Slim down docker cache size (#3190)

* slim down docker cache size

Signed-off-by: Alex Goodman <wagoodman@users.noreply.github.com>

* remove old centos images

Signed-off-by: Alex Goodman <wagoodman@users.noreply.github.com>

* troubleshoot test failure

Signed-off-by: Alex Goodman <wagoodman@users.noreply.github.com>

* fix wget version ref

Signed-off-by: Alex Goodman <wagoodman@users.noreply.github.com>

* refactor caching mechanisms

Signed-off-by: Alex Goodman <wagoodman@users.noreply.github.com>

* add cache cleanup steps

Signed-off-by: Alex Goodman <wagoodman@users.noreply.github.com>

* simplify deleting cache

Signed-off-by: Alex Goodman <wagoodman@users.noreply.github.com>

* fix first clone issue

Signed-off-by: Alex Goodman <wagoodman@users.noreply.github.com>

* add tool dep

Signed-off-by: Alex Goodman <wagoodman@users.noreply.github.com>

---------

Signed-off-by: Alex Goodman <wagoodman@users.noreply.github.com>

.binny.yaml

@@ -115,3 +115,19 @@ tools:
     method: github-release
     with:
       repo: cli/cli
+
+  # used to upload test fixture cache
+  - name: oras
+    version:
+      want: v1.2.0
+    method: github-release
+    with:
+      repo: oras-project/oras
+
+  # used to upload test fixture cache
+  - name: yq
+    version:
+      want: v4.44.3
+    method: github-release
+    with:
+      repo: mikefarah/yq

.github/actions/bootstrap/action.yaml

@@ -13,16 +13,15 @@ inputs:
   cache-key-prefix:
     description: "Prefix all cache keys with this value"
     required: true
-    default: "1ac8281053"
-  compute-fingerprints:
-    description: "Compute test fixture fingerprints"
+    default: "181053ac82"
+  download-test-fixture-cache:
+    description: "Download test fixture cache from OCI and github actions"
     required: true
-    default: "true"
+    default: "false"
   bootstrap-apt-packages:
     description: "Space delimited list of tools to install via apt"
     default: "libxml2-utils"
 
-
 runs:
   using: "composite"
   steps:
@@ -54,8 +53,14 @@ runs:
       run: |
         DEBIAN_FRONTEND=noninteractive sudo apt update && sudo -E apt install -y ${{ inputs.bootstrap-apt-packages }}
 
-    - name: Create all cache fingerprints
-      if: inputs.compute-fingerprints == 'true'
-      shell: bash
-      run: make fingerprints
+    - name: Restore ORAS cache from github actions
+      if: inputs.download-test-fixture-cache == 'true'
+      uses: actions/cache@704facf57e6136b1bc63b828d79edcd491f0ee84 # v3.3.2
+      with:
+        path: ${{ github.workspace }}/.tmp/oras-cache
+        key: ${{ inputs.cache-key-prefix }}-oras-cache
 
+    - name: Download test fixture cache
+      if: inputs.download-test-fixture-cache == 'true'
+      shell: bash
+      run: make download-test-fixture-cache

.github/scripts/ci-check.sh

@@ -1,11 +0,0 @@
-#!/usr/bin/env bash
-
-red=$(tput setaf 1)
-bold=$(tput bold)
-normal=$(tput sgr0)
-
-# assert we are running in CI (or die!)
-if [[ -z "$CI" ]]; then
-    echo "${bold}${red}This step should ONLY be run in CI. Exiting...${normal}"
-    exit 1
-fi

.github/scripts/find_cache_paths.py

@@ -0,0 +1,135 @@
+#!/usr/bin/env python3
+from __future__ import annotations
+
+import os
+import glob
+import sys
+import json
+import hashlib
+
+
+IGNORED_PREFIXES = []
+
+
+def find_fingerprints_and_check_dirs(base_dir):
+    all_fingerprints = set(glob.glob(os.path.join(base_dir, '**', 'test*', '**', '*.fingerprint'), recursive=True))
+
+    all_fingerprints = {os.path.relpath(fp) for fp in all_fingerprints
+                        if not any(fp.startswith(prefix) for prefix in IGNORED_PREFIXES)}
+
+    if not all_fingerprints:
+        show("No .fingerprint files or cache directories found.")
+        exit(1)
+
+    missing_content = []
+    valid_paths = set()
+    fingerprint_contents = []
+
+    for fingerprint in all_fingerprints:
+        path = fingerprint.replace('.fingerprint', '')
+
+        if not os.path.exists(path):
+            missing_content.append(path)
+            continue
+
+        if not os.path.isdir(path):
+            valid_paths.add(path)
+            continue
+
+        if os.listdir(path):
+            valid_paths.add(path)
+        else:
+            missing_content.append(path)
+
+        with open(fingerprint, 'r') as f:
+            content = f.read().strip()
+            fingerprint_contents.append((fingerprint, content))
+
+    return sorted(valid_paths), missing_content, fingerprint_contents
+
+
+def parse_fingerprint_contents(fingerprint_content):
+    input_map = {}
+    for line in fingerprint_content.splitlines():
+        digest, path = line.split()
+        input_map[path] = digest
+    return input_map
+
+
+def calculate_sha256(fingerprint_contents):
+    sorted_fingerprint_contents = sorted(fingerprint_contents, key=lambda x: x[0])
+
+    concatenated_contents = ''.join(content for _, content in sorted_fingerprint_contents)
+
+    sha256_hash = hashlib.sha256(concatenated_contents.encode()).hexdigest()
+
+    return sha256_hash
+
+
+def calculate_file_sha256(file_path):
+    sha256_hash = hashlib.sha256()
+    with open(file_path, 'rb') as f:
+        for byte_block in iter(lambda: f.read(4096), b""):
+            sha256_hash.update(byte_block)
+    return sha256_hash.hexdigest()
+
+
+def show(*s: str):
+    print(*s, file=sys.stderr)
+
+
+def main(file_path: str | None):
+    base_dir = '.'
+    valid_paths, missing_content, fingerprint_contents = find_fingerprints_and_check_dirs(base_dir)
+
+    if missing_content:
+        show("The following paths are missing or have no content, but have corresponding .fingerprint files:")
+        for path in sorted(missing_content):
+            show(f"- {path}")
+        show("Please ensure these paths exist and have content if they are directories.")
+        exit(1)
+
+    sha256_hash = calculate_sha256(fingerprint_contents)
+
+    paths_with_digests = []
+    for path in sorted(valid_paths):
+        fingerprint_file = f"{path}.fingerprint"
+        try:
+            if os.path.exists(fingerprint_file):
+                file_digest = calculate_file_sha256(fingerprint_file)
+
+                # Parse the fingerprint file to get the digest/path tuples
+                with open(fingerprint_file, 'r') as f:
+                    fingerprint_content = f.read().strip()
+                    input_map = parse_fingerprint_contents(fingerprint_content)
+
+                paths_with_digests.append({
+                    "path": path,
+                    "digest": file_digest,
+                    "input": input_map
+                })
+
+        except Exception as e:
+            show(f"Error processing {fingerprint_file}: {e}")
+            raise e
+
+
+    output = {
+        "digest": sha256_hash,
+        "paths": paths_with_digests
+    }
+
+    content = json.dumps(output, indent=2, sort_keys=True)
+
+    if file_path:
+        with open(file_path, 'w') as f:
+            f.write(content)
+
+    print(content)
+
+
+if __name__ == "__main__":
+    file_path = None
+    if len(sys.argv) > 1:
+        file_path = sys.argv[1]
+    main(file_path)

.github/scripts/fingerprint_docker_fixtures.py

@@ -0,0 +1,70 @@
+#!/usr/bin/env python3
+
+import os
+import subprocess
+import hashlib
+
+BOLD = '\033[1m'
+YELLOW = '\033[0;33m'
+RESET = '\033[0m'
+
+
+def print_message(message):
+    print(f"{YELLOW}{message}{RESET}")
+
+
+def sha256sum(filepath):
+    h = hashlib.sha256()
+    with open(filepath, 'rb') as f:
+        for chunk in iter(lambda: f.read(4096), b""):
+            h.update(chunk)
+    return h.hexdigest()
+
+
+def is_git_tracked_or_untracked(directory):
+    """Returns a sorted list of files in the directory that are tracked or not ignored by Git."""
+    result = subprocess.run(
+        ["git", "ls-files", "--cached", "--others", "--exclude-standard"],
+        cwd=directory,
+        stdout=subprocess.PIPE,
+        text=True
+    )
+    return sorted(result.stdout.strip().splitlines())
+
+
+def find_test_fixture_dirs_with_images(base_dir):
+    """Find directories that contain 'test-fixtures' and at least one 'image-*' directory."""
+    for root, dirs, files in os.walk(base_dir):
+        if 'test-fixtures' in root:
+            image_dirs = [d for d in dirs if d.startswith('image-')]
+            if image_dirs:
+                yield os.path.realpath(root)
+
+
+def generate_fingerprints():
+    print_message("creating fingerprint files for docker fixtures...")
+
+    for test_fixture_dir in find_test_fixture_dirs_with_images('.'):
+        cache_fingerprint_path = os.path.join(test_fixture_dir, 'cache.fingerprint')
+
+        with open(cache_fingerprint_path, 'w') as fingerprint_file:
+            for image_dir in find_image_dirs(test_fixture_dir):
+                for file in is_git_tracked_or_untracked(image_dir):
+                    file_path = os.path.join(image_dir, file)
+                    checksum = sha256sum(file_path)
+                    path_from_fixture_dir = os.path.relpath(file_path, test_fixture_dir)
+                    fingerprint_file.write(f"{checksum}  {path_from_fixture_dir}\n")
+
+
+def find_image_dirs(test_fixture_dir):
+    """Find all 'image-*' directories inside a given test-fixture directory."""
+    result = []
+    for root, dirs, files in os.walk(test_fixture_dir):
+        for dir_name in dirs:
+            if dir_name.startswith('image-'):
+                result.append(os.path.join(root, dir_name))
+    return sorted(result)
+
+
+if __name__ == "__main__":
+    generate_fingerprints()

.github/scripts/labeler.py

@@ -1,3 +1,5 @@
+#!/usr/bin/env python3
+
 from __future__ import annotations
 
 import sys

.github/scripts/labeler_test.py

@@ -1,3 +1,5 @@
+#!/usr/bin/env python3
+
 import unittest
 from unittest.mock import patch
 import subprocess

.github/workflows/release-version-file.yaml

@@ -1,4 +1,4 @@
-name: "Release"
+name: "Release: version file"
 
 on:
 

.github/workflows/test-fixture-cache-publish.yaml

@@ -0,0 +1,39 @@
+name: "Test fixture cache: publish"
+
+on:
+  workflow_dispatch:
+  schedule:
+    # run nightly at 4AM UTC
+    - cron: "0 4 * * *"
+
+permissions:
+  contents: read
+
+jobs:
+
+  Publish:
+    name: "Publish test fixture image cache"
+    # we use this runner to get enough storage space for docker images and fixture cache
+    runs-on: ubuntu-22.04-4core-16gb
+    permissions:
+      packages: write
+    steps:
+      - uses: actions/checkout@692973e3d937129bcbf40652eb9f2f61becf3332 #v4.1.7
+
+      - name: Bootstrap environment
+        uses: ./.github/actions/bootstrap
+        with:
+          # we want to rebuild the cache with no previous state
+          download-test-fixture-cache: false
+
+      - name: Run all tests
+        run: make test
+        env:
+          # we want to rebuild the cache with no previous state
+          DOWNLOAD_TEST_FIXTURE_CACHE: "false"
+
+      - name: Login to GitHub Container Registry (ORAS)
+        run: echo "${{ secrets.GITHUB_TOKEN }}" | .tool/oras login ghcr.io -u ${{ github.actor }} --password-stdin
+
+      - name: Publish test fixture cache
+        run: make upload-test-fixture-cache

