chore(deps): update github artifact actions

Co-authored-by: renovate[bot] <29139614+renovate[bot]@users.noreply.github.com>

.github/workflows/GnuComment.yml

@@ -0,0 +1,100 @@
+name: GnuComment
+
+on:
+  workflow_run:
+    workflows: ["GnuTests"]
+    types:
+      - completed
+
+permissions: {}
+
+jobs:
+  post-comment:
+    permissions:
+      actions: read  #  to list workflow runs artifacts
+      pull-requests: write  #  to comment on pr
+
+    runs-on: ubuntu-latest
+    if: >
+      github.event.workflow_run.event == 'pull_request'
+    steps:
+      - name: 'Download artifact'
+        uses: actions/github-script@v9
+        with:
+          script: |
+            // List all artifacts from GnuTests
+            var artifacts = await github.rest.actions.listWorkflowRunArtifacts({
+              owner: context.repo.owner,
+              repo: context.repo.repo,
+              run_id: ${{ github.event.workflow_run.id }},
+            });
+
+            // Download the "comment" artifact, which contains a PR number (NR) and result.txt
+            var matchArtifact = artifacts.data.artifacts.filter((artifact) => {
+              return artifact.name == "comment"
+            })[0];
+
+            if (!matchArtifact) {
+              console.log('No comment artifact found');
+              return;
+            }
+
+            var download = await github.rest.actions.downloadArtifact({
+              owner: context.repo.owner,
+              repo: context.repo.repo,
+              artifact_id: matchArtifact.id,
+              archive_format: 'zip',
+            });
+            var fs = require('fs');
+            fs.writeFileSync('${{ github.workspace }}/comment.zip', Buffer.from(download.data));
+
+      - run: unzip comment.zip || echo "Failed to unzip comment artifact"
+
+      - name: 'Comment on PR'
+        uses: actions/github-script@v9
+        with:
+          github-token: ${{ secrets.GITHUB_TOKEN }}
+          script: |
+            var fs = require('fs');
+
+            // Check if files exist
+            if (!fs.existsSync('./NR')) {
+              console.log('No NR file found, skipping comment');
+              return;
+            }
+            if (!fs.existsSync('./result.txt')) {
+              console.log('No result.txt file found, skipping comment');
+              return;
+            }
+
+            var issue_number = Number(fs.readFileSync('./NR'));
+            var content = fs.readFileSync('./result.txt');
+
+            if (content.toString().trim().length > 7) { // 7 because we have backquote + \n
+              // Update existing comment if present, otherwise create a new one
+              var marker = '<!-- gnu-tests-bot -->';
+              var body = marker + '\nGNU diffutils testsuite comparison:\n```\n' + content + '```';
+              var comments = await github.rest.issues.listComments({
+                owner: context.repo.owner,
+                repo: context.repo.repo,
+                issue_number: issue_number,
+              });
+              var existing = comments.data.filter(c => c.body.includes(marker))[0];
+              if (existing) {
+                await github.rest.issues.updateComment({
+                  owner: context.repo.owner,
+                  repo: context.repo.repo,
+                  comment_id: existing.id,
+                  body: body,
+                });
+              } else {
+                await github.rest.issues.createComment({
+                  owner: context.repo.owner,
+                  repo: context.repo.repo,
+                  issue_number: issue_number,
+                  body: body,
+                });
+              }
+            } else {
+              console.log('Comment content too short, skipping');
+            }

.github/workflows/GnuTests.yml

@@ -0,0 +1,231 @@
+name: GnuTests
+
+# Run GNU diffutils testsuite against the Rust diffutils implementation
+# and compare results against the main branch to catch regressions
+
+on:
+  pull_request:
+  push:
+    branches:
+      - '*'
+
+permissions:
+  contents: write # Publish diffutils instead of discarding
+
+# End the current execution if there is a new changeset in the PR
+concurrency:
+  group: ${{ github.workflow }}-${{ github.ref }}
+  cancel-in-progress: ${{ github.ref != 'refs/heads/main' }}
+
+env:
+  DEFAULT_BRANCH: ${{ github.event.repository.default_branch }}
+  TEST_FULL_SUMMARY_FILE: 'diffutils-gnu-full-result.json'
+
+jobs:
+  native:
+    name: Run GNU diffutils testsuite
+    runs-on: ubuntu-24.04
+    steps:
+      - name: Checkout code
+        uses: actions/checkout@v4
+        with:
+          persist-credentials: false
+
+      - uses: dtolnay/rust-toolchain@master
+        with:
+          toolchain: stable
+
+      - uses: Swatinem/rust-cache@v2
+
+      ### Build
+      - name: Build Rust diffutils binary
+        shell: bash
+        run: |
+          ## Build Rust diffutils binary
+          cargo build --config=profile.release.strip=true --profile=release
+          zstd -19 target/release/diffutils -o diffutils-x86_64-unknown-linux-gnu.zst
+
+      - name: Publish latest commit
+        uses: softprops/action-gh-release@v3
+        if: github.event_name == 'push' && github.ref == 'refs/heads/main'
+        with:
+          tag_name: latest-commit
+          body: |
+            commit: ${{ github.sha }}
+          draft: false
+          prerelease: true
+          files: |
+            diffutils-x86_64-unknown-linux-gnu.zst
+        env:
+          GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
+
+      ### Run tests
+      - name: Run GNU diffutils testsuite
+        shell: bash
+        run: |
+          ## Run GNU diffutils testsuite
+          ./tests/run-upstream-testsuite.sh release || true
+        env:
+          TERM: xterm
+
+      - name: Upload full json results
+        uses: actions/upload-artifact@v7
+        with:
+          name: diffutils-gnu-full-result
+          path: tests/test-results.json
+          if-no-files-found: warn
+
+  aggregate:
+    needs: [native]
+    permissions:
+      actions: read
+      contents: read
+      pull-requests: read
+    name: Aggregate GNU test results
+    runs-on: ubuntu-24.04
+    steps:
+      - name: Initialize workflow variables
+        id: vars
+        shell: bash
+        run: |
+          ## VARs setup
+          outputs() { step_id="${{ github.action }}"; for var in "$@" ; do echo steps.${step_id}.outputs.${var}="${!var}"; echo "${var}=${!var}" >> $GITHUB_OUTPUT; done; }
+
+          TEST_SUMMARY_FILE='diffutils-gnu-result.json'
+          outputs TEST_SUMMARY_FILE
+
+      - name: Checkout code
+        uses: actions/checkout@v4
+        with:
+          persist-credentials: false
+
+      - name: Retrieve reference artifacts
+        uses: dawidd6/action-download-artifact@v21
+        continue-on-error: true
+        with:
+          workflow: GnuTests.yml
+          branch: "${{ env.DEFAULT_BRANCH }}"
+          workflow_conclusion: completed
+          path: "reference"
+          if_no_artifact_found: warn
+
+      - name: Download full json results
+        uses: actions/download-artifact@v8
+        with:
+          name: diffutils-gnu-full-result
+          path: results
+
+      - name: Extract/summarize testing info
+        id: summary
+        shell: bash
+        run: |
+          ## Extract/summarize testing info
+          outputs() { step_id="${{ github.action }}"; for var in "$@" ; do echo steps.${step_id}.outputs.${var}="${!var}"; echo "${var}=${!var}" >> $GITHUB_OUTPUT; done; }
+
+          RESULT_FILE="results/test-results.json"
+          if [[ ! -f "$RESULT_FILE" ]]; then
+            echo "::error ::Missing test results at $RESULT_FILE"
+            exit 1
+          fi
+
+          TOTAL=$(jq '[.tests[]] | length' "$RESULT_FILE")
+          PASS=$(jq '[.tests[] | select(.result=="PASS")] | length' "$RESULT_FILE")
+          FAIL=$(jq '[.tests[] | select(.result=="FAIL")] | length' "$RESULT_FILE")
+          SKIP=$(jq '[.tests[] | select(.result=="SKIP")] | length' "$RESULT_FILE")
+          ERROR=0
+
+          output="GNU diffutils tests summary = TOTAL: $TOTAL / PASS: $PASS / FAIL: $FAIL / SKIP: $SKIP"
+          echo "${output}"
+
+          if [[ "$FAIL" -gt 0 ]]; then
+            echo "::warning ::${output}"
+          fi
+
+          jq -n \
+            --arg date "$(date --rfc-email)" \
+            --arg sha "$GITHUB_SHA" \
+            --arg total "$TOTAL" \
+            --arg pass "$PASS" \
+            --arg skip "$SKIP" \
+            --arg fail "$FAIL" \
+            --arg error "$ERROR" \
+            '{($date): { sha: $sha, total: $total, pass: $pass, skip: $skip, fail: $fail, error: $error }}' > '${{ steps.vars.outputs.TEST_SUMMARY_FILE }}'
+
+          HASH=$(sha1sum '${{ steps.vars.outputs.TEST_SUMMARY_FILE }}' | cut --delim=" " -f 1)
+          outputs HASH TOTAL PASS FAIL SKIP
+
+      - name: Upload SHA1/ID of 'test-summary'
+        uses: actions/upload-artifact@v7
+        with:
+          name: "${{ steps.summary.outputs.HASH }}"
+          path: "${{ steps.vars.outputs.TEST_SUMMARY_FILE }}"
+
+      - name: Upload test results summary
+        uses: actions/upload-artifact@v7
+        with:
+          name: test-summary
+          path: "${{ steps.vars.outputs.TEST_SUMMARY_FILE }}"
+
+      - name: Compare test failures VS reference
+        shell: bash
+        run: |
+          ## Compare test failures VS reference
+          REF_SUMMARY_FILE='reference/diffutils-gnu-full-result/test-results.json'
+          CURRENT_SUMMARY_FILE="results/test-results.json"
+
+          IGNORE_INTERMITTENT=".github/workflows/ignore-intermittent.txt"
+
+          COMMENT_DIR="reference/comment"
+          mkdir -p ${COMMENT_DIR}
+          echo ${{ github.event.number }} > ${COMMENT_DIR}/NR
+          COMMENT_LOG="${COMMENT_DIR}/result.txt"
+
+          COMPARISON_RESULT=0
+          if test -f "${CURRENT_SUMMARY_FILE}"; then
+            if test -f "${REF_SUMMARY_FILE}"; then
+              echo "Reference summary SHA1/ID: $(sha1sum -- "${REF_SUMMARY_FILE}")"
+              echo "Current summary SHA1/ID: $(sha1sum -- "${CURRENT_SUMMARY_FILE}")"
+
+              python3 util/compare_test_results.py \
+                --ignore-file "${IGNORE_INTERMITTENT}" \
+                --output "${COMMENT_LOG}" \
+                "${CURRENT_SUMMARY_FILE}" "${REF_SUMMARY_FILE}"
+
+              COMPARISON_RESULT=$?
+            else
+              echo "::warning ::Skipping test comparison; no prior reference summary is available at '${REF_SUMMARY_FILE}'."
+            fi
+          else
+            echo "::error ::Failed to find summary of test results (missing '${CURRENT_SUMMARY_FILE}'); failing early"
+            exit 1
+          fi
+
+          if [ ${COMPARISON_RESULT} -eq 1 ]; then
+            echo "::error ::Found new non-intermittent test failures"
+            exit 1
+          else
+            echo "::notice ::No new test failures detected"
+          fi
+
+      - name: Upload comparison log (for GnuComment workflow)
+        if: success() || failure()
+        uses: actions/upload-artifact@v7
+        with:
+          name: comment
+          path: reference/comment/
+
+      - name: Report test results
+        if: success() || failure()
+        shell: bash
+        run: |
+          ## Report final results
+          echo "::notice ::GNU diffutils testsuite results:"
+          echo "::notice ::  Total tests: ${{ steps.summary.outputs.TOTAL }}"
+          echo "::notice ::  Passed: ${{ steps.summary.outputs.PASS }}"
+          echo "::notice ::  Failed: ${{ steps.summary.outputs.FAIL }}"
+          echo "::notice ::  Skipped: ${{ steps.summary.outputs.SKIP }}"
+
+          if [[ "${{ steps.summary.outputs.FAIL }}" -gt 0 ]]; then
+            PASS_RATE=$(( ${{ steps.summary.outputs.PASS }} * 100 / (${{ steps.summary.outputs.PASS }} + ${{ steps.summary.outputs.FAIL }}) ))
+            echo "::notice ::  Pass rate: ${PASS_RATE}%"
+          fi

.github/workflows/audit.yml

@@ -0,0 +1,15 @@
+name: Security audit
+
+on:
+  schedule:
+    - cron: "0 0 * * *"
+jobs:
+  audit:
+    runs-on: ubuntu-latest
+    steps:
+      - uses: actions/checkout@v6
+        with:
+          persist-credentials: false
+      - uses: rustsec/audit-check@v2
+        with:
+          token: ${{ secrets.GITHUB_TOKEN }}

.github/workflows/ci.yml

@@ -4,6 +4,7 @@ name: Basic CI
 
 env:
   CARGO_TERM_COLOR: always
+  CARGO_INCREMENTAL: 0
 
 jobs:
   check:
@@ -15,7 +16,6 @@ jobs:
         os: [ubuntu-latest, macOS-latest, windows-latest]
     steps:
       - uses: actions/checkout@v4
-      - uses: dtolnay/rust-toolchain@stable
       - run: cargo check
 
   test:
@@ -27,12 +27,10 @@ jobs:
         os: [ubuntu-latest, macOS-latest, windows-latest]
     steps:
       - uses: actions/checkout@v4
-      - uses: dtolnay/rust-toolchain@stable
       - name: install GNU patch on MacOS
         if: runner.os == 'macOS'
         run: |
           brew install gpatch
-          echo "/opt/homebrew/opt/gpatch/libexec/gnubin" >> "$GITHUB_PATH"
       - name: set up PATH on Windows
         # Needed to use GNU's patch.exe instead of Strawberry Perl patch
         if: runner.os == 'Windows'
@@ -44,8 +42,6 @@ jobs:
     runs-on: ubuntu-latest
     steps:
       - uses: actions/checkout@v4
-      - uses: dtolnay/rust-toolchain@stable
-      - run: rustup component add rustfmt
       - run: cargo fmt --all -- --check
 
   clippy:
@@ -57,29 +53,12 @@ jobs:
         os: [ubuntu-latest, macOS-latest, windows-latest]
     steps:
       - uses: actions/checkout@v4
-      - uses: dtolnay/rust-toolchain@stable
-      - run: rustup component add clippy
       - run: cargo clippy -- -D warnings
 
