /// /// Also note that this function will read the entire entry if the entry /// itself is a list of extensions. pub async fn pax_extensions(&mut self) -> io::Result>> { self.fields.pax_extensions().await } /// Returns access to the header of this entry in the archive. /// /// This provides access to the metadata for this entry in the archive. pub fn header(&self) -> &Header { &self.fields.header } /// Returns the starting position, in bytes, of the header of this entry in /// the archive. /// /// The header is always a contiguous section of 512 bytes, so if the /// underlying reader implements `Seek`, then the slice from `header_pos` to /// `header_pos + 512` contains the raw header bytes. pub fn raw_header_position(&self) -> u64 { self.fields.header_pos } /// Returns the starting position, in bytes, of the file of this entry in /// the archive. /// /// If the file of this entry is continuous (e.g. not a sparse file), and /// if the underlying reader implements `Seek`, then the slice from /// `file_pos` to `file_pos + entry_size` contains the raw file bytes. pub fn raw_file_position(&self) -> u64 { self.fields.file_pos } /// Writes this file to the specified location. /// /// This function will write the entire contents of this file into the /// location specified by `dst`. Metadata will also be propagated to the /// path `dst`. /// /// This function will create a file at the path `dst`, and it is required /// that the intermediate directories are created. Any existing file at the /// location `dst` will be overwritten. /// /// > **Note**: This function does not have as many sanity checks as /// > `Archive::unpack` or `Entry::unpack_in`. As a result if you're /// > thinking of unpacking untrusted tarballs you may want to review the /// > implementations of the previous two functions and perhaps implement /// > similar logic yourself. /// /// # Examples /// /// ```no_run /// # fn main() -> Result<(), Box> { tokio::runtime::Runtime::new().unwrap().block_on(async { /// # /// use tokio::fs::File; /// use tokio_tar::Archive; /// use tokio_stream::*; /// /// let mut ar = Archive::new(File::open("foo.tar").await?); /// let mut entries = ar.entries()?; /// let mut i = 0; /// while let Some(file) = entries.next().await { /// let mut file = file?; /// file.unpack(format!("file-{}", i)).await?; /// i += 1; /// } /// # /// # Ok(()) }) } /// ``` pub async fn unpack>(&mut self, dst: P) -> io::Result { self.fields.unpack(None, dst.as_ref()).await } /// Extracts this file under the specified path, avoiding security issues. /// /// This function will write the entire contents of this file into the /// location obtained by appending the path of this file in the archive to /// `dst`, creating any intermediate directories if needed. Metadata will /// also be propagated to the path `dst`. Any existing file at the location /// `dst` will be overwritten. /// /// This function carefully avoids writing outside of `dst`. If the file has /// a '..' in its path, this function will skip it and return false. /// /// # Examples /// /// ```no_run /// # fn main() -> Result<(), Box> { tokio::runtime::Runtime::new().unwrap().block_on(async { /// # /// use tokio::{fs::File, stream::*}; /// use tokio_tar::Archive; /// use tokio_stream::*; /// /// let mut ar = Archive::new(File::open("foo.tar").await?); /// let mut entries = ar.entries()?; /// let mut i = 0; /// while let Some(file) = entries.next().await { /// let mut file = file.unwrap(); /// file.unpack_in("target").await?; /// i += 1; /// } /// # /// # Ok(()) }) } /// ``` pub async fn unpack_in>(&mut self, dst: P) -> io::Result> { let dst = dst.as_ref().canonicalize()?; let mut memo = FxHashSet::default(); self.fields.unpack_in(&dst, &mut memo).await } /// Extracts this file under the specified path, avoiding security issues. /// /// Like [`unpack_in`], but memoizes the set of validated paths to avoid /// redundant filesystem operations and assumes that the destination path /// is already canonicalized. pub async fn unpack_in_raw>( &mut self, dst: P, memo: &mut FxHashSet, ) -> io::Result> { self.fields.unpack_in(dst.as_ref(), memo).await } /// Indicate whether extended file attributes (xattrs on Unix) are preserved /// when unpacking this entry. /// /// This flag is disabled by default and is currently only implemented on /// Unix using xattr support. This may eventually be implemented for /// Windows, however, if other archive implementations are found which do /// this as well. pub fn set_unpack_xattrs(&mut self, unpack_xattrs: bool) { self.fields.unpack_xattrs = unpack_xattrs; } /// Indicate whether extended permissions (like suid on Unix) are preserved /// when unpacking this entry. /// /// This flag is disabled by default and is currently only implemented on /// Unix. pub fn set_preserve_permissions(&mut self, preserve: bool) { self.fields.preserve_permissions = preserve; } /// Indicate whether access time information is preserved when unpacking /// this entry. /// /// This flag is enabled by default. pub fn set_preserve_mtime(&mut self, preserve: bool) { self.fields.preserve_mtime = preserve; } } impl Read for Entry { fn poll_read( mut self: Pin<&mut Self>, cx: &mut Context<'_>, into: &mut io::ReadBuf<'_>, ) -> Poll> { Pin::new(&mut self.as_mut().fields).poll_read(cx, into) } } impl EntryFields { pub fn from(entry: Entry) -> Self { entry.fields } pub fn into_entry(self) -> Entry { Entry { fields: self, _ignored: marker::PhantomData, } } pub(crate) fn poll_read_all( self: Pin<&mut Self>, cx: &mut Context<'_>, out: &mut Vec, ) -> Poll> { // Copied from futures::ReadToEnd match poll_read_all_internal(self, cx, out) { Poll::Ready(t) => Poll::Ready(t.map(|_| ())), Poll::Pending => Poll::Pending, } } pub async fn read_all(&mut self) -> io::Result> { // Preallocate some data but don't let ourselves get too crazy now. let cap = cmp::min(self.size, 128 * 1024); let mut v = Vec::with_capacity(cap as usize); self.read_to_end(&mut v).await.map(|_| v) } fn path(&self) -> io::Result> { bytes2path(self.path_bytes()) } fn path_bytes(&self) -> Cow<[u8]> { match self.long_pathname { Some(ref bytes) => { if let Some(&0) = bytes.last() { Cow::Borrowed(&bytes[..bytes.len() - 1]) } else { Cow::Borrowed(bytes) } } None => { if let Some(ref pax) = self.pax_extensions { let pax = pax_extensions(pax) .filter_map(|f| f.ok()) .find(|f| f.key_bytes() == b"path") .map(|f| f.value_bytes()); if let Some(field) = pax { return Cow::Borrowed(field); } } self.header.path_bytes() } } } /// Gets the path in a "lossy" way, used for error reporting ONLY. fn path_lossy(&self) -> String { String::from_utf8_lossy(&self.path_bytes()).to_string() } fn link_name(&self) -> io::Result>> { match self.link_name_bytes() { Some(bytes) => bytes2path(bytes).map(Some), None => Ok(None), } } fn link_name_bytes(&self) -> Option> { match self.long_linkname { Some(ref bytes) => { if let Some(&0) = bytes.last() { Some(Cow::Borrowed(&bytes[..bytes.len() - 1])) } else { Some(Cow::Borrowed(bytes)) } } None => { if let Some(ref pax) = self.pax_extensions { let pax = pax_extensions(pax) .filter_map(|f| f.ok()) .find(|f| f.key_bytes() == b"linkpath") .map(|f| f.value_bytes()); if let Some(field) = pax { return Some(Cow::Borrowed(field)); } } self.header.link_name_bytes() } } } async fn pax_extensions(&mut self) -> io::Result>> { if self.pax_extensions.is_none() { if !self.header.entry_type().is_pax_global_extensions() && !self.header.entry_type().is_pax_local_extensions() { return Ok(None); } self.pax_extensions = Some(self.read_all().await?); } Ok(Some(pax_extensions(self.pax_extensions.as_ref().unwrap()))) } /// Unpack the [`Entry`] into the specified destination. /// /// It's assumed that `dst` is already canonicalized, and that the memoized set of validated /// paths are tied to `dst`. async fn unpack_in( &mut self, dst: &Path, memo: &mut FxHashSet, ) -> io::Result> { // It's assumed that `dst` is already canonicalized. if cfg!(debug_assertions) { let canon_target = dst.canonicalize()?; assert_eq!(canon_target, dst, "Destination path must be canonicalized"); } // Notes regarding bsdtar 2.8.3 / libarchive 2.8.3: // * Leading '/'s are trimmed. For example, `///test` is treated as // `test`. // * If the filename contains '..', then the file is skipped when // extracting the tarball. // * '//' within a filename is effectively skipped. An error is // logged, but otherwise the effect is as if any two or more // adjacent '/'s within the filename were consolidated into one // '/'. // // Most of this is handled by the `path` module of the standard // library, but we specially handle a few cases here as well. let mut file_dst = dst.to_path_buf(); { let path = self.path().map_err(|e| { TarError::new( format!("invalid path in entry header: {}", self.path_lossy()), e, ) })?; for part in path.components() { match part { // Leading '/' characters, root paths, and '.' // components are just ignored and treated as "empty // components" Component::Prefix(..) | Component::RootDir | Component::CurDir => continue, // If any part of the filename is '..', then skip over // unpacking the file to prevent directory traversal // security issues. See, e.g.: CVE-2001-1267, // CVE-2002-0399, CVE-2005-1918, CVE-2007-4131 Component::ParentDir => return Ok(None), Component::Normal(part) => file_dst.push(part), } } } // Skip cases where only slashes or '.' parts were seen, because // this is effectively an empty filename. if *dst == *file_dst { return Ok(None); } // Skip entries without a parent (i.e. outside of FS root) let parent = match file_dst.parent() { Some(p) => p, None => return Ok(None), }; // Validate the parent, if we haven't seen it yet. if !memo.contains(parent) { self.ensure_dir_created(dst, parent).await.map_err(|e| { TarError::new(format!("failed to create `{}`", parent.display()), e) })?; self.validate_inside_dst(dst, parent).await?; memo.insert(parent.to_path_buf()); } self.unpack(Some(dst), &file_dst) .await .map_err(|e| TarError::new(format!("failed to unpack `{}`", file_dst.display()), e))?; Ok(Some(file_dst)) } /// Unpack as destination directory `dst`. async fn unpack_dir(&mut self, dst: &Path) -> io::Result<()> { // If the directory already exists just let it slide match fs::create_dir(dst).await { Ok(()) => Ok(()), Err(err) => { if err.kind() == ErrorKind::AlreadyExists { let prev = fs::metadata(dst).await; if prev.map(|m| m.is_dir()).unwrap_or(false) { return Ok(()); } } Err(Error::new( err.kind(), format!("{} when creating dir {}", err, dst.display()), )) } } } /// Returns access to the header of this entry in the archive. async fn unpack(&mut self, target_base: Option<&Path>, dst: &Path) -> io::Result { fn get_mtime(header: &Header) -> Option { header.mtime().ok().map(|mtime| { // For some more information on this see the comments in // `Header::fill_platform_from`, but the general idea is that // we're trying to avoid 0-mtime files coming out of archives // since some tools don't ingest them well. Perhaps one day // when Cargo stops working with 0-mtime archives we can remove // this. let mtime = if mtime == 0 { 1 } else { mtime }; FileTime::from_unix_time(mtime as i64, 0) }) } let kind = self.header.entry_type(); if kind.is_dir() { self.unpack_dir(dst).await?; if self.preserve_permissions { if let Ok(mode) = self.header.mode() { set_perms(dst, None, mode).await?; } } return Ok(Unpacked::Other); } else if kind.is_hard_link() || kind.is_symlink() { let src = match self.link_name()? { Some(name) => name, None => { return Err(other(&format!( "hard link listed for {} but no link name found", String::from_utf8_lossy(self.header.as_bytes()) ))); } }; if src.iter().count() == 0 { return Err(other(&format!( "symlink destination for {} is empty", String::from_utf8_lossy(self.header.as_bytes()) ))); } if kind.is_hard_link() { let link_src = match target_base { // If we're unpacking within a directory then ensure that // the destination of this hard link is both present and // inside our own directory. This is needed because we want // to make sure to not overwrite anything outside the root. // // Note that this logic is only needed for hard links // currently. With symlinks the `validate_inside_dst` which // happens before this method as part of `unpack_in` will // use canonicalization to ensure this guarantee. For hard // links though they're canonicalized to their existing path // so we need to validate at this time. Some(p) => { let link_src = p.join(src); self.validate_inside_dst(p, &link_src).await?; link_src } None => src.into_owned(), }; fs::hard_link(&link_src, dst).await.map_err(|err| { Error::new( err.kind(), format!( "{} when hard linking {} to {}", err, link_src.display(), dst.display() ), ) })?; } else { match symlink(&src, dst).await { Ok(()) => Ok(()), Err(err) => { if err.kind() == io::ErrorKind::AlreadyExists && self.overwrite { match remove_file(dst).await { Ok(()) => symlink(&src, dst).await, Err(ref e) if e.kind() == io::ErrorKind::NotFound => { symlink(&src, dst).await } Err(e) => Err(e), } } else { Err(err) } } }?; if self.preserve_mtime { if let Some(mtime) = get_mtime(&self.header) { filetime::set_symlink_file_times(dst, mtime, mtime).map_err(|e| { TarError::new(format!("failed to set mtime for `{}`", dst.display()), e) })?; } } }; return Ok(Unpacked::Other); #[cfg(target_arch = "wasm32")] #[allow(unused_variables)] async fn symlink(src: &Path, dst: &Path) -> io::Result<()> { Err(io::Error::new(io::ErrorKind::Other, "Not implemented")) } #[cfg(windows)] async fn symlink(src: &Path, dst: &Path) -> io::Result<()> { let (src, dst) = (src.to_owned(), dst.to_owned()); tokio::task::spawn_blocking(|| std::os::windows::fs::symlink_file(src, dst)) .await .unwrap() } #[cfg(unix)] async fn symlink(src: &Path, dst: &Path) -> io::Result<()> { tokio::fs::symlink(src, dst).await } } else if kind.is_pax_global_extensions() || kind.is_pax_local_extensions() || kind.is_gnu_longname() || kind.is_gnu_longlink() { return Ok(Unpacked::Other); }; // Old BSD-tar compatibility. // Names that have a trailing slash should be treated as a directory. // Only applies to old headers. if self.header.as_ustar().is_none() && self.path_bytes().ends_with(b"/") { self.unpack_dir(dst).await?; if self.preserve_permissions { if let Ok(mode) = self.header.mode() { set_perms(dst, None, mode).await?; } } return Ok(Unpacked::Other); } // Note the lack of `else` clause above. According to the FreeBSD // documentation: // // > A POSIX-compliant implementation must treat any unrecognized // > typeflag value as a regular file. // // As a result if we don't recognize the kind we just write out the file // as we would normally. // Ensure we write a new file rather than overwriting in-place which // is attackable; if an existing file is found unlink it. async fn open(dst: &Path) -> io::Result { OpenOptions::new() .write(true) .create_new(true) .open(dst) .await } let mut f = async { let mut f = match open(dst).await { Ok(f) => Ok(f), Err(err) => { if err.kind() == ErrorKind::AlreadyExists && self.overwrite { match fs::remove_file(dst).await { Ok(()) => open(dst).await, Err(ref e) if e.kind() == io::ErrorKind::NotFound => open(dst).await, Err(e) => Err(e), } } else { Err(err) } } }?; let size = usize::try_from(self.size).unwrap_or(usize::MAX); let capacity = cmp::min(size, 128 * 1024); let mut writer = io::BufWriter::with_capacity(capacity, &mut f); for io in self.data.drain(..) { match io { EntryIo::Data(mut d) => { let expected = d.limit(); if io::copy(&mut d, &mut writer).await? != expected { return Err(other("failed to write entire file")); } } EntryIo::Pad(d) => { // TODO: checked cast to i64 let pad_len = d.limit() as i64; writer.flush().await?; let f = writer.get_mut(); let new_size = f.seek(SeekFrom::Current(pad_len)).await?; f.set_len(new_size).await?; } } } writer.flush().await?; Ok::(f) } .await .map_err(|e| { let header = self.header.path_bytes(); TarError::new( format!( "failed to unpack `{}` into `{}`", String::from_utf8_lossy(&header), dst.display() ), e, ) })?; if self.preserve_mtime { if let Some(mtime) = get_mtime(&self.header) { filetime::set_file_times(dst, mtime, mtime).map_err(|e| { TarError::new(format!("failed to set mtime for `{}`", dst.display()), e) })?; } } if self.preserve_permissions { if let Ok(mode) = self.header.mode() { set_perms(dst, Some(&mut f), mode).await?; } } if self.unpack_xattrs { set_xattrs(self, dst).await?; } return Ok(Unpacked::File(f)); async fn set_perms( dst: &Path, f: Option<&mut fs::File>, mode: u32, ) -> Result<(), TarError> { _set_perms(dst, f, mode).await.map_err(|e| { TarError::new( format!