.github/workflows/update-bootstrap-tools.yml

@@ -19,7 +19,6 @@ jobs:
         uses: ./.github/actions/bootstrap
         with:
           bootstrap-apt-packages: ""
-          compute-fingerprints: "false"
           go-dependencies: false
 
       - name: "Update tool versions"

.github/workflows/validations.yaml

@@ -35,48 +35,8 @@ jobs:
 
       - name: Bootstrap environment
         uses: ./.github/actions/bootstrap
-
-      - name: Restore file executable test-fixture cache
-        uses: actions/cache@0c45773b623bea8c8e75f6c82b208c3cf94ea4f9 #v4.0.2
         with:
-          path: syft/file/cataloger/executable/test-fixtures/elf/bin
-          key: ${{ runner.os }}-unit-file-executable-elf-cache-${{ hashFiles( 'syft/file/cataloger/executable/test-fixtures/elf/cache.fingerprint' ) }}
-
-      - name: Restore file executable shared-info test-fixture cache
-        uses: actions/cache@0c45773b623bea8c8e75f6c82b208c3cf94ea4f9 #v4.0.2
-        with:
-          path: syft/file/cataloger/executable/test-fixtures/shared-info/bin
-          key: ${{ runner.os }}-unit-file-executable-shared-info-cache-${{ hashFiles( 'syft/file/cataloger/executable/test-fixtures/shared-info/cache.fingerprint' ) }}
-
-      - name: Restore Java test-fixture cache
-        uses: actions/cache@0c45773b623bea8c8e75f6c82b208c3cf94ea4f9 #v4.0.2
-        with:
-          path: syft/pkg/cataloger/java/test-fixtures/java-builds/packages
-          key: ${{ runner.os }}-unit-java-cache-${{ hashFiles( 'syft/pkg/cataloger/java/test-fixtures/java-builds/cache.fingerprint' ) }}
-
-      - name: Restore RPM test-fixture cache
-        uses: actions/cache@0c45773b623bea8c8e75f6c82b208c3cf94ea4f9 #v4.0.2
-        with:
-          path: syft/pkg/cataloger/redhat/test-fixtures/rpms
-          key: ${{ runner.os }}-unit-rpm-cache-${{ hashFiles( 'syft/pkg/cataloger/redhat/test-fixtures/rpms.fingerprint' ) }}
-
-      - name: Restore go binary test-fixture cache
-        uses: actions/cache@0c45773b623bea8c8e75f6c82b208c3cf94ea4f9 #v4.0.2
-        with:
-          path: syft/pkg/cataloger/golang/test-fixtures/archs/binaries
-          key: ${{ runner.os }}-unit-go-binaries-cache-${{ hashFiles( 'syft/pkg/cataloger/golang/test-fixtures/archs/binaries.fingerprint' ) }}
-
-      - name: Restore binary cataloger test-fixture cache
-        uses: actions/cache@0c45773b623bea8c8e75f6c82b208c3cf94ea4f9 #v4.0.2
-        with:
-          path: syft/pkg/cataloger/binary/test-fixtures/classifiers/bin
-          key: ${{ runner.os }}-unit-binary-cataloger-cache-${{ hashFiles( 'syft/pkg/cataloger/binary/test-fixtures/cache.fingerprint' ) }}
-
-      - name: Restore Kernel test-fixture cache
-        uses: actions/cache@0c45773b623bea8c8e75f6c82b208c3cf94ea4f9 #v4.0.2
-        with:
-          path: syft/pkg/cataloger/kernel/test-fixtures/cache
-          key: ${{ runner.os }}-unit-kernel-cache-${{ hashFiles( 'syft/pkg/cataloger/kernel/test-fixtures/cache.fingerprint' ) }}
+          download-test-fixture-cache: true
 
       - name: Run unit tests
         run: make unit
@@ -91,16 +51,12 @@ jobs:
 
       - name: Bootstrap environment
         uses: ./.github/actions/bootstrap
+        with:
+          download-test-fixture-cache: true
 
       - name: Validate syft output against the CycloneDX schema
         run: make validate-cyclonedx-schema
 
-      - name: Restore integration test cache
-        uses: actions/cache@0c45773b623bea8c8e75f6c82b208c3cf94ea4f9 #v4.0.2
-        with:
-          path: ${{ github.workspace }}/cmd/syft/internal/test/integration/test-fixtures/cache
-          key: ${{ runner.os }}-integration-test-cache-${{ hashFiles('/cmd/syft/internal/test/integration/test-fixtures/cache.fingerprint') }}
-
       - name: Run integration tests
         run: make integration
 
@@ -143,6 +99,8 @@ jobs:
 
       - name: Bootstrap environment
         uses: ./.github/actions/bootstrap
+        with:
+          download-test-fixture-cache: true
 
       - name: Download snapshot build
         id: snapshot-cache
@@ -162,13 +120,6 @@ jobs:
       - name: Run comparison tests (Linux)
         run: make compare-linux
 
-      - name: Restore install.sh test image cache
-        id: install-test-image-cache
-        uses: actions/cache@0c45773b623bea8c8e75f6c82b208c3cf94ea4f9 #v4.0.2
-        with:
-          path: ${{ github.workspace }}/test/install/cache
-          key: ${{ runner.os }}-install-test-image-cache-${{ hashFiles('test/install/cache.fingerprint') }}
-
       - name: Load test image cache
         if: steps.install-test-image-cache.outputs.cache-hit == 'true'
         run: make install-test-cache-load
@@ -196,8 +147,8 @@ jobs:
         uses: ./.github/actions/bootstrap
         with:
           bootstrap-apt-packages: ""
-          compute-fingerprints: "false"
           go-dependencies: false
+          download-test-fixture-cache: true
 
       - name: Download snapshot build
         id: snapshot-cache
@@ -214,13 +165,6 @@ jobs:
         if: steps.snapshot-cache.outputs.cache-hit != 'true'
         run: echo "unable to download snapshots from previous job" && false
 
-      - name: Restore docker image cache for compare testing
-        id: mac-compare-testing-cache
-        uses: actions/cache@0c45773b623bea8c8e75f6c82b208c3cf94ea4f9 #v4.0.2
-        with:
-          path: image.tar
-          key: ${{ runner.os }}-${{ hashFiles('test/compare/mac.sh') }}
-
       - name: Run comparison tests (Mac)
         run: make compare-mac
 
@@ -238,12 +182,8 @@ jobs:
 
       - name: Bootstrap environment
         uses: ./.github/actions/bootstrap
-
-      - name: Restore CLI test-fixture cache
-        uses: actions/cache@0c45773b623bea8c8e75f6c82b208c3cf94ea4f9 #v4.0.2
         with:
-          path: ${{ github.workspace }}/test/cli/test-fixtures/cache
-          key: ${{ runner.os }}-cli-test-cache-${{ hashFiles('test/cli/test-fixtures/cache.fingerprint') }}
+          download-test-fixture-cache: true
 
       - name: Download snapshot build
         id: snapshot-cache
@@ -262,3 +202,22 @@ jobs:
 
       - name: Run CLI Tests (Linux)
         run: make cli
+
+
+  Cleanup-Cache:
+    name: "Cleanup snapshot cache"
+    if: always()
+    runs-on: ubuntu-20.04
+    permissions:
+      actions: write
+    needs:
+      - Acceptance-Linux
+      - Acceptance-Mac
+      - Cli-Linux
+    steps:
+      - uses: actions/checkout@692973e3d937129bcbf40652eb9f2f61becf3332 #v4.1.7
+
+      - name: Delete snapshot cache
+        run: gh cache delete "snapshot-build-${{ github.run_id }}"
+        env:
+          GH_TOKEN: ${{ github.token }}

Makefile

@@ -25,8 +25,8 @@ ci-bootstrap-go:
 
 # this is a bootstrapping catch-all, where if the target doesn't exist, we'll ensure the tools are installed and then try again
 %:
-	make $(TASK)
-	$(TASK) $@
+	@make --silent $(TASK)
+	@$(TASK) $@
 
 ## Shim targets #################################
 

Taskfile.yaml

@@ -4,9 +4,19 @@ vars:
   OWNER: anchore
   PROJECT: syft
 
+  CACHE_IMAGE: ghcr.io/{{ .OWNER }}/{{ .PROJECT }}/test-fixture-cache:latest
+
   # static file dirs
   TOOL_DIR: .tool
   TMP_DIR: .tmp
+  ORAS_CACHE: "{{ .TMP_DIR }}/oras-cache"
+  CACHE_PATHS_FILE: "{{ .TMP_DIR }}/cache_paths.json"
+  LAST_CACHE_PULL_FILE: "{{ .TMP_DIR }}/last_cache_paths.json"
+
+  # TOOLS
+  ORAS: "{{ .TOOL_DIR }}/oras"
+  YQ: "{{ .TOOL_DIR }}/yq"
+  TASK: "{{ .TOOL_DIR }}/task"
 
   # used for changelog generation
   CHANGELOG: CHANGELOG.md
@@ -33,6 +43,9 @@ vars:
   COMPARE_DIR: ./test/compare
   COMPARE_TEST_IMAGE: centos:8.2.2004
 
+env:
+  GNUMAKEFLAGS: '--no-print-directory'
+
 tasks:
 
   ## High-level tasks #################################
@@ -65,6 +78,7 @@ tasks:
       - task: benchmark
       - task: test-utils
       - task: cli
+      - task: check-docker-cache
 
   ## Bootstrap tasks #################################
 
@@ -212,10 +226,6 @@ tasks:
     # that the cache being restored with the correct binary will be rebuilt since the timestamps
     # and local checksums will not line up.
     deps: [tools, snapshot]
-    sources:
-      - "{{ .SNAPSHOT_BIN }}"
-      - ./test/cli/**
-      - ./**/*.go
     cmds:
       - cmd: "echo 'testing binary: {{ .SNAPSHOT_BIN }}'"
         silent: true
@@ -229,18 +239,14 @@ tasks:
 
   test-utils:
     desc: Run tests for pipeline utils
-    sources:
-      - .github/scripts/labeler*.py
     cmds:
-      - cmd: python .github/scripts/labeler_test.py
+      - cmd: .github/scripts/labeler_test.py
 
 
   ## Benchmark test targets #################################
 
   benchmark:
     deps: [tmpdir]
-    sources:
-      - ./**/*.go
     generates:
       - "{{ .TMP_DIR }}/benchmark-main.txt"
     cmds:
@@ -253,8 +259,6 @@ tasks:
 
   show-benchstat:
     deps: [benchmark, tmpdir]
-    sources:
-      - "{{ .TMP_DIR }}/benchstat.txt"
     cmds:
       - cmd: "cat {{ .TMP_DIR }}/benchstat.txt"
         silent: true
@@ -263,56 +267,188 @@ tasks:
   ## Test-fixture-related targets #################################
 
   fingerprints:
-    desc: Generate test fixture fingerprints
+    desc: Generate fingerprints for all non-docker test fixture
+    silent: true
+    # this will look for `test-fixtures/Makefile` and invoke the `fingerprint` target to calculate all cache input fingerprint files
     generates:
-      - cmd/syft/internal/test/integration/test-fixtures/cache.fingerprint
-      - syft/file/cataloger/executable/test-fixtures/elf/cache.fingerprint
-      - syft/file/cataloger/executable/test-fixtures/shared-info/cache.fingerprint
-      - syft/pkg/cataloger/binary/test-fixtures/cache.fingerprint
-      - syft/pkg/cataloger/java/test-fixtures/java-builds/cache.fingerprint
-      - syft/pkg/cataloger/golang/test-fixtures/archs/binaries.fingerprint
-      - syft/pkg/cataloger/redhat/test-fixtures/rpms.fingerprint
-      - syft/pkg/cataloger/kernel/test-fixtures/cache.fingerprint
+      - '**/test-fixtures/**/*.fingerprint'
       - test/install/cache.fingerprint
-      - test/cli/test-fixtures/cache.fingerprint
     cmds:
-      # for EXECUTABLE unit test fixtures
-      - "cd syft/file/cataloger/executable/test-fixtures/elf && make cache.fingerprint"
-      - "cd syft/file/cataloger/executable/test-fixtures/shared-info && make cache.fingerprint"
-      # for IMAGE integration test fixtures
-      - "cd cmd/syft/internal/test/integration/test-fixtures && make cache.fingerprint"
-      # for BINARY unit test fixtures
-      - "cd syft/pkg/cataloger/binary/test-fixtures && make cache.fingerprint"
-      # for JAVA BUILD unit test fixtures
-      - "cd syft/pkg/cataloger/java/test-fixtures/java-builds && make cache.fingerprint"
-      # for GO BINARY unit test fixtures
-      - "cd syft/pkg/cataloger/golang/test-fixtures/archs && make binaries.fingerprint"
-      # for RPM unit test fixtures
-      - "cd syft/pkg/cataloger/redhat/test-fixtures && make rpms.fingerprint"
-      # for Kernel unit test fixtures
-      - "cd syft/pkg/cataloger/kernel/test-fixtures && make cache.fingerprint"
-      # for INSTALL test fixtures
-      - "cd test/install && make cache.fingerprint"
-      # for CLI test fixtures
-      - "cd test/cli/test-fixtures && make cache.fingerprint"
+      - |
+        BOLD='\033[1m'
+        YELLOW='\033[0;33m'
+        RESET='\033[0m'
 