-  gnu-testsuite:
-    name: GNU test suite
-    runs-on: ubuntu-latest
-    steps:
-      - uses: actions/checkout@v4
-      - uses: dtolnay/rust-toolchain@stable
-      - run: cargo build --release
-      # do not fail, the report is merely informative (at least until all tests pass reliably)
-      - run: ./tests/run-upstream-testsuite.sh release || true
-        env:
-          TERM: xterm
-      - uses: actions/upload-artifact@v4
-        with:
-          name: test-results.json
-          path: tests/test-results.json
-      - run: ./tests/print-test-results.sh tests/test-results.json
-
   coverage:
     name: Code Coverage
+    env:
+      RUSTC_BOOTSTRAP: 1
     runs-on: ${{ matrix.job.os }}
     strategy:
       fail-fast: false
@@ -96,26 +75,15 @@ jobs:
       run: |
         ## VARs setup
         outputs() { step_id="vars"; for var in "$@" ; do echo steps.${step_id}.outputs.${var}="${!var}"; echo "${var}=${!var}" >> $GITHUB_OUTPUT; done; }
-        # toolchain
-        TOOLCHAIN="nightly" ## default to "nightly" toolchain (required for certain required unstable compiler flags) ## !maint: refactor when stable channel has needed support
-        # * specify gnu-type TOOLCHAIN for windows; `grcov` requires gnu-style code coverage data files
-        case ${{ matrix.job.os }} in windows-*) TOOLCHAIN="$TOOLCHAIN-x86_64-pc-windows-gnu" ;; esac;
-        # * use requested TOOLCHAIN if specified
-        if [ -n "${{ matrix.job.toolchain }}" ]; then TOOLCHAIN="${{ matrix.job.toolchain }}" ; fi
-        outputs TOOLCHAIN
         # target-specific options
         # * CODECOV_FLAGS
         CODECOV_FLAGS=$( echo "${{ matrix.job.os }}" | sed 's/[^[:alnum:]]/_/g' )
         outputs CODECOV_FLAGS
 
-    - name: rust toolchain ~ install
-      uses: dtolnay/rust-toolchain@nightly
-    - run: rustup component add llvm-tools-preview
     - name: install GNU patch on MacOS
       if: runner.os == 'macOS'
       run: |
         brew install gpatch
-        echo "/opt/homebrew/opt/gpatch/libexec/gnubin" >> "$GITHUB_PATH"
     - name: set up PATH on Windows
       # Needed to use GNU's patch.exe instead of Strawberry Perl patch
       if: runner.os == 'Windows'
@@ -123,7 +91,6 @@ jobs:
     - name: Test
       run: cargo test --all-features --no-fail-fast
       env:
-        CARGO_INCREMENTAL: "0"
         RUSTC_WRAPPER: ""
         RUSTFLAGS: "-Cinstrument-coverage -Zcoverage-options=branch -Ccodegen-units=1 -Copt-level=0 -Coverflow-checks=off -Zpanic_abort_tests -Cpanic=abort"
         RUSTDOCFLAGS: "-Cpanic=abort"
@@ -160,7 +127,7 @@ jobs:
         grcov . --output-type lcov --output-path "${COVERAGE_REPORT_FILE}" --binary-path "${COVERAGE_REPORT_DIR}" --branch
         echo "report=${COVERAGE_REPORT_FILE}" >> $GITHUB_OUTPUT
     - name: Upload coverage results (to Codecov.io)
-      uses: codecov/codecov-action@v5
+      uses: codecov/codecov-action@v7
       with:
         token: ${{ secrets.CODECOV_TOKEN }}
         files: ${{ steps.coverage.outputs.report }}
@@ -168,4 +135,3 @@ jobs:
         flags: ${{ steps.vars.outputs.CODECOV_FLAGS }}
         name: codecov-umbrella
         fail_ci_if_error: false
-

.github/workflows/codspeed.yml

@@ -0,0 +1,37 @@
+name: CodSpeed
+
+on:
+  push:
+    branches:
+      - "main"
+  pull_request:
+  # `workflow_dispatch` allows CodSpeed to trigger backtest
+  # performance analysis in order to generate initial data.
+  workflow_dispatch:
+
+permissions:
+  contents: read
+  id-token: write
+
+jobs:
+  codspeed:
+    name: Run benchmarks
+    runs-on: ubuntu-latest
+    steps:
+      - uses: actions/checkout@v4
+
+      - name: Setup rust toolchain, cache and cargo-codspeed binary
+        uses: moonrepo/setup-rust@v1
+        with:
+          channel: stable
+          cache-target: release
+          bins: cargo-codspeed
+
+      - name: Build the benchmark target(s)
+        run: cargo codspeed build -m simulation
+
+      - name: Run the benchmarks
+        uses: CodSpeedHQ/action@v4
+        with:
+          mode: simulation
+          run: cargo codspeed run

.github/workflows/fuzzing.yml

@@ -2,6 +2,10 @@ name: Fuzzing
 
 # spell-checker:ignore fuzzer
 
+env:
+  CARGO_INCREMENTAL: 0
+  RUSTC_BOOTSTRAP: 1
+
 on:
   pull_request:
   push:
@@ -21,15 +25,15 @@ jobs:
     runs-on: ubuntu-latest
     steps:
       - uses: actions/checkout@v4
-      - uses: dtolnay/rust-toolchain@nightly
       - name: Install `cargo-fuzz`
-        run: cargo install cargo-fuzz
+        run: |
+          cargo install cargo-fuzz --locked
       - uses: Swatinem/rust-cache@v2
         with:
           shared-key: "cargo-fuzz-cache-key"
           cache-directories: "fuzz/target"
       - name: Run `cargo-fuzz build`
-        run: cargo +nightly fuzz build
+        run: cargo fuzz build
 
   fuzz-run:
     needs: fuzz-build
@@ -49,9 +53,9 @@ jobs:
           - { name: fuzz_side, should_pass: true }
     steps:
       - uses: actions/checkout@v4
-      - uses: dtolnay/rust-toolchain@nightly
       - name: Install `cargo-fuzz`
-        run: cargo install cargo-fuzz
+        run: |
+          cargo install cargo-fuzz --locked
       - uses: Swatinem/rust-cache@v2
         with:
           shared-key: "cargo-fuzz-cache-key"
@@ -64,9 +68,9 @@ jobs:
             fuzz/corpus/${{ matrix.test-target.name }}
       - name: Run ${{ matrix.test-target.name }} for XX seconds
         shell: bash
-        continue-on-error: ${{ !matrix.test-target.name.should_pass }}
+        continue-on-error: ${{ !matrix.test-target.should_pass }}
         run: |
-          cargo +nightly fuzz run ${{ matrix.test-target.name }} -- -max_total_time=${{ env.RUN_FOR }} -detect_leaks=0
+          cargo fuzz run ${{ matrix.test-target.name }} -- -max_total_time=${{ env.RUN_FOR }} -detect_leaks=0
       - name: Save Corpus Cache
         uses: actions/cache/save@v5
         with:

.github/workflows/release.yml

@@ -66,7 +66,7 @@ jobs:
         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"
       - name: Cache dist
-        uses: actions/upload-artifact@v4
+        uses: actions/upload-artifact@v7
         with:
           name: cargo-dist-cache
           path: ~/.cargo/bin/dist
@@ -82,7 +82,7 @@ jobs:
           cat plan-dist-manifest.json
           echo "manifest=$(jq -c "." plan-dist-manifest.json)" >> "$GITHUB_OUTPUT"
       - name: "Upload dist-manifest.json"
-        uses: actions/upload-artifact@v4
+        uses: actions/upload-artifact@v7
         with:
           name: artifacts-plan-dist-manifest
           path: plan-dist-manifest.json
@@ -131,7 +131,7 @@ jobs:
         run: ${{ matrix.install_dist.run }}
       # Get the dist-manifest
       - name: Fetch local artifacts
-        uses: actions/download-artifact@v4
+        uses: actions/download-artifact@v8
         with:
           pattern: artifacts-*
           path: target/distrib/
@@ -158,7 +158,7 @@ jobs:
 
           cp dist-manifest.json "$BUILD_MANIFEST_NAME"
       - name: "Upload artifacts"
-        uses: actions/upload-artifact@v4
+        uses: actions/upload-artifact@v7
         with:
           name: artifacts-build-local-${{ join(matrix.targets, '_') }}
           path: |
@@ -180,14 +180,14 @@ jobs:
           persist-credentials: false
           submodules: recursive
       - name: Install cached dist
-        uses: actions/download-artifact@v4
+        uses: actions/download-artifact@v8
         with:
           name: cargo-dist-cache
           path: ~/.cargo/bin/
       - run: chmod +x ~/.cargo/bin/dist
       # Get all the local artifacts for the global tasks to use (for e.g. checksums)
       - name: Fetch local artifacts
-        uses: actions/download-artifact@v4
+        uses: actions/download-artifact@v8
         with:
           pattern: artifacts-*
           path: target/distrib/
@@ -205,7 +205,7 @@ jobs:
 
           cp dist-manifest.json "$BUILD_MANIFEST_NAME"
       - name: "Upload artifacts"
-        uses: actions/upload-artifact@v4
+        uses: actions/upload-artifact@v7
         with:
           name: artifacts-build-global
           path: |
@@ -230,14 +230,14 @@ jobs:
           persist-credentials: false
           submodules: recursive
       - name: Install cached dist
-        uses: actions/download-artifact@v4
+        uses: actions/download-artifact@v8
         with:
           name: cargo-dist-cache
           path: ~/.cargo/bin/
       - run: chmod +x ~/.cargo/bin/dist
       # Fetch artifacts from scratch-storage
       - name: Fetch artifacts
-        uses: actions/download-artifact@v4
+        uses: actions/download-artifact@v8
         with:
           pattern: artifacts-*
           path: target/distrib/
@@ -250,14 +250,14 @@ jobs:
           cat dist-manifest.json
           echo "manifest=$(jq -c "." dist-manifest.json)" >> "$GITHUB_OUTPUT"
       - name: "Upload dist-manifest.json"
-        uses: actions/upload-artifact@v4
+        uses: actions/upload-artifact@v7
         with:
           # Overwrite the previous copy
           name: artifacts-dist-manifest
           path: dist-manifest.json
       # Create a GitHub Release while uploading all files to it
       - name: "Download GitHub Artifacts"
-        uses: actions/download-artifact@v4
+        uses: actions/download-artifact@v8
         with:
           pattern: artifacts-*
           path: artifacts

.github/workflows/wasi.yml

@@ -0,0 +1,28 @@
+# spell-checker:ignore wasip
+name: WASI
+
+on:
+  pull_request:
+  push:
+    branches:
+      - main
+
+permissions:
+  contents: read
+
+# End the current execution if there is a new changeset in the PR.
+concurrency:
+  group: ${{ github.workflow }}-${{ github.ref }}
+  cancel-in-progress: ${{ github.ref != 'refs/heads/main' }}
+
+jobs:
+  test_wasi:
+    name: Tests
+    runs-on: ubuntu-latest
+    steps:
+    - uses: actions/checkout@v6
+    - uses: dtolnay/rust-toolchain@stable
+      with:
+        targets: wasm32-wasip1
+    - name: check
+      run: cargo check --target wasm32-wasip1

.pre-commit-config.yaml

@@ -0,0 +1,48 @@
+# See https://pre-commit.com for more information
+# See https://pre-commit.com/hooks.html for more hooks
+exclude: ^tests/fixtures/
+repos:
+  - repo: https://github.com/pre-commit/pre-commit-hooks
+    rev: v6.0.0
+    hooks:
+      - id: check-added-large-files
+      - id: check-executables-have-shebangs
+      - id: check-json
+        exclude: '\.vscode/(cSpell|extensions)\.json' # cSpell.json and extensions.json use comments
+      - id: check-shebang-scripts-are-executable
+        exclude: '.+\.rs' # would be triggered by #![some_attribute]
+      - id: check-symlinks
+      - id: check-toml
+      - id: check-yaml
+        args: [ --allow-multiple-documents ]
+      - id: destroyed-symlinks
+      - id: end-of-file-fixer
+      - id: mixed-line-ending
+        args: [ --fix=lf ]
+      - id: trailing-whitespace
+
+  - repo: local
+    hooks:
+      - id: rust-linting
+        name: Rust linting
+        description: Run cargo fmt on files included in the commit.
+        entry: cargo +stable fmt --
+        pass_filenames: true
+        types: [file, rust]
+        language: system
+      - id: rust-clippy
+        name: Rust clippy
+        description: Run cargo clippy on files included in the commit.
+        entry: cargo +stable clippy --workspace --all-targets --all-features -- -D warnings
+        pass_filenames: false
+        types: [file, rust]
+        language: system
+      - id: cspell
+        name: Code spell checker (cspell)
+        description: Run cspell to check for spelling errors (if available).
+        entry: bash -c 'if command -v cspell >/dev/null 2>&1; then cspell --no-must-find-files -- "$@"; else echo "cspell not found, skipping spell check"; exit 0; fi' --
+        pass_filenames: true
+        language: system
+
+ci:
+  skip: [rust-linting, rust-clippy, cspell]

CONTRIBUTING.md

@@ -0,0 +1,32 @@
+# Contributing to diffutils
+
+Hi! Welcome to uutils/diffutils, and thanks for wanting to contribute!
+
+This project follows the shared conventions of the [uutils](https://github.com/uutils)
+organization. Before opening a pull request, please read:
+
+- Our **[Review Guidelines](https://uutils.github.io/reviews/)** — what we expect
+  from a pull request and how reviews are carried out.
+- Our community's [CODE_OF_CONDUCT.md](./CODE_OF_CONDUCT.md), if present.
+
+Finally, feel free to join our [Discord](https://discord.gg/wQVJbvJ)!
+
+> [!WARNING]
+> uutils is original code and cannot contain any code from GNU or other
+> strongly-licensed (GPL/LGPL) implementations. We **cannot** accept changes
+> based on the GNU source code, and you **must not link** to it either. You may
+> look at permissively-licensed implementations (MIT/BSD) and read the GNU
+> *manuals* — never the GNU *source*.
+
+## In short
+
+- Discuss non-trivial changes in an issue **before** writing the code.
+- Keep pull requests **small, self-contained, and descriptively titled**
+  (e.g. `diffutils: fix ...`).
+- Make sure CI passes: tests are green, `rustfmt` is satisfied, and there are
+  no `clippy` warnings.
+- Add tests for new behavior; don't let coverage regress.
+- Write small, atomic commits annotated with the component you touched.
+
+See the [Review Guidelines](https://uutils.github.io/reviews/) for the full
+details.

