chore(deps): update rust crate divan to v5.0.3

When a range command is first evaluated on a line after its numeric first
address (inside a block, or after d, n, N, c or a branch skipped it), start
the range on the current line, as GNU sed does, unless a numeric second
address has also been passed. The range can start this way only once.

Fixes #541

.github/workflows/GnuTests.yml

@@ -49,6 +49,19 @@ jobs:
 
     # Omit installing part of https://github.com/actions/runner-images/tree/main/images/ubuntu
 
+    ### Locales
+    # Several testsuite scripts guard themselves with require_*_locale_ and
+    # silently skip when the locale is missing, which hides both regressions
+    # and improvements in the multibyte code paths.
+    - name: Generate the locales required by the testsuite
+      shell: bash
+      run: |
+        ## Generate the locales required by the testsuite
+        sudo localedef -i en_US -f UTF-8 en_US.UTF-8
+        sudo localedef -i ru_RU -f UTF-8 ru_RU.UTF-8
+        sudo localedef -i el_GR -f ISO-8859-7 el_GR.iso88597
+        locale -a
+
     ### Build
     - name: Build Rust sed binary
       shell: bash
@@ -134,7 +147,7 @@ jobs:
         persist-credentials: false
 
     - name: Retrieve reference artifacts
-      uses: dawidd6/action-download-artifact@v21
+      uses: dawidd6/action-download-artifact@v27
       continue-on-error: true
       with:
         workflow: GnuTests.yml

.github/workflows/benchmarks.yml

@@ -33,7 +33,7 @@ jobs:
       - uses: Swatinem/rust-cache@v2
 
       - name: Run sccache-cache
-        uses: mozilla-actions/sccache-action@v0.0.10
+        uses: mozilla-actions/sccache-action@v0.0.11
 
       - name: Install tools
         uses: taiki-e/install-action@v2
@@ -47,7 +47,7 @@ jobs:
           cargo codspeed build -p ${{ matrix.benchmark-target.package }}
 
       - name: Run benchmarks for ${{ matrix.benchmark-target.package }}
-        uses: CodSpeedHQ/action@v4
+        uses: CodSpeedHQ/action@v5
         env:
           CODSPEED_LOG: debug
         with:

.github/workflows/ci.yml

@@ -97,12 +97,12 @@ jobs:
         RUSTC_WRAPPER: ""
         RUSTFLAGS: "-Cinstrument-coverage -Zcoverage-options=branch -Ccodegen-units=1 -Copt-level=0 -Coverflow-checks=off -Zpanic_abort_tests -Cpanic=abort"
         RUSTDOCFLAGS: "-Cpanic=abort"
-        LLVM_PROFILE_FILE: "sed-%p-%m.profraw"
+        LLVM_PROFILE_FILE: "${{ github.workspace }}/sed-%p-%m.profraw"
     - name: "`grcov` ~ install"
       id: build_grcov
       shell: bash
       run: |
-        git clone https://github.com/mozilla/grcov.git ~/grcov/
+        git clone --depth=1 https://github.com/mozilla/grcov.git ~/grcov/
         cd ~/grcov
         cargo install --path .
         cd -

.github/workflows/code-quality.yml

@@ -51,7 +51,7 @@ jobs:
           components: clippy
       - uses: Swatinem/rust-cache@v2
       - name: Run sccache-cache
-        uses: mozilla-actions/sccache-action@v0.0.10
+        uses: mozilla-actions/sccache-action@v0.0.11
       - name: Initialize workflow variables
         id: vars
         shell: bash

.github/workflows/documentation.yml

@@ -0,0 +1,41 @@
+name: Documentation
+
+on:
+  pull_request:
+    paths:
+      - 'docs/**'
+      - '.github/workflows/documentation.yml'
+  push:
+    branches:
+      - main
+    paths:
+      - 'docs/**'
+      - '.github/workflows/documentation.yml'
+
+permissions:
+  contents: read
+
+jobs:
+  build-doc:
+    name: Build the mdBook documentation
+    runs-on: ubuntu-latest
+
+    steps:
+      - name: Checkout repository
+        uses: actions/checkout@v7
+        with:
+          persist-credentials: false
+
+      - name: Install mdbook
+        uses: taiki-e/install-action@v2
+        with:
+          tool: mdbook@0.5.0
+
+      - name: Build documentation
+        run: mdbook build docs
+
+      - name: Upload documentation
+        uses: actions/upload-artifact@v7
+        with:
+          name: documentation
+          path: docs/book