-  fixtures:
-    desc: Generate test fixtures
+        echo -e "${YELLOW}creating fingerprint files for non-docker fixtures...${RESET}"        
+        for dir in $(find . -type d -name 'test-fixtures'); do
+          if [ -f "$dir/Makefile" ]; then
+            # for debugging...
+            #echo -e "${YELLOW}• calculating fingerprints in $dir... ${RESET}"
+            
+            (make -C "$dir" fingerprint)
+          fi
+        done
+
+        # for debugging...
+        # echo -e "generated all fixture fingerprints"
+
+      - .github/scripts/fingerprint_docker_fixtures.py
+      - |
+        # if DOWNLOAD_TEST_FIXTURE_CACHE is set to 'false', then we don't need to calculate the fingerprint for the cache
+        if [ "$DOWNLOAD_TEST_FIXTURE_CACHE" = "false" ]; then
+          exit 0
+        fi
+        .github/scripts/find_cache_paths.py {{ .CACHE_PATHS_FILE }} > /dev/null
+
+
+  refresh-fixtures:
+    desc: Clear and fetch all test fixture cache
+    aliases:
+      - fixtures
+    silent: true
+    deps:
+      - tools
     cmds:
-      - "cd syft/file/cataloger/executable/test-fixtures/elf && make"
-      - "cd syft/file/cataloger/executable/test-fixtures/shared-info && make"
-      - "cd syft/pkg/cataloger/java/test-fixtures/java-builds && make"
-      - "cd syft/pkg/cataloger/redhat/test-fixtures && make"
-      - "cd syft/pkg/cataloger/binary/test-fixtures && make"
+      - |
+        BOLD='\033[1m'
+        PURPLE='\033[0;35m'
+        RESET='\033[0m'
+        
+        # if DOWNLOAD_TEST_FIXTURE_CACHE is set to 'false', then skip the cache download and always build
+        if [ "$DOWNLOAD_TEST_FIXTURE_CACHE" = "false" ]; then
+            echo -e "${BOLD}${PURPLE}skipping cache download, rebuilding cache...${RESET}"
+            {{ .TASK }} build-fixtures
+            exit 0
+        fi
+        
+        LATEST_FINGERPRINT=$(docker manifest inspect {{ .CACHE_IMAGE }} | {{ .YQ }} -r '.annotations.fingerprint')
+
+        echo "latest cache:      $LATEST_FINGERPRINT"
+        
+        if [ -f {{ .LAST_CACHE_PULL_FILE }} ]; then
+          LAST_PULL_FINGERPRINT=$(cat {{ .LAST_CACHE_PULL_FILE }} | {{ .YQ }} -r '.digest')
+        else
+          echo -e "${BOLD}${PURPLE}empty cache, downloading cache...${RESET}"
+          {{ .TASK }} download-test-fixture-cache
+          exit 0
+        fi
+        
+        {{ .TASK }} fingerprints
+
+        WANT_FINGERPRINT=$(cat {{ .CACHE_PATHS_FILE }} | {{ .YQ }} -r '.digest')        
+
+        echo "desired cache:     $WANT_FINGERPRINT"
+        echo "last pulled cache: $LAST_PULL_FINGERPRINT"
+  
+        # if we already have the latest cache, skip the refresh
+        if [ "$LAST_PULL_FINGERPRINT" = "$WANT_FINGERPRINT" ]; then
+            echo -e "${BOLD}${PURPLE}already have the latest cache (skipping cache download)${RESET}"
+            exit 0
+        fi
+  
+        # at this point we only refresh the cache if we want the same cache that is currently available.
+        # we don't by default refresh the cache if the cache if it is simply different from what we have,
+        # because we may be working on a code change that doesn't require a cache refresh (but could trigger one,
+        # which would be annoying to deal with in a development workflow).
+  
+        if [ "$LATEST_FINGERPRINT" = "$WANT_FINGERPRINT" ]; then
+          echo -e "${BOLD}${PURPLE}found newer cache! downloading cache...${RESET}"
+          {{ .TASK }} download-test-fixture-cache
+        else
+          echo -e "${BOLD}${PURPLE}found different cache, but isn't clear if it's newer (skipping cache download and manually building)${RESET}"
+        
+          {{ .YQ }} eval '.paths[] | "\(.digest) \(.path)"' {{ .LAST_CACHE_PULL_FILE }} > .tmp/last_cache_lines
+          {{ .YQ }} eval '.paths[] | "\(.digest) \(.path)"' {{ .CACHE_PATHS_FILE }} > .tmp/cache_lines
+          diff .tmp/last_cache_lines .tmp/cache_lines || true
+        
+          echo -e "${BOLD}${PURPLE}diff with more context...${RESET}"
+        
+          diff -U10000 {{ .LAST_CACHE_PULL_FILE }} {{ .CACHE_PATHS_FILE }} || true
+        
+          echo -e "${BOLD}${PURPLE}detected changes to input material, manually building fixtures...${RESET}"
+        
+          {{ .TASK }} build-fixtures
+        fi
+
+  build-fixtures:
+    desc: Generate all non-docker test fixtures
+    silent: true
+    # this will look for `test-fixtures/Makefile` and invoke the `fixtures` target to generate any and all test fixtures
+    cmds:
+      - |
+        BOLD='\033[1m'
+        YELLOW='\033[0;33m'
+        RESET='\033[0m'
+        
+        # Use a for loop with command substitution to avoid subshell issues
+        for dir in $(find . -type d -name 'test-fixtures'); do
+          if [ -f "$dir/Makefile" ]; then
+            echo -e "${YELLOW}${BOLD}generating fixtures in $dir${RESET}"
+            (make -C "$dir" fixtures)
+          fi
+        done
+        echo -e "${BOLD}generated all fixtures${RESET}"
+
+  download-test-fixture-cache:
+    desc: Download test fixture cache from ghcr.io
+    deps: [tools, clean-cache]
+    vars:
+      CACHE_DIGEST:
+        sh: docker manifest inspect {{ .CACHE_IMAGE }} | {{ .YQ }} -r '.annotations.fingerprint'
+    cmds:
+      - silent: true
+        cmd: |
+          # if oras cache is > 4 GB, delete it
+          if [ -d {{ .ORAS_CACHE }} ]; then
+            total_size=$(du -c {{ .ORAS_CACHE }} | grep total | awk '{print $1}')
+            if [ "$total_size" -gt 4194304 ]; then
+              echo 'deleting oras cache'
+              rm -rf {{ .ORAS_CACHE }}
+            fi
+          fi
+      - "ORAS_CACHE={{ .ORAS_CACHE }} {{ .ORAS }} pull {{ .CACHE_IMAGE }}"
+      - "cp {{ .CACHE_PATHS_FILE }} {{ .LAST_CACHE_PULL_FILE }}"
+
+  upload-test-fixture-cache:
+    desc: Upload the test fixture cache to ghcr.io
+    deps: [tools, fingerprints]
+    silent: true
+    cmd: |
+      set -eu
+      oras_command="{{ .ORAS }} push {{ .CACHE_IMAGE }}"
+
+      paths=$(cat {{ .CACHE_PATHS_FILE }} | {{ .YQ }} -r '.paths[].path')      
+      for path in $paths; do
+        oras_command+=" $path"
+      done
+      oras_command+=" {{ .CACHE_PATHS_FILE }}"
+      
+      oras_command+=" --annotation org.opencontainers.image.source=https://github.com/{{ .OWNER }}/{{ .PROJECT }}"
+      oras_command+=" --annotation fingerprint=$(cat {{ .CACHE_PATHS_FILE }} | {{ .YQ }} -r '.digest')"
+    
+      echo "Executing: $oras_command"
+      eval $oras_command
 
   show-test-image-cache:
     silent: true
     cmds:
-      - "echo '\nDocker daemon cache:'"
+      - "echo 'Docker daemon cache:'"
       - "docker images --format '{{`{{.ID}}`}} {{`{{.Repository}}`}}:{{`{{.Tag}}`}}' | grep stereoscope-fixture- | sort"
       - "echo '\nTar cache:'"
-      - 'find . -type f -wholename "**/test-fixtures/snapshot/*" | sort'
+      - 'find . -type f -wholename "**/test-fixtures/cache/stereoscope-fixture-*.tar" | sort'
 
+  check-docker-cache:
+    desc: Ensure docker caches aren't using too much disk space
+    silent: true
+    cmd: |
+      total_size=$(find . | grep cache | grep tar | xargs du -c | grep total | awk '{print $1}')
+      find . | grep cache | grep tar | xargs du
+      echo "total $total_size KB"
+      
+      if [ "$total_size" -gt 1048576 ]; then
+        echo 'docker cache is larger than 1GB'
+        exit 1
+      fi
 
   ## install.sh testing targets #################################
 
@@ -457,7 +593,16 @@ tasks:
   ci-check:
     # desc: "[CI only] Are you in CI?"
     cmds:
-      - cmd: .github/scripts/ci-check.sh
+      - cmd: |
+          red=$(tput setaf 1)
+          bold=$(tput bold)
+          normal=$(tput sgr0)
+  
+          # assert we are running in CI (or die!)
+          if [[ -z "$CI" ]]; then
+              echo "${bold}${red}This step should ONLY be run in CI. Exiting...${normal}"
+              exit 1
+          fi
         silent: true
 
   ci-release:
@@ -489,8 +634,31 @@ tasks:
       - "rm -rf {{ .SNAPSHOT_DIR }}"
       - "rm -rf {{ .TMP_DIR }}/goreleaser.yaml"
 