COPYRIGHT

@@ -0,0 +1,8 @@
+Copyright (c) Michael Howell
+Copyright (c) uutils developers
+
+Licensed under the Apache License, Version 2.0 <LICENSE-APACHE or
+http://www.apache.org/licenses/LICENSE-2.0> or the MIT license
+<LICENSE-MIT or http://opensource.org/licenses/MIT>, at your
+option. All files in the project carrying such notice may not be
+copied, modified, or distributed except according to those terms.

Cargo.toml

@@ -23,20 +23,30 @@ same-file = "1.0.6"
 unicode-width = "0.2.0"
 
 [dev-dependencies]
-pretty_assertions = "1.4.0"
 assert_cmd = "2.0.14"
+divan = { version = "4.3.0", package = "codspeed-divan-compat" }
+pretty_assertions = "1.4.0"
 predicates = "3.1.0"
-tempfile = "3.10.1"
+rand = "0.10.0"
+tempfile = "3.26.0"
 
 [profile.release]
 lto = "thin"
 codegen-units = 1
-
-[profile.release-fast]
-inherits = "release"
 panic = "abort"
 
-# The profile that 'dist' will build with
+# alias profile for 'dist'
 [profile.dist]
 inherits = "release"
-lto = "thin"
+
+[[bench]]
+name = "bench_diffutils"
+path = "benches/bench-diffutils.rs"
+harness = false
+
+[features]
+# default = ["feat_bench_not_diff"]
+# Turn bench for diffutils cmp off
+feat_bench_not_cmp = []
+# Turn bench for diffutils diff off
+feat_bench_not_diff = []

LICENSE-APACHE

@@ -1,6 +1,3 @@
-Copyright (c) Michael Howell
-Copyright (c) uutils developers
-
                               Apache License
                         Version 2.0, January 2004
                      http://www.apache.org/licenses/

LICENSE-MIT

@@ -1,6 +1,3 @@
-Copyright (c) Michael Howell
-Copyright (c) uutils developers
-
 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

README.md

@@ -2,6 +2,7 @@
 [![Discord](https://img.shields.io/badge/discord-join-7289DA.svg?logo=discord&longCache=true&style=flat)](https://discord.gg/wQVJbvJ)
 [![License](http://img.shields.io/badge/license-MIT-blue.svg)](https://github.com/uutils/diffutils/blob/main/LICENSE)
 [![dependency status](https://deps.rs/repo/github/uutils/diffutils/status.svg)](https://deps.rs/repo/github/uutils/diffutils)
+[![CodSpeed](https://img.shields.io/endpoint?url=https://codspeed.io/badge.json)](https://codspeed.io/uutils/diffutils?utm_source=badge)
 
 [![CodeCov](https://codecov.io/gh/uutils/diffutils/branch/main/graph/badge.svg)](https://codecov.io/gh/uutils/diffutils)
 
@@ -53,4 +54,8 @@ $ cargo run -- -u fruits_old.txt fruits_new.txt
 
 ## License
 
-diffutils is licensed under the MIT and Apache Licenses - see the `LICENSE-MIT` or `LICENSE-APACHE` files for details
+This project is distributed under the terms of both the MIT license and the
+Apache License (Version 2.0).
+
+See [LICENSE-APACHE](LICENSE-APACHE), [LICENSE-MIT](LICENSE-MIT), and
+[COPYRIGHT](COPYRIGHT) for details.

SECURITY.md

@@ -0,0 +1,44 @@
+# Security Policy
+
+## Supported Versions
+
+We provide security updates only for the latest released version of `uutils/diffutils`.
+Older versions may not receive patches.
+If you are using a version packaged by your Linux distribution, please check with your distribution maintainers for their update policy.
+
+---
+
+## Reporting a Vulnerability
+
+**Do not open public GitHub issues for security vulnerabilities.**
+This prevents accidental disclosure before a fix is available.
+
+Instead, please use the following method:
+
+- **Email:** [sylvestre@debian.org](mailto:Sylvestre@debian.org)
+- **Encryption (optional):** You may encrypt your report using our PGP key:
+Fingerprint: B60D B599 4D39 BEC4 D1A9 5CCF 7E65 28DA 752F 1BE1
+---
+
+### What to Include in Your Report
+
+To help us investigate and resolve the issue quickly, please include as much detail as possible:
+
+- **Type of issue:** e.g. privilege escalation, information disclosure.
+- **Location in the source:** file path, commit hash, branch, or tag.
+- **Steps to reproduce:** exact commands, test cases, or scripts.
+- **Special configuration:** any flags, environment variables, or system setup required.
+- **Affected systems:** OS/distribution and version(s) where the issue occurs.
+- **Impact:** your assessment of the potential severity (DoS, RCE, data leak, etc.).
+
+---
+
+## Disclosure Policy
+
+We follow a **Coordinated Vulnerability Disclosure (CVD)** process:
+
+1. We will acknowledge receipt of your report within **10 days**.
+2. We will investigate, reproduce, and assess the issue.
+3. We will provide a timeline for developing and releasing a fix.
+4. Once a fix is available, we will publish a GitHub Security Advisory.
+5. You will be credited in the advisory unless you request anonymity.

benches/bench-diffutils.rs

@@ -0,0 +1,377 @@
+// This file is part of the uutils diffutils package.
+//
+// For the full copyright and license information, please view the LICENSE-*
+// files that was distributed with this source code.
+
+//! Benches for all utils in diffutils.
+//!
+//! There is a file generator included to create files of different sizes for comparison. \
+//! Set the TEMP_DIR const to keep the files. df_to_ files have small changes in them, search for '#'. \
+//! File generation up to 1 GB is really fast, Benchmarking above 100 MB takes very long.
+
+/// Generate test files with these sizes in KB.
+const FILE_SIZE_KILO_BYTES: [u64; 4] = [100, 1 * MB, 10 * MB, 25 * MB];
+// const FILE_SIZE_KILO_BYTES: [u64; 3] = [100, 1 * MB, 5 * MB];
+// Empty String to use TempDir (files will be removed after test) or specify dir to keep generated files
+const TEMP_DIR: &str = "";
+const NUM_DIFF: u64 = 4;
+// just for FILE_SIZE_KILO_BYTES
+const MB: u64 = 1_000;
+const CHANGE_CHAR: u8 = b'#';
+
+#[cfg(not(feature = "feat_bench_not_cmp"))]
+mod diffutils_cmp {
+    use std::hint::black_box;
+
+    use diffutilslib::cmp;
+    use divan::Bencher;
+
+    use crate::{binary, prepare::*, FILE_SIZE_KILO_BYTES};
+
+    #[divan::bench(args = FILE_SIZE_KILO_BYTES)]
+    fn cmp_compare_files_equal(bencher: Bencher, kb: u64) {
+        let (from, to) = get_context().get_test_files_equal(kb);
+        let cmd = format!("cmp {from} {to}");
+        let opts = str_to_options(&cmd).into_iter().peekable();
+        let params = cmp::parse_params(opts).unwrap();
+
+        bencher
+            // .with_inputs(|| prepare::cmp_params_identical_testfiles(lines))
+            .with_inputs(|| params.clone())
+            .bench_refs(|params| black_box(cmp::cmp(&params).unwrap()));
+    }
+
+    // bench the actual compare; cmp exits on first difference
+    #[divan::bench(args = FILE_SIZE_KILO_BYTES)]
+    fn cmp_compare_files_different(bencher: Bencher, bytes: u64) {
+        let (from, to) = get_context().get_test_files_different(bytes);
+        let cmd = format!("cmp {from} {to} -s");
+        let opts = str_to_options(&cmd).into_iter().peekable();
+        let params = cmp::parse_params(opts).unwrap();
+
+        bencher
+            // .with_inputs(|| prepare::cmp_params_identical_testfiles(lines))
+            .with_inputs(|| params.clone())
+            .bench_refs(|params| black_box(cmp::cmp(&params).unwrap()));
+    }
+
+    // bench original GNU cmp
+    #[divan::bench(args = FILE_SIZE_KILO_BYTES)]
+    fn cmd_cmp_gnu_equal(bencher: Bencher, bytes: u64) {
+        let (from, to) = get_context().get_test_files_equal(bytes);
+        let args_str = format!("{from} {to}");
+        bencher
+            // .with_inputs(|| prepare::cmp_params_identical_testfiles(lines))
+            .with_inputs(|| args_str.clone())
+            .bench_refs(|cmd_args| binary::bench_binary("cmp", cmd_args));
+    }
+
+    // bench the compiled release version
+    #[divan::bench(args = FILE_SIZE_KILO_BYTES)]
+    fn cmd_cmp_release_equal(bencher: Bencher, bytes: u64) {
+        let (from, to) = get_context().get_test_files_equal(bytes);
+        let args_str = format!("cmp {from} {to}");
+
+        bencher
+            // .with_inputs(|| prepare::cmp_params_identical_testfiles(lines))
+            .with_inputs(|| args_str.clone())
+            .bench_refs(|cmd_args| binary::bench_binary("target/release/diffutils", cmd_args));
+    }
+}
+
+#[cfg(not(feature = "feat_bench_not_diff"))]
+mod diffutils_diff {
+    // use std::hint::black_box;
+
+    use crate::{binary, prepare::*, FILE_SIZE_KILO_BYTES};
+    // use diffutilslib::params;
+    use divan::Bencher;
+
+    // bench the actual compare
+    // TODO diff does not have a diff function
+    //     #[divan::bench(args = [100_000,10_000])]
+    //     fn diff_compare_files(bencher: Bencher, bytes: u64) {
+    //         let (from, to) = gen_testfiles(lines, 0, "id");
+    //         let cmd = format!("cmp {from} {to}");
+    //         let opts = str_to_options(&cmd).into_iter().peekable();
+    //         let params = params::parse_params(opts).unwrap();
+    //
+    //         bencher
+    //             // .with_inputs(|| prepare::cmp_params_identical_testfiles(lines))
+    //             .with_inputs(|| params.clone())
+    //             .bench_refs(|params| diff::diff(&params).unwrap());
+    //     }
+
+    // bench original GNU diff
+    #[divan::bench(args = FILE_SIZE_KILO_BYTES)]
+    fn cmd_diff_gnu_equal(bencher: Bencher, bytes: u64) {
+        let (from, to) = get_context().get_test_files_equal(bytes);
+        let args_str = format!("{from} {to}");
+        bencher
+            // .with_inputs(|| prepare::cmp_params_identical_testfiles(lines))
+            .with_inputs(|| args_str.clone())
+            .bench_refs(|cmd_args| binary::bench_binary("diff", cmd_args));
+    }
+
+    // bench the compiled release version
+    #[divan::bench(args = FILE_SIZE_KILO_BYTES)]
+    fn cmd_diff_release_equal(bencher: Bencher, bytes: u64) {
+        let (from, to) = get_context().get_test_files_equal(bytes);
+        let args_str = format!("diff {from} {to}");
+
+        bencher
+            // .with_inputs(|| prepare::cmp_params_identical_testfiles(lines))
+            .with_inputs(|| args_str.clone())
+            .bench_refs(|cmd_args| binary::bench_binary("target/release/diffutils", cmd_args));
+    }
+}
+
+mod parser {
+    use std::hint::black_box;
+
+    use diffutilslib::{cmp, params};
+    use divan::Bencher;
+
+    use crate::prepare::str_to_options;
+
+    // bench the time it takes to parse the command line arguments
+    #[divan::bench]
+    fn cmp_parser(bencher: Bencher) {
+        let cmd = "cmd file_1.txt file_2.txt -bl n10M --ignore-initial=100KiB:1MiB";
+        let args = str_to_options(&cmd).into_iter().peekable();
+        bencher
+            .with_inputs(|| args.clone())
+            .bench_values(|data| black_box(cmp::parse_params(data)));
+    }
+
+    // // test the impact on the benchmark if not converting the cmd to Vec<OsString> (doubles for parse)
+    // #[divan::bench]
+    // fn cmp_parser_no_prepare() {
+    //     let cmd = "cmd file_1.txt file_2.txt -bl n10M --ignore-initial=100KiB:1MiB";
+    //     let args = str_to_options(&cmd).into_iter().peekable();
+    //     let _ = cmp::parse_params(args);
+    // }
+
+    // bench the time it takes to parse the command line arguments
+    #[divan::bench]
+    fn diff_parser(bencher: Bencher) {
+        let cmd = "diff file_1.txt file_2.txt -s --brief --expand-tabs --width=100";
+        let args = str_to_options(&cmd).into_iter().peekable();
+        bencher
+            .with_inputs(|| args.clone())
+            .bench_values(|data| black_box(params::parse_params(data)));
+    }
+}
+
+mod prepare {
+    use std::{
+        ffi::OsString,
+        fs::{self, File},
+        io::{BufWriter, Write},
+        path::Path,
+        sync::OnceLock,
+    };
+
+    use rand::RngExt;
+    use tempfile::TempDir;
+
+    use crate::{CHANGE_CHAR, FILE_SIZE_KILO_BYTES, NUM_DIFF, TEMP_DIR};
+
+    // file lines and .txt will be added
+    const FROM_FILE: &str = "from_file";
+    const TO_FILE: &str = "to_file";
+    const LINE_LENGTH: usize = 60;
+
+    /// Contains test data (file names) which only needs to be created once.
+    #[derive(Debug, Default)]
+    pub struct BenchContext {
+        pub tmp_dir: Option<TempDir>,
+        pub dir: String,
+        pub files_equal: Vec<(String, String)>,
+        pub files_different: Vec<(String, String)>,
+    }
+
+    impl BenchContext {
+        pub fn get_path(&self) -> &Path {
+            match &self.tmp_dir {
+                Some(tmp) => tmp.path(),
+                None => Path::new(&self.dir),
+            }
+        }
+
+        pub fn get_test_files_equal(&self, kb: u64) -> &(String, String) {
+            let p = FILE_SIZE_KILO_BYTES.iter().position(|f| *f == kb).unwrap();
+            &self.files_equal[p]
+        }
+
+        #[allow(unused)]
+        pub fn get_test_files_different(&self, kb: u64) -> &(String, String) {
+            let p = FILE_SIZE_KILO_BYTES.iter().position(|f| *f == kb).unwrap();
+            &self.files_different[p]
+        }
+    }
+
+    // Since each bench function is separate in Divan it is more difficult to dynamically create test data.
+    // This keeps the TempDir alive until the program exits and generates the files only once.
+    static SHARED_CONTEXT: OnceLock<BenchContext> = OnceLock::new();
+    /// Creates the test files once and provides them to all tests.
+    pub fn get_context() -> &'static BenchContext {
+        SHARED_CONTEXT.get_or_init(|| {
+            let mut ctx = BenchContext::default();
+            if TEMP_DIR.is_empty() {
+                let tmp_dir = TempDir::new().expect("Failed to create temp dir");
+                ctx.tmp_dir = Some(tmp_dir);
+            } else {
+                // uses current directory, the generated files are kept
+                let path = Path::new(TEMP_DIR);
+                if !path.exists() {
+                    fs::create_dir_all(path).expect("Path {path} could not be created");
+                }
+                ctx.dir = TEMP_DIR.to_string();
+            };
+
+            // generate test bytes
+            for kb in FILE_SIZE_KILO_BYTES {
+                let f = generate_test_files_bytes(ctx.get_path(), kb * 1000, 0, "eq")
+                    .expect("generate_test_files failed");
+                ctx.files_equal.push(f);
+                let f = generate_test_files_bytes(ctx.get_path(), kb * 1000, NUM_DIFF, "df")
+                    .expect("generate_test_files failed");
+                ctx.files_different.push(f);
+            }
+
+            ctx
+        })
+    }
+
+    pub fn str_to_options(opt: &str) -> Vec<OsString> {
+        let s: Vec<OsString> = opt
+            .split(" ")
+            .into_iter()
+            .filter(|s| !s.is_empty())
+            .map(|s| OsString::from(s))
+            .collect();
+
+        s
+    }
+
+    /// Generates two test files for comparison with <bytes> size.
+    ///
+    /// Each line consists of 10 words with 5 letters, giving a line length of 60 bytes.
+    /// If num_differences is set, '#' will be inserted between the first two words of a line,
+    /// evenly spaced in the file. 1 will add the change in the last line, so the comparison takes longest.
+    fn generate_test_files_bytes(
+        dir: &Path,
+        bytes: u64,
+        num_differences: u64,
+        id: &str,
+    ) -> std::io::Result<(String, String)> {
+        let id = if id.is_empty() {
+            "".to_string()
+        } else {
+            format!("{id}_")
+        };
+        let f1 = format!("{id}{FROM_FILE}_{bytes}.txt");
+        let f2 = format!("{id}{TO_FILE}_{bytes}.txt");
+        let from_path = dir.join(f1);
+        let to_path = dir.join(f2);
+
+        generate_file_bytes(&from_path, &to_path, bytes, num_differences)?;
+
+        Ok((
+            from_path.to_string_lossy().to_string(),
+            to_path.to_string_lossy().to_string(),
+        ))
+    }
+
+    fn generate_file_bytes(
+        from_name: &Path,
+        to_name: &Path,
+        bytes: u64,
+        num_differences: u64,
+    ) -> std::io::Result<()> {
+        let file_from = File::create(from_name)?;
+        let file_to = File::create(to_name)?;
+        // for int division, lines will be smaller than requested bytes
+        let n_lines = bytes / LINE_LENGTH as u64;
+        let change_every_n_lines = if num_differences == 0 {
+            0
+        } else {
+            let c = n_lines / num_differences;
+            if c == 0 {
+                1
+            } else {
+                c
+            }
+        };
+        // Use a larger 128KB buffer for massive files
+        let mut writer_from = BufWriter::with_capacity(128 * 1024, file_from);
+        let mut writer_to = BufWriter::with_capacity(128 * 1024, file_to);
+        let mut rng = rand::rng();
+
+        // Each line: (5 chars * 10 words) + 9 spaces + 1 newline = 60 bytes
+        let mut line_buffer = [b' '; 60];
+        line_buffer[59] = b'\n'; // Set the newline once at the end
+
+        for i in (0..n_lines).rev() {
+            // Fill only the letter positions, skipping spaces and the newline
+            for word_idx in 0..10 {
+                let start = word_idx * 6; // Each word + space block is 6 bytes
+                for i in 0..5 {
+                    line_buffer[start + i] = rng.random_range(b'a'..b'z' + 1);
+                }
+            }
+
+            // Write the raw bytes directly to both files
+            writer_from.write_all(&line_buffer)?;
+            // make changes in the file
+            if num_differences == 0 {
+                writer_to.write_all(&line_buffer)?;
+            } else {
+                if i % change_every_n_lines == 0 && n_lines - i > 2 {
+                    line_buffer[5] = CHANGE_CHAR;
+                }
+                writer_to.write_all(&line_buffer)?;
+                line_buffer[5] = b' ';
+            }
+        }
+
+        // create last line
+        let missing = (bytes - n_lines as u64 * LINE_LENGTH as u64) as usize;
+        if missing > 0 {
+            for word_idx in 0..10 {
+                let start = word_idx * 6; // Each word + space block is 6 bytes
+                for i in 0..5 {
+                    line_buffer[start + i] = rng.random_range(b'a'..b'z' + 1);
+                }
+            }
+            line_buffer[missing - 1] = b'\n';
+            writer_from.write_all(&line_buffer[0..missing])?;
+            writer_to.write_all(&line_buffer[0..missing])?;
+        }
+
+        writer_from.flush()?;
+        writer_to.flush()?;
+
+        Ok(())
+    }
+}
+
+mod binary {
+    use std::process::Command;
+
+    use crate::prepare::str_to_options;
+
+    pub fn bench_binary(program: &str, cmd_args: &str) -> std::process::ExitStatus {
+        let args = str_to_options(cmd_args);
+        Command::new(program)
+            .args(args)
+            .status()
+            .expect("Failed to execute binary")
+    }
+}
+
+fn main() {
+    // Run registered benchmarks.
+    divan::main();
+}