.github/workflows/release.yml

@@ -64,7 +64,7 @@ jobs:
         # we specify bash to get pipefail; it guards against the `curl` command
         # failing. otherwise `sh` won't catch that `curl` returned non-0
         shell: bash
-        run: "curl --proto '=https' --tlsv1.2 -LsSf https://github.com/axodotdev/cargo-dist/releases/download/v0.30.3/cargo-dist-installer.sh | sh"
+        run: "curl --proto '=https' --tlsv1.2 -LsSf https://github.com/axodotdev/cargo-dist/releases/download/v0.33.0/cargo-dist-installer.sh | sh"
       - name: Cache dist
         uses: actions/upload-artifact@v7
         with:

Cargo.toml

@@ -1,12 +1,12 @@
 # sed (uutils)
 # * see the repository LICENSE, README, and CONTRIBUTING files for more information
 
-# spell-checker:ignore (libs) mangen sed
+# spell-checker:ignore (libs) mangen sed binstall codspeed libc memchr memmap rlimit rustix textwrap uucore uutests xattr
+# spell-checker:ignore (misc) clippy codegen msvc startswith uninlined uutils
 
 [package]
 name = "sed"
-version = "0.1.1"
-authors = ["uutils developers"]
+version = "0.3.0"
 license = "MIT"
 description = "sed ~ implemented as universal (cross-platform) utils, written in Rust"
 default-run = "sed"
@@ -21,6 +21,19 @@ rust-version = "1.88"
 
 build = "build.rs"
 
+# cargo-binstall: fetch the archives published by `dist` in the GitHub releases
+# instead of building from source. The layout below has to stay in sync with
+# `dist-workspace.toml` (targets) and the archives it produces:
+#   sed-<target>.tar.xz (zip on Windows), containing a sed-<target>/ directory.
+# Note that the release tags carry no `v` prefix, hence `{ version }`.
+[package.metadata.binstall]
+pkg-url = "{ repo }/releases/download/{ version }/{ name }-{ target }.{ archive-suffix }"
+bin-dir = "{ name }-{ target }/{ bin }{ binary-ext }"
+pkg-fmt = "txz"
+
+[package.metadata.binstall.overrides.x86_64-pc-windows-msvc]
+pkg-fmt = "zip"
+
 [features]
 default = ["feat_common_core"]
 
@@ -29,10 +42,6 @@ sed = []
 
 [workspace.dependencies]
 assert_fs = "1.1.3"
-bytesize = "2.0.0"
-chrono = { version = "0.4.37", default-features = false, features = [
-  "clock",
-] }
 clap = { version = "4.4", features = ["wrap_help", "cargo"] }
 clap_complete = "4.5"
 clap_mangen = "0.3"
@@ -47,12 +56,13 @@ phf_codegen = "0.14.0"
 predicates = "3.1.3"
 rand = { version = "0.10.0" }
 regex = "1.10.4"
-sha2 = "0.11"
-sysinfo = "0.38"
+sha2 = { version = "0.11.0", default-features = false, features = ["alloc"] }
 tempfile = "3.10.1"
 terminal_size = "0.4.2"
 textwrap = { version = "0.16.1", features = ["terminal_size"] }
-uucore = { version = "0.9.0", features = ["libc"] }
+unicode-width = "0.2.2"
+uucore = { version = "0.12.0", features = ["diagnostics", "libc"] }
+rustix = "1.1.4"
 xattr = "1.3.1"
 
 
@@ -66,15 +76,15 @@ memmap2.workspace = true
 phf = { workspace = true }
 predicates = { workspace = true }
 regex = { workspace = true }
-sysinfo = { workspace = true }
 tempfile = { workspace = true }
 terminal_size = { workspace = true }
 textwrap = { workspace = true }
+unicode-width = { workspace = true }
 uucore = { workspace = true }
+rustix = { workspace = true }
 
 [dev-dependencies]
 assert_fs = { workspace = true }
-chrono = { workspace = true }
 ctor = "1.0.0"
 divan = { workspace = true }
 hex = { workspace = true }
@@ -85,7 +95,7 @@ regex = { workspace = true }
 sha2 = { workspace = true }
 tempfile = { workspace = true }
 uucore = { workspace = true, features = ["entries", "process", "signals", "benchmark"] }
-uutests = "0.9.0"
+uutests = "0.12.0"
 