-  clean-cache:
-    desc: Remove all docker cache and local image tar cache
+  clean-docker-cache:
+    desc: Remove all docker cache tars and images from the daemon
     cmds:
-      - 'find . -type f -wholename "**/test-fixtures/cache/stereoscope-fixture-*.tar" -delete'
-      - "docker images --format '{{`{{.ID}}`}} {{`{{.Repository}}`}}' | grep stereoscope-fixture- | awk '{print $$1}' | uniq | xargs -r docker rmi --force"
+      - find . -type d -wholename "**/test-fixtures/cache" | xargs rm -rf
+      - docker images --format '{{`{{.ID}}`}} {{`{{.Repository}}`}}' | grep stereoscope-fixture- | awk '{print $1}' | uniq | xargs -r docker rmi --force
+
+  clean-oras-cache:
+    desc: Remove all cache for oras commands
+    cmd: rm -rf {{ .ORAS_CACHE }}
+
+  clean-cache:
+    desc: Remove all image docker tar cache, images from the docker daemon, and ephemeral test fixtures
+    cmds:
+      - task: clean-docker-cache
+      - |
+        BOLD='\033[1m'
+        YELLOW='\033[0;33m'
+        RESET='\033[0m'
+        
+        # Use a for loop with command substitution to avoid subshell issues
+        for dir in $(find . -type d -name 'test-fixtures'); do
+        if [ -f "$dir/Makefile" ]; then
+          echo -e "${YELLOW}${BOLD}deleting ephemeral test fixtures in $dir${RESET}"
+          (make -C "$dir" clean)
+        fi
+        done
+        echo -e "${BOLD}Deleted all ephemeral test fixtures${RESET}"
+      - rm -f {{ .LAST_CACHE_PULL_FILE }} {{ .CACHE_PATHS_FILE }}

cmd/syft/internal/test/integration/.gitignore

@@ -0,0 +1 @@
+results

cmd/syft/internal/test/integration/all_layers_squashed_comparison_test.go

@@ -7,7 +7,7 @@ import (
 )
 
 func Test_AllLayersIncludesSquashed(t *testing.T) {
-	// This is a verification test for issue #894 (https://github.com/anchore/syft/issues/894)
+	// This is a verification test for issue grype/#894 (https://github.com/anchore/grype/issues/894)
 	allLayers, _ := catalogFixtureImage(t, "image-suse-all-layers", source.AllLayersScope)
 	squashed, _ := catalogFixtureImage(t, "image-suse-all-layers", source.SquashedScope)
 

cmd/syft/internal/test/integration/encode_decode_cycle_test.go

@@ -2,7 +2,9 @@ package integration
 
 import (
 	"bytes"
-	"regexp"
+	"os"
+	"path/filepath"
+	"strings"
 	"testing"
 
 	"github.com/google/go-cmp/cmp"
@@ -12,8 +14,6 @@ import (
 
 	"github.com/anchore/syft/cmd/syft/internal/options"
 	"github.com/anchore/syft/syft/format"
-	"github.com/anchore/syft/syft/format/cyclonedxjson"
-	"github.com/anchore/syft/syft/format/cyclonedxxml"
 	"github.com/anchore/syft/syft/format/syftjson"
 	"github.com/anchore/syft/syft/source"
 )
@@ -43,26 +43,27 @@ func TestEncodeDecodeEncodeCycleComparison(t *testing.T) {
 			},
 			json: true,
 		},
-		{
-			name: cyclonedxjson.ID.String(),
-			redactor: func(in []byte) []byte {
-				// unstable values
-				in = regexp.MustCompile(`"(timestamp|serialNumber|bom-ref|ref)":\s*"(\n|[^"])+"`).ReplaceAll(in, []byte(`"$1": "redacted"`))
-				in = regexp.MustCompile(`"(dependsOn)":\s*\[(?:\s|[^]])+]`).ReplaceAll(in, []byte(`"$1": []`))
-				return in
-			},
-			json: true,
-		},
-		{
-			name: cyclonedxxml.ID.String(),
-			redactor: func(in []byte) []byte {
-				// unstable values
-				in = regexp.MustCompile(`(serialNumber|bom-ref|ref)="[^"]+"`).ReplaceAll(in, []byte{})
-				in = regexp.MustCompile(`<timestamp>[^<]+</timestamp>`).ReplaceAll(in, []byte{})
-
-				return in
-			},
-		},
+		// TODO: ignoring the `ref` field though does create stable results to compare, but the SBOM is fundamentally gutted and not worth comparing (find a better redaction or compare method)
+		//{
+		//	name: cyclonedxjson.ID.String(),
+		//	redactor: func(in []byte) []byte {
+		//		// unstable values
+		//		in = regexp.MustCompile(`"(timestamp|serialNumber|bom-ref|ref)":\s*"(\n|[^"])+"`).ReplaceAll(in, []byte(`"$1": "redacted"`))
+		//		in = regexp.MustCompile(`"(dependsOn)":\s*\[(?:\s|[^]])+]`).ReplaceAll(in, []byte(`"$1": []`))
+		//		return in
+		//	},
+		//	json: true,
+		//},
+		//{
+		//	name: cyclonedxxml.ID.String(),
+		//	redactor: func(in []byte) []byte {
+		//		// unstable values
+		//		in = regexp.MustCompile(`(serialNumber|bom-ref|ref)="[^"]+"`).ReplaceAll(in, []byte{})
+		//		in = regexp.MustCompile(`<timestamp>[^<]+</timestamp>`).ReplaceAll(in, []byte{})
+		//
+		//		return in
+		//	},
+		//},
 	}
 
 	opts := options.DefaultOutput()
@@ -112,6 +113,21 @@ func TestEncodeDecodeEncodeCycleComparison(t *testing.T) {
 					diffs := dmp.DiffMain(string(by1), string(by2), true)
 					t.Errorf("diff: %s", dmp.DiffPrettyText(diffs))
 				}
+
+				// write raw IMAGE@NAME-start and IMAGE@NAME-finish to files within the results dir
+				// ... this is helpful for debugging
+				require.NoError(t, os.MkdirAll("results", 0700))
+
+				suffix := "sbom"
+				switch {
+				case strings.Contains(test.name, "json"):
+					suffix = "json"
+				case strings.Contains(test.name, "xml"):
+					suffix = "xml"
+				}
+
+				require.NoError(t, os.WriteFile(filepath.Join("results", image+"@"+test.name+"-start."+suffix), by1, 0600))
+				require.NoError(t, os.WriteFile(filepath.Join("results", image+"@"+test.name+"-finish."+suffix), by2, 0600))
 			}
 		})
 	}

cmd/syft/internal/test/integration/go_compiler_detection_test.go

@@ -35,8 +35,8 @@ func TestGolangCompilerDetection(t *testing.T) {
 			for _, pkg := range packages {
 				foundCompilerVersions[pkg.Version] = struct{}{}
 				foundPURL[pkg.PURL] = struct{}{}
-				for _, cpe := range pkg.CPEs {
-					foundCPE[cpe] = struct{}{}
+				for _, c := range pkg.CPEs {
+					foundCPE[c] = struct{}{}
 				}
 			}
 

cmd/syft/internal/test/integration/java_purl_test.go

@@ -1,10 +1,9 @@
 package integration
 
 import (
-	"fmt"
 	"testing"
 
-	"github.com/stretchr/testify/assert"
+	"github.com/google/go-cmp/cmp"
 
 	"github.com/anchore/syft/syft/pkg"
 	"github.com/anchore/syft/syft/source"
@@ -26,13 +25,9 @@ func TestJavaPURLs(t *testing.T) {
 			found[metadata.VirtualPath] = p.PURL
 		}
 	}
-	for key, expectedPURL := range expectedPURLs {
-		purl := found[key]
-		assert.Equal(t, expectedPURL, purl, fmt.Sprintf("found wrong or missing PURL for %s want %s, got %s", key, expectedPURL, purl))
-	}
-	for key, foundPURL := range found {
-		expectedPURL := expectedPURLs[key]
-		assert.Equal(t, expectedPURL, foundPURL, fmt.Sprintf("found extra purl for %s want %s, got %s", key, expectedPURL, foundPURL))
+
+	if d := cmp.Diff(expectedPURLs, found); d != "" {
+		t.Errorf("unexpected purl values:\n%s", d)
 	}
 }
 

cmd/syft/internal/test/integration/package_deduplication_test.go

@@ -1,5 +1,3 @@
-//go:build !arm64
-
 package integration
 
 import (
@@ -7,7 +5,6 @@ import (
 	"testing"
 
 	"github.com/stretchr/testify/assert"
-	"github.com/stretchr/testify/require"
 
 	"github.com/anchore/syft/syft/pkg"
 	"github.com/anchore/syft/syft/source"
@@ -22,41 +19,39 @@ func TestPackageDeduplication(t *testing.T) {
 	}{
 		{
 			scope:        source.AllLayersScope,
-			packageCount: 172, // without deduplication this would be 618
+			packageCount: 178, // without deduplication this would be ~600
 			instanceCount: map[string]int{
-				"basesystem": 1,
-				"wget":       1,
-				"curl":       2, // upgraded in the image
-				"vsftpd":     1,
-				"httpd":      1, // rpm, - we exclude binary
+				"basesystem":   1,
+				"wget":         1,
+				"curl-minimal": 2, // upgraded in the image
+				"vsftpd":       1,
+				"httpd":        1, // rpm, - we exclude binary
 			},
 			locationCount: map[string]int{
-				"basesystem-10.0-7.el7.centos": 4,
-				"curl-7.29.0-59.el7":           1, // from base image
-				"curl-7.29.0-59.el7_9.1":       3, // upgrade
-				"wget-1.14-18.el7_6.1":         3,
-				"vsftpd-3.0.2-29.el7_9":        2,
-				"httpd-2.4.6-97.el7.centos.5":  1,
-				// "httpd-2.4.6":                  1, // binary
+				"basesystem-11-13.el9":               5, // in all layers
+				"curl-minimal-7.76.1-26.el9_3.2.0.1": 2, // base + wget layer
+				"curl-minimal-7.76.1-29.el9_4.1":     3, // curl upgrade layer + all above layers
+				"wget-1.21.1-8.el9_4":                4, // wget + all above layers
+				"vsftpd-3.0.5-5.el9":                 2, // vsftpd + all above layers
+				"httpd-2.4.57-11.el9_4.1":            1, // last layer
 			},
 		},
 		{
 			scope:        source.SquashedScope,
-			packageCount: 170,
+			packageCount: 172,
 			instanceCount: map[string]int{
-				"basesystem": 1,
-				"wget":       1,
-				"curl":       1, // upgraded, but the most recent
-				"vsftpd":     1,
-				"httpd":      1, // rpm, binary is now excluded by overlap
+				"basesystem":   1,
+				"wget":         1,
+				"curl-minimal": 1, // upgraded, but the most recent
+				"vsftpd":       1,
+				"httpd":        1, // rpm, binary is now excluded by overlap
 			},
 			locationCount: map[string]int{
-				"basesystem-10.0-7.el7.centos": 1,
-				"curl-7.29.0-59.el7_9.1":       1, // upgrade
-				"wget-1.14-18.el7_6.1":         1,
-				"vsftpd-3.0.2-29.el7_9":        1,
-				"httpd-2.4.6-97.el7.centos.5":  1,
-				// "httpd-2.4.6":                  1, // binary (excluded)
+				"basesystem-11-13.el9":           1,
+				"curl-minimal-7.76.1-29.el9_4.1": 1, // upgrade
+				"wget-1.21.1-8.el9_4":            1,
+				"vsftpd-3.0.5-5.el9":             1,
+				"httpd-2.4.57-11.el9_4.1":        1,
 			},
 		},
 	}
@@ -75,20 +70,21 @@ func TestPackageDeduplication(t *testing.T) {
 				pkgs := sbom.Artifacts.Packages.PackagesByName(name)
 
 				// with multiple packages with the same name, something is wrong (or this is the wrong fixture)
-				require.Len(t, pkgs, expectedInstanceCount)
+				if assert.Len(t, pkgs, expectedInstanceCount, "unexpected package count for %s", name) {
+					for _, p := range pkgs {
+						nameVersion := fmt.Sprintf("%s-%s", name, p.Version)
+						expectedLocationCount, ok := tt.locationCount[nameVersion]
+						if !ok {
+							t.Errorf("missing name-version: %s", nameVersion)
+							continue
+						}
 