fuzz/Cargo.lock

@@ -34,15 +34,15 @@ checksum = "c08606f8c3cbf4ce6ec8e28fb0014a2c086708fe954eaa885384a6165172e7e8"
 
 [[package]]
 name = "bumpalo"
-version = "3.19.1"
+version = "3.20.2"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "5dd9dc738b7a8311c7ade152424974d8115f2cdad61e8dab8dac9f2362298510"
+checksum = "5d20789868f4b01b2f2caec9f5c4e0213b41e3e5702a50157d699ae31ced2fcb"
 
 [[package]]
 name = "cc"
-version = "1.2.51"
+version = "1.2.60"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "7a0aeaff4ff1a90589618835a598e545176939b97874f7abc7851caa0618f203"
+checksum = "43c5703da9466b66a946814e1adf53ea2c90f10063b86290cc9eb67ce3478a20"
 dependencies = [
  "find-msvc-tools",
  "jobserver",
@@ -58,9 +58,9 @@ checksum = "9330f8b2ff13f34540b44e946ef35111825727b38d33286ef986142615121801"
 
 [[package]]
 name = "chrono"
-version = "0.4.42"
+version = "0.4.44"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "145052bdd345b87320e369255277e3fb5152762ad123a901ef5c262dd38fe8d2"
+checksum = "c673075a2e0e5f4a1dde27ce9dee1ea4558c7ffe648f576438a20ca1d2acc4b0"
 dependencies = [
  "iana-time-zone",
  "js-sys",
@@ -95,9 +95,9 @@ dependencies = [
 
 [[package]]
 name = "find-msvc-tools"
-version = "0.1.6"
+version = "0.1.9"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "645cbb3a84e60b7531617d5ae4e57f7e27308f6445f5abf653209ea76dec8dff"
+checksum = "5baebc0774151f905a1a2cc41989300b1e6fbb29aff0ceffa1064fdd3088d582"
 
 [[package]]
 name = "getrandom"
@@ -113,9 +113,9 @@ dependencies = [
 
 [[package]]
 name = "iana-time-zone"
-version = "0.1.64"
+version = "0.1.65"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "33e57f83510bb73707521ebaffa789ec8caf86f9657cad665b092b581d40e9fb"
+checksum = "e31bc9ad994ba00e440a8aa5c9ef0ec67d5cb5e5cb0cc7f8b744a35b389cc470"
 dependencies = [
  "android_system_properties",
  "core-foundation-sys",
@@ -137,9 +137,9 @@ dependencies = [
 
 [[package]]
 name = "itoa"
-version = "1.0.17"
+version = "1.0.18"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "92ecc6618181def0457392ccd0ee51198e065e016d1d527a7ac1b6dc7c1f09d2"
+checksum = "8f42a60cbdf9a97f5d2305f08a87dc4e09308d1276d28c869c684d7777685682"
 
 [[package]]
 name = "jobserver"
@@ -153,9 +153,9 @@ dependencies = [
 
 [[package]]
 name = "js-sys"
-version = "0.3.83"
+version = "0.3.94"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "464a3709c7f55f1f721e5389aa6ea4e3bc6aba669353300af094b29ffbdde1d8"
+checksum = "2e04e2ef80ce82e13552136fabeef8a5ed1f985a96805761cbb9a2c34e7664d9"
 dependencies = [
  "once_cell",
  "wasm-bindgen",
@@ -163,15 +163,15 @@ dependencies = [
 
 [[package]]
 name = "libc"
-version = "0.2.178"
+version = "0.2.184"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "37c93d8daa9d8a012fd8ab92f088405fb202ea0b6ab73ee2482ae66af4f42091"
+checksum = "48f5d2a454e16a5ea0f4ced81bd44e4cfc7bd3a507b61887c99fd3538b28e4af"
 
 [[package]]
 name = "libfuzzer-sys"
-version = "0.4.10"
+version = "0.4.13"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "5037190e1f70cbeef565bd267599242926f724d3b8a9f510fd7e0b540cfa4404"
+checksum = "a9fd2f41a1cba099f79a0b6b6c35656cf7c03351a7bae8ff0f28f25270f929d2"
 dependencies = [
  "arbitrary",
  "cc",
@@ -185,9 +185,9 @@ checksum = "5e5032e24019045c762d3c0f28f5b6b8bbf38563a65908389bf7978758920897"
 
 [[package]]
 name = "memchr"
-version = "2.7.6"
+version = "2.8.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "f52b00d39961fc5b2736ea853c9cc86238e165017a493d1d5c8eac6bdc4cc273"
+checksum = "f8ca58f447f06ed17d5fc4043ce1b10dd205e060fb3ce5b979b8ed8e59ff3f79"
 
 [[package]]
 name = "num-traits"
@@ -200,24 +200,24 @@ dependencies = [
 
 [[package]]
 name = "once_cell"
-version = "1.21.3"
+version = "1.21.4"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "42f5e15c9953c5e4ccceeb2e7382a716482c34515315f7b03532b8b4e8393d2d"
+checksum = "9f7c3e4beb33f85d45ae3e3a1792185706c8e16d043238c593331cc7cd313b50"
 
 [[package]]
 name = "proc-macro2"
-version = "1.0.104"
+version = "1.0.106"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "9695f8df41bb4f3d222c95a67532365f569318332d03d5f3f67f37b20e6ebdf0"
+checksum = "8fd00f0bb2e90d81d1044c2b32617f68fcb9fa3bb7640c23e9c748e53fb30934"
 dependencies = [
  "unicode-ident",
 ]
 
 [[package]]
 name = "quote"
-version = "1.0.42"
+version = "1.0.45"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "a338cc41d27e6cc6dce6cefc13a0729dfbb81c262b1f519331575dd80ef3067f"
+checksum = "41f2619966050689382d2b44f664f4bc593e129785a36d6ee376ddf37259b924"
 dependencies = [
  "proc-macro2",
 ]
@@ -230,9 +230,9 @@ checksum = "69cdb34c158ceb288df11e18b4bd39de994f6657d83847bdffdbd7f346754b0f"
 
 [[package]]
 name = "regex"
-version = "1.12.2"
+version = "1.12.3"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "843bc0191f75f3e22651ae5f1e72939ab2f72a4bc30fa80a066bd66edefc24d4"
+checksum = "e10754a14b9137dd7b1e3e5b0493cc9171fdd105e0ab477f51b72e7f3ac0e276"
 dependencies = [
  "aho-corasick",
  "memchr",
@@ -242,9 +242,9 @@ dependencies = [
 
 [[package]]
 name = "regex-automata"
-version = "0.4.13"
+version = "0.4.14"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "5276caf25ac86c8d810222b3dbb938e512c55c6831a10f3e6ed1c93b84041f1c"
+checksum = "6e1dd4122fc1595e8162618945476892eefca7b88c52820e74af6262213cae8f"
 dependencies = [
  "aho-corasick",
  "memchr",
@@ -253,9 +253,9 @@ dependencies = [
 
 [[package]]
 name = "regex-syntax"
-version = "0.8.8"
+version = "0.8.10"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "7a2d987857b319362043e95f5353c0535c1f58eec5336fdfcf626430af7def58"
+checksum = "dc897dd8d9e8bd1ed8cdad82b5966c3e0ecae09fb1907d58efaa013543185d0a"
 
 [[package]]
 name = "rustversion"
@@ -280,9 +280,9 @@ checksum = "0fda2ff0d084019ba4d7c6f371c95d8fd75ce3524c3cb8fb653a3023f6323e64"
 
 [[package]]
 name = "syn"
-version = "2.0.112"
+version = "2.0.117"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "21f182278bf2d2bcb3c88b1b08a37df029d71ce3d3ae26168e3c653b213b99d4"
+checksum = "e665b8803e7b1d2a727f4023456bbbbe74da67099c585258af0ad9c5013b9b99"
 dependencies = [
  "proc-macro2",
  "quote",
@@ -291,9 +291,9 @@ dependencies = [
 