 [target.'cfg(unix)'.dev-dependencies]
 xattr = { workspace = true }
@@ -119,6 +129,10 @@ inherits = "release"
 opt-level = "z"
 strip = true
 
+# The profile that 'dist' will build with
+[profile.dist]
+inherits = "release"
+
 [lints.clippy]
 all = { level = "warn", priority = -1 }
 
@@ -135,7 +149,7 @@ must_use_candidate = "allow"                 # 56
 needless_raw_string_hashes = "allow"         # 20
 needless_pass_by_value = "allow"             # 15
 missing_panics_doc = "allow"                 # 12
-cast_possible_truncation = "allow"           # 7
+# cast_possible_truncation = "allow"           # 7
 unnecessary_wraps = "allow"                  # 6
 match_wildcard_for_single_variants = "allow" # 4
 cast_sign_loss = "allow"                     # 4

README.md

@@ -15,22 +15,30 @@ and other extensions.
 ## Status
 
 At this state _sed_ implements all [POSIX features](https://pubs.opengroup.org/onlinepubs/9799919799/)
-and can run correctly the two complex scripts of its integration tests:
-[hanoi.sed](https://github.com/uutils/sed/blob/main/tests/fixtures/sed/script/hanoi.sed) (solves the Towers of Hanoi puzzle) and
+and can run correctly the three complex scripts of its integration tests:
+[hanoi.sed](https://github.com/uutils/sed/blob/main/tests/fixtures/sed/script/hanoi.sed) (solves the Towers of Hanoi puzzle),
+[mandelbrot.sed](https://github.com/uutils/sed/blob/main/tests/fixtures/sed/script/mandelbrot.sed), (draws the Mandelbrot set) and
 [math.sed](https://github.com/uutils/sed/blob/main/tests/fixtures/sed/script/math.sed)  (implements an arbitrary precision integer math calculator).
 
 The performance of this Rust implementation is now better than the GNU and FreeBSD implementations for most benchmarked cases.
 
 Further work aims to:
-* Adjust buffering on terminal output to match current implementations,
-* Implement more GNU extensions,
-* Improve performance where possible.
+* improve GNU _sed_ compatibility, especially on the regular expression front,
+* implement more GNU extensions, and
+* improve performance where possible.
 
 ## Installation and Use
 
 We provide a Linux x86_64 binary archive from the main branch at
 https://github.com/uutils/sed/releases/tag/latest-commit .
 
+If you have [cargo-binstall](https://github.com/cargo-bins/cargo-binstall),
+the released binaries can be installed directly with:
+
+```bash
+cargo binstall sed
+```
+
 For other platforms, ensure you have Rust installed on your system. You can install Rust through [rustup](https://rustup.rs/).
 