-				for _, p := range pkgs {
-					nameVersion := fmt.Sprintf("%s-%s", name, p.Version)
-					expectedLocationCount, ok := tt.locationCount[nameVersion]
-					if !ok {
-						t.Fatalf("missing name-version: %s", nameVersion)
+						// we should see merged locations (assumption, there was 1 location for each package)
+						assert.Len(t, p.Locations.ToSlice(), expectedLocationCount, "unexpected location count for %s", nameVersion)
+
+						// all paths should match
+						assert.Len(t, p.Locations.CoordinateSet().Paths(), 1, "unexpected location count for %s", nameVersion)
 					}
-
-					// we should see merged locations (assumption, there was 1 location for each package)
-					assert.Len(t, p.Locations.ToSlice(), expectedLocationCount)
-
-					// all paths should match
-					assert.Len(t, p.Locations.CoordinateSet().Paths(), 1)
 				}
 			}
 

cmd/syft/internal/test/integration/test-fixtures/Makefile

@@ -1,6 +1,21 @@
-# change these if you want CI to not use previous stored cache
-INTEGRATION_CACHE_BUSTER := "894d8ca"
+FINGERPRINT_FILE := cache.fingerprint
 
-.PHONY: cache.fingerprint
-cache.fingerprint:
-	find image-* -type f -exec md5sum {} + | awk '{print $1}' | sort | tee /dev/stderr | md5sum | tee cache.fingerprint && echo "$(INTEGRATION_CACHE_BUSTER)" >> cache.fingerprint
+.DEFAULT_GOAL := fixtures
+
+# requirement 1: 'fixtures' goal to generate any and all test fixtures
+fixtures:
+	@echo "nothing to do"
+
+# requirement 2: 'fingerprint' goal to determine if the fixture input that indicates any existing cache should be busted
+fingerprint: $(FINGERPRINT_FILE)
+
+# requirement 3: we always need to recalculate the fingerprint based on source regardless of any existing fingerprint
+.PHONY: $(FINGERPRINT_FILE)
+$(FINGERPRINT_FILE):
+	@find image-* -type f -exec sha256sum {} \; | sort -k2 > $(FINGERPRINT_FILE)
+	@#cat $(FINGERPRINT_FILE) | sha256sum | awk '{print $$1}'
+
+# requirement 4: 'clean' goal to remove all generated test fixtures
+.PHONY: clean
+clean:
+	rm -f $(FINGERPRINT_FILE)

cmd/syft/internal/test/integration/test-fixtures/image-golang-compiler/Dockerfile

@@ -1 +1,6 @@
-FROM golang:1.18.10-alpine
+FROM --platform=linux/amd64 golang:1.18.10-alpine
+
+FROM scratch
+
+# we don't need the entire golang toolchain, just a single binary with the stdlib baked in
+COPY --from=0 /usr/local/go/bin/gofmt bin/gofmt

cmd/syft/internal/test/integration/test-fixtures/image-java-no-main-package/Dockerfile

@@ -1,4 +1,4 @@
-FROM jenkins/jenkins:2.346.3-slim-jdk17@sha256:028fbbd9112c60ed086f5197fcba71992317864d27644e5949cf9c52ff4b65f0
+FROM jenkins/jenkins:2.346.3-slim-jdk17@sha256:028fbbd9112c60ed086f5197fcba71992317864d27644e5949cf9c52ff4b65f0 AS base
 
 USER root
 
@@ -12,7 +12,7 @@ RUN apt-get update 2>&1 > /dev/null && apt-get install -y less zip 2>&1 > /dev/n
 
 RUN unzip ../jenkins.war 2>&1 > /dev/null
 
-RUN rm -f ./META-INF/MANIFEST.MF
+RUN rm -rf ./META-INF/MANIFEST.MF ./WEB-INF ./jsbundles ./scripts ./css
 
 WORKDIR /usr/share/jenkins
 
@@ -21,3 +21,7 @@ RUN rm -rf jenkins.war
 RUN cd ./tmp && zip -r ../jenkins.war . && cd ..
 
 RUN rm -rf ./tmp
+
+FROM scratch
+
+COPY --from=base /usr/share/jenkins/jenkins.war /jenkins.war

cmd/syft/internal/test/integration/test-fixtures/image-java-virtualpath-regression/Dockerfile

@@ -1,7 +1,15 @@
-FROM alpine:3.18.3@sha256:7144f7bab3d4c2648d7e59409f15ec52a18006a128c733fcff20d3a4a54ba44a
+FROM alpine:3.18.3@sha256:7144f7bab3d4c2648d7e59409f15ec52a18006a128c733fcff20d3a4a54ba44a AS base
 
 RUN wget https://repo1.maven.org/maven2/org/jvnet/hudson/main/hudson-war/2.2.1/hudson-war-2.2.1.war
 
 RUN mv hudson-war-2.2.1.war hudson.war
 
+# let's make this image a little smaller as to not take up so much disk space
+# we'll only keep the jar metadata files (pom data + manifest) and throw away the rest
+RUN apk add --no-cache python3 py3-pip
+COPY extract.py /extract.py
+RUN python extract.py
 
+FROM scratch
+
+COPY --from=base /slim /

cmd/syft/internal/test/integration/test-fixtures/image-java-virtualpath-regression/extract.py

@@ -0,0 +1,69 @@
+import os
+import zipfile
+import io
+
+ARCHIVE_EXTENSIONS = ('.jar', '.war', '.ear', '.hpi', '.war', '.sar', '.nar', '.par')
+METADATA_FILES = ('pom.xml', 'pom.properties', 'MANIFEST.MF')
+
+
+def slim_archive(archive, output_dir, base_path="", archive_name=""):
+    """
+    extracts metadata files from the archive and creates a slim JAR file
+    containing only these files. handles nested JARs by preserving them.
+    """
+    slim_buffer = io.BytesIO()
+    with zipfile.ZipFile(archive, 'r') as zip_file:
+        with zipfile.ZipFile(slim_buffer, 'w', zipfile.ZIP_DEFLATED) as slim_zip:
+            for file_name in zip_file.namelist():
+                # check for metadata files or nested JARs
+                if file_name.endswith(METADATA_FILES):
+                    # add metadata files directly to the slimmed archive
+                    file_data = zip_file.read(file_name)
+                    slim_zip.writestr(file_name, file_data)
+                elif file_name.endswith(ARCHIVE_EXTENSIONS):
+                    # if it's a nested archive, recursively slim it
+                    nested_archive = io.BytesIO(zip_file.read(file_name))
+                    nested_slim_buffer = io.BytesIO()
+                    slim_archive(
+                        nested_archive,
+                        nested_slim_buffer,
+                        base_path=os.path.join(base_path, os.path.dirname(file_name)),
+                        archive_name=os.path.basename(file_name)
+                    )
+                    # add the slimmed nested archive back to the parent archive
+                    nested_slim_buffer.seek(0)
+                    slim_zip.writestr(file_name, nested_slim_buffer.read())
+
+    # write out the slimmed JAR to the output directory if output_dir is a directory
+    if isinstance(output_dir, str):
+        output_path = os.path.join(output_dir, base_path, archive_name)
+        os.makedirs(os.path.dirname(output_path), exist_ok=True)
+        with open(output_path, 'wb') as f:
+            slim_buffer.seek(0)
+            f.write(slim_buffer.read())
+    else:
+        # if output_dir is a BytesIO buffer (for nested archives), just write to it
+        output_dir.seek(0)
+        output_dir.write(slim_buffer.getvalue())
+
+
+def walk_directory_and_slim_jars(base_dir, output_dir):
+    """
+    recursively walks through a directory tree looking for .jar, .war, .ear,
+    .hpi files and slims them down by keeping only metadata files.
+    """
+    for dirpath, _, filenames in os.walk(base_dir):
+        for filename in filenames:
+            if filename.endswith(ARCHIVE_EXTENSIONS):
+                archive_path = os.path.join(dirpath, filename)
+                print(f"Processing {archive_path}")
+                slim_archive(archive_path, output_dir, os.path.relpath(dirpath, base_dir), filename)
+
+
+# a helper script for slimming down JAR files by keeping only metadata files but still keeping the jar packaging,
+# including nested JARs! Useful for testing purposes.
+if __name__ == "__main__":
+    BASE_DIR = "."
+    OUTPUT_DIR = "./slim"
+    os.makedirs(OUTPUT_DIR, exist_ok=True)
+    walk_directory_and_slim_jars(BASE_DIR, OUTPUT_DIR)

cmd/syft/internal/test/integration/test-fixtures/image-large-apk-data/Dockerfile

@@ -1,4 +1,4 @@
-FROM alpine@sha256:d9a7354e3845ea8466bb00b22224d9116b183e594527fb5b6c3d30bc01a20378
+FROM alpine@sha256:d9a7354e3845ea8466bb00b22224d9116b183e594527fb5b6c3d30bc01a20378 AS base
 
 # we keep these unpinned so that if alpine
 # changes our integration tests can adapt
@@ -6,3 +6,8 @@ RUN apk add --no-cache \
             tzdata \
             vim \
             alpine-sdk
+
+# we don't need the installed bins for this test, only the APK installed metadata
+FROM scratch
+
+COPY --from=base /lib/apk/db/installed /lib/apk/db/installed

cmd/syft/internal/test/integration/test-fixtures/image-mariner-distroless/Dockerfile

@@ -1 +1,8 @@
-FROM mcr.microsoft.com/cbl-mariner/distroless/base:2.0.202205275@sha256:f550c5428df17b145851ad75983aca6d613ad4b51ca7983b2a83e67d0ac91a5d
+FROM mcr.microsoft.com/cbl-mariner/distroless/base:2.0.202205275@sha256:f550c5428df17b145851ad75983aca6d613ad4b51ca7983b2a83e67d0ac91a5d AS base
+
+# let's shoot for smaller test fixtures
+FROM scratch
+
+COPY --from=base /var/lib/rpmmanifest/container-manifest-2 /var/lib/rpmmanifest/container-manifest-2
+COPY --from=base /usr/bin/gencat /usr/bin/gencat
+COPY --from=base /usr/bin/openssl /usr/bin/openssl

cmd/syft/internal/test/integration/test-fixtures/image-owning-package/Dockerfile

@@ -1,3 +1,8 @@
-FROM ubuntu:20.04@sha256:33a5cc25d22c45900796a1aca487ad7a7cb09f09ea00b779e3b2026b4fc2faba
+FROM ubuntu:20.04@sha256:33a5cc25d22c45900796a1aca487ad7a7cb09f09ea00b779e3b2026b4fc2faba AS base
 # this covers rpm-python
 RUN apt-get update && apt-get install -y python-pil=6.2.1-3
+
+# let's save some space...
+FROM scratch
+
+COPY --from=base /var/lib/dpkg/status /var/lib/dpkg/status

cmd/syft/internal/test/integration/test-fixtures/image-photon-all-layers/Dockerfile

@@ -1 +1,5 @@
-FROM photon:5.0-20230729@sha256:4cf2a1ce0a3f4625f13a0becb6b9bccfdb014c565be6e9a2ec4c4aad1ff8a5d9
+FROM photon:5.0-20230729@sha256:4cf2a1ce0a3f4625f13a0becb6b9bccfdb014c565be6e9a2ec4c4aad1ff8a5d9 AS base
+
+FROM scratch
+
+COPY --from=base /usr/lib/sysimage/rpm /usr/lib/sysimage/rpm

cmd/syft/internal/test/integration/test-fixtures/image-sqlite-rpmdb/Dockerfile

@@ -1 +1,6 @@
-FROM fedora:35@sha256:36af84ba69e21c9ef86a0424a090674c433b2b80c2462e57503886f1d823abe8
+FROM fedora:35@sha256:36af84ba69e21c9ef86a0424a090674c433b2b80c2462e57503886f1d823abe8 AS base
+
+# lets save some space
+FROM scratch
+
+COPY --from=base /var/lib/rpm /var/lib/rpm

cmd/syft/internal/test/integration/test-fixtures/image-suse-all-layers/Dockerfile