 [[package]]
 name = "unicode-ident"
-version = "1.0.22"
+version = "1.0.24"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "9312f7c4f6ff9069b165498234ce8be658059c6728633667c526e27dc2cf1df5"
+checksum = "e6e4313cd5fcd3dad5cafa179702e2b244f760991f45397d14d4ebf38247da75"
 
 [[package]]
 name = "unicode-width"
@@ -311,18 +311,18 @@ dependencies = [
 
 [[package]]
 name = "wasip2"
-version = "1.0.1+wasi-0.2.4"
+version = "1.0.2+wasi-0.2.9"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "0562428422c63773dad2c345a1882263bbf4d65cf3f42e90921f787ef5ad58e7"
+checksum = "9517f9239f02c069db75e65f174b3da828fe5f5b945c4dd26bd25d89c03ebcf5"
 dependencies = [
  "wit-bindgen",
 ]
 
 [[package]]
 name = "wasm-bindgen"
-version = "0.2.106"
+version = "0.2.117"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "0d759f433fa64a2d763d1340820e46e111a7a5ab75f993d1852d70b03dbb80fd"
+checksum = "0551fc1bb415591e3372d0bc4780db7e587d84e2a7e79da121051c5c4b89d0b0"
 dependencies = [
  "cfg-if",
  "once_cell",
@@ -333,9 +333,9 @@ dependencies = [
 
 [[package]]
 name = "wasm-bindgen-macro"
-version = "0.2.106"
+version = "0.2.117"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "48cb0d2638f8baedbc542ed444afc0644a29166f1595371af4fecf8ce1e7eeb3"
+checksum = "7fbdf9a35adf44786aecd5ff89b4563a90325f9da0923236f6104e603c7e86be"
 dependencies = [
  "quote",
  "wasm-bindgen-macro-support",
@@ -343,9 +343,9 @@ dependencies = [
 
 [[package]]
 name = "wasm-bindgen-macro-support"
-version = "0.2.106"
+version = "0.2.117"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "cefb59d5cd5f92d9dcf80e4683949f15ca4b511f4ac0a6e14d4e1ac60c6ecd40"
+checksum = "dca9693ef2bab6d4e6707234500350d8dad079eb508dca05530c85dc3a529ff2"
 dependencies = [
  "bumpalo",
  "proc-macro2",
@@ -356,9 +356,9 @@ dependencies = [
 
 [[package]]
 name = "wasm-bindgen-shared"
-version = "0.2.106"
+version = "0.2.117"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "cbc538057e648b67f72a982e708d485b2efa771e1ac05fec311f9f63e5800db4"
+checksum = "39129a682a6d2d841b6c429d0c51e5cb0ed1a03829d8b3d1e69a011e62cb3d3b"
 dependencies = [
  "unicode-ident",
 ]
@@ -442,6 +442,6 @@ dependencies = [
 
 [[package]]
 name = "wit-bindgen"
-version = "0.46.0"
+version = "0.51.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "f17a85883d4e6d00e8a97c586de764dabcc06133f7f1d55dce5cdc070ad7fe59"
+checksum = "d7249219f66ced02969388cf2bb044a09756a083d0fab1e566056b04d9fbcaa5"

fuzz/Cargo.toml

@@ -51,4 +51,4 @@ doc = false
 name = "fuzz_side"
 path = "fuzz_targets/fuzz_side.rs"
 test = false
-doc = false
+doc = false

fuzz/fuzz_targets/fuzz_cmp.rs

@@ -4,7 +4,7 @@ extern crate libfuzzer_sys;
 use diffutilslib::cmp::{self, Cmp};
 
 use std::ffi::OsString;
-use std::fs::File;
+use std::fs::{self, File};
 use std::io::Write;
 
 fn os(s: &str) -> OsString {
@@ -18,7 +18,7 @@ fuzz_target!(|x: (Vec<u8>, Vec<u8>)| {
         .peekable();
 
     let (from, to) = x;
-
+    fs::create_dir_all("target").unwrap();
     File::create("target/fuzz.cmp.a")
         .unwrap()
         .write_all(&from)

fuzz/fuzz_targets/fuzz_cmp_args.rs

@@ -11,6 +11,9 @@ fn os(s: &str) -> OsString {
 }
 
 fuzz_target!(|x: Vec<OsString>| -> Corpus {
+    if x.iter().any(|a| a == "--help") {
+        return Corpus::Reject;
+    }
     if x.len() > 6 {
         // Make sure we try to parse an option when we get longer args. x[0] will be
         // the executable name.

fuzz/fuzz_targets/fuzz_ed.rs

@@ -38,6 +38,7 @@ fuzz_target!(|x: (Vec<u8>, Vec<u8>)| {
     } else {
         return;
     }
+    fs::create_dir_all("target").unwrap();
     let diff = diff_w(&from, &to, "target/fuzz.file").unwrap();
     File::create("target/fuzz.file.original")
         .unwrap()

fuzz/fuzz_targets/fuzz_normal.rs

@@ -23,6 +23,7 @@ fuzz_target!(|x: (Vec<u8>, Vec<u8>)| {
         return
     }*/
     let diff = normal_diff::diff(&from, &to, &Params::default());
+    fs::create_dir_all("target").unwrap();
     File::create("target/fuzz.file.original")
         .unwrap()
         .write_all(&from)

fuzz/fuzz_targets/fuzz_patch.rs

@@ -21,15 +21,17 @@ fuzz_target!(|x: (Vec<u8>, Vec<u8>, u8)| {
     } else {
         return
     }*/
+    fs::create_dir_all("target").unwrap();
+    let patched = "target/fuzz.file";
     let diff = unified_diff::diff(
         &from,
         &to,
         &Params {
-            from: "a/fuzz.file".into(),
-            to: "target/fuzz.file".into(),
+            from: patched.into(),
+            to: patched.into(),
             context_count: context as usize,
             ..Default::default()
-        }
+        },
     );
     File::create("target/fuzz.file.original")
         .unwrap()

fuzz/fuzz_targets/fuzz_side.rs

@@ -4,9 +4,9 @@ extern crate libfuzzer_sys;
 
 use diffutilslib::side_diff;
 
-use std::fs::File;
-use std::io::Write;
 use diffutilslib::params::Params;
+use std::fs::{self, File};
+use std::io::Write;
 
 fuzz_target!(|x: (Vec<u8>, Vec<u8>, /* usize, usize */ bool)| {
     let (original, new, /* width, tabsize, */ expand) = x;
@@ -21,6 +21,7 @@ fuzz_target!(|x: (Vec<u8>, Vec<u8>, /* usize, usize */ bool)| {
         expand_tabs: expand,
         ..Default::default()
     };
+    fs::create_dir_all("target").unwrap();
     let mut output_buf = vec![];
     side_diff::diff(&original, &new, &mut output_buf, &params);
     File::create("target/fuzz.file.original")
@@ -39,4 +40,4 @@ fuzz_target!(|x: (Vec<u8>, Vec<u8>, /* usize, usize */ bool)| {
         .unwrap()
         .write_all(&output_buf)
         .unwrap();
-});
+});

src/cmp.rs

@@ -11,24 +11,29 @@ use std::iter::Peekable;
 use std::process::ExitCode;
 use std::{cmp, fs, io};
 
-#[cfg(not(target_os = "windows"))]
+#[cfg(unix)]
 use std::os::fd::{AsRawFd, FromRawFd};
 
-#[cfg(not(target_os = "windows"))]
+#[cfg(unix)]
 use std::os::unix::fs::MetadataExt;
 
 #[cfg(target_os = "windows")]
 use std::os::windows::fs::MetadataExt;
 
+/// for --bytes, so really large number limits can be expressed, like 1Y.
+pub type BytesLimitU64 = u64;
+// ignore initial is currently limited to u64, as take(skip) is used.
+pub type SkipU64 = u64;
+
 #[derive(Clone, Debug, Default, Eq, PartialEq)]
 pub struct Params {
     executable: OsString,
     from: OsString,
     to: OsString,
     print_bytes: bool,
-    skip_a: Option<usize>,
-    skip_b: Option<usize>,
-    max_bytes: Option<usize>,
+    skip_a: Option<SkipU64>,
+    skip_b: Option<SkipU64>,
+    max_bytes: Option<BytesLimitU64>,
     verbose: bool,
     quiet: bool,
 }
@@ -38,7 +43,7 @@ fn usage_string(executable: &str) -> String {
     format!("Usage: {executable} <from> <to>")
 }
 
-#[cfg(not(target_os = "windows"))]
+#[cfg(unix)]
 fn is_stdout_dev_null() -> bool {
     let Ok(dev_null) = fs::metadata("/dev/null") else {
         return false;
@@ -60,19 +65,22 @@ fn is_stdout_dev_null() -> bool {
     is_dev_null
 }
 
+#[cfg(not(any(unix, target_os = "windows")))]
+fn is_stdout_dev_null() -> bool {
+    false
+}
+
 pub fn parse_params<I: Iterator<Item = OsString>>(mut opts: Peekable<I>) -> Result<Params, String> {
-    let Some(executable) = opts.next() else {
-        return Err("Usage: <exe> <from> <to>".to_string());
-    };
+    let executable = opts.next().ok_or("Usage: <exe> <from> <to>".to_string())?;
     let executable_str = executable.to_string_lossy().to_string();
 
-    let parse_skip = |param: &str, skip_desc: &str| -> Result<usize, String> {
+    let parse_skip = |param: &str, skip_desc: &str| -> Result<SkipU64, String> {
         let suffix_start = param
             .find(|b: char| !b.is_ascii_digit())
             .unwrap_or(param.len());
-        let mut num = match param[..suffix_start].parse::<usize>() {
+        let mut num = match param[..suffix_start].parse::<SkipU64>() {
             Ok(num) => num,
-            Err(e) if *e.kind() == std::num::IntErrorKind::PosOverflow => usize::MAX,
+            Err(e) if *e.kind() == std::num::IntErrorKind::PosOverflow => SkipU64::MAX,
             Err(_) => {
                 return Err(format!(
                     "{executable_str}: invalid --ignore-initial value '{skip_desc}'"
@@ -83,33 +91,24 @@ pub fn parse_params<I: Iterator<Item = OsString>>(mut opts: Peekable<I>) -> Resu
         if suffix_start != param.len() {
             // Note that GNU cmp advertises supporting up to Y, but fails if you try
             // to actually use anything beyond E.
-            let multiplier: usize = match &param[suffix_start..] {
+            let multiplier: SkipU64 = match &param[suffix_start..] {
                 "kB" => 1_000,
                 "K" => 1_024,
                 "MB" => 1_000_000,
                 "M" => 1_048_576,
                 "GB" => 1_000_000_000,
                 "G" => 1_073_741_824,
-                // This only generates a warning when compiling for target_pointer_width < 64
-                #[allow(unused_variables)]
-                suffix @ ("TB" | "T" | "PB" | "P" | "EB" | "E") => {
-                    #[cfg(target_pointer_width = "64")]
-                    match suffix {
-                        "TB" => 1_000_000_000_000,
-                        "T" => 1_099_511_627_776,
-                        "PB" => 1_000_000_000_000_000,
-                        "P" => 1_125_899_906_842_624,
-                        "EB" => 1_000_000_000_000_000_000,
-                        "E" => 1_152_921_504_606_846_976,
-                        _ => unreachable!(),
-                    }
-                    #[cfg(not(target_pointer_width = "64"))]
-                    usize::MAX
-                }
-                "ZB" => usize::MAX, // 1_000_000_000_000_000_000_000,
-                "Z" => usize::MAX,  // 1_180_591_620_717_411_303_424,
-                "YB" => usize::MAX, // 1_000_000_000_000_000_000_000_000,
-                "Y" => usize::MAX,  // 1_208_925_819_614_629_174_706_176,
+                "TB" => 1_000_000_000_000,
+                "T" => 1_099_511_627_776,
+                "PB" => 1_000_000_000_000_000,
+                "P" => 1_125_899_906_842_624,
+                "EB" => 1_000_000_000_000_000_000,
+                "E" => 1_152_921_504_606_846_976,
+                // TODO setting usize:MAX does not mimic GNU cmp behavior, it should be an error.
+                "ZB" => SkipU64::MAX, // 1_000_000_000_000_000_000_000,
+                "Z" => SkipU64::MAX,  // 1_180_591_620_717_411_303_424,
+                "YB" => SkipU64::MAX, // 1_000_000_000_000_000_000_000_000,
+                "Y" => SkipU64::MAX,  // 1_208_925_819_614_629_174_706_176,
                 _ => {
                     return Err(format!(
                         "{executable_str}: invalid --ignore-initial value '{skip_desc}'"
@@ -119,7 +118,7 @@ pub fn parse_params<I: Iterator<Item = OsString>>(mut opts: Peekable<I>) -> Resu
 
             num = match num.overflowing_mul(multiplier) {
                 (n, false) => n,
-                _ => usize::MAX,
+                _ => SkipU64::MAX,
             }
         }
 
@@ -173,9 +172,10 @@ pub fn parse_params<I: Iterator<Item = OsString>>(mut opts: Peekable<I>) -> Resu
                 let (_, arg) = param_str.split_once('=').unwrap();
                 arg.to_string()
             };
-            let max_bytes = match max_bytes.parse::<usize>() {
+            let max_bytes = match max_bytes.parse::<BytesLimitU64>() {
                 Ok(num) => num,
-                Err(e) if *e.kind() == std::num::IntErrorKind::PosOverflow => usize::MAX,
+                // TODO limit to MAX is dangerous, this should become an error like in GNU cmp.
+                Err(e) if *e.kind() == std::num::IntErrorKind::PosOverflow => BytesLimitU64::MAX,
                 Err(_) => {
                     return Err(format!(
                         "{executable_str}: invalid --bytes value '{max_bytes}'"
@@ -217,7 +217,7 @@ pub fn parse_params<I: Iterator<Item = OsString>>(mut opts: Peekable<I>) -> Resu
             std::process::exit(0);
         }
         if param_str.starts_with('-') {
-            return Err(format!("Unknown option: {param:?}"));
+            return Err(format!("unrecognized option '{}'", param.to_string_lossy()));
         }
         if from.is_none() {
             from = Some(param);
@@ -233,7 +233,7 @@ pub fn parse_params<I: Iterator<Item = OsString>>(mut opts: Peekable<I>) -> Resu
     }
 