( "failed to set permissions to {:o} \ for `{}`", mode, dst.display() ), e, ) }) } #[cfg(unix)] async fn _set_perms(dst: &Path, f: Option<&mut fs::File>, mode: u32) -> io::Result<()> { use std::os::unix::prelude::*; let perm = std::fs::Permissions::from_mode(mode as _); match f { Some(f) => f.set_permissions(perm).await, None => fs::set_permissions(dst, perm).await, } } #[cfg(windows)] async fn _set_perms(dst: &Path, f: Option<&mut fs::File>, mode: u32) -> io::Result<()> { if mode & 0o200 == 0o200 { return Ok(()); } match f { Some(f) => { let mut perm = f.metadata().await?.permissions(); perm.set_readonly(true); f.set_permissions(perm).await } None => { let mut perm = fs::metadata(dst).await?.permissions(); perm.set_readonly(true); fs::set_permissions(dst, perm).await } } } #[cfg(target_arch = "wasm32")] #[allow(unused_variables)] async fn _set_perms(dst: &Path, f: Option<&mut fs::File>, mode: u32) -> io::Result<()> { Err(io::Error::new(io::ErrorKind::Other, "Not implemented")) } #[cfg(all(unix, feature = "xattr"))] async fn set_xattrs( me: &mut EntryFields, dst: &Path, ) -> io::Result<()> { use std::{ffi::OsStr, os::unix::prelude::*}; let exts = match me.pax_extensions().await { Ok(Some(e)) => e, _ => return Ok(()), }; let exts = exts .filter_map(|e| e.ok()) .filter_map(|e| { let key = e.key_bytes(); let prefix = b"SCHILY.xattr."; key.strip_prefix(prefix).map(|rest| (rest, e)) }) .map(|(key, e)| (OsStr::from_bytes(key), e.value_bytes())); for (key, value) in exts { xattr::set(dst, key, value).map_err(|e| { TarError::new( format!( "failed to set extended \ attributes to {}. \ Xattrs: key={:?}, value={:?}.", dst.display(), key, String::from_utf8_lossy(value) ), e, ) })?; } Ok(()) } // Windows does not completely support posix xattrs // https://en.wikipedia.org/wiki/Extended_file_attributes#Windows_NT #[cfg(any(windows, not(feature = "xattr"), target_arch = "wasm32"))] async fn set_xattrs(_: &mut EntryFields, _: &Path) -> io::Result<()> { Ok(()) } } async fn ensure_dir_created(&self, dst: &Path, dir: &Path) -> io::Result<()> { let mut ancestor = dir; let mut dirs_to_create = Vec::new(); while tokio::fs::symlink_metadata(ancestor).await.is_err() { dirs_to_create.push(ancestor); if let Some(parent) = ancestor.parent() { ancestor = parent; } else { break; } } for ancestor in dirs_to_create.into_iter().rev() { if let Some(parent) = ancestor.parent() { self.validate_inside_dst(dst, parent).await?; } fs::create_dir_all(ancestor).await?; } Ok(()) } async fn validate_inside_dst(&self, dst: &Path, file_dst: &Path) -> io::Result<()> { // Abort if target (canonical) parent is outside of `dst` let canon_parent = file_dst.canonicalize().map_err(|err| { Error::new( err.kind(), format!("{} while canonicalizing {}", err, file_dst.display()), ) })?; if !canon_parent.starts_with(dst) { let err = TarError::new( format!( "trying to unpack outside of destination path: {}", dst.display() ), // TODO: use ErrorKind::InvalidInput here? (minor breaking change) Error::new(ErrorKind::Other, "Invalid argument"), ); return Err(err.into()); } Ok(()) } } impl Read for EntryFields { fn poll_read( self: Pin<&mut Self>, cx: &mut Context<'_>, into: &mut io::ReadBuf<'_>, ) -> Poll> { let this = self.get_mut(); loop { if this.read_state.is_none() { this.read_state = this.data.pop_front(); } if let Some(ref mut io) = &mut this.read_state { let start = into.filled().len(); let ret = Pin::new(io).poll_read(cx, into); match ret { Poll::Ready(Ok(())) if into.filled().len() == start => { this.read_state = None; if this.data.is_empty() { return Poll::Ready(Ok(())); } continue; } Poll::Ready(Ok(())) => { return Poll::Ready(Ok(())); } Poll::Ready(Err(err)) => { return Poll::Ready(Err(err)); } Poll::Pending => { return Poll::Pending; } } } else { // Unable to pull another value from `data`, so we are done. return Poll::Ready(Ok(())); } } } } impl Read for EntryIo { fn poll_read( self: Pin<&mut Self>, cx: &mut Context<'_>, into: &mut io::ReadBuf<'_>, ) -> Poll> { match self.get_mut() { EntryIo::Pad(ref mut io) => Pin::new(io).poll_read(cx, into), EntryIo::Data(ref mut io) => Pin::new(io).poll_read(cx, into), } } } struct Guard<'a> { buf: &'a mut Vec, len: usize, } impl Drop for Guard<'_> { fn drop(&mut self) { unsafe { self.buf.set_len(self.len); } } } fn poll_read_all_internal( mut rd: Pin<&mut R>, cx: &mut Context<'_>, buf: &mut Vec, ) -> Poll> { let mut g = Guard { len: buf.len(), buf, }; let ret; loop { if g.len == g.buf.len() { unsafe { g.buf.reserve(32); let capacity = g.buf.capacity(); g.buf.set_len(capacity); let buf = &mut g.buf[g.len..]; std::ptr::write_bytes(buf.as_mut_ptr(), 0, buf.len()); } } let mut read_buf = io::ReadBuf::new(&mut g.buf[g.len..]); match futures_core::ready!(rd.as_mut().poll_read(cx, &mut read_buf)) { Ok(()) if read_buf.filled().is_empty() => { ret = Poll::Ready(Ok(g.len)); break; } Ok(()) => g.len += read_buf.filled().len(), Err(e) => { ret = Poll::Ready(Err(e)); break; } } } ret } astral-tokio-tar-0.5.1/src/entry_type.rs000064400000000000000000000130131046102023000163610ustar 00000000000000// See https://en.wikipedia.org/wiki/Tar_%28computing%29#UStar_format /// Indicate for the type of file described by a header. /// /// Each `Header` has an `entry_type` method returning an instance of this type /// which can be used to inspect what the header is describing. /// /// A non-exhaustive enum representing the possible entry types #[derive(Clone, Copy, PartialEq, Eq, Debug)] #[non_exhaustive] pub enum EntryType { /// Regular file Regular, /// Hard link Link, /// Symbolic link Symlink, /// Character device Char, /// Block device Block, /// Directory Directory, /// Named pipe (fifo) Fifo, /// Implementation-defined 'high-performance' type, treated as regular file Continuous, /// GNU extension - long file name GNULongName, /// GNU extension - long link name (link target) GNULongLink, /// GNU extension - sparse file GNUSparse, /// Global extended header XGlobalHeader, /// Extended Header XHeader, /// Unknown header, Other(u8), } impl EntryType { /// Creates a new entry type from a raw byte. /// /// Note that the other named constructors of entry type may be more /// appropriate to create a file type from. pub fn new(byte: u8) -> EntryType { match byte { b'\x00' | b'0' => EntryType::Regular, b'1' => EntryType::Link, b'2' => EntryType::Symlink, b'3' => EntryType::Char, b'4' => EntryType::Block, b'5' => EntryType::Directory, b'6' => EntryType::Fifo, b'7' => EntryType::Continuous, b'x' => EntryType::XHeader, b'g' => EntryType::XGlobalHeader, b'L' => EntryType::GNULongName, b'K' => EntryType::GNULongLink, b'S' => EntryType::GNUSparse, other => EntryType::Other(other), } } /// Returns the raw underlying byte that this entry type represents. pub fn as_byte(self) -> u8 { match self { EntryType::Regular => b'0', EntryType::Link => b'1', EntryType::Symlink => b'2', EntryType::Char => b'3', EntryType::Block => b'4', EntryType::Directory => b'5', EntryType::Fifo => b'6', EntryType::Continuous => b'7', EntryType::XHeader => b'x', EntryType::XGlobalHeader => b'g', EntryType::GNULongName => b'L', EntryType::GNULongLink => b'K', EntryType::GNUSparse => b'S', EntryType::Other(other) => other, } } /// Creates a new entry type representing a regular file. pub fn file() -> EntryType { EntryType::Regular } /// Creates a new entry type representing a hard link. pub fn hard_link() -> EntryType { EntryType::Link } /// Creates a new entry type representing a symlink. pub fn symlink() -> EntryType { EntryType::Symlink } /// Creates a new entry type representing a character special device. pub fn character_special() -> EntryType { EntryType::Char } /// Creates a new entry type representing a block special device. pub fn block_special() -> EntryType { EntryType::Block } /// Creates a new entry type representing a directory. pub fn dir() -> EntryType { EntryType::Directory } /// Creates a new entry type representing a FIFO. pub fn fifo() -> EntryType { EntryType::Fifo } /// Creates a new entry type representing a contiguous file. pub fn contiguous() -> EntryType { EntryType::Continuous } /// Returns whether this type represents a regular file. pub fn is_file(self) -> bool { self == EntryType::Regular } /// Returns whether this type represents a hard link. pub fn is_hard_link(self) -> bool { self == EntryType::Link } /// Returns whether this type represents a symlink. pub fn is_symlink(self) -> bool { self == EntryType::Symlink } /// Returns whether this type represents a character special device. pub fn is_character_special(self) -> bool { self == EntryType::Char } /// Returns whether this type represents a block special device. pub fn is_block_special(self) -> bool { self == EntryType::Block } /// Returns whether this type represents a directory. pub fn is_dir(self) -> bool { self == EntryType::Directory } /// Returns whether this type represents a FIFO. pub fn is_fifo(self) -> bool { self == EntryType::Fifo } /// Returns whether this type represents a contiguous file. pub fn is_contiguous(self) -> bool { self == EntryType::Continuous } /// Returns whether this type represents a GNU long name header. pub fn is_gnu_longname(self) -> bool { self == EntryType::GNULongName } /// Returns whether this type represents a GNU sparse header. pub fn is_gnu_sparse(self) -> bool { self == EntryType::GNUSparse } /// Returns whether this type represents a GNU long link header. pub fn is_gnu_longlink(self) -> bool { self == EntryType::GNULongLink } /// Returns whether this type represents a GNU long name header. pub fn is_pax_global_extensions(self) -> bool { self == EntryType::XGlobalHeader } /// Returns whether this type represents a GNU long link header. pub fn is_pax_local_extensions(self) -> bool { self == EntryType::XHeader } } astral-tokio-tar-0.5.1/src/error.rs000064400000000000000000000014161046102023000153140ustar 00000000000000use std::borrow::Cow; use std::error; use std::fmt; use std::io::{self, Error}; #[derive(Debug)] pub struct TarError { desc: Cow<'static, str>, io: io::Error, } impl TarError { pub fn new(desc: impl Into>, err: Error) -> TarError { TarError { desc: desc.into(), io: err, } } } impl error::Error for TarError { fn description(&self) -> &str { &self.desc } fn source(&self) -> Option<&(dyn error::Error + 'static)> { Some(&self.io) } } impl fmt::Display for TarError { fn fmt(&self, f: &mut fmt::Formatter) -> fmt::Result { self.desc.fmt(f) } } impl From for Error { fn from(t: TarError) -> Error { Error::new(t.io.kind(), t) } } astral-tokio-tar-0.5.1/src/header.rs000064400000000000000000001545541046102023000154270ustar 00000000000000#[cfg(unix)] use std::os::unix::prelude::*; #[cfg(windows)] use std::os::windows::prelude::*; use std::{ borrow::Cow, fmt, fs::Metadata, iter, iter::{once, repeat}, mem, path::{Component, Path, PathBuf}, str, }; use tokio::io; use crate::{other, EntryType}; /// A deterministic, arbitrary, non-zero timestamp that use used as `mtime` /// of headers when [`HeaderMode::Deterministic`] is used. /// /// This value, chosen after careful deliberation, corresponds to _Jul 23, 2006_, /// which is the date of the first commit for what would become Rust. #[cfg(any(unix, windows))] const DETERMINISTIC_TIMESTAMP: u64 = 1153704088; pub(crate) const BLOCK_SIZE: u64 = 512; /// Representation of the header of an entry in an archive #[repr(C)] #[allow(missing_docs)] pub struct Header { bytes: [u8; BLOCK_SIZE as usize], } /// Declares the information that should be included when filling a Header /// from filesystem metadata. #[derive(Clone, Copy, PartialEq, Eq, Debug)] #[non_exhaustive] pub enum HeaderMode { /// All supported metadata, including mod/access times and ownership will /// be included. Complete, /// Only metadata that is directly relevant to the identity of a file will /// be included. In particular, ownership and mod/access times are excluded. Deterministic, } /// Representation of the header of an entry in an archive #[repr(C)] #[allow(missing_docs)] pub struct OldHeader { pub name: [u8; 100], pub mode: [u8; 8], pub uid: [u8; 8], pub gid: [u8; 8], pub size: [u8; 12], pub mtime: [u8; 12], pub cksum: [u8; 8], pub linkflag: [u8; 1], pub linkname: [u8; 100], pub pad: [u8; 255], } /// Representation of the header of an entry in an archive #[repr(C)] #[allow(missing_docs)] pub struct UstarHeader { pub name: [u8; 100], pub mode: [u8; 8], pub uid: [u8; 8], pub gid: [u8; 8], pub size: [u8; 12], pub mtime: [u8; 12], pub cksum: [u8; 8], pub typeflag: [u8; 1], pub linkname: [u8; 100], // UStar format pub magic: [u8; 6], pub version: [u8; 2], pub uname: [u8; 32], pub gname: [u8; 32], pub dev_major: [u8; 8], pub dev_minor: [u8; 8], pub prefix: [u8; 155], pub pad: [u8; 12], } /// Representation of the header of an entry in an archive #[repr(C)] #[allow(missing_docs)] pub struct GnuHeader { pub name: [u8; 100], pub mode: [u8; 8], pub uid: [u8; 8], pub gid: [u8; 8], pub size: [u8; 12], pub mtime: [u8; 12], pub cksum: [u8; 8], pub typeflag: [u8; 1], pub linkname: [u8; 100], // GNU format pub magic: [u8; 6], pub version: [u8; 2], pub uname: [u8; 32], pub gname: [u8; 32], pub dev_major: [u8; 8], pub dev_minor: [u8; 8], pub atime: [u8; 12], pub ctime: [u8; 12], pub offset: [u8; 12], pub longnames: [u8; 4], pub unused: [u8; 1], pub sparse: [GnuSparseHeader; 4], pub isextended: [u8; 1], pub realsize: [u8; 12], pub pad: [u8; 17], } /// Description of the header of a spare entry. /// /// Specifies the offset/number of bytes of a chunk of data in octal. #[repr(C)] #[allow(missing_docs)] pub struct GnuSparseHeader { pub offset: [u8; 12], pub numbytes: [u8; 12], } /// Representation of the entry found to represent extended GNU sparse files. /// /// When a `GnuHeader` has the `isextended` flag set to `1` then the contents of /// the next entry will be one of these headers. #[repr(C)] #[allow(missing_docs)] pub struct GnuExtSparseHeader { pub sparse: [GnuSparseHeader; 21], pub isextended: [u8; 1], pub padding: [u8; 7], } impl Header { /// Creates a new blank GNU header. /// /// The GNU style header is the default for this library and allows various /// extensions such as long path names, long link names, and setting the /// atime/ctime metadata attributes of files. pub fn new_gnu() -> Header { let mut header = Header { bytes: [0; BLOCK_SIZE as usize], }; unsafe { let gnu = cast_mut::<_, GnuHeader>(&mut header); gnu.magic = *b"ustar "; gnu.version = *b" \0"; } header.set_mtime(0); header } /// Creates a new blank UStar header. /// /// The UStar style header is an extension of the original archive header /// which enables some extra metadata along with storing a longer (but not /// too long) path name. /// /// UStar is also the basis used for pax archives. pub fn new_ustar() -> Header { let mut header = Header { bytes: [0; BLOCK_SIZE as usize], }; unsafe { let gnu = cast_mut::<_, UstarHeader>(&mut header); gnu.magic = *b"ustar\0"; gnu.version = *b"00"; } header.set_mtime(0); header } /// Creates a new blank old header. /// /// This header format is the original archive header format which all other /// versions are compatible with (e.g. they are a superset). This header /// format limits the path name limit and isn't able to contain extra /// metadata like atime/ctime. pub fn new_old() -> Header { let mut header = Header { bytes: [0; BLOCK_SIZE as usize], }; header.set_mtime(0); header } fn is_ustar(&self) -> bool { let ustar = unsafe { cast::<_, UstarHeader>(self) }; ustar.magic[..] == b"ustar\0"[..] && ustar.version[..] == b"00"[..] } fn is_gnu(&self) -> bool { let ustar = unsafe { cast::<_, UstarHeader>(self) }; ustar.magic[..] == b"ustar "[..] && ustar.version[..] == b" \0"[..] } /// View this archive header as a raw "old" archive header. /// /// This view will always succeed as all archive header formats will fill /// out at least the fields specified in the old header format. pub fn as_old(&self) -> &OldHeader { unsafe { cast(self) } } /// Same as `as_old`, but the mutable version. pub fn as_old_mut(&mut self) -> &mut OldHeader { unsafe { cast_mut(self) } } /// View this archive header as a raw UStar archive header. /// /// The UStar format is an extension to the tar archive format which enables /// longer pathnames and a few extra attributes such as the group and user /// name. /// /// This cast may not succeed as this function will test whether the /// magic/version fields of the UStar format have the appropriate values, /// returning `None` if they aren't correct. pub fn as_ustar(&self) -> Option<&UstarHeader> { if self.is_ustar() { Some(unsafe { cast(self) }) } else { None } } /// Same as `as_ustar_mut`, but the mutable version. pub fn as_ustar_mut(&mut self) -> Option<&mut UstarHeader> { if self.is_ustar() { Some(unsafe { cast_mut(self) }) } else { None } } /// View this archive header as a raw GNU archive header. /// /// The GNU format is an extension to the tar archive format which enables /// longer pathnames and a few extra attributes such as the group and user /// name. /// /// This cast may not succeed as this function will test whether the /// magic/version fields of the GNU format have the appropriate values, /// returning `None` if they aren't correct. pub fn as_gnu(&self) -> Option<&GnuHeader> { if self.is_gnu() { Some(unsafe { cast(self) }) } else { None } } /// Same as `as_gnu`, but the mutable version. pub fn as_gnu_mut(&mut self) -> Option<&mut GnuHeader> { if self.is_gnu() { Some(unsafe { cast_mut(self) }) } else { None } } /// Treats the given byte slice as a header. /// /// Panics if the length of the passed slice is not equal to 512. pub fn from_byte_slice(bytes: &[u8]) -> &Header { assert_eq!(bytes.len(), mem::size_of::
()); assert_eq!(mem::align_of_val(bytes), mem::align_of::