@@ -1,2 +1,11 @@
-FROM registry.suse.com/suse/sle15:15.3.17.20.20@sha256:fd657ecbab5ca564d6933e887f6ae8542a9398e6a4b399f352ce10c3a24afc64
+FROM registry.suse.com/suse/sle15:15.3.17.20.20@sha256:fd657ecbab5ca564d6933e887f6ae8542a9398e6a4b399f352ce10c3a24afc64 AS base
 RUN zypper in -y wget
+
+# let's save some space... we really just need an image that has an RPM DB that is linked across layers
+FROM --platform=linux/amd64 busybox:1.36.1
+
+# setup a link /var/lib/rpm -> ../../usr/lib/sysimage/rpm
+RUN mkdir -p /var/lib && ln -s ../../usr/lib/sysimage/rpm /var/lib/rpm
+
+# copy the RPM DB from the SUSE image
+COPY --from=base /usr/lib/sysimage/rpm/Packages.db /usr/lib/sysimage/rpm/Packages.db

cmd/syft/internal/test/integration/test-fixtures/image-test-java-purls/Dockerfile

@@ -1,3 +1,18 @@
-FROM docker.io/anchore/test_images:java-88948cc@sha256:dea0e6c24636937f53bdc997d9960c2a18966d1e38bcd8ebd0c395d4e169b806
+FROM docker.io/anchore/test_images:java-88948cc@sha256:dea0e6c24636937f53bdc997d9960c2a18966d1e38bcd8ebd0c395d4e169b806 AS base
 
-RUN rm /packages/gradle-7.1.1-bin.zip
+# not covered in testing...
+RUN rm /packages/gradle-7.1.1-bin.zip
+
+RUN apk add --no-cache python3 py3-pip
+
+COPY extract.py /extract.py
+
+WORKDIR /
+
+# let's make this image a little smaller as to not take up so much disk space
+# we'll only keep the jar metadata files (pom data + manifest) and throw away the rest
+RUN python extract.py
+
+FROM scratch
+
+COPY --from=base /slim/packages /packages

cmd/syft/internal/test/integration/test-fixtures/image-test-java-purls/extract.py

@@ -0,0 +1,69 @@
+import os
+import zipfile
+import io
+
+ARCHIVE_EXTENSIONS = ('.jar', '.war', '.ear', '.hpi', '.war', '.sar', '.nar', '.par')
+METADATA_FILES = ('pom.xml', 'pom.properties', 'MANIFEST.MF')
+
+
+def slim_archive(archive, output_dir, base_path="", archive_name=""):
+    """
+    extracts metadata files from the archive and creates a slim JAR file
+    containing only these files. handles nested JARs by preserving them.
+    """
+    slim_buffer = io.BytesIO()
+    with zipfile.ZipFile(archive, 'r') as zip_file:
+        with zipfile.ZipFile(slim_buffer, 'w', zipfile.ZIP_DEFLATED) as slim_zip:
+            for file_name in zip_file.namelist():
+                # check for metadata files or nested JARs
+                if file_name.endswith(METADATA_FILES):
+                    # add metadata files directly to the slimmed archive
+                    file_data = zip_file.read(file_name)
+                    slim_zip.writestr(file_name, file_data)
+                elif file_name.endswith(ARCHIVE_EXTENSIONS):
+                    # if it's a nested archive, recursively slim it
+                    nested_archive = io.BytesIO(zip_file.read(file_name))
+                    nested_slim_buffer = io.BytesIO()
+                    slim_archive(
+                        nested_archive,
+                        nested_slim_buffer,
+                        base_path=os.path.join(base_path, os.path.dirname(file_name)),
+                        archive_name=os.path.basename(file_name)
+                    )
+                    # add the slimmed nested archive back to the parent archive
+                    nested_slim_buffer.seek(0)
+                    slim_zip.writestr(file_name, nested_slim_buffer.read())
+
+    # write out the slimmed JAR to the output directory if output_dir is a directory
+    if isinstance(output_dir, str):
+        output_path = os.path.join(output_dir, base_path, archive_name)
+        os.makedirs(os.path.dirname(output_path), exist_ok=True)
+        with open(output_path, 'wb') as f:
+            slim_buffer.seek(0)
+            f.write(slim_buffer.read())
+    else:
+        # if output_dir is a BytesIO buffer (for nested archives), just write to it
+        output_dir.seek(0)
+        output_dir.write(slim_buffer.getvalue())
+
+
+def walk_directory_and_slim_jars(base_dir, output_dir):
+    """
+    recursively walks through a directory tree looking for .jar, .war, .ear,
+    .hpi files and slims them down by keeping only metadata files.
+    """
+    for dirpath, _, filenames in os.walk(base_dir):
+        for filename in filenames:
+            if filename.endswith(ARCHIVE_EXTENSIONS):
+                archive_path = os.path.join(dirpath, filename)
+                print(f"Processing {archive_path}")
+                slim_archive(archive_path, output_dir, os.path.relpath(dirpath, base_dir), filename)
+
+
+# a helper script for slimming down JAR files by keeping only metadata files but still keeping the jar packaging,
+# including nested JARs! Useful for testing purposes.
+if __name__ == "__main__":
+    BASE_DIR = "."
+    OUTPUT_DIR = "./slim"
+    os.makedirs(OUTPUT_DIR, exist_ok=True)
+    walk_directory_and_slim_jars(BASE_DIR, OUTPUT_DIR)

cmd/syft/internal/test/integration/test-fixtures/image-vertical-package-dups/Dockerfile

@@ -1,6 +1,27 @@
-FROM centos:7.9.2009@sha256:be65f488b7764ad3638f236b7b515b3678369a5124c47b8d32916d6487418ea4
+FROM --platform=linux/amd64 rockylinux:9.3.20231119@sha256:d644d203142cd5b54ad2a83a203e1dee68af2229f8fe32f52a30c6e1d3c3a9e0 AS base
+
 # modifying the RPM DB multiple times will result in duplicate packages when using all-layers (if there was no de-dup logic)
 # curl is tricky, it already exists in the image and is being upgraded
-RUN yum install -y wget-1.14-18.el7_6.1 curl-7.29.0-59.el7_9.1
-RUN yum install -y vsftpd-3.0.2-29.el7_9
-RUN yum install -y httpd-2.4.6-97.el7.centos.5
+
+# but... we want to make the test image as small as possible, so we are making the changes in stages and then
+# copying the RPM DB from each stage to a final stage in separate layers. This will result in a much smaller image.
+
+FROM base AS stage1
+RUN dnf install -y wget
+
+FROM stage1 AS stage2
+RUN dnf update -y curl-minimal
+
+FROM stage2 AS stage3
+RUN dnf install -y vsftpd
+
+FROM stage3 AS stage4
+RUN dnf install -y httpd
+
+FROM scratch
+
+COPY --from=base /var/lib/rpm /var/lib/rpm
+COPY --from=stage1 /var/lib/rpm /var/lib/rpm
+COPY --from=stage2 /var/lib/rpm /var/lib/rpm
+COPY --from=stage3 /var/lib/rpm /var/lib/rpm
+COPY --from=stage4 /var/lib/rpm /var/lib/rpm

go.mod

@@ -88,6 +88,7 @@ require google.golang.org/genproto v0.0.0-20231106174013-bbf56f31fb17 // indirec
 
 require (
 	github.com/BurntSushi/toml v1.4.0
+	github.com/OneOfOne/xxhash v1.2.8
 	github.com/adrg/xdg v0.5.0
 	github.com/magiconair/properties v1.8.7
 	golang.org/x/exp v0.0.0-20231108232855-2478ac86f678

go.sum

@@ -79,6 +79,8 @@ github.com/Microsoft/go-winio v0.6.1/go.mod h1:LRdKpFKfdobln8UmuiYcKPot9D2v6svN5
 github.com/Microsoft/hcsshim v0.11.4 h1:68vKo2VN8DE9AdN4tnkWnmdhqdbpUFM8OF3Airm7fz8=
 github.com/Microsoft/hcsshim v0.11.4/go.mod h1:smjE4dvqPX9Zldna+t5FG3rnoHhaB7QYxPRqGcpAD9w=
 github.com/OneOfOne/xxhash v1.2.2/go.mod h1:HSdplMjZKSmBqAxg5vPj2TmRDmfkzw+cTzAElWljhcU=
+github.com/OneOfOne/xxhash v1.2.8 h1:31czK/TI9sNkxIKfaUfGlU47BAxQ0ztGgd9vPyqimf8=
+github.com/OneOfOne/xxhash v1.2.8/go.mod h1:eZbhyaAYD41SGSSsnmcpxVoRiQ/MPUTjUdIIOT9Um7Q=
 github.com/ProtonMail/go-crypto v1.0.0 h1:LRuvITjQWX+WIfr930YHG2HNfjR1uOfyf5vE0kC2U78=
 github.com/ProtonMail/go-crypto v1.0.0/go.mod h1:EjAoLdwvbIOoOQr3ihjnSoLZRtE8azugULFRteWMNc0=
 github.com/acarl005/stripansi v0.0.0-20180116102854-5a71ef0e047d h1:licZJFw2RwpHMqeKTCYkitsPqHNxTmd4SNR5r94FGM8=

internal/task/executor.go

@@ -7,9 +7,9 @@ import (
 	"sync"
 	"time"
 
-	"github.com/anchore/syft/internal/log"
 	"github.com/hashicorp/go-multierror"
 
+	"github.com/anchore/syft/internal/log"
 	"github.com/anchore/syft/internal/sbomsync"
 	"github.com/anchore/syft/syft/event/monitor"
 	"github.com/anchore/syft/syft/file"

syft/file/cataloger/executable/test-fixtures/Makefile

@@ -0,0 +1,15 @@
+.DEFAULT_GOAL := default
+
+default:
+	@for dir in $(shell find . -mindepth 1 -maxdepth 1 -type d); do \
+		if [ -f "$$dir/Makefile" ]; then \
+			$(MAKE) -C $$dir; \
+		fi; \
+	done
+
+%:
+	@for dir in $(shell find . -mindepth 1 -maxdepth 1 -type d); do \
+		if [ -f "$$dir/Makefile" ]; then \
+			$(MAKE) -C $$dir $@; \
+		fi; \
+	done

syft/file/cataloger/executable/test-fixtures/elf/Makefile

@@ -1,8 +1,19 @@
 BIN=./bin
 TOOL_IMAGE=localhost/syft-bin-build-tools:latest
 VERIFY_FILE=actual_verify
+FINGERPRINT_FILE=$(BIN).fingerprint
 
-all: build verify
+ifndef BIN
+    $(error BIN is not set)
+endif
+
+.DEFAULT_GOAL := fixtures
+
+# requirement 1: 'fixtures' goal to generate any and all test fixtures
+fixtures: build verify
+
+# requirement 2: 'fingerprint' goal to determine if the fixture input that indicates any existing cache should be busted
+fingerprint: $(FINGERPRINT_FILE)
 
 tools-check:
 	@sha256sum -c Dockerfile.sha256 || (echo "Tools Dockerfile has changed" && exit 1)
@@ -25,10 +36,14 @@ verify: tools
 debug:
 	docker run -i --rm -v $(shell pwd):/mount -w /mount/project $(TOOL_IMAGE) bash
 
-cache.fingerprint:
-	@find project Dockerfile Makefile -type f -exec md5sum {} + | awk '{print $1}' | sort | tee cache.fingerprint
+# requirement 3: we always need to recalculate the fingerprint based on source regardless of any existing fingerprint
+.PHONY: $(FINGERPRINT_FILE)
+$(FINGERPRINT_FILE):
+	@find project Dockerfile Makefile -type f -exec sha256sum {} \; | sort -k2 > $(FINGERPRINT_FILE)
+	@#cat $(FINGERPRINT_FILE) | sha256sum | awk '{print $$1}'
 
+# requirement 4: 'clean' goal to remove all generated test fixtures
 clean:
-	rm -f $(BIN)/*
+	rm -rf $(BIN) Dockerfile.sha256 $(VERIFY_FILE) $(FINGERPRINT_FILE)
 