     // Do as GNU cmp, and completely disable printing if we are
-    // outputing to /dev/null.
+    // outputting to /dev/null.
     #[cfg(not(target_os = "windows"))]
     if is_stdout_dev_null() {
         params.quiet = true;
@@ -285,32 +285,21 @@ pub fn parse_params<I: Iterator<Item = OsString>>(mut opts: Peekable<I>) -> Resu
 
 fn prepare_reader(
     path: &OsString,
-    skip: &Option<usize>,
+    skip: &Option<SkipU64>,
     params: &Params,
 ) -> Result<Box<dyn BufRead>, String> {
     let mut reader: Box<dyn BufRead> = if path == "-" {
         Box::new(BufReader::new(io::stdin()))
     } else {
-        match fs::File::open(path) {
-            Ok(file) => Box::new(BufReader::new(file)),
-            Err(e) => {
-                return Err(format_failure_to_read_input_file(
-                    &params.executable,
-                    path,
-                    &e,
-                ));
-            }
-        }
+        let file = fs::File::open(path)
+            .map_err(|e| format_failure_to_read_input_file(&params.executable, path, &e))?;
+        Box::new(BufReader::new(file))
     };
 
     if let Some(skip) = skip {
-        if let Err(e) = io::copy(&mut reader.by_ref().take(*skip as u64), &mut io::sink()) {
-            return Err(format_failure_to_read_input_file(
-                &params.executable,
-                path,
-                &e,
-            ));
-        }
+        // cast as u64 must remain, because value of IgnInit data type could be changed.
+        io::copy(&mut reader.by_ref().take(*skip), &mut io::sink())
+            .map_err(|e| format_failure_to_read_input_file(&params.executable, path, &e))?;
     }
 
     Ok(reader)
@@ -326,11 +315,11 @@ pub fn cmp(params: &Params) -> Result<Cmp, String> {
     let mut from = prepare_reader(&params.from, &params.skip_a, params)?;
     let mut to = prepare_reader(&params.to, &params.skip_b, params)?;
 
-    let mut offset_width = params.max_bytes.unwrap_or(usize::MAX);
+    let mut offset_width = params.max_bytes.unwrap_or(BytesLimitU64::MAX);
 
     if let (Ok(a_meta), Ok(b_meta)) = (fs::metadata(&params.from), fs::metadata(&params.to)) {
         #[cfg(not(target_os = "windows"))]
-        let (a_size, b_size) = (a_meta.size(), b_meta.size());
+        let (a_size, b_size) = (a_meta.len(), b_meta.len());
 
         #[cfg(target_os = "windows")]
         let (a_size, b_size) = (a_meta.file_size(), b_meta.file_size());
@@ -341,7 +330,7 @@ pub fn cmp(params: &Params) -> Result<Cmp, String> {
             return Ok(Cmp::Different);
         }
 
-        let smaller = cmp::min(a_size, b_size) as usize;
+        let smaller = cmp::min(a_size, b_size) as BytesLimitU64;
         offset_width = cmp::min(smaller, offset_width);
     }
 
@@ -350,34 +339,20 @@ pub fn cmp(params: &Params) -> Result<Cmp, String> {
     // Capacity calc: at_byte width + 2 x 3-byte octal numbers + 2 x 4-byte value + 4 spaces
     let mut output = Vec::<u8>::with_capacity(offset_width + 3 * 2 + 4 * 2 + 4);
 
-    let mut at_byte = 1;
-    let mut at_line = 1;
+    let mut at_byte: BytesLimitU64 = 1;
+    let mut at_line: u64 = 1;
     let mut start_of_line = true;
     let mut stdout = BufWriter::new(io::stdout().lock());
     let mut compare = Cmp::Equal;
     loop {
         // Fill up our buffers.
-        let from_buf = match from.fill_buf() {
-            Ok(buf) => buf,
-            Err(e) => {
-                return Err(format_failure_to_read_input_file(
-                    &params.executable,
-                    &params.from,
-                    &e,
-                ));
-            }
-        };
+        let from_buf = from
+            .fill_buf()
+            .map_err(|e| format_failure_to_read_input_file(&params.executable, &params.from, &e))?;
 
-        let to_buf = match to.fill_buf() {
-            Ok(buf) => buf,
-            Err(e) => {
-                return Err(format_failure_to_read_input_file(
-                    &params.executable,
-                    &params.to,
-                    &e,
-                ));
-            }
-        };
+        let to_buf = to
+            .fill_buf()
+            .map_err(|e| format_failure_to_read_input_file(&params.executable, &params.to, &e))?;
 
         // Check for EOF conditions.
         if from_buf.is_empty() && to_buf.is_empty() {
@@ -401,8 +376,8 @@ pub fn cmp(params: &Params) -> Result<Cmp, String> {
         if from_buf[..consumed] == to_buf[..consumed] {
             let last = from_buf[..consumed].last().unwrap();
 
-            at_byte += consumed;
-            at_line += from_buf[..consumed].iter().filter(|&c| *c == b'\n').count();
+            at_byte += consumed as BytesLimitU64;
+            at_line += (from_buf[..consumed].iter().filter(|&c| *c == b'\n').count()) as u64;
 
             start_of_line = *last == b'\n';
 
@@ -500,12 +475,6 @@ pub fn main(opts: Peekable<ArgsOs>) -> ExitCode {
     }
 }
 
-#[inline]
-fn is_ascii_printable(byte: u8) -> bool {
-    let c = byte as char;
-    c.is_ascii() && !c.is_ascii_control()
-}
-
 #[inline]
 fn format_octal(byte: u8, buf: &mut [u8; 3]) -> &str {
     *buf = [b' ', b' ', b'0'];
@@ -525,32 +494,68 @@ fn format_octal(byte: u8, buf: &mut [u8; 3]) -> &str {
 }
 
 #[inline]
-fn format_byte(byte: u8) -> String {
-    let mut byte = byte;
-    let mut quoted = vec![];
-
-    if !is_ascii_printable(byte) {
-        if byte >= 128 {
-            quoted.push(b'M');
-            quoted.push(b'-');
-            byte -= 128;
+fn write_visible_byte(output: &mut Vec<u8>, byte: u8) -> usize {
+    match byte {
+        // Control characters: ^@, ^A, ..., ^_
+        0..=31 => {
+            output.push(b'^');
+            output.push(byte + 64);
+            2
         }
-
-        if byte < 32 {
-            quoted.push(b'^');
-            byte += 64;
-        } else if byte == 127 {
-            quoted.push(b'^');
-            byte = b'?';
+        // Printable ASCII (space through ~)
+        32..=126 => {
+            output.push(byte);
+            1
+        }
+        // DEL: ^?
+        127 => {
+            output.extend_from_slice(b"^?");
+            2
+        }
+        // High bytes with control equivalents: M-^@, M-^A, ..., M-^_
+        128..=159 => {
+            output.push(b'M');
+            output.push(b'-');
+            output.push(b'^');
+            output.push(byte - 64);
+            4
+        }
+        // High bytes: M-<space>, M-!, ..., M-~
+        160..=254 => {
+            output.push(b'M');
+            output.push(b'-');
+            output.push(byte - 128);
+            3
+        }
+        // Byte 255: M-^?
+        255 => {
+            output.extend_from_slice(b"M-^?");
+            4
         }
-        assert!((byte as char).is_ascii());
     }
+}
 
-    quoted.push(byte);
+/// Writes a byte in visible form with right-padding to 4 spaces.
+#[inline]
+fn write_visible_byte_padded(output: &mut Vec<u8>, byte: u8) {
+    const SPACES: &[u8] = b"    ";
+    const WIDTH: usize = SPACES.len();
 
-    // SAFETY: the checks and shifts we do above match what cat and GNU
+    let display_width = write_visible_byte(output, byte);
+
+    // Add right-padding spaces
+    let padding = WIDTH.saturating_sub(display_width);
+    output.extend_from_slice(&SPACES[..padding]);
+}
+
+/// Formats a byte as a visible string (for non-performance-critical path)
+#[inline]
+fn format_visible_byte(byte: u8) -> String {
+    let mut result = Vec::with_capacity(4);
+    write_visible_byte(&mut result, byte);
+    // SAFETY: the checks and shifts in write_visible_byte match what cat and GNU
     // cmp do to ensure characters fall inside the ascii range.
-    unsafe { String::from_utf8_unchecked(quoted) }
+    unsafe { String::from_utf8_unchecked(result) }
 }
 
 // This function has been optimized to not use the Rust fmt system, which
@@ -560,7 +565,7 @@ fn format_byte(byte: u8) -> String {
 fn format_verbose_difference(
     from_byte: u8,
     to_byte: u8,
-    at_byte: usize,
+    at_byte: BytesLimitU64,
     offset_width: usize,
     output: &mut Vec<u8>,
     params: &Params,
@@ -588,14 +593,7 @@ fn format_verbose_difference(
 
         output.push(b' ');
 
-        let from_byte_str = format_byte(from_byte);
-        let from_byte_padding = 4 - from_byte_str.len();
-
-        output.extend_from_slice(from_byte_str.as_bytes());
-
-        for _ in 0..from_byte_padding {
-            output.push(b' ')
-        }
+        write_visible_byte_padded(output, from_byte);
 
         output.push(b' ');
 
@@ -603,7 +601,7 @@ fn format_verbose_difference(
 
         output.push(b' ');
 
-        output.extend_from_slice(format_byte(to_byte).as_bytes());
+        write_visible_byte(output, to_byte);
 
         output.push(b'\n');
     } else {
@@ -632,7 +630,13 @@ fn format_verbose_difference(
 }
 
 #[inline]
-fn report_eof(at_byte: usize, at_line: usize, start_of_line: bool, eof_on: &str, params: &Params) {
+fn report_eof(
+    at_byte: BytesLimitU64,
+    at_line: u64,
+    start_of_line: bool,
+    eof_on: &str,
+    params: &Params,
+) {
     if params.quiet {
         return;
     }
@@ -684,7 +688,13 @@ fn is_posix_locale() -> bool {
 }
 
 #[inline]
-fn report_difference(from_byte: u8, to_byte: u8, at_byte: usize, at_line: usize, params: &Params) {
+fn report_difference(
+    from_byte: u8,
+    to_byte: u8,
+    at_byte: BytesLimitU64,
+    at_line: u64,
+    params: &Params,
+) {
     if params.quiet {
         return;
     }
@@ -706,9 +716,9 @@ fn report_difference(from_byte: u8, to_byte: u8, at_byte: usize, at_line: usize,
         print!(
             " is {:>3o} {:char_width$} {:>3o} {:char_width$}",
             from_byte,
-            format_byte(from_byte),
+            format_visible_byte(from_byte),
             to_byte,
-            format_byte(to_byte)
+            format_visible_byte(to_byte)
         );
     }
     println!();
@@ -781,7 +791,7 @@ mod tests {
                 from: os("foo"),
                 to: os("bar"),
                 skip_a: Some(1),
-                skip_b: Some(usize::MAX),
+                skip_b: Some(SkipU64::MAX),
                 ..Default::default()
             }),
             parse_params(
@@ -959,7 +969,7 @@ mod tests {
                 executable: os("cmp"),
                 from: os("foo"),
                 to: os("bar"),
-                max_bytes: Some(usize::MAX),
+                max_bytes: Some(BytesLimitU64::MAX),
                 ..Default::default()
             }),
             parse_params(
@@ -976,6 +986,7 @@ mod tests {
         );
 
         // Failure case
+        // TODO This is actually fine in GNU cmp. --bytes does not have a unit parser yet.
         assert_eq!(
             Err("cmp: invalid --bytes value '1K'".to_string()),
             parse_params(
@@ -1021,8 +1032,8 @@ mod tests {
                 executable: os("cmp"),
                 from: os("foo"),
                 to: os("bar"),
-                skip_a: Some(usize::MAX),
-                skip_b: Some(usize::MAX),
+                skip_a: Some(SkipU64::MAX),
+                skip_b: Some(SkipU64::MAX),
                 ..Default::default()
             }),
             parse_params(
@@ -1062,6 +1073,9 @@ mod tests {
                 from: os("foo"),
                 to: os("bar"),
                 skip_a: Some(1_000_000_000),
+                #[cfg(target_pointer_width = "32")]
+                skip_b: Some((2_147_483_647.5 * 2.0) as usize),
+                #[cfg(target_pointer_width = "64")]
                 skip_b: Some(1_152_921_504_606_846_976 * 2),
                 ..Default::default()
             }),
@@ -1093,8 +1107,12 @@ mod tests {
         .enumerate()
         {
             let values = [
-                1_000usize.checked_pow((i + 1) as u32).unwrap_or(usize::MAX),
-                1024usize.checked_pow((i + 1) as u32).unwrap_or(usize::MAX),
+                (1_000 as SkipU64)
+                    .checked_pow((i + 1) as u32)
+                    .unwrap_or(SkipU64::MAX),
+                (1024 as SkipU64)
+                    .checked_pow((i + 1) as u32)
+                    .unwrap_or(SkipU64::MAX),
             ];
             for (j, v) in values.iter().enumerate() {
                 assert_eq!(