()); unsafe { &*(bytes.as_ptr() as *const Header) } } /// Returns a view into this header as a byte array. pub fn as_bytes(&self) -> &[u8; BLOCK_SIZE as usize] { &self.bytes } /// Returns a view into this header as a byte array. pub fn as_mut_bytes(&mut self) -> &mut [u8; BLOCK_SIZE as usize] { &mut self.bytes } /// Blanket sets the metadata in this header from the metadata argument /// provided. /// /// This is useful for initializing a `Header` from the OS's metadata from a /// file. By default, this will use `HeaderMode::Complete` to include all /// metadata. pub fn set_metadata(&mut self, meta: &Metadata) { self.fill_from(meta, HeaderMode::Complete); } /// Sets only the metadata relevant to the given HeaderMode in this header /// from the metadata argument provided. pub fn set_metadata_in_mode(&mut self, meta: &Metadata, mode: HeaderMode) { self.fill_from(meta, mode); } /// Returns the size of entry's data this header represents. /// /// This is different from `Header::size` for sparse files, which have /// some longer `size()` but shorter `entry_size()`. The `entry_size()` /// listed here should be the number of bytes in the archive this header /// describes. /// /// May return an error if the field is corrupted. pub fn entry_size(&self) -> io::Result { num_field_wrapper_from(&self.as_old().size).map_err(|err| { io::Error::new( err.kind(), format!("{} when getting size for {}", err, self.path_lossy()), ) }) } /// Returns the file size this header represents. /// /// May return an error if the field is corrupted. pub fn size(&self) -> io::Result { if self.entry_type().is_gnu_sparse() { self.as_gnu() .ok_or_else(|| other("sparse header was not a gnu header")) .and_then(|h| h.real_size()) } else { self.entry_size() } } /// Encodes the `size` argument into the size field of this header. pub fn set_size(&mut self, size: u64) { num_field_wrapper_into(&mut self.as_old_mut().size, size); } /// Returns the raw path name stored in this header. /// /// This method may fail if the pathname is not valid Unicode and this is /// called on a Windows platform. /// /// Note that this function will convert any `\` characters to directory /// separators. pub fn path(&self) -> io::Result> { bytes2path(self.path_bytes()) } /// Returns the pathname stored in this header as a byte array. /// /// This function is guaranteed to succeed, but you may wish to call the /// `path` method to convert to a `Path`. /// /// Note that this function will convert any `\` characters to directory /// separators. pub fn path_bytes(&self) -> Cow<[u8]> { if let Some(ustar) = self.as_ustar() { ustar.path_bytes() } else { let name = truncate(&self.as_old().name); Cow::Borrowed(name) } } /// Gets the path in a "lossy" way, used for error reporting ONLY. fn path_lossy(&self) -> String { String::from_utf8_lossy(&self.path_bytes()).to_string() } /// Sets the path name for this header. /// /// This function will set the pathname listed in this header, encoding it /// in the appropriate format. May fail if the path is too long or if the /// path specified is not Unicode and this is a Windows platform. Will /// strip out any "." path component, which signifies the current directory. /// /// Note: This function does not support names over 100 bytes, or paths /// over 255 bytes, even for formats that support longer names. Instead, /// use `Builder` methods to insert a long-name extension at the same time /// as the file content. pub fn set_path>(&mut self, p: P) -> io::Result<()> { self.set_path_inner(p.as_ref(), false) } // Sets the truncated path for GNU header // // Same as set_path but skips some validations. pub(crate) fn set_truncated_path_for_gnu_header>( &mut self, p: P, ) -> std::io::Result<()> { self.set_path_inner(p.as_ref(), true) } fn set_path_inner( &mut self, path: &Path, is_truncated_gnu_long_path: bool, ) -> std::io::Result<()> { if let Some(ustar) = self.as_ustar_mut() { return ustar.set_path(path); } if is_truncated_gnu_long_path { copy_path_into_gnu_long(&mut self.as_old_mut().name, path, false) } else { copy_path_into(&mut self.as_old_mut().name, path, false) } .map_err(|err| { io::Error::new( err.kind(), format!("{} when setting path for {}", err, self.path_lossy()), ) }) } /// Returns the link name stored in this header, if any is found. /// /// This method may fail if the pathname is not valid Unicode and this is /// called on a Windows platform. `Ok(None)` being returned, however, /// indicates that the link name was not present. /// /// Note that this function will convert any `\` characters to directory /// separators. pub fn link_name(&self) -> io::Result>> { match self.link_name_bytes() { Some(bytes) => bytes2path(bytes).map(Some), None => Ok(None), } } /// Returns the link name stored in this header as a byte array, if any. /// /// This function is guaranteed to succeed, but you may wish to call the /// `link_name` method to convert to a `Path`. /// /// Note that this function will convert any `\` characters to directory /// separators. pub fn link_name_bytes(&self) -> Option> { let old = self.as_old(); if old.linkname[0] != 0 { Some(Cow::Borrowed(truncate(&old.linkname))) } else { None } } /// Sets the link name for this header. /// /// This function will set the linkname listed in this header, encoding it /// in the appropriate format. May fail if the link name is too long or if /// the path specified is not Unicode and this is a Windows platform. Will /// strip out any "." path component, which signifies the current directory. pub fn set_link_name>(&mut self, p: P) -> io::Result<()> { self._set_link_name(p.as_ref()) } fn _set_link_name(&mut self, path: &Path) -> io::Result<()> { copy_path_into(&mut self.as_old_mut().linkname, path, true).map_err(|err| { io::Error::new( err.kind(), format!("{} when setting link name for {}", err, self.path_lossy()), ) }) } /// Returns the mode bits for this file /// /// May return an error if the field is corrupted. pub fn mode(&self) -> io::Result { octal_from(&self.as_old().mode) .map(|u| u as u32) .map_err(|err| { io::Error::new( err.kind(), format!("{} when getting mode for {}", err, self.path_lossy()), ) }) } /// Encodes the `mode` provided into this header. pub fn set_mode(&mut self, mode: u32) { octal_into(&mut self.as_old_mut().mode, mode); } /// Returns the value of the owner's user ID field /// /// May return an error if the field is corrupted. pub fn uid(&self) -> io::Result { num_field_wrapper_from(&self.as_old().uid).map_err(|err| { io::Error::new( err.kind(), format!("{} when getting uid for {}", err, self.path_lossy()), ) }) } /// Encodes the `uid` provided into this header. pub fn set_uid(&mut self, uid: u64) { num_field_wrapper_into(&mut self.as_old_mut().uid, uid); } /// Returns the value of the group's user ID field pub fn gid(&self) -> io::Result { num_field_wrapper_from(&self.as_old().gid).map_err(|err| { io::Error::new( err.kind(), format!("{} when getting gid for {}", err, self.path_lossy()), ) }) } /// Encodes the `gid` provided into this header. pub fn set_gid(&mut self, gid: u64) { num_field_wrapper_into(&mut self.as_old_mut().gid, gid); } /// Returns the last modification time in Unix time format pub fn mtime(&self) -> io::Result { num_field_wrapper_from(&self.as_old().mtime).map_err(|err| { io::Error::new( err.kind(), format!("{} when getting mtime for {}", err, self.path_lossy()), ) }) } /// Encodes the `mtime` provided into this header. /// /// Note that this time is typically a number of seconds passed since /// January 1, 1970. pub fn set_mtime(&mut self, mtime: u64) { num_field_wrapper_into(&mut self.as_old_mut().mtime, mtime); } /// Return the user name of the owner of this file. /// /// A return value of `Ok(Some(..))` indicates that the user name was /// present and was valid utf-8, `Ok(None)` indicates that the user name is /// not present in this archive format, and `Err` indicates that the user /// name was present but was not valid utf-8. pub fn username(&self) -> Result, str::Utf8Error> { match self.username_bytes() { Some(bytes) => str::from_utf8(bytes).map(Some), None => Ok(None), } } /// Returns the user name of the owner of this file, if present. /// /// A return value of `None` indicates that the user name is not present in /// this header format. #[allow(clippy::manual_map)] pub fn username_bytes(&self) -> Option<&[u8]> { if let Some(ustar) = self.as_ustar() { Some(ustar.username_bytes()) } else if let Some(gnu) = self.as_gnu() { Some(gnu.username_bytes()) } else { None } } /// Sets the username inside this header. /// /// This function will return an error if this header format cannot encode a /// user name or the name is too long. pub fn set_username(&mut self, name: &str) -> io::Result<()> { if let Some(ustar) = self.as_ustar_mut() { return ustar.set_username(name); } if let Some(gnu) = self.as_gnu_mut() { gnu.set_username(name) } else { Err(other("not a ustar or gnu archive, cannot set username")) } } /// Return the group name of the owner of this file. /// /// A return value of `Ok(Some(..))` indicates that the group name was /// present and was valid utf-8, `Ok(None)` indicates that the group name is /// not present in this archive format, and `Err` indicates that the group /// name was present but was not valid utf-8. pub fn groupname(&self) -> Result, str::Utf8Error> { match self.groupname_bytes() { Some(bytes) => str::from_utf8(bytes).map(Some), None => Ok(None), } } /// Returns the group name of the owner of this file, if present. /// /// A return value of `None` indicates that the group name is not present in /// this header format. #[allow(clippy::manual_map)] pub fn groupname_bytes(&self) -> Option<&[u8]> { if let Some(ustar) = self.as_ustar() { Some(ustar.groupname_bytes()) } else if let Some(gnu) = self.as_gnu() { Some(gnu.groupname_bytes()) } else { None } } /// Sets the group name inside this header. /// /// This function will return an error if this header format cannot encode a /// group name or the name is too long. pub fn set_groupname(&mut self, name: &str) -> io::Result<()> { if let Some(ustar) = self.as_ustar_mut() { return ustar.set_groupname(name); } if let Some(gnu) = self.as_gnu_mut() { gnu.set_groupname(name) } else { Err(other("not a ustar or gnu archive, cannot set groupname")) } } /// Returns the device major number, if present. /// /// This field may not be present in all archives, and it may not be /// correctly formed in all archives. `Ok(Some(..))` means it was present /// and correctly decoded, `Ok(None)` indicates that this header format does /// not include the device major number, and `Err` indicates that it was /// present and failed to decode. pub fn device_major(&self) -> io::Result> { if let Some(ustar) = self.as_ustar() { ustar.device_major().map(Some) } else if let Some(gnu) = self.as_gnu() { gnu.device_major().map(Some) } else { Ok(None) } } /// Encodes the value `major` into the dev_major field of this header. /// /// This function will return an error if this header format cannot encode a /// major device number. pub fn set_device_major(&mut self, major: u32) -> io::Result<()> { if let Some(ustar) = self.as_ustar_mut() { ustar.set_device_major(major); Ok(()) } else if let Some(gnu) = self.as_gnu_mut() { gnu.set_device_major(major); Ok(()) } else { Err(other("not a ustar or gnu archive, cannot set dev_major")) } } /// Returns the device minor number, if present. /// /// This field may not be present in all archives, and it may not be /// correctly formed in all archives. `Ok(Some(..))` means it was present /// and correctly decoded, `Ok(None)` indicates that this header format does /// not include the device minor number, and `Err` indicates that it was /// present and failed to decode. pub fn device_minor(&self) -> io::Result> { if let Some(ustar) = self.as_ustar() { ustar.device_minor().map(Some) } else if let Some(gnu) = self.as_gnu() { gnu.device_minor().map(Some) } else { Ok(None) } } /// Encodes the value `minor` into the dev_minor field of this header. /// /// This function will return an error if this header format cannot encode a /// minor device number. pub fn set_device_minor(&mut self, minor: u32) -> io::Result<()> { if let Some(ustar) = self.as_ustar_mut() { ustar.set_device_minor(minor); Ok(()) } else if let Some(gnu) = self.as_gnu_mut() { gnu.set_device_minor(minor); Ok(()) } else { Err(other("not a ustar or gnu archive, cannot set dev_minor")) } } /// Returns the type of file described by this header. pub fn entry_type(&self) -> EntryType { EntryType::new(self.as_old().linkflag[0]) } /// Sets the type of file that will be described by this header. pub fn set_entry_type(&mut self, ty: EntryType) { self.as_old_mut().linkflag = [ty.as_byte()]; } /// Returns the checksum field of this header. /// /// May return an error if the field is corrupted. pub fn cksum(&self) -> io::Result { octal_from(&self.as_old().cksum) .map(|u| u as u32) .map_err(|err| { io::Error::new( err.kind(), format!("{} when getting cksum for {}", err, self.path_lossy()), ) }) } /// Sets the checksum field of this header based on the current fields in /// this header. pub fn set_cksum(&mut self) { let cksum = self.calculate_cksum(); octal_into(&mut self.as_old_mut().cksum, cksum); } fn calculate_cksum(&self) -> u32 { let old = self.as_old(); let start = old as *const _ as usize; let cksum_start = old.cksum.as_ptr() as *const _ as usize; let offset = cksum_start - start; let len = old.cksum.len(); self.bytes[0..offset] .iter() .chain(iter::repeat(&b' ').take(len)) .chain(&self.bytes[offset + len..]) .fold(0, |a, b| a + (*b as u32)) } fn fill_from(&mut self, meta: &Metadata, mode: HeaderMode) { self.fill_platform_from(meta, mode); // Set size of directories to zero self.set_size(if meta.is_dir() || meta.file_type().is_symlink() { 0 } else { meta.len() }); if let Some(ustar) = self.as_ustar_mut() { ustar.set_device_major(0); ustar.set_device_minor(0); } if let Some(gnu) = self.as_gnu_mut() { gnu.set_device_major(0); gnu.set_device_minor(0); } } #[cfg(target_arch = "wasm32")] #[allow(unused_variables)] fn fill_platform_from(&mut self, meta: &Metadata, mode: HeaderMode) { unimplemented!(); } #[cfg(unix)] fn fill_platform_from(&mut self, meta: &Metadata, mode: HeaderMode) { match mode { HeaderMode::Complete => { self.set_mtime(meta.mtime() as u64); self.set_uid(meta.uid() as u64); self.set_gid(meta.gid() as u64); self.set_mode(meta.mode()); } HeaderMode::Deterministic => { // We could in theory set the mtime to zero here, but not all tools seem to behave // well when ingesting files with a 0 timestamp. // For example, rust-lang/cargo#9512 shows that lldb doesn't ingest files with a // zero timestamp correctly. self.set_mtime(DETERMINISTIC_TIMESTAMP); self.set_uid(0); self.set_gid(0); // Use a default umask value, but propagate the (user) execute bit. let fs_mode = if meta.is_dir() || (0o100 & meta.mode() == 0o100) { 0o755 } else { 0o644 }; self.set_mode(fs_mode); } } // Note that if we are a GNU header we *could* set atime/ctime, except // the `tar` utility doesn't do that by default and it causes problems // with 7-zip [1]. // // It's always possible to fill them out manually, so we just don't fill // it out automatically here. // // [1]: https://github.com/alexcrichton/tar-rs/issues/70 // TODO: need to bind more file types self.set_entry_type(entry_type(meta.mode())); fn entry_type(mode: u32) -> EntryType { match mode as libc::mode_t & libc::S_IFMT { libc::S_IFREG => EntryType::file(), libc::S_IFLNK => EntryType::symlink(), libc::S_IFCHR => EntryType::character_special(), libc::S_IFBLK => EntryType::block_special(), libc::S_IFDIR => EntryType::dir(), libc::S_IFIFO => EntryType::fifo(), _ => EntryType::new(b' '), } } } #[cfg(windows)] fn fill_platform_from(&mut self, meta: &Metadata, mode: HeaderMode) { // There's no concept of a file mode on Windows, so do a best approximation here. match mode { HeaderMode::Complete => { self.set_uid(0); self.set_gid(0); // The dates listed in tarballs are always seconds relative to // January 1, 1970. On Windows, however, the timestamps are returned as // dates relative to January 1, 1601 (in 100ns intervals), so we need to // add in some offset for those dates. let mtime = (meta.last_write_time() / (1_000_000_000 / 100)) - 11644473600; self.set_mtime(mtime); let fs_mode = { const FILE_ATTRIBUTE_READONLY: u32 = 0x00000001; let readonly = meta.file_attributes() & FILE_ATTRIBUTE_READONLY; match (meta.is_dir(), readonly != 0) { (true, false) => 0o755, (true, true) => 0o555, (false, false) => 0o644, (false, true) => 0o444, } }; self.set_mode(fs_mode); } HeaderMode::Deterministic => { self.set_uid(0); self.set_gid(0); self.set_mtime(DETERMINISTIC_TIMESTAMP); // see above in unix let fs_mode = if meta.is_dir() { 0o755 } else { 0o644 }; self.set_mode(fs_mode); } } let ft = meta.file_type(); self.set_entry_type(if ft.is_dir() { EntryType::dir() } else if ft.is_file() { EntryType::file() } else if ft.is_symlink() { EntryType::symlink() } else { EntryType::new(b' ') }); } fn debug_fields(&self, b: &mut fmt::DebugStruct) { if let Ok(entry_size) = self.entry_size() { b.field("entry_size", &entry_size); } if let Ok(size) = self.size() { b.field("size", &size); } if let Ok(path) = self.path() { b.field("path", &path); } if let Ok(link_name) = self.link_name() { b.field("link_name", &link_name); } if let Ok(mode) = self.mode() { b.field("mode", &DebugAsOctal(mode)); } if let Ok(uid) = self.uid() { b.field("uid", &uid); } if let Ok(gid) = self.gid() { b.field("gid", &gid); } if let Ok(mtime) = self.mtime() { b.field("mtime", &mtime); } if let Ok(username) = self.username() { b.field("username", &username); } if let Ok(groupname) = self.groupname() { b.field("groupname", &groupname); } if let Ok(device_major) = self.device_major() { b.field("device_major", &device_major); } if let Ok(device_minor) = self.device_minor() { b.field("device_minor", &device_minor); } if let Ok(cksum) = self.cksum() { b.field("cksum", &cksum); b.field("cksum_valid", &(cksum == self.calculate_cksum())); } } } struct DebugAsOctal(T); impl fmt::Debug for DebugAsOctal { fn fmt(&self, f: &mut fmt::Formatter) -> fmt::Result { fmt::Octal::fmt(&self.0, f) } } unsafe fn cast(a: &T) -> &U { assert_eq!