-.PHONY: build verify debug build-image build-bins clean dockerfile-check cache.fingerprint
+.PHONY: tools tools-check build verify debug clean

syft/file/cataloger/executable/test-fixtures/shared-info/Makefile

@@ -1,8 +1,20 @@
 BIN=./bin
 TOOL_IMAGE=localhost/syft-shared-info-build-tools:latest
 VERIFY_FILE=actual_verify
+FINGERPRINT_FILE=$(BIN).fingerprint
+
+ifndef BIN
+    $(error BIN is not set)
+endif
+
+.DEFAULT_GOAL := fixtures
+
+# requirement 1: 'fixtures' goal to generate any and all test fixtures
+fixtures: build
+
+# requirement 2: 'fingerprint' goal to determine if the fixture input that indicates any existing cache should be busted
+fingerprint: $(FINGERPRINT_FILE)
 
-all: build
 tools-check:
 	@sha256sum -c Dockerfile.sha256 || (echo "Tools Dockerfile has changed" && exit 1)
 
@@ -10,16 +22,20 @@ tools:
 	@(docker inspect $(TOOL_IMAGE) > /dev/null && make tools-check) || (docker build -t $(TOOL_IMAGE) . && sha256sum Dockerfile > Dockerfile.sha256)
 
 build: tools
-	mkdir -p $(BIN)
+	@mkdir -p $(BIN)
 	docker run --platform linux/amd64 -i -v $(shell pwd):/mount -w /mount/project $(TOOL_IMAGE) make
 
 debug:
 	docker run --platform linux/amd64 -i --rm -v $(shell pwd):/mount -w /mount/project $(TOOL_IMAGE) bash
 
-cache.fingerprint:
-	@find project Dockerfile Makefile -type f -exec md5sum {} + | awk '{print $1}' | sort | tee cache.fingerprint
+# requirement 3: we always need to recalculate the fingerprint based on source regardless of any existing fingerprint
+.PHONY: $(FINGERPRINT_FILE)
+$(FINGERPRINT_FILE):
+	@find project Dockerfile Makefile -type f -exec sha256sum {} \; | sort -k2 > $(FINGERPRINT_FILE)
+	@#cat $(FINGERPRINT_FILE) | sha256sum | awk '{print $$1}'
 
+# requirement 4: 'clean' goal to remove all generated test fixtures
 clean:
-	rm -f $(BIN)/*
+	rm -rf $(BIN) Dockerfile.sha256 $(VERIFY_FILE) $(FINGERPRINT_FILE)
 
-.PHONY: build verify debug build-image build-bins clean dockerfile-check cache.fingerprint
+.PHONY: tools tools-check build debug clean

syft/format/text/test-fixtures/image-simple/Dockerfile

@@ -1,4 +0,0 @@
-# Note: changes to this file will result in updating several test values. Consider making a new image fixture instead of editing this one.
-FROM scratch
-ADD file-1.txt /somefile-1.txt
-ADD file-2.txt /somefile-2.txt

syft/format/text/test-fixtures/image-simple/file-1.txt

@@ -1 +0,0 @@
-this file has contents

syft/format/text/test-fixtures/image-simple/file-2.txt

@@ -1 +0,0 @@
-file-2 contents!

syft/format/text/test-fixtures/snapshot/TestTextImageEncoder.golden

@@ -1,11 +1,11 @@
 [Image]
  Layer:		 0
- Digest:	 sha256:fb6beecb75b39f4bb813dbf177e501edd5ddb3e69bb45cedeb78c676ee1b7a59
+ Digest:	 sha256:100d5a55f9032faead28b7427fa3e650e4f0158f86ea89d06e1489df00cb8c6f
  Size:		 22
  MediaType:	 application/vnd.docker.image.rootfs.diff.tar.gzip
 
  Layer:		 1
- Digest:	 sha256:319b588ce64253a87b533c8ed01cf0025e0eac98e7b516e12532957e1244fdec
+ Digest:	 sha256:000fb9200890d3a19138478b20023023c0dce1c54352007c2863716780f049eb
  Size:		 16
  MediaType:	 application/vnd.docker.image.rootfs.diff.tar.gzip
 

syft/pkg/cataloger/binary/test-fixtures/.gitignore

@@ -1,6 +1,5 @@
 classifiers/dynamic
 classifiers/bin
-cache.fingerprint
 
 # allow for lb patterns (rust, pytho, php and more)
 !lib*.so

syft/pkg/cataloger/binary/test-fixtures/Makefile

@@ -1,8 +1,14 @@
-.PHONY: default list download download-all cache.fingerprint
+BIN=classifiers/bin
+FINGERPRINT_FILE=$(BIN).fingerprint
 
-.DEFAULT_GOAL := default
 
-default: download
+.DEFAULT_GOAL := fixtures
+
+# requirement 1: 'fixtures' goal to generate any and all test fixtures
+fixtures: download
+
+# requirement 2: 'fingerprint' goal to determine if the fixture input that indicates any existing cache should be busted
+fingerprint: clean-fingerprint $(FINGERPRINT_FILE)
 
 list: ## list all managed binaries and snippets
 	go run ./manager list
@@ -16,14 +22,23 @@ download-all: ## download all managed binaries
 add-snippet: ## add a new snippet from an existing binary
 	go run ./manager add-snippet
 
-cache.fingerprint: ## prints the sha256sum of the any input to the download command (to determine if there is a cache miss)
-	@cat ./config.yaml | sha256sum | awk '{print $$1}' | tee cache.fingerprint
+# requirement 3: we always need to recalculate the fingerprint based on source regardless of any existing fingerprint
+.PHONY: $(FINGERPRINT_FILE)
+$(FINGERPRINT_FILE): ## prints the sha256sum of the any input to the download command (to determine if there is a cache miss)
+	@sha256sum ./config.yaml > $(FINGERPRINT_FILE)
+
+# requirement 4: 'clean' goal to remove all generated test fixtures
+clean:  ## clean up all downloaded binaries
+	rm -rf $(BIN)
+
+clean-fingerprint: ## clean up all legacy fingerprint files
+	@find $(BIN) -name '*.fingerprint' -delete
 
-clean: ## clean up all downloaded binaries
-	rm -rf ./classifiers/bin
 
 ## Halp! #################################
 
 .PHONY: help
 help:
-	@grep -E '^[a-zA-Z_-]+:.*?## .*$$' $(MAKEFILE_LIST) | sort | awk 'BEGIN {FS = ":.*?## "}; {printf "$(BOLD)$(CYAN)%-25s$(RESET)%s\n", $$1, $$2}'
+	@grep -E '^[a-zA-Z_-]+:.*?## .*$$' $(MAKEFILE_LIST) | sort | awk 'BEGIN {FS = ":.*?## "}; {printf "$(BOLD)$(CYAN)%-25s$(RESET)%s\n", $$1, $$2}'
+
+.PHONY: default list download download-all clean clean-fingerprint add-snippet fingerprint

syft/pkg/cataloger/binary/test-fixtures/config.yaml

@@ -85,6 +85,7 @@ from-images:
     paths:
       - /usr/local/go/bin/go
 
+  # TODO: this is no longer available from dockerhub! (the snippet is vital)
   - version: 1.5.14
     images:
       - ref: haproxy:1.5.14@sha256:3d57e3921cc84e860f764e863ce729dd0765e3d28d444775127bc42d68f98e10

syft/pkg/cataloger/binary/test-fixtures/elf-test-fixtures/Dockerfile

@@ -1,14 +1,31 @@
-FROM rockylinux:8
+FROM rockylinux:8 AS base
+
 RUN dnf update -y; \
 	dnf install make automake gcc gcc-c++ kernel-devel -y; \
 	dnf clean all
 RUN mkdir -p /usr/local/bin/elftests/elfbinwithnestedlib
 RUN mkdir -p /usr/local/bin/elftests/elfbinwithsisterlib
+
 COPY ./elfbinwithnestedlib /usr/local/bin/elftests/elfbinwithnestedlib
 COPY ./elfbinwithsisterlib /usr/local/bin/elftests/elfbinwithsisterlib
+
 ENV LD_LIBRARY_PATH=/usr/local/bin/elftests/elfbinwithnestedlib/bin/lib
+
 WORKDIR /usr/local/bin/elftests/elfbinwithnestedlib/
 RUN make
+
 WORKDIR /usr/local/bin/elftests/elfbinwithsisterlib
 RUN make
 
+# let's make the test image smaller, since we only require the built binaries and supporting libraries
+FROM busybox:1.36.1-musl
+
+COPY --from=base /usr/local/bin/elftests /usr/local/bin/elftests
+COPY --from=base /var/lib/rpm /var/lib/rpm
+COPY --from=base '/usr/lib64/libstdc++.so.6.0.25' '/usr/lib64/libstdc++.so.6.0.25'
+COPY --from=base '/usr/lib64/libstdc++.so.6' '/usr/lib64/libstdc++.so.6'
+COPY --from=base '/usr/lib64/libc.so.6' '/usr/lib64/libc.so.6'
+COPY --from=base '/usr/lib64/libc.so' '/usr/lib64/libc.so'
+
+# prove we can operate over symlinks (/lib64 -> usr/lib64)
+RUN ln -s /usr/lib64 /lib64

syft/pkg/cataloger/binary/test-fixtures/image-fedora-32bit/Dockerfile

@@ -1,4 +1,4 @@
-FROM --platform=linux/arm arm32v7/fedora:36 as build
+FROM --platform=linux/arm arm32v7/fedora:36 AS build
 
 FROM scratch
 COPY --from=build /bin/sha256sum /sha256sum

syft/pkg/cataloger/binary/test-fixtures/image-fedora-64bit/Dockerfile

@@ -1,4 +1,4 @@
-FROM --platform=linux/amd64 fedora:41@sha256:c05bf79137835bf5c521c58f8252d6031780ae865a0379ab57f412e0ac6b42aa as build
+FROM --platform=linux/amd64 fedora:41@sha256:c05bf79137835bf5c521c58f8252d6031780ae865a0379ab57f412e0ac6b42aa AS build
 
 
 FROM scratch

syft/pkg/cataloger/binary/test-fixtures/manager/internal/config/binary_from_image.go

@@ -1,11 +1,11 @@
 package config
 
 import (
-	"crypto/sha256"
 	"fmt"
 	"path/filepath"
 	"strings"
 
+	"github.com/OneOfOne/xxhash"
 	"gopkg.in/yaml.v3"
 )
 
@@ -68,13 +68,13 @@ func PlatformAsValue(platform string) string {
 	return strings.ReplaceAll(platform, "/", "-")
 }
 
-func (c BinaryFromImage) Fingerprint() string {
+func (c BinaryFromImage) Digest() string {
 	by, err := yaml.Marshal(c)
 	if err != nil {
 		panic(err)
 	}
 
-	hasher := sha256.New()
-	hasher.Write(by)
+	hasher := xxhash.New64()
+	_, _ = hasher.Write(by)
 	return fmt.Sprintf("%x", hasher.Sum(nil))
 }

syft/pkg/cataloger/binary/test-fixtures/manager/internal/config/binary_from_image_test.go

@@ -158,7 +158,7 @@ func TestPlatformAsValue(t *testing.T) {
 	}
 }
 
-func TestFingerprint(t *testing.T) {
+func TestDigest(t *testing.T) {
 	tests := []struct {
 		name     string
 		binary   BinaryFromImage
@@ -179,13 +179,13 @@ func TestFingerprint(t *testing.T) {
 					"path/to/test",
 				},
 			},
-			expected: "54ed081c07e4eba031afed4c04315cf96047822196473971be98d0769a0e3645",
+			expected: "fc25c48e3d2f01e3",
 		},
 	}
 
 	for _, tt := range tests {
 		t.Run(tt.name, func(t *testing.T) {
-			assert.Equal(t, tt.expected, tt.binary.Fingerprint())
+			assert.Equal(t, tt.expected, tt.binary.Digest())
 		})
 	}
 }

syft/pkg/cataloger/binary/test-fixtures/manager/internal/download_from_image.go

@@ -2,6 +2,7 @@ package internal
 
 import (
 	"encoding/json"
+	"errors"
 	"fmt"
 	"os"
 	"os/exec"
@@ -14,6 +15,8 @@ import (
 	"github.com/anchore/syft/syft/pkg/cataloger/binary/test-fixtures/manager/internal/ui"
 )
 