 Clone the repository and build the project using Cargo:
@@ -80,58 +88,8 @@ cargo test
 ```
 
 ## Extensions and incompatibilities
-### Supported GNU extensions
-* Command-line arguments can be specified in long (`--`) form.
-* Spaces can precede a regular expression modifier.
-* `I` can be used in as a synonym for the `i` (case insensitive) substitution
-  flag.
-* `M` and `m` substitution flags allow multi-line matching.
-* In addition to `\n`, other escape sequences (octal, hex, C) are supported
-  in the strings of the `y` command.
-  Under POSIX these yield undefined behavior.
-* The `a`, `c`, and `i` commands do not require an initial backslash,
-  allow text to appear on the same line, and support escape sequences
-  in the specified text.
-* The `a`, `i`, `=`, `l`, `q` and `r` commands support address range as an extension to POSIX.
-* The substitution command replacement group `\0` is a synonym for &.
-* A `Q` command (optionally followed by an exit code) quits immediately.
-* The `q` command can be optionally followed by an exit code.
-* The `l` command can be optionally followed by the output width.
-* The `--follow-symlinks` option for in-place editing.
-* The `--sandbox` option that limits potentially destructive commands.
-* Address 0 can be used to specify an address range that is already
-  active on line 1 and can finish with the specified regular expression.
-* Address steps can be specified in the form of start~step and start,~step
-  ranges.
-* Address 0 can be used in the `r` command to prepend a file.
-
-### Supported BSD and GNU extensions
-* The second address in a range can be specified as a relative address with +N.
-* In-place editing of file with the `-i` flag.
-
-### New extensions
-* Unicode characters can be specified in regular expression pattern, replacement
-  and transliteration sequences using `\uXXXX` or `\UXXXXXXXX` sequences.
-* The `l` command lists Unicode characters using the `\uXXXX` and `\UXXXXXXXX`
-  sequences.
-
-### Incompatibilities
-* The input is assumed to be valid UTF-8 (this includes 7-bit ASCII).
-  If the input is in another code page, consider converting it through UTF-8
-  in order to avoid errors on invalid UTF-8 sequences and for the correct
-  handling of regular expressions.
-  This _sed_ program can also handle arbitrary byte sequences if no part of the
-  input is treated as string.
-* The command will report an error and fail if duplicate labels are found
-  in the script.
-  This matches the BSD behavior. The GNU version accepts duplicate labels.
-* The last line (`$`) address is interpreted as the last non-empty line of
-  the last file.  If files specified in subsequent arguments until the last
-  one are empty, then the last line condition will never be triggered.
-  This behavior is consistent with the
-  [original implementation](https://github.com/dspinellis/unix-history-repo/blob/Research-V7/usr/src/cmd/sed/sed1.c#L665).
-* Labels are parsed for alphanumeric characters. The BSD version parses them
-  until the end of the line, preventing ; to be used as a separator.
+The GNU, BSD and new extensions _sed_ supports, and where it differs from GNU
+_sed_, are listed in [docs/src/extensions.md](docs/src/extensions.md).
 
 ## GNU test suite compatibility
 

dist-workspace.toml

@@ -4,7 +4,7 @@ members = ["cargo:."]
 # Config for 'dist'
 [dist]
 # The preferred dist version to use in CI (Cargo.toml SemVer syntax)
-cargo-dist-version = "0.30.3"
+cargo-dist-version = "0.33.0"
 # CI backends to support
 ci = "github"
 # The installers to generate for each app
@@ -13,5 +13,5 @@ installers = []
 targets = ["aarch64-apple-darwin", "aarch64-unknown-linux-gnu", "x86_64-apple-darwin", "x86_64-unknown-linux-gnu", "x86_64-unknown-linux-musl", "x86_64-pc-windows-msvc"]
 # Which actions to run on pull requests
 pr-run-mode = "plan"
-# Allow dirty workflows to ignore out-of-date contents
+# Skip checking whether the specified configuration files are up to date
 allow-dirty = ["ci"]

docs/.gitignore

@@ -0,0 +1 @@
+book

docs/book.toml

@@ -0,0 +1,7 @@
+[book]
+language = "en"
+src = "src"
+title = "uutils sed Documentation"
+
+[output.html]
+git-repository-url = "https://github.com/uutils/sed/tree/main/docs/src"

docs/src/SUMMARY.md

@@ -0,0 +1,5 @@
+# Summary
+
+[Introduction](index.md)
+
+* [Extensions and incompatibilities](extensions.md)

docs/src/extensions.md

@@ -0,0 +1,86 @@
+# Extensions and incompatibilities
+
+The main goal of the project is compatibility with GNU _sed_, but _sed_ also
+supports features that GNU _sed_ does not, and differs from it in a few places.
+Below is a list of these extensions and incompatibilities.
+
+## Supported GNU extensions
+* Command-line arguments can be specified in long (`--`) form.
+* Spaces can precede a regular expression modifier.