(mem::size_of_val(a), mem::size_of::()); assert_eq!(mem::align_of_val(a), mem::align_of::()); &*(a as *const T as *const U) } unsafe fn cast_mut(a: &mut T) -> &mut U { assert_eq!(mem::size_of_val(a), mem::size_of::()); assert_eq!(mem::align_of_val(a), mem::align_of::()); &mut *(a as *mut T as *mut U) } impl Clone for Header { fn clone(&self) -> Header { Header { bytes: self.bytes } } } impl fmt::Debug for Header { fn fmt(&self, f: &mut fmt::Formatter) -> fmt::Result { if let Some(me) = self.as_ustar() { me.fmt(f) } else if let Some(me) = self.as_gnu() { me.fmt(f) } else { self.as_old().fmt(f) } } } impl OldHeader { /// Views this as a normal `Header` pub fn as_header(&self) -> &Header { unsafe { cast(self) } } /// Views this as a normal `Header` pub fn as_header_mut(&mut self) -> &mut Header { unsafe { cast_mut(self) } } } impl fmt::Debug for OldHeader { fn fmt(&self, f: &mut fmt::Formatter) -> fmt::Result { let mut f = f.debug_struct("OldHeader"); self.as_header().debug_fields(&mut f); f.finish() } } impl UstarHeader { /// See `Header::path_bytes` pub fn path_bytes(&self) -> Cow<[u8]> { if self.prefix[0] == 0 && !self.name.contains(&b'\\') { Cow::Borrowed(truncate(&self.name)) } else { let mut bytes = Vec::new(); let prefix = truncate(&self.prefix); if !prefix.is_empty() { bytes.extend_from_slice(prefix); bytes.push(b'/'); } bytes.extend_from_slice(truncate(&self.name)); Cow::Owned(bytes) } } /// Gets the path in a "lossy" way, used for error reporting ONLY. fn path_lossy(&self) -> String { String::from_utf8_lossy(&self.path_bytes()).to_string() } /// See `Header::set_path` pub fn set_path>(&mut self, p: P) -> io::Result<()> { self._set_path(p.as_ref()) } fn _set_path(&mut self, path: &Path) -> io::Result<()> { // This can probably be optimized quite a bit more, but for now just do // something that's relatively easy and readable. // // First up, if the path fits within `self.name` then we just shove it // in there. If not then we try to split it between some existing path // components where it can fit in name/prefix. To do that we peel off // enough until the path fits in `prefix`, then we try to put both // halves into their destination. let bytes = path2bytes(path)?; let (maxnamelen, maxprefixlen) = (self.name.len(), self.prefix.len()); if bytes.len() <= maxnamelen { copy_path_into(&mut self.name, path, false).map_err(|err| { io::Error::new( err.kind(), format!("{} when setting path for {}", err, self.path_lossy()), ) })?; } else { let mut prefix = path; let mut prefixlen; loop { match prefix.parent() { Some(parent) => prefix = parent, None => { return Err(other(&format!( "path cannot be split to be inserted into archive: {}", path.display() ))); } } prefixlen = path2bytes(prefix)?.len(); if prefixlen <= maxprefixlen { break; } } copy_path_into(&mut self.prefix, prefix, false).map_err(|err| { io::Error::new( err.kind(), format!("{} when setting path for {}", err, self.path_lossy()), ) })?; let path = bytes2path(Cow::Borrowed(&bytes[prefixlen + 1..]))?; copy_path_into(&mut self.name, &path, false).map_err(|err| { io::Error::new( err.kind(), format!("{} when setting path for {}", err, self.path_lossy()), ) })?; } Ok(()) } /// See `Header::username_bytes` pub fn username_bytes(&self) -> &[u8] { truncate(&self.uname) } /// See `Header::set_username` pub fn set_username(&mut self, name: &str) -> io::Result<()> { copy_into(&mut self.uname, name.as_bytes()).map_err(|err| { io::Error::new( err.kind(), format!("{} when setting username for {}", err, self.path_lossy()), ) }) } /// See `Header::groupname_bytes` pub fn groupname_bytes(&self) -> &[u8] { truncate(&self.gname) } /// See `Header::set_groupname` pub fn set_groupname(&mut self, name: &str) -> io::Result<()> { copy_into(&mut self.gname, name.as_bytes()).map_err(|err| { io::Error::new( err.kind(), format!("{} when setting groupname for {}", err, self.path_lossy()), ) }) } /// See `Header::device_major` pub fn device_major(&self) -> io::Result { octal_from(&self.dev_major) .map(|u| u as u32) .map_err(|err| { io::Error::new( err.kind(), format!( "{} when getting device_major for {}", err, self.path_lossy() ), ) }) } /// See `Header::set_device_major` pub fn set_device_major(&mut self, major: u32) { octal_into(&mut self.dev_major, major); } /// See `Header::device_minor` pub fn device_minor(&self) -> io::Result { octal_from(&self.dev_minor) .map(|u| u as u32) .map_err(|err| { io::Error::new( err.kind(), format!( "{} when getting device_minor for {}", err, self.path_lossy() ), ) }) } /// See `Header::set_device_minor` pub fn set_device_minor(&mut self, minor: u32) { octal_into(&mut self.dev_minor, minor); } /// Views this as a normal `Header` pub fn as_header(&self) -> &Header { unsafe { cast(self) } } /// Views this as a normal `Header` pub fn as_header_mut(&mut self) -> &mut Header { unsafe { cast_mut(self) } } } impl fmt::Debug for UstarHeader { fn fmt(&self, f: &mut fmt::Formatter) -> fmt::Result { let mut f = f.debug_struct("UstarHeader"); self.as_header().debug_fields(&mut f); f.finish() } } impl GnuHeader { /// See `Header::username_bytes` pub fn username_bytes(&self) -> &[u8] { truncate(&self.uname) } /// Gets the fullname (group:user) in a "lossy" way, used for error reporting ONLY. fn fullname_lossy(&self) -> String { format!( "{}:{}", String::from_utf8_lossy(self.groupname_bytes()), String::from_utf8_lossy(self.username_bytes()), ) } /// See `Header::set_username` pub fn set_username(&mut self, name: &str) -> io::Result<()> { copy_into(&mut self.uname, name.as_bytes()).map_err(|err| { io::Error::new( err.kind(), format!( "{} when setting username for {}", err, self.fullname_lossy() ), ) }) } /// See `Header::groupname_bytes` pub fn groupname_bytes(&self) -> &[u8] { truncate(&self.gname) } /// See `Header::set_groupname` pub fn set_groupname(&mut self, name: &str) -> io::Result<()> { copy_into(&mut self.gname, name.as_bytes()).map_err(|err| { io::Error::new( err.kind(), format!( "{} when setting groupname for {}", err, self.fullname_lossy() ), ) }) } /// See `Header::device_major` pub fn device_major(&self) -> io::Result { octal_from(&self.dev_major) .map(|u| u as u32) .map_err(|err| { io::Error::new( err.kind(), format!( "{} when getting device_major for {}", err, self.fullname_lossy() ), ) }) } /// See `Header::set_device_major` pub fn set_device_major(&mut self, major: u32) { octal_into(&mut self.dev_major, major); } /// See `Header::device_minor` pub fn device_minor(&self) -> io::Result { octal_from(&self.dev_minor) .map(|u| u as u32) .map_err(|err| { io::Error::new( err.kind(), format!( "{} when getting device_minor for {}", err, self.fullname_lossy() ), ) }) } /// See `Header::set_device_minor` pub fn set_device_minor(&mut self, minor: u32) { octal_into(&mut self.dev_minor, minor); } /// Returns the last modification time in Unix time format pub fn atime(&self) -> io::Result { num_field_wrapper_from(&self.atime).map_err(|err| { io::Error::new( err.kind(), format!("{} when getting atime for {}", err, self.fullname_lossy()), ) }) } /// Encodes the `atime` provided into this header. /// /// Note that this time is typically a number of seconds passed since /// January 1, 1970. pub fn set_atime(&mut self, atime: u64) { num_field_wrapper_into(&mut self.atime, atime); } /// Returns the last modification time in Unix time format pub fn ctime(&self) -> io::Result { num_field_wrapper_from(&self.ctime).map_err(|err| { io::Error::new( err.kind(), format!("{} when getting ctime for {}", err, self.fullname_lossy()), ) }) } /// Encodes the `ctime` provided into this header. /// /// Note that this time is typically a number of seconds passed since /// January 1, 1970. pub fn set_ctime(&mut self, ctime: u64) { num_field_wrapper_into(&mut self.ctime, ctime); } /// Returns the "real size" of the file this header represents. /// /// This is applicable for sparse files where the returned size here is the /// size of the entire file after the sparse regions have been filled in. pub fn real_size(&self) -> io::Result { num_field_wrapper_from(&self.realsize).map_err(|err| { io::Error::new( err.kind(), format!( "{} when getting real_size for {}", err, self.fullname_lossy() ), ) }) } /// Indicates whether this header will be followed by additional /// sparse-header records. /// /// Note that this is handled internally by this library, and is likely only /// interesting if a `raw` iterator is being used. pub fn is_extended(&self) -> bool { self.isextended[0] == 1 } /// Views this as a normal `Header` pub fn as_header(&self) -> &Header { unsafe { cast(self) } } /// Views this as a normal `Header` pub fn as_header_mut(&mut self) -> &mut Header { unsafe { cast_mut(self) } } } impl fmt::Debug for GnuHeader { fn fmt(&self, f: &mut fmt::Formatter) -> fmt::Result { let mut f = f.debug_struct("GnuHeader"); self.as_header().debug_fields(&mut f); if let Ok(atime) = self.atime() { f.field("atime", &atime); } if let Ok(ctime) = self.ctime() { f.field("ctime", &ctime); } f.field("is_extended", &self.is_extended()) .field("sparse", &DebugSparseHeaders(&self.sparse)) .finish() } } struct DebugSparseHeaders<'a>(&'a [GnuSparseHeader]); impl fmt::Debug for DebugSparseHeaders<'_> { fn fmt(&self, f: &mut fmt::Formatter) -> fmt::Result { let mut f = f.debug_list(); for header in self.0 { if !header.is_empty() { f.entry(header); } } f.finish() } } impl GnuSparseHeader { /// Returns true if block is empty pub fn is_empty(&self) -> bool { self.offset[0] == 0 || self.numbytes[0] == 0 } /// Offset of the block from the start of the file /// /// Returns `Err` for a malformed `offset` field. pub fn offset(&self) -> io::Result { num_field_wrapper_from(&self.offset).map_err(|err| { io::Error::new( err.kind(), format!("{} when getting offset from sparse header", err), ) }) } /// Length of the block /// /// Returns `Err` for a malformed `numbytes` field. pub fn length(&self) -> io::Result { num_field_wrapper_from(&self.numbytes).map_err(|err| { io::Error::new( err.kind(), format!("{} when getting length from sparse header", err), ) }) } } impl fmt::Debug for GnuSparseHeader { fn fmt(&self, f: &mut fmt::Formatter) -> fmt::Result { let mut f = f.debug_struct("GnuSparseHeader"); if let Ok(offset) = self.offset() { f.field("offset", &offset); } if let Ok(length) = self.length() { f.field("length", &length); } f.finish() } } impl GnuExtSparseHeader { /// Crates a new zero'd out sparse header entry. pub fn new() -> GnuExtSparseHeader { unsafe { mem::zeroed() } } /// Returns a view into this header as a byte array. pub fn as_bytes(&self) -> &[u8; BLOCK_SIZE as usize] { debug_assert_eq!(mem::size_of_val(self), BLOCK_SIZE as usize); unsafe { &*(self as *const GnuExtSparseHeader as *const [u8; BLOCK_SIZE as usize]) } } /// Returns a view into this header as a byte array. pub fn as_mut_bytes(&mut self) -> &mut [u8; BLOCK_SIZE as usize] { debug_assert_eq!(mem::size_of_val(self), BLOCK_SIZE as usize); unsafe { &mut *(self as *mut GnuExtSparseHeader as *mut [u8; BLOCK_SIZE as usize]) } } /// Returns a slice of the underlying sparse headers. /// /// Some headers may represent empty chunks of both the offset and numbytes /// fields are 0. pub fn sparse(&self) -> &[GnuSparseHeader; 21] { &self.sparse } /// Indicates if another sparse header should be following this one. pub fn is_extended(&self) -> bool { self.isextended[0] == 1 } } impl Default for GnuExtSparseHeader { fn default() -> Self { Self::new() } } fn octal_from(slice: &[u8]) -> io::Result { let trun = truncate(slice); let num = match str::from_utf8(trun) { Ok(n) => n, Err(_) => { return Err(other(&format!( "numeric field did not have utf-8 text: {}", String::from_utf8_lossy(trun) ))); } }; match u64::from_str_radix(num.trim(), 8) { Ok(n) => Ok(n), Err(_) => Err(other(&format!("numeric field was not a number: {}", num))), } } fn octal_into(dst: &mut [u8], val: T) { let o = format!("{:o}", val); let value = once(b'\0').chain(o.bytes().rev().chain(repeat(b'0'))); for (slot, value) in dst.iter_mut().rev().zip(value) { *slot = value; } } // Wrapper to figure out if we should fill the header field using tar's numeric // extension (binary) or not (octal). fn num_field_wrapper_into(dst: &mut [u8], src: u64) { if src >= 8_589_934_592 || (src >= 2_097_152 && dst.len() == 8) { numeric_extended_into(dst, src); } else { octal_into(dst, src); } } // Wrapper to figure out if we should read the header field in binary (numeric // extension) or octal (standard encoding). fn num_field_wrapper_from(src: &[u8]) -> io::Result { if src[0] & 0x80 != 0 { Ok(numeric_extended_from(src)) } else { octal_from(src) } } // When writing numeric fields with is the extended form, the high bit of the // first byte is set to 1 and the remainder of the field is treated as binary // instead of octal ascii. // This handles writing u64 to 8 (uid, gid) or 12 (size, *time) bytes array. fn numeric_extended_into(dst: &mut [u8], src: u64) { let len: usize = dst.len(); for (slot, val) in dst.iter_mut().zip( repeat(0) .take(len - 8) // to zero init extra bytes .chain((0..8).rev().map(|x| ((src >> (8 * x)) & 0xff) as u8)), ) { *slot = val; } dst[0] |= 0x80; } fn numeric_extended_from(src: &[u8]) -> u64 { let mut dst: u64 = 0; let mut b_to_skip = 1; if src.len() == 8 { // read first byte without extension flag bit dst = (src[0] ^ 0x80) as u64; } else { // only read last 8 bytes b_to_skip = src.len() - 8; } for byte in src.iter().skip(b_to_skip) { dst <<= 8; dst |= *byte as u64; } dst } fn truncate(slice: &[u8]) -> &[u8] { match slice.iter().position(|i| *i == 0) { Some(i) => &slice[..i], None => slice, } } /// Copies `bytes` into the `slot` provided, returning an error if the `bytes` /// array is too long or if it contains any nul bytes. fn copy_into(slot: &mut [u8], bytes: &[u8]) -> io::Result<()> { if bytes.len() > slot.len() { Err(other("provided value is too long")) } else if bytes.iter().any(|b| *b == 0) { Err(other("provided value contains a nul byte")) } else { for (slot, val) in slot.iter_mut().zip(bytes.iter().chain(Some(&0))) { *slot = *val; } Ok(()) } } fn copy_path_into_inner( mut slot: &mut [u8], path: &Path, is_link_name: bool, is_truncated_gnu_long_path: bool, ) -> io::Result<()> { let mut emitted = false; let mut needs_slash = false; let mut iter = path.components().peekable(); while let Some(component) = iter.next() { let bytes = path2bytes(Path::new(component.as_os_str()))?; match (component, is_link_name) { (Component::Prefix(..), false) | (Component::RootDir, false) => { return Err(other("paths in archives must be relative")); } (Component::ParentDir, false) => { // If it's last component of a gnu long path we know that there might be more // to the component than .. (the rest is stored elsewhere) // Otherwise it's a clear error if !is_truncated_gnu_long_path || iter.peek().is_some() { return Err(other("paths in archives must not have `..`")); } } // Allow "./" as the path (Component::CurDir, false) if path.components().count() == 1 => {} (Component::CurDir, false) => continue, (Component::Normal(_), _) | (_, true) => {} }; if needs_slash { copy(&mut slot, b"/")?; } if bytes.contains(&b'/') { if let Component::Normal(..) = component { return Err(other("path component in archive cannot contain `/`")); } } copy(&mut slot, &bytes)?; if &*bytes != b"/" { needs_slash = true; } emitted = true; } if !emitted { return Err(other("paths in archives must have at least one component")); } if ends_with_slash(path) { copy(&mut slot, b"/")?; } return Ok(()); fn copy(slot: &mut &mut [u8], bytes: &[u8]) -> io::Result<()> { copy_into(slot, bytes)?; let tmp = std::mem::take(slot); *slot = &mut tmp[bytes.len()..]; Ok(()) } } /// Copies `path` into the `slot` provided /// /// Returns an error if: /// /// * the path is too long to fit /// * a nul byte was found /// * an invalid path component is encountered (e.g. a root path or parent dir) /// * the path itself is empty fn copy_path_into(slot: &mut [u8], path: &Path, is_link_name: bool) -> io::Result<()> { copy_path_into_inner(slot, path, is_link_name, false) } /// Copies `path` into the `slot` provided /// /// Returns an error if: /// /// * the path is too long to fit /// * a nul byte was found /// * an invalid path component is encountered (e.g. a root path or parent dir) /// * the path itself is empty /// /// This is less restrictive version meant to be used for truncated GNU paths. fn copy_path_into_gnu_long(slot: &mut [u8], path: &Path, is_link_name: bool) -> io::Result<()> { copy_path_into_inner(slot, path, is_link_name, true) } #[cfg(target_arch = "wasm32")] fn ends_with_slash(p: &Path) -> bool { p.to_string_lossy().ends_with('/') } #[cfg(windows)] fn ends_with_slash(p: &Path) -> bool { let last = p.as_os_str().encode_wide().last(); last == Some(b'/' as u16) || last == Some(b'\\' as u16) } #[cfg(unix)] fn ends_with_slash(p: &Path) -> bool { p.as_os_str().as_bytes().ends_with(b"/") } #[cfg(any(windows, target_arch = "wasm32"))] pub fn path2bytes(p: &Path) -> io::Result> { p.as_os_str() .to_str() .map(|s| s.as_bytes()) .ok_or_else(|| other(&format!