src/context_diff.rs

@@ -381,6 +381,9 @@ pub fn diff(expected: &[u8], actual: &[u8], params: &Params) -> Vec<u8> {
 mod tests {
     use super::*;
     use pretty_assertions::assert_eq;
+
+    use crate::utils::testcmds::PATCH_CMD;
+
     #[test]
     fn test_permutations() {
         // test all possible six-line files.
@@ -394,7 +397,6 @@ mod tests {
                             for &f in &[0, 1, 2] {
                                 use std::fs::{self, File};
                                 use std::io::Write;
-                                use std::process::Command;
                                 let mut alef = Vec::new();
                                 let mut bet = Vec::new();
                                 alef.write_all(if a == 0 { b"a\n" } else { b"b\n" })
@@ -429,12 +431,13 @@ mod tests {
                                 }
                                 // This test diff is intentionally reversed.
                                 // We want it to turn the alef into bet.
+                                let patched = &format!("{target}/alef");
                                 let diff = diff(
                                     &alef,
                                     &bet,
                                     &Params {
-                                        from: "a/alef".into(),
-                                        to: (&format!("{target}/alef")).into(),
+                                        from: patched.into(),
+                                        to: patched.into(),
                                         context_count: 2,
                                         ..Default::default()
                                     },
@@ -449,7 +452,8 @@ mod tests {
                                 fb.write_all(&bet[..]).unwrap();
                                 let _ = fa;
                                 let _ = fb;
-                                let output = Command::new("patch")
+                                let output = PATCH_CMD
+                                    .new()
                                     .arg("-p0")
                                     .arg("--context")
                                     .stdin(File::open(format!("{target}/ab.diff")).unwrap())
@@ -481,7 +485,6 @@ mod tests {
                             for &f in &[0, 1, 2] {
                                 use std::fs::{self, File};
                                 use std::io::Write;
-                                use std::process::Command;
                                 let mut alef = Vec::new();
                                 let mut bet = Vec::new();
                                 alef.write_all(if a == 0 { b"\n" } else { b"b\n" }).unwrap();
@@ -510,12 +513,13 @@ mod tests {
                                 }
                                 // This test diff is intentionally reversed.
                                 // We want it to turn the alef into bet.
+                                let patched = &format!("{target}/alef_");
                                 let diff = diff(
                                     &alef,
                                     &bet,
                                     &Params {
-                                        from: "a/alef_".into(),
-                                        to: (&format!("{target}/alef_")).into(),
+                                        from: patched.into(),
+                                        to: patched.into(),
                                         context_count: 2,
                                         ..Default::default()
                                     },
@@ -530,7 +534,8 @@ mod tests {
                                 fb.write_all(&bet[..]).unwrap();
                                 let _ = fa;
                                 let _ = fb;
-                                let output = Command::new("patch")
+                                let output = PATCH_CMD
+                                    .new()
                                     .arg("-p0")
                                     .arg("--context")
                                     .stdin(File::open(format!("{target}/ab_.diff")).unwrap())
@@ -562,7 +567,6 @@ mod tests {
                             for &f in &[0, 1, 2] {
                                 use std::fs::{self, File};
                                 use std::io::Write;
-                                use std::process::Command;
                                 let mut alef = Vec::new();
                                 let mut bet = Vec::new();
                                 alef.write_all(if a == 0 { b"a\n" } else { b"" }).unwrap();
@@ -594,12 +598,13 @@ mod tests {
                                 };
                                 // This test diff is intentionally reversed.
                                 // We want it to turn the alef into bet.
+                                let patched = &format!("{target}/alefx");
                                 let diff = diff(
                                     &alef,
                                     &bet,
                                     &Params {
-                                        from: "a/alefx".into(),
-                                        to: (&format!("{target}/alefx")).into(),
+                                        from: patched.into(),
+                                        to: patched.into(),
                                         context_count: 2,
                                         ..Default::default()
                                     },
@@ -614,7 +619,8 @@ mod tests {
                                 fb.write_all(&bet[..]).unwrap();
                                 let _ = fa;
                                 let _ = fb;
-                                let output = Command::new("patch")
+                                let output = PATCH_CMD
+                                    .new()
                                     .arg("-p0")
                                     .arg("--context")
                                     .stdin(File::open(format!("{target}/abx.diff")).unwrap())
@@ -646,7 +652,6 @@ mod tests {
                             for &f in &[0, 1, 2] {
                                 use std::fs::{self, File};
                                 use std::io::Write;
-                                use std::process::Command;
                                 let mut alef = Vec::new();
                                 let mut bet = Vec::new();
                                 alef.write_all(if a == 0 { b"a\n" } else { b"f\n" })
@@ -681,12 +686,13 @@ mod tests {
                                 }
                                 // This test diff is intentionally reversed.
                                 // We want it to turn the alef into bet.
+                                let alefr_path = &format!("{target}/alefr");
                                 let diff = diff(
                                     &alef,
                                     &bet,
                                     &Params {
-                                        from: "a/alefr".into(),
-                                        to: (&format!("{target}/alefr")).into(),
+                                        from: alefr_path.into(),
+                                        to: alefr_path.into(),
                                         context_count: 2,
                                         ..Default::default()
                                     },
@@ -701,7 +707,8 @@ mod tests {
                                 fb.write_all(&bet[..]).unwrap();
                                 let _ = fa;
                                 let _ = fb;
-                                let output = Command::new("patch")
+                                let output = PATCH_CMD
+                                    .new()
                                     .arg("-p0")
                                     .arg("--context")
                                     .stdin(File::open(format!("{target}/abr.diff")).unwrap())

src/ed_diff.rs

@@ -162,6 +162,9 @@ pub fn diff(expected: &[u8], actual: &[u8], params: &Params) -> Result<Vec<u8>,
 mod tests {
     use super::*;
     use pretty_assertions::assert_eq;
+
+    use crate::utils::testcmds::ED_CMD;
+
     pub fn diff_w(expected: &[u8], actual: &[u8], filename: &str) -> Result<Vec<u8>, DiffError> {
         let mut output = diff(expected, actual, &Params::default())?;
         writeln!(&mut output, "w {filename}").unwrap();
@@ -237,8 +240,8 @@ mod tests {
                                 let _ = fb;
                                 #[cfg(not(windows))] // there's no ed on windows
                                 {
-                                    use std::process::Command;
-                                    let output = Command::new("ed")
+                                    let output = ED_CMD
+                                        .new()
                                         .arg(format!("{target}/alef"))
                                         .stdin(File::open(format!("{target}/ab.ed")).unwrap())
                                         .output()
@@ -311,8 +314,8 @@ mod tests {
                                 let _ = fb;
                                 #[cfg(not(windows))] // there's no ed on windows
                                 {
-                                    use std::process::Command;
-                                    let output = Command::new("ed")
+                                    let output = ED_CMD
+                                        .new()
                                         .arg(format!("{target}/alef_"))
                                         .stdin(File::open(format!("{target}/ab_.ed")).unwrap())
                                         .output()
@@ -391,8 +394,8 @@ mod tests {
                                 let _ = fb;
                                 #[cfg(not(windows))] // there's no ed on windows
                                 {
-                                    use std::process::Command;
-                                    let output = Command::new("ed")
+                                    let output = ED_CMD
+                                        .new()
                                         .arg(format!("{target}/alefr"))
                                         .stdin(File::open(format!("{target}/abr.ed")).unwrap())
                                         .output()

src/main.rs

@@ -58,7 +58,7 @@ fn main() -> ExitCode {
     let exe_path = binary_path(&mut args);
     let exe_name = name(&exe_path);
 
-    let util_name = if exe_name == "diffutils" {
+    let util_name = if exe_name.as_encoded_bytes().ends_with(b"diffutils") {
         // Discard the item we peeked.
         let _ = args.next();
 
@@ -69,13 +69,17 @@ fn main() -> ExitCode {
         OsString::from(exe_name)
     };
 
-    match util_name.to_str() {
-        Some("diff") => diff::main(args),
-        Some("cmp") => cmp::main(args),
-        Some(name) => {
-            eprintln!("{name}: utility not supported");
+    match util_name.as_encoded_bytes() {
+        name if name.ends_with(b"diff") => diff::main(args),
+        name if name.ends_with(b"cmp") => cmp::main(args),
+        name => {
+            use std::io::{stderr, Write as _};
+            let _ = writeln!(
+                stderr(),
+                "{}: utility not supported",
+                String::from_utf8_lossy(name)
+            );
             ExitCode::from(2)
         }
-        None => second_arg_error(exe_name),
     }
 }

src/normal_diff.rs

@@ -215,6 +215,8 @@ mod tests {
     use super::*;
     use pretty_assertions::assert_eq;
 
+    use crate::utils::testcmds::PATCH_CMD;
+
     #[test]
     fn test_basic() {
         let mut a = Vec::new();
@@ -239,7 +241,6 @@ mod tests {
                             for &f in &[0, 1, 2] {
                                 use std::fs::{self, File};
                                 use std::io::Write;
-                                use std::process::Command;
                                 let mut alef = Vec::new();
                                 let mut bet = Vec::new();
                                 alef.write_all(if a == 0 { b"a\n" } else { b"b\n" })
@@ -285,7 +286,8 @@ mod tests {
                                 fb.write_all(&bet[..]).unwrap();
                                 let _ = fa;
                                 let _ = fb;
-                                let output = Command::new("patch")
+                                let output = PATCH_CMD
+                                    .new()
                                     .arg("-p0")
                                     .arg(format!("{target}/alef"))
                                     .stdin(File::open(format!("{target}/ab.diff")).unwrap())
@@ -318,7 +320,6 @@ mod tests {
                                 for &g in &[0, 1, 2] {
                                     use std::fs::{self, File};
                                     use std::io::Write;
-                                    use std::process::Command;
                                     let mut alef = Vec::new();
                                     let mut bet = Vec::new();
                                     alef.write_all(if a == 0 { b"a\n" } else { b"b\n" })
@@ -377,7 +378,8 @@ mod tests {
                                     fb.write_all(&bet[..]).unwrap();
                                     let _ = fa;
                                     let _ = fb;
-                                    let output = Command::new("patch")
+                                    let output = PATCH_CMD
+                                        .new()
                                         .arg("-p0")
                                         .arg("--normal")
                                         .arg(format!("{target}/alefn"))
@@ -411,7 +413,6 @@ mod tests {
                             for &f in &[0, 1, 2] {
                                 use std::fs::{self, File};
                                 use std::io::Write;
-                                use std::process::Command;
                                 let mut alef = Vec::new();
                                 let mut bet = Vec::new();
                                 alef.write_all(if a == 0 { b"\n" } else { b"b\n" }).unwrap();
@@ -451,7 +452,8 @@ mod tests {
                                 fb.write_all(&bet[..]).unwrap();
                                 let _ = fa;
                                 let _ = fb;
-                                let output = Command::new("patch")
+                                let output = PATCH_CMD
+                                    .new()
                                     .arg("-p0")
                                     .arg(format!("{target}/alef_"))
                                     .stdin(File::open(format!("{target}/ab_.diff")).unwrap())
@@ -483,7 +485,6 @@ mod tests {
                             for &f in &[0, 1, 2] {
                                 use std::fs::{self, File};
                                 use std::io::Write;
-                                use std::process::Command;
                                 let mut alef = Vec::new();
                                 let mut bet = Vec::new();
                                 alef.write_all(if a == 0 { b"a\n" } else { b"f\n" })
@@ -529,7 +530,8 @@ mod tests {
                                 fb.write_all(&bet[..]).unwrap();
                                 let _ = fa;
                                 let _ = fb;
-                                let output = Command::new("patch")
+                                let output = PATCH_CMD
+                                    .new()
                                     .arg("-p0")
                                     .arg(format!("{target}/alefr"))
                                     .stdin(File::open(format!("{target}/abr.diff")).unwrap())

src/params.rs

@@ -195,7 +195,7 @@ pub fn parse_params<I: Iterator<Item = OsString>>(mut opts: Peekable<I>) -> Resu
             Err(error) => return Err(error),
         }
         if param.to_string_lossy().starts_with('-') {
-            return Err(format!("Unknown option: {param:?}"));
+            return Err(format!("unrecognized option '{}'", param.to_string_lossy()));
         }
         if from.is_none() {
             from = Some(param);

src/unified_diff.rs

@@ -408,6 +408,8 @@ mod tests {
     use super::*;
     use pretty_assertions::assert_eq;
 
+    use crate::utils::testcmds::PATCH_CMD;
+
     #[test]
     fn test_permutations() {
         let target = "target/unified-diff/";
@@ -421,7 +423,6 @@ mod tests {
                             for &f in &[0, 1, 2] {
                                 use std::fs::{self, File};
                                 use std::io::Write;
-                                use std::process::Command;
                                 let mut alef = Vec::new();
                                 let mut bet = Vec::new();
                                 alef.write_all(if a == 0 { b"a\n" } else { b"b\n" })
@@ -456,12 +457,13 @@ mod tests {
                                 }
                                 // This test diff is intentionally reversed.
                                 // We want it to turn the alef into bet.
+                                let patched = &format!("{target}/alef");
                                 let diff = diff(
                                     &alef,
                                     &bet,
                                     &Params {
-                                        from: "a/alef".into(),
-                                        to: (&format!("{target}/alef")).into(),
+                                        from: patched.into(),
+                                        to: patched.into(),
                                         context_count: 2,
                                         ..Default::default()
                                     },
@@ -492,7 +494,10 @@ mod tests {
                                         .unwrap_or_else(|_| String::from("[Invalid UTF-8]"))
                                 );
 