+* `I` can be used in as a synonym for the `i` (case insensitive) substitution
+  flag.
+* `M` and `m` substitution flags allow multi-line matching.
+* In addition to `\n`, other escape sequences (octal, hex, C) are supported
+  in the strings of the `y` command.
+  Under POSIX these yield undefined behavior.
+* The `a`, `c`, and `i` commands do not require an initial backslash,
+  allow text to appear on the same line, and support escape sequences
+  in the specified text.
+* The `a`, `i`, `=`, `l`, `q` and `r` commands support address range as an extension to POSIX.
+* The substitution command replacement group `\0` is a synonym for &.
+* An `F` command outputs the name of the file currently being processed.
+* A `Q` command (optionally followed by an exit code) quits immediately.
+* The `q` command can be optionally followed by an exit code.
+* A `W` command writes to a file the pattern's first line.
+* An `R` command reads one line at a time from a file.
+* The `l` command can be optionally followed by the output width.
+* The `--follow-symlinks` option for in-place editing.
+* The `--sandbox` option that limits potentially destructive commands.
+* Address 0 can be used to specify an address range that is already
+  active on line 1 and can finish with the specified regular expression.
+* Address steps can be specified in the form of start~step and start,~step
+  ranges.
+* Address 0 can be used in the `r` command to prepend a file.
+
+## Supported BSD and GNU extensions
+* The second address in a range can be specified as a relative address with +N.
+* In-place editing of file with the `-i` flag.
+  As in GNU _sed_, a backup suffix must be attached (`-i.bak`,
+  `--in-place=.bak`); BSD's `-i .bak` and `-i ''` are not supported.
+
+## New extensions
+* Unicode characters can be specified in regular expression pattern, replacement
+  and transliteration sequences using `\uXXXX` or `\UXXXXXXXX` sequences.
+* On a terminal, script errors also quote the script line and underline the
+  character at fault (`UUTILS_DIAG=always` or `never` overrides this):
+  ```
+  $ sed 's/a/b/q'
+  sed: <script argument 1>:1:7: error: invalid substitute flag: 'q'
+     ╭─[ <script argument 1>:1:7 ]
+     │
+   1 │ s/a/b/q
+     │       ─
+  ───╯
+  ```
+
+## Incompatible extensions
+The `-U` or `--uutil-extensions` option enables useful extensions or bug fixes
+that aren't compatible with GNU sed or POSIX.
+
+* The `l` command lists Unicode characters using the `\uXXXX` and `\UXXXXXXXX`
+  escapes rather than as octal UTF-8 byte sequences.
+
+## Incompatibilities
+* Similarly to GNU _sed_, input is processed as raw bytes or as valid UTF-8
+  (this includes 7-bit ASCII) based on the locale as specified by the
+  `LC_ALL`, `LC_CTYPE`, and `LANG` environment variables,
+  with the default being byte processing.
+  However, in contrast with GNU _sed_, other locales (e.g. ISO-8859-1)
+  are not supported. If the input is in another code page or encoding
+  and requires locale-specific processing (e.g. ignore/map case,
+  character classes), consider converting it through UTF-8 to ensure
+  the correct handling of locale-specific regular expressions.
+  This _sed_ program can also handle arbitrary byte sequences
+  if no part of the input requires treating it as a Rust String.
+* Back-references aren't supported when input is processed as bytes
+  (`LC_ALL=C`).
+* The command will report an error and fail if duplicate labels are found
+  in the script.
+  This matches the BSD behavior. The GNU version accepts duplicate labels.
+* The last line (`$`) address is interpreted as the last non-empty line of
+  the last file.  If files specified in subsequent arguments until the last
+  one are empty, then the last line condition will never be triggered.
+  This behavior is consistent with the
+  [original implementation](https://github.com/dspinellis/unix-history-repo/blob/Research-V7/usr/src/cmd/sed/sed1.c#L665).
+* Labels are parsed for alphanumeric characters. The BSD version parses them
+  until the end of the line, preventing ; to be used as a separator.

docs/src/index.md

@@ -0,0 +1,10 @@
+# uutils sed
+
+_sed_ is a Rust reimplementation of the
+[sed utility](https://pubs.opengroup.org/onlinepubs/9799919799/utilities/sed.html)
+with some [GNU sed](https://www.gnu.org/software/sed/manual/sed.html),
+[FreeBSD sed](https://man.freebsd.org/cgi/man.cgi?sed(1)),
+and other extensions.
+
+It is part of the [uutils](https://uutils.github.io/) project. The source code
+is on [GitHub](https://github.com/uutils/sed).

fuzz/Cargo.toml

@@ -14,8 +14,8 @@ libc = "0.2.153"
 tempfile = "3.15.0"
 rand = { version = "0.10.0" }
 similar = "3.0.0"
-uucore = { version = "0.9.0", features = ["libc"] }
-uufuzz = "0.9.0"
+uucore = { version = "0.12.0", features = ["libc"] }
+uufuzz = "0.12.0"
 
 sed = { path = ".." }