("path {} was not valid Unicode", p.display()))) .map(|bytes| { if bytes.contains(&b'\\') { // Normalize to Unix-style path separators let mut bytes = bytes.to_owned(); for b in &mut bytes { if *b == b'\\' { *b = b'/'; } } Cow::Owned(bytes) } else { Cow::Borrowed(bytes) } }) } #[cfg(unix)] /// On unix this will never fail pub fn path2bytes(p: &Path) -> io::Result> { Ok(Cow::Borrowed(p.as_os_str().as_bytes())) } #[cfg(windows)] /// On windows we cannot accept non-Unicode bytes because it /// is impossible to convert it to UTF-16. pub fn bytes2path(bytes: Cow<[u8]>) -> io::Result> { return match bytes { Cow::Borrowed(bytes) => { let s = str::from_utf8(bytes).map_err(|_| not_unicode(bytes))?; Ok(Cow::Borrowed(Path::new(s))) } Cow::Owned(bytes) => { let s = String::from_utf8(bytes).map_err(|uerr| not_unicode(&uerr.into_bytes()))?; Ok(Cow::Owned(PathBuf::from(s))) } }; fn not_unicode(v: &[u8]) -> io::Error { other(&format!( "only Unicode paths are supported on Windows: {}", String::from_utf8_lossy(v) )) } } #[cfg(unix)] /// On unix this operation can never fail. pub fn bytes2path(bytes: Cow<'_, [u8]>) -> io::Result> { use std::ffi::{OsStr, OsString}; Ok(match bytes { Cow::Borrowed(bytes) => Cow::Borrowed(Path::new(OsStr::from_bytes(bytes))), Cow::Owned(bytes) => Cow::Owned(PathBuf::from(OsString::from_vec(bytes))), }) } #[cfg(target_arch = "wasm32")] pub fn bytes2path(bytes: Cow<[u8]>) -> io::Result> { Ok(match bytes { Cow::Borrowed(bytes) => { Cow::Borrowed({ Path::new(str::from_utf8(bytes).map_err(invalid_utf8)?) }) } Cow::Owned(bytes) => { Cow::Owned({ PathBuf::from(String::from_utf8(bytes).map_err(invalid_utf8)?) }) } }) } #[cfg(target_arch = "wasm32")] fn invalid_utf8(_: T) -> io::Error { io::Error::new(io::ErrorKind::InvalidData, "Invalid utf-8") } astral-tokio-tar-0.5.1/src/lib.rs000064400000000000000000000030251046102023000147270ustar 00000000000000//! A library for reading and writing TAR archives in an async fashion. //! //! This library provides utilities necessary to manage [TAR archives][1] //! abstracted over a reader or writer. Great strides are taken to ensure that //! an archive is never required to be fully resident in memory, and all objects //! provide largely a streaming interface to read bytes from. //! //! [1]: http://en.wikipedia.org/wiki/Tar_%28computing%29 // More docs about the detailed tar format can also be found here: // http://www.freebsd.org/cgi/man.cgi?query=tar&sektion=5&manpath=FreeBSD+8-current // NB: some of the coding patterns and idioms here may seem a little strange. // This is currently attempting to expose a super generic interface while // also not forcing clients to codegen the entire crate each time they use // it. To that end lots of work is done to ensure that concrete // implementations are all found in this crate and the generic functions are // all just super thin wrappers (e.g. easy to codegen). #![deny(missing_docs)] use std::io::{Error, ErrorKind}; pub use crate::{ archive::{Archive, ArchiveBuilder, Entries}, builder::Builder, entry::{Entry, Unpacked}, entry_type::EntryType, header::{ GnuExtSparseHeader, GnuHeader, GnuSparseHeader, Header, HeaderMode, OldHeader, UstarHeader, }, pax::{PaxExtension, PaxExtensions}, }; mod archive; mod builder; mod entry; mod entry_type; mod error; mod header; mod pax; fn other(msg: &str) -> Error { Error::new(ErrorKind::Other, msg) } astral-tokio-tar-0.5.1/src/pax.rs000064400000000000000000000051601046102023000147530ustar 00000000000000use std::{slice, str}; use tokio::io; use crate::other; /// An iterator over the pax extensions in an archive entry. /// /// This iterator yields structures which can themselves be parsed into /// key/value pairs. pub struct PaxExtensions<'entry> { data: slice::Split<'entry, u8, fn(&u8) -> bool>, } /// A key/value pair corresponding to a pax extension. pub struct PaxExtension<'entry> { key: &'entry [u8], value: &'entry [u8], } pub fn pax_extensions(a: &[u8]) -> PaxExtensions { PaxExtensions { data: a.split(|a| *a == b'\n'), } } impl<'entry> Iterator for PaxExtensions<'entry> { type Item = io::Result>; fn next(&mut self) -> Option>> { let line = match self.data.next() { Some([]) => return None, Some(line) => line, None => return None, }; Some( line.iter() .position(|b| *b == b' ') .and_then(|i| { str::from_utf8(&line[..i]) .ok() .and_then(|len| len.parse::().ok().map(|j| (i + 1, j))) }) .and_then(|(kvstart, reported_len)| { if line.len() + 1 == reported_len { line[kvstart..] .iter() .position(|b| *b == b'=') .map(|equals| (kvstart, equals)) } else { None } }) .map(|(kvstart, equals)| PaxExtension { key: &line[kvstart..kvstart + equals], value: &line[kvstart + equals + 1..], }) .ok_or_else(|| other("malformed pax extension")), ) } } impl<'entry> PaxExtension<'entry> { /// Returns the key for this key/value pair parsed as a string. /// /// May fail if the key isn't actually utf-8. pub fn key(&self) -> Result<&'entry str, str::Utf8Error> { str::from_utf8(self.key) } /// Returns the underlying raw bytes for the key of this key/value pair. pub fn key_bytes(&self) -> &'entry [u8] { self.key } /// Returns the value for this key/value pair parsed as a string. /// /// May fail if the value isn't actually utf-8. pub fn value(&self) -> Result<&'entry str, str::Utf8Error> { str::from_utf8(self.value) } /// Returns the underlying raw bytes for this value of this key/value pair. pub fn value_bytes(&self) -> &'entry [u8] { self.value } } astral-tokio-tar-0.5.1/tests/all.rs000064400000000000000000001401261046102023000153100ustar 00000000000000extern crate tokio_tar as async_tar; extern crate filetime; extern crate tempfile; #[cfg(all(unix, feature = "xattr"))] extern crate xattr; use std::{ io::Cursor, path::{Path, PathBuf}, }; use tokio::{ fs::{self, File}, io::{self, AsyncRead, AsyncReadExt, AsyncWrite, AsyncWriteExt}, }; use tokio_stream::*; use async_tar::{Archive, ArchiveBuilder, Builder, EntryType, Header}; use filetime::FileTime; use tempfile::{Builder as TempBuilder, TempDir}; macro_rules! t { ($e:expr) => { match $e { Ok(v) => v, Err(e) => panic!("{} returned {}", stringify!($e), e), } }; } macro_rules! tar { ($e:expr) => { &include_bytes!(concat!("archives/", $e))[..] }; } mod header; /// test that we can concatenate the simple.tar archive and extract the same entries twice when we /// use the ignore_zeros option. #[tokio::test] async fn simple_concat() { let bytes = tar!("simple.tar"); let mut archive_bytes = Vec::new(); archive_bytes.extend(bytes); let original_names: Vec = decode_names(&mut Archive::new(Cursor::new(&archive_bytes))).await; let expected: Vec<&str> = original_names.iter().map(|n| n.as_str()).collect(); // concat two archives (with null in-between); archive_bytes.extend(bytes); // test now that when we read the archive, it stops processing at the first zero header. let actual = decode_names(&mut Archive::new(Cursor::new(&archive_bytes))).await; assert_eq!(expected, actual); // extend expected by itself. let expected: Vec<&str> = { let mut o = Vec::new(); o.extend(&expected); o.extend(&expected); o }; let builder = ArchiveBuilder::new(Cursor::new(&archive_bytes)).set_ignore_zeros(true); let mut ar = builder.build(); let actual = decode_names(&mut ar).await; assert_eq!(expected, actual); async fn decode_names(ar: &mut Archive) -> Vec where R: AsyncRead + Unpin, { let mut names = Vec::new(); let mut entries = t!(ar.entries()); while let Some(entry) = entries.next().await { let e = t!(entry); names.push(t!(::std::str::from_utf8(&e.path_bytes())).to_string()); } names } } #[tokio::test] async fn header_impls() { let mut ar = Archive::new(Cursor::new(tar!("simple.tar"))); let hn = Header::new_old(); let hnb = hn.as_bytes(); let mut entries = t!(ar.entries()); while let Some(file) = entries.next().await { let file = t!(file); let h1 = file.header(); let h1b = h1.as_bytes(); let h2 = h1.clone(); let h2b = h2.as_bytes(); assert!(h1b[..] == h2b[..] && h2b[..] != hnb[..]) } } #[tokio::test] async fn header_impls_missing_last_header() { let mut ar = Archive::new(Cursor::new(tar!("simple_missing_last_header.tar"))); let hn = Header::new_old(); let hnb = hn.as_bytes(); let mut entries = t!(ar.entries()); while let Some(file) = entries.next().await { let file = t!(file); let h1 = file.header(); let h1b = h1.as_bytes(); let h2 = h1.clone(); let h2b = h2.as_bytes(); assert!(h1b[..] == h2b[..] && h2b[..] != hnb[..]) } } #[tokio::test] async fn reading_files() { let rdr = Cursor::new(tar!("reading_files.tar")); let mut ar = Archive::new(rdr); let mut entries = t!(ar.entries()); let mut a = t!(entries.next().await.unwrap()); assert_eq!(&*a.header().path_bytes(), b"a"); let mut s = String::new(); t!(a.read_to_string(&mut s).await); assert_eq!(s, "a\na\na\na\na\na\na\na\na\na\na\n"); let mut b = t!(entries.next().await.unwrap()); assert_eq!(&*b.header().path_bytes(), b"b"); s.truncate(0); t!(b.read_to_string(&mut s).await); assert_eq!(s, "b\nb\nb\nb\nb\nb\nb\nb\nb\nb\nb\n"); assert!(entries.next().await.is_none()); } #[tokio::test] async fn writing_files() { let mut ar = Builder::new(Vec::new()); let td = t!(TempBuilder::new().prefix("async-tar").tempdir()); let path = td.path().join("test"); let mut file = t!(File::create(&path).await); t!(file.write_all(b"test").await); t!(file.flush().await); t!(ar .append_file("test2", &mut t!(File::open(&path).await)) .await); let data = t!(ar.into_inner().await); let mut ar = Archive::new(Cursor::new(data)); let mut entries = t!(ar.entries()); let mut f = t!(entries.next().await.unwrap()); assert_eq!(&*f.header().path_bytes(), b"test2"); assert_eq!(f.header().size().unwrap(), 4); let mut s = String::new(); t!(f.read_to_string(&mut s).await); assert_eq!(s, "test"); assert!(entries.next().await.is_none()); } #[tokio::test] async fn large_filename() { let mut ar = Builder::new(Vec::new()); let td = t!(TempBuilder::new().prefix("async-tar").tempdir()); let path = td.path().join("test"); let mut file = t!(File::create(&path).await); t!(file.write_all(b"test").await); t!(file.flush().await); let filename = "abcd/".repeat(50); let mut header = Header::new_ustar(); header.set_path(&filename).unwrap(); header.set_metadata(&t!(fs::metadata(&path).await)); header.set_cksum(); t!(ar.append(&header, &b"test"[..]).await); let too_long = "abcd".repeat(200); t!(ar .append_file(&too_long, &mut t!(File::open(&path).await)) .await); t!(ar.append_data(&mut header, &too_long, &b"test"[..]).await); let rd = Cursor::new(t!(ar.into_inner().await)); let mut ar = Archive::new(rd); let mut entries = t!(ar.entries()); // The short entry added with `append` let mut f = entries.next().await.unwrap().unwrap(); assert_eq!(&*f.header().path_bytes(), filename.as_bytes()); assert_eq!(f.header().size().unwrap(), 4); let mut s = String::new(); t!(f.read_to_string(&mut s).await); assert_eq!(s, "test"); // The long entry added with `append_file` let mut f = entries.next().await.unwrap().unwrap(); assert_eq!(&*f.path_bytes(), too_long.as_bytes()); assert_eq!(f.header().size().unwrap(), 4); let mut s = String::new(); t!(f.read_to_string(&mut s).await); assert_eq!(s, "test"); // The long entry added with `append_data` let mut f = entries.next().await.unwrap().unwrap(); assert!(f.header().path_bytes().len() < too_long.len()); assert_eq!(&*f.path_bytes(), too_long.as_bytes()); assert_eq!(f.header().size().unwrap(), 4); let mut s = String::new(); t!(f.read_to_string(&mut s).await); assert_eq!(s, "test"); assert!(entries.next().await.is_none()); } // This test checks very particular scenario where a path component starting // with ".." of a long path gets split at 100-byte mark so that ".." part goes // into header and gets interpreted as parent dir (and rejected) . #[tokio::test] async fn large_filename_with_dot_dot_at_100_byte_mark() { let mut ar = Builder::new(Vec::new()); let mut header = Header::new_gnu(); header.set_mode(0o644); header.set_size(4); let mut long_name_with_dot_dot = "tdir/".repeat(19); long_name_with_dot_dot.push_str("tt/..file"); t!(ar .append_data(&mut header, &long_name_with_dot_dot, b"test".as_slice()) .await); let rd = Cursor::new(t!(ar.into_inner().await)); let mut ar = Archive::new(rd); let mut entries = t!(ar.entries()); let mut f = t!(entries.next().await.unwrap()); assert_eq!(&*f.path_bytes(), long_name_with_dot_dot.as_bytes()); assert_eq!(f.header().size().unwrap(), 4); let mut s = String::new(); t!(f.read_to_string(&mut s).await); assert_eq!(s, "test"); assert!(entries.next().await.is_none()); } #[tokio::test] async fn reading_entries() { let rdr = Cursor::new(tar!("reading_files.tar")); let mut ar = Archive::new(rdr); let mut entries = t!(ar.entries()); let mut a = t!(entries.next().await.unwrap()); assert_eq!(&*a.header().path_bytes(), b"a"); let mut s = String::new(); t!(a.read_to_string(&mut s).await); assert_eq!(s, "a\na\na\na\na\na\na\na\na\na\na\n"); s.truncate(0); t!(a.read_to_string(&mut s).await); assert_eq!(s, ""); let mut b = t!(entries.next().await.unwrap()); assert_eq!(&*b.header().path_bytes(), b"b"); s.truncate(0); t!(b.read_to_string(&mut s).await); assert_eq!(s, "b\nb\nb\nb\nb\nb\nb\nb\nb\nb\nb\n"); assert!(entries.next().await.is_none()); } async fn check_dirtree(td: &TempDir) { let dir_a = td.path().join("a"); let dir_b = td.path().join("a/b"); let file_c = td.path().join("a/c"); assert!(fs::metadata(&dir_a) .await .map(|m| m.is_dir()) .unwrap_or(false)); assert!(fs::metadata(&dir_b) .await .map(|m| m.is_dir()) .unwrap_or(false)); assert!(fs::metadata(&file_c) .await .map(|m| m.is_file()) .unwrap_or(false)); } #[tokio::test] async fn extracting_directories() { let td = t!(TempBuilder::new().prefix("async-tar").tempdir()); let rdr = Cursor::new(tar!("directory.tar")); let mut ar = Archive::new(rdr); t!(ar.unpack(td.path()).await); check_dirtree(&td).await; } #[tokio::test] async fn extracting_duplicate_file_fail() { let td = t!(TempBuilder::new().prefix("async-tar").tempdir()); let path_present = td.path().join("a"); let mut file = t!(File::create(path_present).await); t!(file.write_all(b"").await); t!(file.flush().await); let rdr = Cursor::new(tar!("reading_files.tar")); let builder = ArchiveBuilder::new(rdr).set_overwrite(false); let mut ar = builder.build(); if let Err(err) = ar.unpack(td.path()).await { if err.kind() == std::io::ErrorKind::AlreadyExists { // as expected with overwrite false return; } panic!("unexpected error: {:?}", err); } panic!( "unpack() should have returned an error of kind {:?}, returned Ok", std::io::ErrorKind::AlreadyExists ) } #[tokio::test] async fn extracting_duplicate_file_succeed() { let td = t!(TempBuilder::new().prefix("async-tar").tempdir()); let path_present = td.path().join("a"); let mut file = t!(File::create(path_present).await); t!(file.write_all(b"").await); t!(file.flush().await); let rdr = Cursor::new(tar!("reading_files.tar")); let builder = ArchiveBuilder::new(rdr).set_overwrite(true); let mut ar = builder.build(); t!(ar.unpack(td.path()).await); } #[tokio::test] #[cfg(unix)] async fn extracting_duplicate_link_fail() { let td = t!(TempBuilder::new().prefix("async-tar").tempdir()); let path_present = td.path().join("lnk"); t!(std::os::unix::fs::symlink("file", path_present)); let rdr = Cursor::new(tar!("link.tar")); let builder = ArchiveBuilder::new(rdr).set_overwrite(false); let mut ar = builder.build(); if let Err(err) = ar.unpack(td.path()).await { if err.kind() == std::io::ErrorKind::AlreadyExists { // as expected with overwrite false return; } panic!("unexpected error: {:?}", err); } panic!( "unpack() should have returned an error of kind {:?}, returned Ok", std::io::ErrorKind::AlreadyExists ) } #[tokio::test] #[cfg(unix)] async fn extracting_duplicate_link_succeed() { let td = t!(TempBuilder::new().prefix("async-tar").tempdir()); let path_present = td.path().join("lnk"); t!(std::os::unix::fs::symlink("file", path_present)); let rdr = Cursor::new(tar!("link.tar")); let builder = ArchiveBuilder::new(rdr).set_overwrite(true); let mut ar = builder.build(); t!(ar.unpack(td.path()).await); } #[tokio::test] #[cfg(all(unix, feature = "xattr"))] async fn xattrs() { // If /tmp is a tmpfs, xattr will fail // The xattr crate's unit tests also use /var/tmp for this reason let td = t!(TempBuilder::new() .prefix("async-tar") .tempdir_in("/var/tmp")); let rdr = Cursor::new(tar!("xattrs.tar")); let builder = ArchiveBuilder::new(rdr).set_unpack_xattrs(true); let mut ar = builder.build(); t!