-                                let output = Command::new("patch")
+                                use crate::utils::testcmds::PATCH_CMD;
+
+                                let output = PATCH_CMD
+                                    .new()
                                     .arg("-p0")
                                     .stdin(File::open(format!("{target}/ab.diff")).unwrap())
                                     .output()
@@ -524,7 +529,6 @@ mod tests {
                                 for &g in &[0, 1, 2] {
                                     use std::fs::{self, File};
                                     use std::io::Write;
-                                    use std::process::Command;
                                     let mut alef = Vec::new();
                                     let mut bet = Vec::new();
                                     alef.write_all(if a == 0 { b"a\n" } else { b"b\n" })
@@ -572,12 +576,13 @@ mod tests {
                                     }
                                     // This test diff is intentionally reversed.
                                     // We want it to turn the alef into bet.
+                                    let patched = &format!("{target}/alefn");
                                     let diff = diff(
                                         &alef,
                                         &bet,
                                         &Params {
-                                            from: "a/alefn".into(),
-                                            to: (&format!("{target}/alefn")).into(),
+                                            from: patched.into(),
+                                            to: patched.into(),
                                             context_count: 2,
                                             ..Default::default()
                                         },
@@ -592,7 +597,8 @@ mod tests {
                                     fb.write_all(&bet[..]).unwrap();
                                     let _ = fa;
                                     let _ = fb;
-                                    let output = Command::new("patch")
+                                    let output = PATCH_CMD
+                                        .new()
                                         .arg("-p0")
                                         .stdin(File::open(format!("{target}/abn.diff")).unwrap())
                                         .output()
@@ -625,7 +631,6 @@ mod tests {
                                 for &g in &[0, 1, 2, 3] {
                                     use std::fs::{self, File};
                                     use std::io::Write;
-                                    use std::process::Command;
                                     let mut alef = Vec::new();
                                     let mut bet = Vec::new();
                                     alef.write_all(if a == 0 { b"\n" } else { b"b\n" }).unwrap();
@@ -668,12 +673,13 @@ mod tests {
                                     }
                                     // This test diff is intentionally reversed.
                                     // We want it to turn the alef into bet.
+                                    let patched = &format!("{target}/alef_");
                                     let diff = diff(
                                         &alef,
                                         &bet,
                                         &Params {
-                                            from: "a/alef_".into(),
-                                            to: (&format!("{target}/alef_")).into(),
+                                            from: patched.into(),
+                                            to: patched.into(),
                                             context_count: 2,
                                             ..Default::default()
                                         },
@@ -688,7 +694,8 @@ mod tests {
                                     fb.write_all(&bet[..]).unwrap();
                                     let _ = fa;
                                     let _ = fb;
-                                    let output = Command::new("patch")
+                                    let output = PATCH_CMD
+                                        .new()
                                         .arg("-p0")
                                         .stdin(File::open(format!("{target}/ab_.diff")).unwrap())
                                         .output()
@@ -720,7 +727,6 @@ mod tests {
                             for &f in &[0, 1, 2] {
                                 use std::fs::{self, File};
                                 use std::io::Write;
-                                use std::process::Command;
                                 let mut alef = Vec::new();
                                 let mut bet = Vec::new();
                                 alef.write_all(if a == 0 { b"a\n" } else { b"" }).unwrap();
@@ -749,12 +755,13 @@ mod tests {
                                 }
                                 // This test diff is intentionally reversed.
                                 // We want it to turn the alef into bet.
+                                let patched = &format!("{target}/alefx");
                                 let diff = diff(
                                     &alef,
                                     &bet,
                                     &Params {
-                                        from: "a/alefx".into(),
-                                        to: (&format!("{target}/alefx")).into(),
+                                        from: patched.into(),
+                                        to: patched.into(),
                                         context_count: 2,
                                         ..Default::default()
                                     },
@@ -769,7 +776,8 @@ mod tests {
                                 fb.write_all(&bet[..]).unwrap();
                                 let _ = fa;
                                 let _ = fb;
-                                let output = Command::new("patch")
+                                let output = PATCH_CMD
+                                    .new()
                                     .arg("-p0")
                                     .stdin(File::open(format!("{target}/abx.diff")).unwrap())
                                     .output()
@@ -800,7 +808,6 @@ mod tests {
                             for &f in &[0, 1, 2] {
                                 use std::fs::{self, File};
                                 use std::io::Write;
-                                use std::process::Command;
                                 let mut alef = Vec::new();
                                 let mut bet = Vec::new();
                                 alef.write_all(if a == 0 { b"a\n" } else { b"f\n" })
@@ -835,12 +842,13 @@ mod tests {
                                 }
                                 // This test diff is intentionally reversed.
                                 // We want it to turn the alef into bet.
+                                let patched = &format!("{target}/alefr");
                                 let diff = diff(
                                     &alef,
                                     &bet,
                                     &Params {
-                                        from: "a/alefr".into(),
-                                        to: (&format!("{target}/alefr")).into(),
+                                        from: patched.into(),
+                                        to: patched.into(),
                                         context_count: 2,
                                         ..Default::default()
                                     },
@@ -855,7 +863,8 @@ mod tests {
                                 fb.write_all(&bet[..]).unwrap();
                                 let _ = fa;
                                 let _ = fb;
-                                let output = Command::new("patch")
+                                let output = PATCH_CMD
+                                    .new()
                                     .arg("-p0")
                                     .stdin(File::open(format!("{target}/abr.diff")).unwrap())
                                     .output()

src/utils.rs

@@ -98,6 +98,99 @@ pub fn report_failure_to_read_input_file(
     );
 }
 
+#[cfg(test)]
+pub mod testcmds {
+    // Command construction wrapper that provides some validation and non-obscure, "fail fast"
+    // feedback and error messages.
+    use std::any::Any;
+    use std::io::Write;
+    use std::panic::catch_unwind;
+    use std::process::{Command, Stdio};
+    use std::sync::LazyLock;
+
+    pub struct CmdFactory {
+        cmd: &'static str,
+        validated_once: LazyLock<Result<(), String>>,
+        validate: fn(&CmdFactory) -> (),
+    }
+
+    impl CmdFactory {
+        pub fn new(&self) -> Command {
+            match &*self.validated_once {
+                Ok(()) => Command::new(self.cmd),
+                Err(errmsg) => panic!(
+                    "'{}' validation failed in earlier thread/test: {}",
+                    self.cmd, errmsg
+                ),
+            }
+        }
+        // "self" is not compatible with static initialization
+        fn try_catch_validate(cmd: &CmdFactory) -> Result<(), String> {
+            // Note catch_unwind() does _not_ hide error messages, stack traces, etc.
+            catch_unwind(|| {
+                let _ = (cmd.validate)(cmd);
+            })
+            .map_err(find_panic_message)
+        }
+    }
+
+    fn find_panic_message(payload: Box<dyn Any + Send>) -> String {
+        // https://github.com/rust-lang/rust/blob/1.95.0/library/std/src/panicking.rs#L771
+        if let Some(&s) = payload.downcast_ref::<&'static str>() {
+            String::from(s)
+        } else if let Some(s) = payload.downcast_ref::<String>() {
+            s.clone()
+        } else {
+            format!(
+                "Unusual panic payload type {:?}, look for the first thread/test that failed",
+                payload.type_id(),
+            )
+        }
+    }
+
+    pub static PATCH_CMD: CmdFactory = CmdFactory {
+        cmd: if cfg!(target_os = "macos") {
+            "gpatch" // brew install gpatch
+        } else {
+            "patch"
+        },
+        validated_once: LazyLock::new(|| CmdFactory::try_catch_validate(&PATCH_CMD)),
+
+        validate: (|myself| {
+            let output = Command::new(myself.cmd)
+                .arg("--version")
+                .output()
+                .expect(format!("`{} --version` failed", myself.cmd).as_str());
+            // Non-GNU versions have subtle differences. When some newlines are missing in some test
+            // patches, the macOS version can even stall the whole test run.
+            assert!(output.stdout.starts_with(b"GNU patch"));
+            assert!(output.status.success());
+        }),
+    };
+
+    pub static ED_CMD: CmdFactory = CmdFactory {
+        cmd: "ed",
+        validated_once: LazyLock::new(|| CmdFactory::try_catch_validate(&ED_CMD)),
+
+        validate: (|myself| {
+            let mut child = Command::new(myself.cmd)
+                .arg("!echo hello_ed")
+                .stdin(Stdio::piped())
+                .stdout(Stdio::piped())
+                .spawn()
+                .expect("Failed to start 'ed' command");
+
+            let mut stdin = child.stdin.take().unwrap();
+            writeln!(stdin, "1p\nq").expect("Failed to send command to 'ed'");
+
+            let output = child
+                .wait_with_output()
+                .expect("Failed to read 'ed' stdout");
+            assert_eq!(String::from_utf8_lossy(&output.stdout), "9\nhello_ed\n");
+        }),
+    };
+}
+
 #[cfg(test)]
 mod tests {
     use super::*;

tests/integration.rs

@@ -32,14 +32,14 @@ mod common {
                 "Expected utility name as second argument, got nothing.\n",
             ));
 
-        for subcmd in ["diff", "cmp"] {
+        for subcmd in ["diff", "cmp", "uu-diff", "uucmp"] {
             let mut cmd = cargo_bin_cmd!("diffutils");
             cmd.arg(subcmd);
             cmd.arg("--foobar");
             cmd.assert()
                 .code(predicate::eq(2))
                 .failure()
-                .stderr(predicate::str::starts_with("Unknown option: \"--foobar\""));
+                .stderr(predicate::str::contains("unrecognized option '--foobar'"));
         }
         Ok(())
     }

tests/run-upstream-testsuite.sh

@@ -62,7 +62,7 @@ cd ../tests
 
 # Fetch tests/init.sh from the gnulib repository (needed since
 # https://git.savannah.gnu.org/cgit/diffutils.git/commit/tests?id=1d2456f539)
-curl -s "$gitserver/gitweb/?p=gnulib.git;a=blob_plain;f=tests/init.sh;hb=HEAD" -o init.sh
+curl -sL "$gitserver/gitweb/?p=gnulib.git;a=blob_plain;f=tests/init.sh;hb=HEAD" -o init.sh
 
 if [[ -n "$TESTS" ]]
 then

util/compare_test_results.py

@@ -0,0 +1,158 @@
+#!/usr/bin/env python3
+
+"""
+Compare the current GNU test results to the last results gathered from the main branch to
+highlight if a PR is making the results better/worse.
+Don't exit with error code if all failing tests are in the ignore-intermittent.txt list.
+"""
+
+import json
+import sys
+import argparse
+from pathlib import Path
+
+
+def load_ignore_list(ignore_file):
+    """Load list of intermittent test names to ignore from file."""
+    ignore_set = set()
+    if ignore_file and Path(ignore_file).exists():
+        with open(ignore_file, "r") as f:
+            for line in f:
+                line = line.strip()
+                if line and not line.startswith("#"):
+                    ignore_set.add(line)
+    return ignore_set
+
+
+def extract_test_results(json_data):
+    """Extract test results from a diffutils test-results.json.
+
+    Note: unlike sed, diffutils JSON has no 'summary' object — results are
+    computed from the 'tests' array using the 'result' and 'test' fields.
+    """
+    tests = json_data.get("tests", [])
+    passed  = sum(1 for t in tests if t.get("result") == "PASS")
+    failed  = sum(1 for t in tests if t.get("result") == "FAIL")
+    skipped = sum(1 for t in tests if t.get("result") == "SKIP")
+    summary = {"total": len(tests), "passed": passed, "failed": failed, "skipped": skipped}
+    failed_tests = [t["test"] for t in tests if t.get("result") == "FAIL"]
+    return summary, failed_tests
+
+
+def compare_results(current_file, reference_file, ignore_file=None, output_file=None):
+    """Compare current results with reference results."""
+    ignore_set = load_ignore_list(ignore_file)
+
+    try:
+        with open(current_file, "r") as f:
+            current_data = json.load(f)
+        current_summary, current_failed = extract_test_results(current_data)
+    except Exception as e:
+        print(f"Error loading current results: {e}")
+        return 1
+
+    try:
+        with open(reference_file, "r") as f:
+            reference_data = json.load(f)
+        reference_summary, reference_failed = extract_test_results(reference_data)
+    except Exception as e:
+        print(f"Error loading reference results: {e}")
+        return 1
+
+    # Calculate differences
+    pass_diff  = int(current_summary.get("passed",  0)) - int(reference_summary.get("passed",  0))
+    fail_diff  = int(current_summary.get("failed",  0)) - int(reference_summary.get("failed",  0))
+    total_diff = int(current_summary.get("total",   0)) - int(reference_summary.get("total",   0))
+
+    # Find new failures and improvements
+    current_failed_set   = set(current_failed)
+    reference_failed_set = set(reference_failed)
+
+    new_failures  = current_failed_set   - reference_failed_set
+    improvements  = reference_failed_set - current_failed_set
+
+    # Filter out intermittent failures
+    non_intermittent_new_failures = new_failures - ignore_set
+
+    # Check if results are identical (no changes)
+    no_changes = (
+        pass_diff == 0
+        and fail_diff == 0
+        and total_diff == 0
+        and not new_failures
+        and not improvements
+    )
+
+    # If no changes, write empty output to prevent comment posting
+    if no_changes:
+        if output_file:
+            with open(output_file, "w") as f:
+                f.write("")
+        return 0
+
+    # Prepare output message
+    output_lines = []
+
+    output_lines.append("Test results comparison:")
+    output_lines.append(
+        f"  Current:   TOTAL: {current_summary.get('total', 0)} / PASSED: {current_summary.get('passed', 0)} / FAILED: {current_summary.get('failed', 0)} / SKIPPED: {current_summary.get('skipped', 0)}"
+    )
+    output_lines.append(
+        f"  Reference: TOTAL: {reference_summary.get('total', 0)} / PASSED: {reference_summary.get('passed', 0)} / FAILED: {reference_summary.get('failed', 0)} / SKIPPED: {reference_summary.get('skipped', 0)}"
+    )
+    output_lines.append("")
+
+    if pass_diff != 0 or fail_diff != 0 or total_diff != 0:
+        output_lines.append("Changes from main branch:")
+        output_lines.append(f"  TOTAL: {total_diff:+d}")
+        output_lines.append(f"  PASSED: {pass_diff:+d}")
+        output_lines.append(f"  FAILED: {fail_diff:+d}")
+        output_lines.append("")
+
+    if new_failures:
+        output_lines.append(f"New test failures ({len(new_failures)}):")
+        for test in sorted(new_failures):
+            if test in ignore_set:
+                output_lines.append(f"  - {test} (intermittent)")
+            else:
+                output_lines.append(f"  - {test}")
+        output_lines.append("")
+
+    if improvements:
+        output_lines.append(f"Test improvements ({len(improvements)}):")
+        for test in sorted(improvements):
+            output_lines.append(f"  + {test}")
+        output_lines.append("")
+
+    output_text = "\n".join(output_lines)
+    if output_file:
+        with open(output_file, "w") as f:
+            f.write(output_text)
+    else:
+        print(output_text)
+
+    if non_intermittent_new_failures:
+        print(
+            f"ERROR: Found {len(non_intermittent_new_failures)} new non-intermittent test failures"
+        )
+        return 1
+
+    return 0
+
+
+def main():
+    parser = argparse.ArgumentParser(description="Compare GNU diffutils test results")
+    parser.add_argument("current",   help="Current test results JSON file")
+    parser.add_argument("reference", help="Reference test results JSON file")
+    parser.add_argument(
+        "--ignore-file", help="File containing intermittent test names to ignore"
+    )
+    parser.add_argument("--output",  help="Output file for comparison results")
+
+    args = parser.parse_args()
+
+    return compare_results(args.current, args.reference, args.ignore_file, args.output)
+
+
+if __name__ == "__main__":
+    sys.exit(main())