(ar.unpack(td.path()).await); let val = xattr::get(td.path().join("a/b"), "user.pax.flags").unwrap(); assert_eq!(val.unwrap(), b"epm"); } #[tokio::test] #[cfg(all(unix, feature = "xattr"))] async fn no_xattrs() { // If /tmp is a tmpfs, xattr will fail // The xattr crate's unit tests also use /var/tmp for this reason let td = t!(TempBuilder::new() .prefix("async-tar") .tempdir_in("/var/tmp")); let rdr = Cursor::new(tar!("xattrs.tar")); let builder = ArchiveBuilder::new(rdr).set_unpack_xattrs(false); let mut ar = builder.build(); t!(ar.unpack(td.path()).await); assert_eq!( xattr::get(td.path().join("a/b"), "user.pax.flags").unwrap(), None ); } #[tokio::test] async fn writing_and_extracting_directories() { let td = t!(TempBuilder::new().prefix("async-tar").tempdir()); let mut ar = Builder::new(Vec::new()); let tmppath = td.path().join("tmpfile"); let mut file = t!(File::create(&tmppath).await); t!(file.write_all(b"c").await); t!(file.flush().await); t!(ar.append_dir("a", ".").await); t!(ar.append_dir("a/b", ".").await); t!(ar .append_file("a/c", &mut t!(File::open(&tmppath).await)) .await); t!(ar.finish().await); let rdr = Cursor::new(t!(ar.into_inner().await)); let mut ar = Archive::new(rdr); t!(ar.unpack(td.path()).await); check_dirtree(&td).await; } #[tokio::test] async fn writing_directories_recursively() { let td = t!(TempBuilder::new().prefix("async-tar").tempdir()); let base_dir = td.path().join("base"); t!(fs::create_dir(&base_dir).await); let mut file1 = t!(File::create(base_dir.join("file1")).await); t!(file1.write_all(b"file1").await); t!(file1.flush().await); let sub_dir = base_dir.join("sub"); t!(fs::create_dir(&sub_dir).await); let mut file2 = t!(File::create(sub_dir.join("file2")).await); t!(file2.write_all(b"file2").await); t!(file2.flush().await); let mut ar = Builder::new(Vec::new()); t!(ar.append_dir_all("foobar", base_dir).await); let data = t!(ar.into_inner().await); let mut ar = Archive::new(Cursor::new(data)); t!(ar.unpack(td.path()).await); let base_dir = td.path().join("foobar"); assert!(fs::metadata(&base_dir) .await .map(|m| m.is_dir()) .unwrap_or(false)); let file1_path = base_dir.join("file1"); assert!(fs::metadata(&file1_path) .await .map(|m| m.is_file()) .unwrap_or(false)); let sub_dir = base_dir.join("sub"); assert!(fs::metadata(&sub_dir) .await .map(|m| m.is_dir()) .unwrap_or(false)); let file2_path = sub_dir.join("file2"); assert!(fs::metadata(&file2_path) .await .map(|m| m.is_file()) .unwrap_or(false)); } #[tokio::test] async fn append_dir_all_blank_dest() { let td = t!(TempBuilder::new().prefix("async-tar").tempdir()); let base_dir = td.path().join("base"); t!(fs::create_dir(&base_dir).await); let mut file1 = t!(File::create(base_dir.join("file1")).await); t!(file1.write_all(b"file1").await); t!(file1.flush().await); let sub_dir = base_dir.join("sub"); t!(fs::create_dir(&sub_dir).await); let mut file2 = t!(File::create(sub_dir.join("file2")).await); t!(file2.write_all(b"file2").await); t!(file2.flush().await); let mut ar = Builder::new(Vec::new()); t!(ar.append_dir_all("", base_dir).await); let data = t!(ar.into_inner().await); let mut ar = Archive::new(Cursor::new(data)); t!(ar.unpack(td.path()).await); let base_dir = td.path(); assert!(fs::metadata(&base_dir) .await .map(|m| m.is_dir()) .unwrap_or(false)); let file1_path = base_dir.join("file1"); assert!(fs::metadata(&file1_path) .await .map(|m| m.is_file()) .unwrap_or(false)); let sub_dir = base_dir.join("sub"); assert!(fs::metadata(&sub_dir) .await .map(|m| m.is_dir()) .unwrap_or(false)); let file2_path = sub_dir.join("file2"); assert!(fs::metadata(&file2_path) .await .map(|m| m.is_file()) .unwrap_or(false)); } #[tokio::test] async fn append_dir_all_does_not_work_on_non_directory() { let td = t!(TempBuilder::new().prefix("async-tar").tempdir()); let path = td.path().join("test"); let mut file = t!(File::create(&path).await); t!(file.write_all(b"test").await); t!(file.flush().await); let mut ar = Builder::new(Vec::new()); let result = ar.append_dir_all("test", path).await; assert!(result.is_err()); } #[tokio::test] async fn extracting_duplicate_dirs() { let td = t!(TempBuilder::new().prefix("async-tar").tempdir()); let rdr = Cursor::new(tar!("duplicate_dirs.tar")); let mut ar = Archive::new(rdr); t!(ar.unpack(td.path()).await); let some_dir = td.path().join("some_dir"); assert!(fs::metadata(&some_dir) .await .map(|m| m.is_dir()) .unwrap_or(false)); } #[tokio::test] async fn unpack_old_style_bsd_dir() { let td = t!(TempBuilder::new().prefix("async-tar").tempdir()); let mut ar = Builder::new(Vec::new()); let mut header = Header::new_old(); header.set_entry_type(EntryType::Regular); t!(header.set_path("testdir/")); header.set_size(0); header.set_cksum(); t!(ar.append(&header, &mut io::empty()).await); // Extracting let rdr = Cursor::new(t!(ar.into_inner().await)); let mut ar = Archive::new(rdr); t!(ar.unpack(td.path()).await); // Iterating let rdr = Cursor::new(ar.into_inner().map_err(|_| ()).unwrap().into_inner()); let mut ar = Archive::new(rdr); let mut entries = t!(ar.entries()); while let Some(e) = entries.next().await { assert!(e.is_ok()); } assert!(td.path().join("testdir").is_dir()); } #[tokio::test] async fn handling_incorrect_file_size() { let td = t!(TempBuilder::new().prefix("async-tar").tempdir()); let mut ar = Builder::new(Vec::new()); let path = td.path().join("tmpfile"); let mut file = t!(File::create(&path).await); t!(file.write_all(b"").await); t!(file.flush().await); let mut file = t!(File::open(&path).await); let mut header = Header::new_old(); t!(header.set_path("somepath")); header.set_metadata(&t!(file.metadata().await)); header.set_size(2048); // past the end of file null blocks header.set_cksum(); t!(ar.append(&header, &mut file).await); // Extracting let rdr = Cursor::new(t!(ar.into_inner().await)); let mut ar = Archive::new(rdr); assert!(ar.unpack(td.path()).await.is_err()); // Iterating let rdr = Cursor::new(ar.into_inner().map_err(|_| ()).unwrap().into_inner()); let mut ar = Archive::new(rdr); let mut entries = t!(ar.entries()); while let Some(fr) = entries.next().await { if fr.is_err() { return; } } panic!("Should have errorred"); } #[tokio::test] async fn extracting_malicious_tarball() { let td = t!(TempBuilder::new().prefix("async-tar").tempdir()); let mut evil_tar = Vec::new(); evil_tar = { let mut a = Builder::new(evil_tar); async fn append(a: &mut Builder, path: &'static str) { let mut header = Header::new_gnu(); assert!(header.set_path(path).is_err(), "was ok: {:?}", path); { let h = header.as_gnu_mut().unwrap(); for (a, b) in h.name.iter_mut().zip(path.as_bytes()) { *a = *b; } } header.set_size(1); header.set_cksum(); t!(a.append(&header, io::repeat(1).take(1)).await); } append(&mut a, "/tmp/abs_evil.txt").await; // std parse `//` as UNC path, see rust-lang/rust#100833 append( &mut a, #[cfg(not(windows))] "//tmp/abs_evil2.txt", #[cfg(windows)] "C://tmp/abs_evil2.txt", ) .await; append(&mut a, "///tmp/abs_evil3.txt").await; append(&mut a, "/./tmp/abs_evil4.txt").await; append( &mut a, #[cfg(not(windows))] "//./tmp/abs_evil5.txt", #[cfg(windows)] "C://./tmp/abs_evil5.txt", ) .await; append(&mut a, "///./tmp/abs_evil6.txt").await; append(&mut a, "/../tmp/rel_evil.txt").await; append(&mut a, "../rel_evil2.txt").await; append(&mut a, "./../rel_evil3.txt").await; append(&mut a, "some/../../rel_evil4.txt").await; append(&mut a, "").await; append(&mut a, "././//./..").await; append(&mut a, "..").await; append(&mut a, "/////////..").await; append(&mut a, "/////////").await; a.into_inner().await.unwrap() }; let mut ar = Archive::new(&evil_tar[..]); t!(ar.unpack(td.path()).await); assert!(fs::metadata("/tmp/abs_evil.txt").await.is_err()); assert!(fs::metadata("/tmp/abs_evil.txt2").await.is_err()); assert!(fs::metadata("/tmp/abs_evil.txt3").await.is_err()); assert!(fs::metadata("/tmp/abs_evil.txt4").await.is_err()); assert!(fs::metadata("/tmp/abs_evil.txt5").await.is_err()); assert!(fs::metadata("/tmp/abs_evil.txt6").await.is_err()); assert!(fs::metadata("/tmp/rel_evil.txt").await.is_err()); assert!(fs::metadata("/tmp/rel_evil.txt").await.is_err()); assert!(fs::metadata(td.path().join("../tmp/rel_evil.txt")) .await .is_err()); assert!(fs::metadata(td.path().join("../rel_evil2.txt")) .await .is_err()); assert!(fs::metadata(td.path().join("../rel_evil3.txt")) .await .is_err()); assert!(fs::metadata(td.path().join("../rel_evil4.txt")) .await .is_err()); // The `some` subdirectory should not be created because the only // filename that references this has '..'. assert!(fs::metadata(td.path().join("some")).await.is_err()); // The `tmp` subdirectory should be created and within this // subdirectory, there should be files named `abs_evil.txt` through // `abs_evil6.txt`. assert!(fs::metadata(td.path().join("tmp")) .await .map(|m| m.is_dir()) .unwrap_or(false)); assert!(fs::metadata(td.path().join("tmp/abs_evil.txt")) .await .map(|m| m.is_file()) .unwrap_or(false)); assert!(fs::metadata(td.path().join("tmp/abs_evil2.txt")) .await .map(|m| m.is_file()) .unwrap_or(false)); assert!(fs::metadata(td.path().join("tmp/abs_evil3.txt")) .await .map(|m| m.is_file()) .unwrap_or(false)); assert!(fs::metadata(td.path().join("tmp/abs_evil4.txt")) .await .map(|m| m.is_file()) .unwrap_or(false)); assert!(fs::metadata(td.path().join("tmp/abs_evil5.txt")) .await .map(|m| m.is_file()) .unwrap_or(false)); assert!(fs::metadata(td.path().join("tmp/abs_evil6.txt")) .await .map(|m| m.is_file()) .unwrap_or(false)); // Paths "//tmp/abs_evil2.txt" and "//./tmp/abs_evil5.txt" are not absolute for Windows, // hence this test part does not work as expected on this OS. if cfg!(not(windows)) { assert!(fs::metadata(td.path().join("tmp/abs_evil2.txt")) .await .map(|m| m.is_file()) .unwrap_or(false)); assert!(fs::metadata(td.path().join("tmp/abs_evil5.txt")) .await .map(|m| m.is_file()) .unwrap_or(false)); } } #[tokio::test] async fn octal_spaces() { let rdr = Cursor::new(tar!("spaces.tar")); let mut ar = Archive::new(rdr); let entry = ar.entries().unwrap().next().await.unwrap().unwrap(); assert_eq!(entry.header().mode().unwrap() & 0o777, 0o777); assert_eq!(entry.header().uid().unwrap(), 0); assert_eq!(entry.header().gid().unwrap(), 0); assert_eq!(entry.header().size().unwrap(), 2); assert_eq!(entry.header().mtime().unwrap(), 0o12_440_016_664); assert_eq!(entry.header().cksum().unwrap(), 0o4253); } #[tokio::test] async fn extracting_malformed_tar_null_blocks() { let td = t!(TempBuilder::new().prefix("async-tar").tempdir()); let mut ar = Builder::new(Vec::new()); let path1 = td.path().join("tmpfile1"); let path2 = td.path().join("tmpfile2"); t!(File::create(&path1).await); t!(File::create(&path2).await); t!(ar .append_file("tmpfile1", &mut t!(File::open(&path1).await)) .await); let mut data = t!(ar.into_inner().await); let amt = data.len(); data.truncate(amt - 512); let mut ar = Builder::new(data); t!(ar .append_file("tmpfile2", &mut t!(File::open(&path2).await)) .await); t!(ar.finish().await); let data = t!(ar.into_inner().await); let mut ar = Archive::new(&data[..]); assert!(ar.unpack(td.path()).await.is_ok()); } #[tokio::test] async fn empty_filename() { let td = t!(TempBuilder::new().prefix("async-tar").tempdir()); let rdr = Cursor::new(tar!("empty_filename.tar")); let mut ar = Archive::new(rdr); assert!(ar.unpack(td.path()).await.is_ok()); } #[tokio::test] async fn file_times() { let td = t!(TempBuilder::new().prefix("async-tar").tempdir()); let rdr = Cursor::new(tar!("file_times.tar")); let mut ar = Archive::new(rdr); t!(ar.unpack(td.path()).await); let meta = fs::metadata(td.path().join("a")).await.unwrap(); let mtime = FileTime::from_last_modification_time(&meta); let atime = FileTime::from_last_access_time(&meta); assert_eq!(mtime.unix_seconds(), 1_000_000_000); assert_eq!(mtime.nanoseconds(), 0); assert_eq!(atime.unix_seconds(), 1_000_000_000); assert_eq!(atime.nanoseconds(), 0); } #[tokio::test] async fn backslash_treated_well() { // Insert a file into an archive with a backslash let td = t!(TempBuilder::new().prefix("async-tar").tempdir()); let mut ar = Builder::new(Vec::::new()); t!(ar.append_dir("foo\\bar", td.path()).await); let mut ar = Archive::new(Cursor::new(t!(ar.into_inner().await))); let f = t!(t!(ar.entries()).next().await.unwrap()); if cfg!(unix) { assert_eq!(t!(f.header().path()).to_str(), Some("foo\\bar")); } else { assert_eq!(t!(f.header().path()).to_str(), Some("foo/bar")); } // Unpack an archive with a backslash in the name let mut ar = Builder::new(Vec::::new()); let mut header = Header::new_gnu(); header.set_metadata(&t!(fs::metadata(td.path()).await)); header.set_size(0); for (a, b) in header.as_old_mut().name.iter_mut().zip(b"foo\\bar\x00") { *a = *b; } header.set_cksum(); t!(ar.append(&header, &mut io::empty()).await); let data = t!(ar.into_inner().await); let mut ar = Archive::new(&data[..]); let f = t!(t!(ar.entries()).next().await.unwrap()); assert_eq!(t!(f.header().path()).to_str(), Some("foo\\bar")); let mut ar = Archive::new(&data[..]); t!(ar.unpack(td.path()).await); assert!(fs::metadata(td.path().join("foo\\bar")).await.is_ok()); } #[cfg(unix)] #[tokio::test] async fn nul_bytes_in_path() { use std::{ffi::OsStr, os::unix::prelude::*}; let nul_path = OsStr::from_bytes(b"foo\0"); let td = t!(TempBuilder::new().prefix("async-tar").tempdir()); let mut ar = Builder::new(Vec::::new()); let err = ar.append_dir(nul_path, td.path()).await.unwrap_err(); assert!(err.to_string().contains("contains a nul byte")); } #[tokio::test] async fn links() { let mut ar = Archive::new(Cursor::new(tar!("link.tar"))); let mut entries = t!(ar.entries()); let link = t!(entries.next().await.unwrap()); assert_eq!( t!(link.header().link_name()).as_ref().map(|p| &**p), Some(Path::new("file")) ); let other = t!(entries.next().await.unwrap()); assert!(t!(other.header().link_name()).is_none()); } #[tokio::test] #[cfg(unix)] // making symlinks on windows is hard async fn unpack_links() { let td = t!(TempBuilder::new().prefix("async-tar").tempdir()); let mut ar = Archive::new(Cursor::new(tar!("link.tar"))); t!(ar.unpack(td.path()).await); let md = t!(fs::symlink_metadata(td.path().join("lnk")).await); assert!(md.file_type().is_symlink()); let mtime = FileTime::from_last_modification_time(&md); assert_eq!(mtime.unix_seconds(), 1448291033); assert_eq!( &*t!(fs::read_link(td.path().join("lnk")).await), Path::new("file") ); t!(File::open(td.path().join("lnk")).await); } #[tokio::test] async fn pax_simple() { let mut ar = Archive::new(tar!("pax.tar")); let mut entries = t!(ar.entries()); let mut first = t!(entries.next().await.unwrap()); let mut attributes = t!(first.pax_extensions().await).unwrap(); let first = t!(attributes.next().unwrap()); let second = t!(attributes.next().unwrap()); let third = t!(attributes.next().unwrap()); assert!(attributes.next().is_none()); assert_eq!(first.key(), Ok("mtime")); assert_eq!(first.value(), Ok("1453146164.953123768")); assert_eq!(second.key(), Ok("atime")); assert_eq!(second.value(), Ok("1453251915.24892486")); assert_eq!(third.key(), Ok("ctime")); assert_eq!(third.value(), Ok("1453146164.953123768")); } #[tokio::test] async fn pax_pending_interrupted() { use std::pin::Pin; /// A [`AsyncRead`] that returns `Pending` on every other poll. struct PendingReader { inner: R, n: usize, } impl PendingReader where R: AsyncRead + Unpin, { fn new(reader: R) -> Self { Self { inner: reader, n: 0, } } fn project(self: Pin<&mut Self>) -> (Pin<&mut R>, &mut usize) { let Self { inner, n } = std::pin::Pin::into_inner(self); (Pin::new(inner), n) } } impl AsyncRead for PendingReader where R: AsyncRead + Unpin, { fn poll_read( self: Pin<&mut Self>, cx: &mut std::task::Context<'_>, buf: &mut io::ReadBuf<'_>, ) -> std::task::Poll> { use std::task::Poll; let (inner, n) = self.project(); let pend = *n % 2 == 0; *n += 1; if pend { cx.waker().wake_by_ref(); return Poll::Pending; } inner.poll_read(cx, buf) } } let ar = tar!("paxlongname.tar"); let ar = PendingReader::new(ar); let mut ar = Archive::new(ar); let mut entries = t!(ar.entries()); let entry = t!(entries.next().await.unwrap()); let path = t!(entry.path()); let path = path.to_str().unwrap(); assert_eq!(path, "this_file_name_will_be_one_hundred_and_one_characters_long_once_i_add_some_more_characters_at_the_end"); } #[tokio::test] async fn pax_path() { let mut ar = Archive::new(tar!("pax2.tar")); let mut entries = t!(ar.entries()); let first = t!(entries.next().await.unwrap()); assert!(first.path().unwrap().ends_with("aaaaaaaaaaaaaaa")); } #[tokio::test] async fn long_name_trailing_nul() { let mut b = Builder::new(Vec::::new()); let mut h = Header::new_gnu(); t!(h.set_path("././@LongLink")); h.set_size(4); h.set_entry_type(EntryType::new(b'L')); h.set_cksum(); t!(b.append(&h, b"foo\0" as &[u8]).await); let mut h = Header::new_gnu(); t!(h.set_path("bar")); h.set_size(6); h.set_entry_type(EntryType::file()); h.set_cksum(); t!(b.append(&h, b"foobar" as &[u8]).await); let contents = t!(b.into_inner().await); let mut a = Archive::new(&contents[..]); let e = t!(t!(a.entries()).next().await.unwrap()); assert_eq!(&*e.path_bytes(), b"foo"); } #[tokio::test] async fn long_linkname_trailing_nul() { let mut b = Builder::new(Vec::::new()); let mut h = Header::new_gnu(); t!(h.set_path("././@LongLink")); h.set_size(4); h.set_entry_type(EntryType::new(b'K')); h.set_cksum(); t!(b.append(&h, b"foo\0" as &[u8]).await); let mut h = Header::new_gnu(); t!(h.set_path("bar")); h.set_size(6); h.set_entry_type(EntryType::file()); h.set_cksum(); t!(b.append(&h, b"foobar" as &[u8]).await); let contents = t!(b.into_inner().await); let mut a = Archive::new(&contents[..]); let e = t!(t!(a.entries()).next().await.unwrap()); assert_eq!(&*e.link_name_bytes().unwrap(), b"foo"); } #[tokio::test] async fn encoded_long_name_has_trailing_nul() { let td = t!(TempBuilder::new().prefix("async-tar").tempdir()); let path = td.path().join("foo"); t!(t!(File::create(&path).await).write_all(b"test").await); let mut b = Builder::new(Vec::::new()); let long = "abcd".repeat(200); t!(b.append_file(&long, &mut t!(File::open(&path).await)).await); let contents = t!(b.into_inner().await); let mut a = Archive::new(&contents[..]); let mut e = t!(t!(a.entries_raw()).next().await.unwrap()); let mut name = Vec::new(); t!(e.read_to_end(&mut name).await); assert_eq!(name[name.len() - 1], 0); let header_name = &e.header().as_gnu().unwrap().name; assert!(header_name.starts_with(b"././@LongLink\x00")); } #[tokio::test] async fn reading_sparse() { let rdr = Cursor::new(tar!("sparse.tar")); let mut ar = Archive::new(rdr); let mut entries = t!(ar.entries()); let mut a = t!(entries.next().await.unwrap()); let mut s = String::new(); assert_eq!(&*a.header().path_bytes(), b"sparse_begin.txt"); t!(a.read_to_string(&mut s).await); assert_eq!(&s[..5], "test\n"); assert!(s[5..].chars().all(|x| x == '\u{0}')); let mut a = t!(entries.next().await.unwrap()); let mut s = String::new(); assert_eq!(&*a.header().path_bytes(), b"sparse_end.txt"); t!(a.read_to_string(&mut s).await); assert!(s[..s.len() - 9].chars().all(|x| x == '\u{0}')); assert_eq!(&s[s.len() - 9..], "test_end\n"); let mut a = t!(entries.next().await.unwrap()); let mut s = String::new(); assert_eq!(&*a.header().path_bytes(), b"sparse_ext.txt"); t!(a.read_to_string(&mut s).await); assert!(s[..0x1000].chars().all(|x| x == '\u{0}')); assert_eq!(&s[0x1000..0x1000 + 5], "text\n"); assert!(s[0x1000 + 5..0x3000].chars().all(|x| x == '\u{0}')); assert_eq!(&s[0x3000..0x3000 + 5], "text\n"); assert!(s[0x3000 + 5..0x5000].chars().all(|x| x == '\u{0}')); assert_eq!(&s[0x5000..0x5000 + 5], "text\n"); assert!(s[0x5000 + 5..0x7000].chars().all(|x| x == '\u{0}')); assert_eq!(&s[0x7000..0x7000 + 5], "text\n"); assert!(s[0x7000 + 5..0x9000].chars().all(|x| x == '\u{0}')); assert_eq!(&s[0x9000..0x9000 + 5], "text\n"); assert!(s[0x9000 + 5..0xb000].chars().all(|x| x == '\u{0}')); assert_eq!(&s[0xb000..0xb000 + 5], "text\n"); let mut a = t!(entries.next().await.unwrap()); let mut s = String::new(); assert_eq!(&*a.header().path_bytes(), b"sparse.txt"); t!(a.read_to_string(&mut s).await); assert!(s[..0x1000].chars().all(|x| x == '\u{0}')); assert_eq!(&s[0x1000..0x1000 + 6], "hello\n"); assert!(s[0x1000 + 6..0x2fa0].chars().all(|x| x == '\u{0}')); assert_eq!(&s[0x2fa0..0x2fa0 + 6], "world\n"); assert!(s[0x2fa0 + 6..0x4000].chars().all(|x| x == '\u{0}')); assert!(entries.next().await.is_none()); } #[tokio::test] async fn extract_sparse() { let rdr = Cursor::new(tar!("sparse.tar")); let mut ar = Archive::new(rdr); let td = t!(TempBuilder::new().prefix("async-tar").tempdir()); t!(ar.unpack(td.path()).await); let mut s = String::new(); t!(t!(File::open(td.path().join("sparse_begin.txt")).await) .read_to_string(&mut s) .await); assert_eq!(&s[..5], "test\n"); assert!(s[5..].chars().all(|x| x == '\u{0}')); s.truncate(0); t!(t!(File::open(td.path().join("sparse_end.txt")).await) .read_to_string(&mut s) .await); assert!(s[..s.len() - 9].chars().all(|x| x == '\u{0}')); assert_eq!(&s[s.len() - 9..], "test_end\n"); s.truncate(0); t!(t!(File::open(td.path().join("sparse_ext.txt")).await) .read_to_string(&mut s) .await); assert!(s[..0x1000].chars().all(|x| x == '\u{0}')); assert_eq!(&s[0x1000..0x1000 + 5], "text\n"); assert!(s[0x1000 + 5..0x3000].chars().all(|x| x == '\u{0}')); assert_eq!(&s[0x3000..0x3000 + 5], "text\n"); assert!(s[0x3000 + 5..0x5000].chars().all(|x| x == '\u{0}')); assert_eq!(&s[0x5000..0x5000 + 5], "text\n"); assert!(s[0x5000 + 5..0x7000].chars().all(|x| x == '\u{0}')); assert_eq!(&s[0x7000..0x7000 + 5], "text\n"); assert!(s[0x7000 + 5..0x9000].chars().all(|x| x == '\u{0}')); assert_eq!(&s[0x9000..0x9000 + 5], "text\n"); assert!(s[0x9000 + 5..0xb000].chars().all(|x| x == '\u{0}')); assert_eq!(&s[0xb000..0xb000 + 5], "text\n"); s.truncate(0); t!(t!(File::open(td.path().join("sparse.txt")).await) .read_to_string(&mut s) .await); assert!(s[..0x1000].chars().all(|x| x == '\u{0}')); assert_eq!(&s[0x1000..0x1000 + 6], "hello\n"); assert!(s[0x1000 + 6..0x2fa0].chars().all(|x| x == '\u{0}')); assert_eq!(&s[0x2fa0..0x2fa0 + 6], "world\n"); assert!(s[0x2fa0 + 6..0x4000].chars().all(|x| x == '\u{0}')); } #[tokio::test] async fn large_sparse() { let rdr = Cursor::new(tar!("sparse-large.tar")); let mut ar = Archive::new(rdr); let mut entries = t!(ar.entries()); // Only check the header info without extracting, as the file is very large, // and not all filesystems support sparse files. let a = t!(entries.next().await.unwrap()); let h = a.header().as_gnu().unwrap(); assert_eq!(h.real_size().unwrap(), 12626929280); } #[tokio::test] async fn sparse_with_trailing() { let rdr = Cursor::new(tar!("sparse-1.tar")); let mut ar = Archive::new(rdr); let mut entries = t!(ar.entries()); let mut a = t!(entries.next().await.unwrap()); let mut s = String::new(); t!(a.read_to_string(&mut s).await); assert_eq!(0x100_00c, s.len()); assert_eq!(&s[..0xc], "0MB through\n"); assert!(s[0xc..0x100_000].chars().all(|x| x == '\u{0}')); assert_eq!(&s[0x100_000..], "1MB through\n"); } #[tokio::test] async fn path_separators() { let mut ar = Builder::new(Vec::new()); let td = t!(TempBuilder::new().prefix("async-tar").tempdir()); let path = td.path().join("test"); t!(t!(File::create(&path).await).write_all(b"test").await); let short_path: PathBuf = std::iter::repeat("abcd").take(2).collect(); let long_path: PathBuf = std::iter::repeat("abcd").take(50).collect(); // Make sure UStar headers normalize to Unix path separators let mut header = Header::new_ustar(); t!(header.set_path(&short_path)); assert_eq!(t!(header.path()), short_path); assert!(!header.path_bytes().contains(&b'\\')); t!(header.set_path(&long_path)); assert_eq!(t!(header.path()), long_path); assert!(!header.path_bytes().contains(&b'\\')); // Make sure GNU headers normalize to Unix path separators, // including the `@LongLink` fallback used by `append_file`. t!(ar .append_file(&short_path, &mut t!(File::open(&path).await)) .await); t!(ar .append_file(&long_path, &mut t!(File::open(&path).await)) .await); let rd = Cursor::new(t!(ar.into_inner().await)); let mut ar = Archive::new(rd); let mut entries = t!(ar.entries()); let entry = t!(entries.next().await.unwrap()); assert_eq!(t!(entry.path()), short_path); assert!(!entry.path_bytes().contains(&b'\\')); let entry = t!(entries.next().await.unwrap()); assert_eq!(t!(entry.path()), long_path); assert!(!entry.path_bytes().contains(&b'\\')); assert!(entries.next().await.is_none()); } #[tokio::test] #[cfg(unix)] async fn append_path_symlink() { use std::{borrow::Cow, env, os::unix::fs::symlink}; let mut ar = Builder::new(Vec::new()); ar.follow_symlinks(false); let td = t!(TempBuilder::new().prefix("async-tar").tempdir()); let long_linkname = "abcd".repeat(30); let long_pathname = "dcba".repeat(30); t!(env::set_current_dir(td.path())); // "short" path name / short link name t!(symlink("testdest", "test")); t!(ar.append_path("test").await); // short path name / long link name t!(symlink(&long_linkname, "test2")); t!(ar.append_path("test2").await); // long path name / long link name t!(symlink(&long_linkname, &long_pathname)); t!(ar.append_path(&long_pathname).await); let rd = Cursor::new(t!(ar.into_inner().await)); let mut ar = Archive::new(rd); let mut entries = t!(ar.entries()); let entry = t!(entries.next().await.unwrap()); assert_eq!(t!(entry.path()), Path::new("test")); assert_eq!( t!(entry.link_name()), Some(Cow::from(Path::new("testdest"))) ); assert_eq!(t!(entry.header().size()), 0); let entry = t!(entries.next().await.unwrap()); assert_eq!(t!(entry.path()), Path::new("test2")); assert_eq!( t!(entry.link_name()), Some(Cow::from(Path::new(&long_linkname))) ); assert_eq!(t!(entry.header().size()), 0); let entry = t!(entries.next().await.unwrap()); assert_eq!(t!(entry.path()), Path::new(&long_pathname)); assert_eq!( t!(entry.link_name()), Some(Cow::from(Path::new(&long_linkname))) ); assert_eq!(t!(entry.header().size()), 0); assert!(entries.next().await.is_none()); } #[tokio::test] async fn name_with_slash_doesnt_fool_long_link_and_bsd_compat() { let td = t!(TempBuilder::new().prefix("async-tar").tempdir()); let mut ar = Builder::new(Vec::new()); let mut h = Header::new_gnu(); t!(h.set_path("././@LongLink")); h.set_size(4); h.set_entry_type(EntryType::new(b'L')); h.set_cksum(); t!(ar.append(&h, b"foo\0" as &[u8]).await); let mut header = Header::new_gnu(); header.set_entry_type(EntryType::Regular); t!(header.set_path("testdir/")); header.set_size(0); header.set_cksum(); t!(ar.append(&header, &mut io::empty()).await); // Extracting let rdr = Cursor::new(t!(ar.into_inner().await)); let mut ar = Archive::new(rdr); t!(ar.unpack(td.path()).await); // Iterating let rdr = Cursor::new(ar.into_inner().map_err(|_| ()).unwrap().into_inner()); let mut ar = Archive::new(rdr); let mut entries = t!(ar.entries()); while let Some(entry) = entries.next().await { assert!(entry.is_ok()); } assert!(td.path().join("foo").is_file()); } #[tokio::test] async fn insert_local_file_different_name() { let mut ar = Builder::new(Vec::new()); let td = t!(TempBuilder::new().prefix("async-tar").tempdir()); let path = td.path().join("directory"); t!(fs::create_dir(&path).await); ar.append_path_with_name(&path, "archive/dir") .await .unwrap(); let path = td.path().join("file"); let mut file = t!(File::create(&path).await); t!(file.write_all(b"test").await); t!(file.flush().await); ar.append_path_with_name(&path, "archive/dir/f") .await .unwrap(); let rd = Cursor::new(t!(ar.into_inner().await)); let mut ar = Archive::new(rd); let mut entries = t!(ar.entries()); let entry = t!(entries.next().await.unwrap()); assert_eq!(t!(entry.path()), Path::new("archive/dir")); let entry = t!(entries.next().await.unwrap()); assert_eq!(t!(entry.path()), Path::new("archive/dir/f")); assert!(entries.next().await.is_none()); } #[tokio::test] #[cfg(unix)] async fn tar_directory_containing_symlink_to_directory() { use std::os::unix::fs::symlink; let td = t!(TempBuilder::new().prefix("async-tar").tempdir()); let dummy_src = t!(TempBuilder::new().prefix("dummy_src").tempdir()); let dummy_dst = td.path().join("dummy_dst"); let mut ar = Builder::new(Vec::new()); t!(symlink(dummy_src.path().display().to_string(), &dummy_dst)); assert!(dummy_dst.read_link().is_ok()); assert!(dummy_dst.read_link().unwrap().is_dir()); ar.append_dir_all("symlinks", td.path()).await.unwrap(); ar.finish().await.unwrap(); } #[tokio::test] async fn long_path() { let td = t!(TempBuilder::new().prefix("async-tar").tempdir()); let rdr = Cursor::new(tar!("7z_long_path.tar")); let mut ar = Archive::new(rdr); ar.unpack(td.path()).await.unwrap(); } #[tokio::test] async fn unpack_path_larger_than_windows_max_path() { let dir_name = "iamaprettylongnameandtobepreciseiam91characterslongwhichsomethinkisreallylongandothersdonot"; // 183 character directory name let really_long_path = format!("{}{}", dir_name, dir_name); let td = t!(TempBuilder::new().prefix(&really_long_path).tempdir()); // directory in 7z_long_path.tar is over 100 chars let rdr = Cursor::new(tar!("7z_long_path.tar")); let mut ar = Archive::new(rdr); // should unpack path greater than windows MAX_PATH length of 260 characters assert!(ar.unpack(td.path()).await.is_ok()); } #[tokio::test] async fn append_long_multibyte() { let mut x = Builder::new(Vec::new()); let mut name = String::new(); let data: &[u8] = &[]; for _ in 0..512 { name.push('a'); name.push('ð‘¢®'); x.append_data(&mut Header::new_gnu(), &name, data) .await .unwrap(); name.pop(); } } #[tokio::test] async fn read_only_directory_containing_files() { let td = t!(TempBuilder::new().prefix("async-tar").tempdir()); let mut b = Builder::new(Vec::::new()); let mut h = Header::new_gnu(); t!(h.set_path("dir/")); h.set_size(0); h.set_entry_type(EntryType::dir()); h.set_mode(0o444); h.set_cksum(); t!(b.append(&h, "".as_bytes()).await); let mut h = Header::new_gnu(); t!(h.set_path("dir/file")); h.set_size(2); h.set_entry_type(EntryType::file()); h.set_cksum(); t!(b.append(&h, "hi".as_bytes()).await); let contents = t!(b.into_inner().await); let mut ar = Archive::new(&contents[..]); assert!(ar.unpack(td.path()).await.is_ok()); } // This test was marked linux only due to macOS CI can't handle `set_current_dir` correctly #[tokio::test] #[cfg(target_os = "linux")] async fn tar_directory_containing_special_files() { use std::env; use std::ffi::CString; let td = t!(TempBuilder::new().prefix("async-tar").tempdir()); let fifo = td.path().join("fifo"); unsafe { let fifo_path = t!(CString::new(fifo.to_str().unwrap())); let ret = libc::mknod(fifo_path.as_ptr(), libc::S_IFIFO | 0o644, 0); if ret != 0 { libc::perror(fifo_path.as_ptr()); panic!("Failed to create a FIFO file"); } } t!(env::set_current_dir(td.path())); let mut ar = Builder::new(Vec::new()); // append_path has a different logic for processing files, so we need to test it as well t!(ar.append_path("fifo").await); t!(ar.append_dir_all("special", td.path()).await); t!(env::set_current_dir("/dev/")); // CI systems seem to have issues with creating a chr device t!(ar.append_path("null").await); t!(ar.finish().await); } #[tokio::test] async fn header_size_overflow() { // maximal file size doesn't overflow anything let mut ar = Builder::new(Vec::new()); let mut header = Header::new_gnu(); header.set_size(u64::MAX); header.set_cksum(); t!(ar.append(&header, "x".as_bytes()).await); let result = t!(ar.into_inner().await); let mut ar = Archive::new(&result[..]); let mut e = t!(ar.entries()); let entry = e.next().await.unwrap(); assert!(entry.is_err(), "expected error for size overflow"); let err = entry.unwrap_err(); assert!( err.to_string().contains("size overflow"), "bad error: {}", err ); // back-to-back entries that would overflow also don't panic let mut ar = Builder::new(Vec::new()); let mut header = Header::new_gnu(); header.set_size(1_000); header.set_cksum(); t!(ar.append(&header, &[0u8; 1_000][..]).await); let mut header = Header::new_gnu(); header.set_size(u64::MAX - 513); header.set_cksum(); t!(ar.append(&header, "x".as_bytes()).await); let result = t!(ar.into_inner().await); let mut ar = Archive::new(&result[..]); let mut e = t!(ar.entries()); let first = e.next().await.unwrap(); t!(first); // First entry should be ok let second = e.next().await.unwrap(); assert!(second.is_err(), "expected error for size overflow"); let err = second.unwrap_err(); assert!( err.to_string().contains("size overflow"), "bad error: {}", err ); } astral-tokio-tar-0.5.1/tests/entry.rs000064400000000000000000000304221046102023000156760ustar 00000000000000extern crate tokio_tar as async_tar; extern crate tempfile; #[cfg_attr(windows, allow(unused_imports))] use tokio::{fs, fs::File, io::AsyncReadExt}; use tokio_stream::*; use tempfile::Builder; macro_rules! t { ($e:expr) => { match $e { Ok(v) => v, Err(e) => panic!("{} returned {}", stringify!($e), e), } }; } #[tokio::test] async fn absolute_symlink() { let mut ar = async_tar::Builder::new(Vec::new()); let mut header = async_tar::Header::new_gnu(); header.set_size(0); header.set_entry_type(async_tar::EntryType::Symlink); t!(header.set_path("foo")); t!(header.set_link_name("/bar")); header.set_cksum(); t!(ar.append(&header, &[][..]).await); let bytes = t!(ar.into_inner().await); let mut ar = async_tar::Archive::new(&bytes[..]); let td = t!(Builder::new().prefix("tar").tempdir()); t!(ar.unpack(td.path()).await); t!(td.path().join("foo").symlink_metadata()); let mut ar = async_tar::Archive::new(&bytes[..]); let mut entries = t!(ar.entries()); let entry = t!(entries.next().await.unwrap()); assert_eq!(&*entry.link_name_bytes().unwrap(), b"/bar"); } #[tokio::test] async fn absolute_hardlink() { let td = t!(Builder::new().prefix("tar").tempdir()); let mut ar = async_tar::Builder::new(Vec::new()); let mut header = async_tar::Header::new_gnu(); header.set_size(0); header.set_entry_type(async_tar::EntryType::Regular); t!(header.set_path("foo")); header.set_cksum(); t!(ar.append(&header, &[][..]).await); let mut header = async_tar::Header::new_gnu(); header.set_size(0); header.set_entry_type(async_tar::EntryType::Link); t!(header.set_path("bar")); // This absolute path under tempdir will be created at unpack time t!(header.set_link_name(td.path().join("foo"))); header.set_cksum(); t!(ar.append(&header, &[][..]).await); let bytes = t!(ar.into_inner().await); let mut ar = async_tar::Archive::new(&bytes[..]); t!(ar.unpack(td.path()).await); t!(td.path().join("foo").metadata()); t!(td.path().join("bar").metadata()); } #[tokio::test] async fn relative_hardlink() { let mut ar = async_tar::Builder::new(Vec::new()); let mut header = async_tar::Header::new_gnu(); header.set_size(0); header.set_entry_type(async_tar::EntryType::Regular); t!(header.set_path("foo")); header.set_cksum(); t!(ar.append(&header, &[][..]).await); let mut header = async_tar::Header::new_gnu(); header.set_size(0); header.set_entry_type(async_tar::EntryType::Link); t!(header.set_path("bar")); t!(header.set_link_name("foo")); header.set_cksum(); t!(ar.append(&header, &[][..]).await); let bytes = t!(ar.into_inner().await); let mut ar = async_tar::Archive::new(&bytes[..]); let td = t!(Builder::new().prefix("tar").tempdir()); t!(ar.unpack(td.path()).await); t!(td.path().join("foo").metadata()); t!(td.path().join("bar").metadata()); } #[tokio::test] async fn absolute_link_deref_error() { let mut ar = async_tar::Builder::new(Vec::new()); let mut header = async_tar::Header::new_gnu(); header.set_size(0); header.set_entry_type(async_tar::EntryType::Symlink); t!(header.set_path("foo")); t!(header.set_link_name("/")); header.set_cksum(); t!(ar.append(&header, &[][..]).await); let mut header = async_tar::Header::new_gnu(); header.set_size(0); header.set_entry_type(async_tar::EntryType::Regular); t!(header.set_path("foo/bar")); header.set_cksum(); t!(ar.append(&header, &[][..]).await); let bytes = t!(ar.into_inner().await); let mut ar = async_tar::Archive::new(&bytes[..]); let td = t!(Builder::new().prefix("tar").tempdir()); assert!(ar.unpack(td.path()).await.is_err()); t!(td.path().join("foo").symlink_metadata()); assert!(File::open(td.path().join("foo").join("bar")).await.is_err()); } #[tokio::test] async fn relative_link_deref_error() { let mut ar = async_tar::Builder::new(Vec::new()); let mut header = async_tar::Header::new_gnu(); header.set_size(0); header.set_entry_type(async_tar::EntryType::Symlink); t!(header.set_path("foo")); t!(header.set_link_name("../../../../")); header.set_cksum(); t!(ar.append(&header, &[][..]).await); let mut header = async_tar::Header::new_gnu(); header.set_size(0); header.set_entry_type(async_tar::EntryType::Regular); t!(header.set_path("foo/bar")); header.set_cksum(); t!(ar.append(&header, &[][..]).await); let bytes = t!(ar.into_inner().await); let mut ar = async_tar::Archive::new(&bytes[..]); let td = t!(Builder::new().prefix("tar").tempdir()); assert!(ar.unpack(td.path()).await.is_err()); t!(td.path().join("foo").symlink_metadata()); assert!(File::open(td.path().join("foo").join("bar")).await.is_err()); } #[tokio::test] #[cfg(unix)] async fn directory_permissions_are_cleared() { use ::std::os::unix::fs::PermissionsExt; let mut ar = async_tar::Builder::new(Vec::new()); let mut header = async_tar::Header::new_gnu(); header.set_size(0); header.set_entry_type(async_tar::EntryType::Directory); t!(header.set_path("foo")); header.set_mode(0o777); header.set_cksum(); t!(ar.append(&header, &[][..]).await); let bytes = t!(ar.into_inner().await); let mut ar = async_tar::Archive::new(&bytes[..]); let td = t!(Builder::new().prefix("tar").tempdir()); t!(ar.unpack(td.path()).await); let f = t!(File::open(td.path().join("foo")).await); let from_archive_md = t!(f.metadata().await); assert!(from_archive_md.is_dir()); // To determine the default umask, create a fresh directory that gets it assigned by the OS. let manually_created = td.path().join("bar"); t!(fs::create_dir(&manually_created).await); let f = t!(File::open(&manually_created).await); let loca_md = t!(f.metadata().await); assert_eq!( from_archive_md.permissions().mode(), loca_md.permissions().mode() ); } #[tokio::test] #[cfg(not(windows))] // dangling symlinks have weird permissions async fn modify_link_just_created() { let mut ar = async_tar::Builder::new(Vec::new()); let mut header = async_tar::Header::new_gnu(); header.set_size(0); header.set_entry_type(async_tar::EntryType::Symlink); t!(header.set_path("foo")); t!(header.set_link_name("bar")); header.set_cksum(); t!(ar.append(&header, &[][..]).await); let mut header = async_tar::Header::new_gnu(); header.set_size(0); header.set_entry_type(async_tar::EntryType::Regular); t!(header.set_path("bar/foo")); header.set_cksum(); t!(ar.append(&header, &[][..]).await); let mut header = async_tar::Header::new_gnu(); header.set_size(0); header.set_entry_type(async_tar::EntryType::Regular); t!(header.set_path("foo/bar")); header.set_cksum(); t!(ar.append(&header, &[][..]).await); let bytes = t!(ar.into_inner().await); let mut ar = async_tar::Archive::new(&bytes[..]); let td = t!(Builder::new().prefix("tar").tempdir()); t!(ar.unpack(td.path()).await); t!(File::open(td.path().join("bar/foo")).await); t!(File::open(td.path().join("bar/bar")).await); t!(File::open(td.path().join("foo/foo")).await); t!(File::open(td.path().join("foo/bar")).await); } #[tokio::test] #[cfg(not(windows))] // dangling symlinks have weird permissions async fn modify_outside_with_relative_symlink() { let mut ar = async_tar::Builder::new(Vec::new()); let mut header = async_tar::Header::new_gnu(); header.set_size(0); header.set_entry_type(async_tar::EntryType::Symlink); t!(header.set_path("symlink")); t!(header.set_link_name("..")); header.set_cksum(); t!(ar.append(&header, &[][..]).await); let mut header = async_tar::Header::new_gnu(); header.set_size(0); header.set_entry_type(async_tar::EntryType::Regular); t!(header.set_path("symlink/foo/bar")); header.set_cksum(); t!(ar.append(&header, &[][..]).await); let bytes = t!(ar.into_inner().await); let mut ar = async_tar::Archive::new(&bytes[..]); let td = t!(Builder::new().prefix("tar").tempdir()); let tar_dir = td.path().join("tar"); assert!(ar.unpack(tar_dir).await.is_err()); assert!(!td.path().join("foo").exists()); } #[tokio::test] async fn parent_paths_error() { let mut ar = async_tar::Builder::new(Vec::new()); let mut header = async_tar::Header::new_gnu(); header.set_size(0); header.set_entry_type(async_tar::EntryType::Symlink); t!(header.set_path("foo")); t!(header.set_link_name("..")); header.set_cksum(); t!(ar.append(&header, &[][..]).await); let mut header = async_tar::Header::new_gnu(); header.set_size(0); header.set_entry_type(async_tar::EntryType::Regular); t!(header.set_path("foo/bar")); header.set_cksum(); t!(ar.append(&header, &[][..]).await); let bytes = t!(ar.into_inner().await); let mut ar = async_tar::Archive::new(&bytes[..]); let td = t!(Builder::new().prefix("tar").tempdir()); assert!(ar.unpack(td.path()).await.is_err()); t!(td.path().join("foo").symlink_metadata()); assert!(File::open(td.path().join("foo").join("bar")).await.is_err()); } #[tokio::test] #[cfg(unix)] async fn good_parent_paths_ok() { use std::path::PathBuf; let mut ar = async_tar::Builder::new(Vec::new()); let mut header = async_tar::Header::new_gnu(); header.set_size(0); header.set_entry_type(async_tar::EntryType::Symlink); t!(header.set_path(PathBuf::from("foo").join("bar"))); t!(header.set_link_name(PathBuf::from("..").join("bar"))); header.set_cksum(); t!(ar.append(&header, &[][..]).await); let mut header = async_tar::Header::new_gnu(); header.set_size(0); header.set_entry_type(async_tar::EntryType::Regular); t!(header.set_path("bar")); header.set_cksum(); t!(ar.append(&header, &[][..]).await); let bytes = t!(ar.into_inner().await); let mut ar = async_tar::Archive::new(&bytes[..]); let td = t!(Builder::new().prefix("tar").tempdir()); t!(ar.unpack(td.path()).await); t!(td.path().join("foo").join("bar").read_link()); let dst = t!(td.path().join("foo").join("bar").canonicalize()); t!(File::open(dst).await); } #[tokio::test] async fn modify_hard_link_just_created() { let mut ar = async_tar::Builder::new(Vec::new()); let mut header = async_tar::Header::new_gnu(); header.set_size(0); header.set_entry_type(async_tar::EntryType::Link); t!(header.set_path("foo")); t!(header.set_link_name("../test")); header.set_cksum(); t!(ar.append(&header, &[][..]).await); let mut header = async_tar::Header::new_gnu(); header.set_size(1); header.set_entry_type(async_tar::EntryType::Regular); t!(header.set_path("foo")); header.set_cksum(); t!(ar.append(&header, &b"x"[..]).await); let bytes = t!(ar.into_inner().await); let mut ar = async_tar::Archive::new(&bytes[..]); let td = t!(Builder::new().prefix("tar").tempdir()); let test = td.path().join("test"); t!(File::create(&test).await); let dir = td.path().join("dir"); assert!(ar.unpack(&dir).await.is_err()); let mut contents = Vec::new(); t!(t!(File::open(&test).await).read_to_end(&mut contents).await); assert_eq!(contents.len(), 0); } #[tokio::test] async fn modify_symlink_just_created() { let mut ar = async_tar::Builder::new(Vec::new()); let mut header = async_tar::Header::new_gnu(); header.set_size(0); header.set_entry_type(async_tar::EntryType::Symlink); t!(header.set_path("foo")); t!(header.set_link_name("../test")); header.set_cksum(); t!(ar.append(&header, &[][..]).await); let mut header = async_tar::Header::new_gnu(); header.set_size(1); header.set_entry_type(async_tar::EntryType::Regular); t!(header.set_path("foo")); header.set_cksum(); t!(ar.append(&header, &b"x"[..]).await); let bytes = t!(ar.into_inner().await); let mut ar = async_tar::Archive::new(&bytes[..]); let td = t!(Builder::new().prefix("tar").tempdir()); let test = td.path().join("test"); t!(File::create(&test).await); let dir = td.path().join("dir"); t!(ar.unpack(&dir).await); let mut contents = Vec::new(); t!(t!(File::open(&test).await).read_to_end(&mut contents).await); assert_eq!(contents.len(), 0); } astral-tokio-tar-0.5.1/tests/header/mod.rs000064400000000000000000000204131046102023000165430ustar 00000000000000#![allow(clippy::cognitive_complexity)] use std::{ fs::{self, File}, io::Write, mem, path::Path, thread, time, }; use tempfile::Builder; use async_tar::{GnuHeader, Header, HeaderMode}; #[test] fn default_gnu() { let mut h = Header::new_gnu(); assert!(h.as_gnu().is_some()); assert!(h.as_gnu_mut().is_some()); assert!(h.as_ustar().is_none()); assert!(h.as_ustar_mut().is_none()); } #[test] fn goto_old() { let mut h = Header::new_old(); assert!(h.as_gnu().is_none()); assert!(h.as_gnu_mut().is_none()); assert!(h.as_ustar().is_none()); assert!(h.as_ustar_mut().is_none()); } #[test] fn goto_ustar() { let mut h = Header::new_ustar(); assert!(h.as_gnu().is_none()); assert!(h.as_gnu_mut().is_none()); assert!(h.as_ustar().is_some()); assert!(h.as_ustar_mut().is_some()); } #[test] fn link_name() { let mut h = Header::new_gnu(); t!(h.set_link_name("foo")); assert_eq!(t!(h.link_name()).unwrap().to_str(), Some("foo")); t!(h.set_link_name("../foo")); assert_eq!(t!(h.link_name()).unwrap().to_str(), Some("../foo")); t!(h.set_link_name("foo/bar")); assert_eq!(t!(h.link_name()).unwrap().to_str(), Some("foo/bar")); t!(h.set_link_name("foo\\ba")); if cfg!(windows) { assert_eq!(t!(h.link_name()).unwrap().to_str(), Some("foo/ba")); } else { assert_eq!(t!(h.link_name()).unwrap().to_str(), Some("foo\\ba")); } let name = "foo\\bar\0"; for (slot, val) in h.as_old_mut().linkname.iter_mut().zip(name.as_bytes()) { *slot = *val; } assert_eq!(t!(h.link_name()).unwrap().to_str(), Some("foo\\bar")); assert!(h.set_link_name("\0").is_err()); } #[test] fn mtime() { let h = Header::new_gnu(); assert_eq!(t!(h.mtime()), 0); let h = Header::new_ustar(); assert_eq!(t!(h.mtime()), 0); let h = Header::new_old(); assert_eq!(t!(h.mtime()), 0); } #[test] fn user_and_group_name() { let mut h = Header::new_gnu(); t!(h.set_username("foo")); t!(h.set_groupname("bar")); assert_eq!(t!(h.username()), Some("foo")); assert_eq!(t!(h.groupname()), Some("bar")); h = Header::new_ustar(); t!(h.set_username("foo")); t!(h.set_groupname("bar")); assert_eq!(t!(h.username()), Some("foo")); assert_eq!(t!(h.groupname()), Some("bar")); h = Header::new_old(); assert_eq!(t!(h.username()), None); assert_eq!(t!(h.groupname()), None); assert!(h.set_username("foo").is_err()); assert!(h.set_groupname("foo").is_err()); } #[test] fn dev_major_minor() { let mut h = Header::new_gnu(); t!(h.set_device_major(1)); t!(h.set_device_minor(2)); assert_eq!(t!(h.device_major()), Some(1)); assert_eq!(t!(h.device_minor()), Some(2)); h = Header::new_ustar(); t!(h.set_device_major(1)); t!(h.set_device_minor(2)); assert_eq!(t!(h.device_major()), Some(1)); assert_eq!(t!(h.device_minor()), Some(2)); h.as_ustar_mut().unwrap().dev_minor[0] = 0x7f; h.as_ustar_mut().unwrap().dev_major[0] = 0x7f; assert!(h.device_major().is_err()); assert!(h.device_minor().is_err()); h.as_ustar_mut().unwrap().dev_minor[0] = b'g'; h.as_ustar_mut().unwrap().dev_major[0] = b'h'; assert!(h.device_major().is_err()); assert!(h.device_minor().is_err()); h = Header::new_old(); assert_eq!(t!(h.device_major()), None); assert_eq!(t!(h.device_minor()), None); assert!(h.set_device_major(1).is_err()); assert!(h.set_device_minor(1).is_err()); } #[test] fn set_path() { let mut h = Header::new_gnu(); t!(h.set_path("foo")); assert_eq!(t!(h.path()).to_str(), Some("foo")); t!(h.set_path("foo/")); assert_eq!(t!(h.path()).to_str(), Some("foo/")); t!(h.set_path("foo/bar")); assert_eq!(t!(h.path()).to_str(), Some("foo/bar")); t!(h.set_path("foo\\bar")); if cfg!(windows) { assert_eq!(t!(h.path()).to_str(), Some("foo/bar")); } else { assert_eq!(t!(h.path()).to_str(), Some("foo\\bar")); } // set_path documentation explictly states it removes any ".", signfying the // current directory, from the path. This test ensures that documented // beavhior occurs t!(h.set_path("./control")); assert_eq!(t!(h.path()).to_str(), Some("control")); let long_name = "foo".repeat(100); let medium1 = "foo".repeat(52); let medium2 = "fo/".repeat(52); assert!(h.set_path(&long_name).is_err()); assert!(h.set_path(&medium1).is_err()); assert!(h.set_path(&medium2).is_err()); assert!(h.set_path("\0").is_err()); assert!(h.set_path("..").is_err()); assert!(h.set_path("foo/..").is_err()); assert!(h.set_path("foo/../bar").is_err()); h = Header::new_ustar(); t!(h.set_path("foo")); assert_eq!(t!(h.path()).to_str(), Some("foo")); assert!(h.set_path(&long_name).is_err()); assert!(h.set_path(&medium1).is_err()); t!(h.set_path(&medium2)); assert_eq!(t!(h.path()).to_str(), Some(&medium2[..])); } #[test] fn set_ustar_path_hard() { let mut h = Header::new_ustar(); let p = Path::new("a").join(vec!["a"; 100].join("")); t!(h.set_path(&p)); let path = t!(h.path()); let actual: &Path = path.as_ref(); assert_eq!(actual, p); } #[test] fn set_metadata_deterministic() { let td = t!(Builder::new().prefix("async-tar").tempdir()); let tmppath = td.path().join("tmpfile"); fn mk_header(path: &Path, readonly: bool) -> Header { let mut file = t!(File::create(path)); t!(file.write_all(b"c")); let mut perms = t!(file.metadata()).permissions(); perms.set_readonly(readonly); t!(fs::set_permissions(path, perms)); let mut h = Header::new_ustar(); h.set_metadata_in_mode(&t!(path.metadata()), HeaderMode::Deterministic); h } // Create "the same" File twice in a row, one second apart, with differing readonly values. let one = mk_header(tmppath.as_path(), false); thread::sleep(time::Duration::from_millis(1050)); let two = mk_header(tmppath.as_path(), true); // Always expected to match. assert_eq!(t!(one.size()), t!(two.size())); assert_eq!(t!(one.path()), t!(two.path())); assert_eq!(t!(one.mode()), t!(two.mode())); // Would not match without `Deterministic`. assert_eq!(t!(one.mtime()), t!(two.mtime())); // TODO: No great way to validate that these would not be filled, but // check them anyway. assert_eq!(t!(one.uid()), t!(two.uid())); assert_eq!(t!(one.gid()), t!(two.gid())); } #[test] fn extended_numeric_format() { let mut h: GnuHeader = unsafe { mem::zeroed() }; h.as_header_mut().set_size(42); assert_eq!(h.size, [48, 48, 48, 48, 48, 48, 48, 48, 48, 53, 50, 0]); h.as_header_mut().set_size(8589934593); assert_eq!(h.size, [0x80, 0, 0, 0, 0, 0, 0, 0x02, 0, 0, 0, 1]); h.as_header_mut().set_size(44); assert_eq!(h.size, [48, 48, 48, 48, 48, 48, 48, 48, 48, 53, 52, 0]); h.size = [0x80, 0, 0, 0, 0, 0, 0, 0x02, 0, 0, 0, 0]; assert_eq!(h.as_header().entry_size().unwrap(), 0x0200000000); h.size = [48, 48, 48, 48, 48, 48, 48, 48, 48, 53, 51, 0]; assert_eq!(h.as_header().entry_size().unwrap(), 43); h.as_header_mut().set_gid(42); assert_eq!(h.gid, [48, 48, 48, 48, 48, 53, 50, 0]); assert_eq!(h.as_header().gid().unwrap(), 42); h.as_header_mut().set_gid(0x7fffffffffffffff); assert_eq!(h.gid, [0xff; 8]); assert_eq!(h.as_header().gid().unwrap(), 0x7fffffffffffffff); h.uid = [0x80, 0x00, 0x00, 0x00, 0x12, 0x34, 0x56, 0x78]; assert_eq!(h.as_header().uid().unwrap(), 0x12345678); h.mtime = [ 0x80, 0, 0, 0, 0x01, 0x23, 0x45, 0x67, 0x89, 0xab, 0xcd, 0xef, ]; assert_eq!(h.as_header().mtime().unwrap(), 0x0123456789abcdef); h.realsize = [0x80, 0, 0, 0, 0, 0x12, 0x34, 0x56, 0x78, 0x9a, 0xbc, 0xde]; assert_eq!(h.real_size().unwrap(), 0x00123456789abcde); h.sparse[0].offset = [0x80, 0, 0, 0, 0, 0x01, 0x23, 0x45, 0x67, 0x89, 0xab, 0xcd]; assert_eq!(h.sparse[0].offset().unwrap(), 0x000123456789abcd); h.sparse[0].numbytes = [0x80, 0, 0, 0, 0, 0, 0x12, 0x34, 0x56, 0x78, 0x9a, 0xbc]; assert_eq!(h.sparse[0].length().unwrap(), 0x0000123456789abc); } #[test] fn byte_slice_conversion() { let h = Header::new_gnu(); let b: &[u8] = h.as_bytes(); let b_conv: &[u8] = Header::from_byte_slice(h.as_bytes()).as_bytes(); assert_eq!(b, b_conv); }