+const digestFileSuffix = ".xxh64"
+
 func DownloadFromImage(dest string, config config.BinaryFromImage) error {
 	t := ui.Title{Name: config.Name(), Version: config.Version}
 	t.Start()
@@ -39,22 +42,22 @@ func DownloadFromImage(dest string, config config.BinaryFromImage) error {
 }
 
 func isDownloadStale(config config.BinaryFromImage, binaryPaths []string) bool {
-	currentFingerprint := config.Fingerprint()
+	currentDigest := config.Digest()
 
 	for _, path := range binaryPaths {
-		fingerprintPath := path + ".fingerprint"
-		if _, err := os.Stat(fingerprintPath); err != nil {
+		digestPath := path + digestFileSuffix
+		if _, err := os.Stat(digestPath); err != nil {
 			// missing a fingerprint file means the download is stale
 			return true
 		}
 
-		writtenFingerprint, err := os.ReadFile(fingerprintPath)
+		writtenDigest, err := os.ReadFile(digestPath)
 		if err != nil {
 			// missing a fingerprint file means the download is stale
 			return true
 		}
 
-		if string(writtenFingerprint) != currentFingerprint {
+		if string(writtenDigest) != currentDigest {
 			// the fingerprint file does not match the current fingerprint, so the download is stale
 			return true
 		}
@@ -103,6 +106,12 @@ func pullDockerImage(imageReference, platform string) error {
 	cmd := exec.Command("docker", "pull", "--platform", platform, imageReference)
 	err := cmd.Run()
 	if err != nil {
+		// attach stderr to output message
+		var exitErr *exec.ExitError
+		if errors.As(err, &exitErr) && len(exitErr.Stderr) > 0 {
+			err = fmt.Errorf("pull failed: %w:\n%s", err, exitErr.Stderr)
+		}
+
 		a.Done(err)
 		return err
 	}
@@ -152,6 +161,12 @@ func copyBinariesFromDockerImage(config config.BinaryFromImage, destination stri
 
 	cmd := exec.Command("docker", "create", "--name", containerName, image.Reference)
 	if err = cmd.Run(); err != nil {
+		// attach stderr to output message
+		var exitErr *exec.ExitError
+		if errors.As(err, &exitErr) && len(exitErr.Stderr) > 0 {
+			err = fmt.Errorf("%w:\n%s", err, exitErr.Stderr)
+		}
+
 		return err
 	}
 
@@ -162,7 +177,7 @@ func copyBinariesFromDockerImage(config config.BinaryFromImage, destination stri
 
 	for i, destinationPath := range config.AllStorePathsForImage(image, destination) {
 		path := config.PathsInImage[i]
-		if err := copyBinaryFromContainer(containerName, path, destinationPath, config.Fingerprint()); err != nil {
+		if err := copyBinaryFromContainer(containerName, path, destinationPath, config.Digest()); err != nil {
 			return err
 		}
 	}
@@ -170,7 +185,7 @@ func copyBinariesFromDockerImage(config config.BinaryFromImage, destination stri
 	return nil
 }
 
-func copyBinaryFromContainer(containerName, containerPath, destinationPath, fingerprint string) (err error) {
+func copyBinaryFromContainer(containerName, containerPath, destinationPath, digest string) (err error) {
 	a := ui.Action{Msg: fmt.Sprintf("extract %s", containerPath)}
 	a.Start()
 
@@ -185,13 +200,24 @@ func copyBinaryFromContainer(containerName, containerPath, destinationPath, fing
 	cmd := exec.Command("docker", "cp", fmt.Sprintf("%s:%s", containerName, containerPath), destinationPath) //nolint:gosec
 	// reason for gosec exception: this is for processing test fixtures only, not used in production
 	if err := cmd.Run(); err != nil {
+		// attach stderr to output message
+		var exitErr *exec.ExitError
+		if errors.As(err, &exitErr) && len(exitErr.Stderr) > 0 {
+			err = fmt.Errorf("%w:\n%s", err, exitErr.Stderr)
+		}
+
 		return err
 	}
 
-	// capture fingerprint file
-	fingerprintPath := destinationPath + ".fingerprint"
-	if err := os.WriteFile(fingerprintPath, []byte(fingerprint), 0600); err != nil {
-		return fmt.Errorf("unable to write fingerprint file: %w", err)
+	// ensure permissions are 600 for destination
+	if err := os.Chmod(destinationPath, 0600); err != nil {
+		return fmt.Errorf("unable to set permissions on file %q: %w", destinationPath, err)
+	}
+
+	// capture digest file
+	digestPath := destinationPath + digestFileSuffix
+	if err := os.WriteFile(digestPath, []byte(digest), 0600); err != nil {
+		return fmt.Errorf("unable to write digest file: %w", err)
 	}
 
 	return nil

syft/pkg/cataloger/binary/test-fixtures/manager/internal/download_from_image_test.go

@@ -14,35 +14,35 @@ import (
 func TestIsDownloadStale(t *testing.T) {
 
 	cases := []struct {
-		name        string
-		fingerprint string
-		expected    bool
+		name     string
+		digest   string
+		expected bool
 	}{
 		{
-			name:        "no fingerprint",
-			fingerprint: "",
-			expected:    true,
+			name:     "no digest",
+			digest:   "",
+			expected: true,
 		},
 		{
-			name: "fingerprint matches",
-			// this is the fingerprint for config in the loop body
-			fingerprint: "5177d458eaca031ea16fa707841043df2e31b89be6bae7ea41290aa32f0251a6",
-			expected:    false,
+			name: "digest matches",
+			// this is the digest for config in the loop body
+			digest:   "c9c8007f9c55c2f1",
+			expected: false,
 		},
 		{
-			name:        "fingerprint does not match",
-			fingerprint: "fingerprint",
-			expected:    true,
+			name:     "digest does not match",
+			digest:   "bogus",
+			expected: true,
 		},
 	}
 
 	for _, tt := range cases {
 		t.Run(tt.name, func(t *testing.T) {
 			binaryPath := filepath.Join(t.TempDir(), "binary")
-			fh, err := os.Create(binaryPath + ".fingerprint")
+			fh, err := os.Create(binaryPath + digestFileSuffix)
 			require.NoError(t, err)
 
-			fh.Write([]byte(tt.fingerprint))
+			fh.Write([]byte(tt.digest))
 			require.NoError(t, fh.Close())
 
 			cfg := config.BinaryFromImage{

syft/pkg/cataloger/binary/test-fixtures/manager/internal/list_entries.go

@@ -170,7 +170,7 @@ func getLogicalKey(managedBinaryPath string) (*LogicalEntryKey, error) {
 func allFilePaths(root string) ([]string, error) {
 	var paths []string
 	err := filepath.Walk(root, func(path string, info os.FileInfo, _ error) error {
-		if info != nil && !info.IsDir() && !strings.HasSuffix(path, ".fingerprint") {
+		if info != nil && !info.IsDir() && !strings.HasSuffix(path, digestFileSuffix) {
 			paths = append(paths, path)
 		}
 		return nil

syft/pkg/cataloger/gentoo/cataloger_test.go

@@ -64,7 +64,7 @@ func TestPortageCataloger(t *testing.T) {
 	var expectedRelationships []artifact.Relationship
 
 	pkgtest.NewCatalogTester().
-		FromDirectory(t, "test-fixtures/image-portage").
+		FromDirectory(t, "test-fixtures/layout").
 		Expects(expectedPkgs, expectedRelationships).
 		TestCataloger(t, NewPortageCataloger())
 

syft/pkg/cataloger/golang/test-fixtures/Makefile

@@ -0,0 +1,15 @@
+.DEFAULT_GOAL := default
+
+default:
+	@for dir in $(shell find . -mindepth 1 -maxdepth 1 -type d); do \
+		if [ -f "$$dir/Makefile" ]; then \
+			$(MAKE) -C $$dir; \
+		fi; \
+	done
+
+%:
+	@for dir in $(shell find . -mindepth 1 -maxdepth 1 -type d); do \
+		if [ -f "$$dir/Makefile" ]; then \
+			$(MAKE) -C $$dir $@; \
+		fi; \
+	done

syft/pkg/cataloger/golang/test-fixtures/archs/Makefile

@@ -1,29 +1,39 @@
 DESTINATION=binaries
+FINGERPRINT_FILE=$(DESTINATION).fingerprint
 
-all: $(DESTINATION)/hello-mach-o-arm64  $(DESTINATION)/hello-linux-arm  $(DESTINATION)/hello-linux-ppc64le  $(DESTINATION)/hello-win-amd64
+ifndef DESTINATION
+    $(error DESTINATION is not set)
+endif
+
+.DEFAULT_GOAL := fixtures
+
+# requirement 1: 'fixtures' goal to generate any and all test fixtures
+fixtures: $(DESTINATION)
+
+# requirement 2: 'fingerprint' goal to determine if the fixture input that indicates any existing cache should be busted
+fingerprint: $(DESTINATION).fingerprint
+
+$(DESTINATION): $(DESTINATION)/hello-mach-o-arm64 $(DESTINATION)/hello-linux-arm $(DESTINATION)/hello-linux-ppc64le $(DESTINATION)/hello-win-amd64
 
 $(DESTINATION)/hello-mach-o-arm64:
-	mkdir -p $(DESTINATION)
 	GOARCH=arm64 GOOS=darwin ./src/build.sh $(DESTINATION)/hello-mach-o-arm64
 
 $(DESTINATION)/hello-linux-arm:
-	mkdir -p $(DESTINATION)
 	GOARCH=arm GOOS=linux ./src/build.sh $(DESTINATION)/hello-linux-arm
 
 $(DESTINATION)/hello-linux-ppc64le:
-	mkdir -p $(DESTINATION)
 	GOARCH=ppc64le GOOS=linux ./src/build.sh $(DESTINATION)/hello-linux-ppc64le
 
 $(DESTINATION)/hello-win-amd64:
-	mkdir -p $(DESTINATION)
 	GOARCH=amd64 GOOS=windows ./src/build.sh $(DESTINATION)/hello-win-amd64
 
-# we need a way to determine if CI should bust the test cache based on the source material
-$(DESTINATION).fingerprint: clean
-	mkdir -p $(DESTINATION)
-	find src -type f -exec sha256sum {} \; | sort | tee /dev/stderr | tee $(DESTINATION).fingerprint
-	sha256sum $(DESTINATION).fingerprint
+# requirement 3: we always need to recalculate the fingerprint based on source regardless of any existing fingerprint
+.PHONY: $(FINGERPRINT_FILE)
+$(FINGERPRINT_FILE):
+	@find src -type f -exec sha256sum {} \; | sort -k2 > $(FINGERPRINT_FILE)
+	@#cat $(FINGERPRINT_FILE) | sha256sum | awk '{print $$1}'
 
+# requirement 4: 'clean' goal to remove all generated test fixtures
 .PHONY: clean
 clean:
-	rm -f $(DESTINATION)/*
+	rm -rf $(DESTINATION)

syft/pkg/cataloger/golang/test-fixtures/archs/src/build.sh

@@ -1,10 +1,13 @@
 #!/usr/bin/env bash
-set -uxe
+set -ue
 
 # note: this can be easily done in a 1-liner, however circle CI does NOT allow volume mounts from the host in docker executors (since they are on remote hosts, where the host files are inaccessible)
 # note: gocache override is so we can run docker build not as root in a container without permission issues
 
 BINARY=$1
+
+mkdir -p "$(dirname "$BINARY")"
+
 CTRID=$(docker create -e GOOS="${GOOS}" -e GOARCH="${GOARCH}" -u "$(id -u):$(id -g)" -e GOCACHE=/tmp -w /src golang:1.17 go build -o main main.go)
 
 function cleanup() {

syft/pkg/cataloger/java/test-fixtures/Makefile

@@ -0,0 +1,15 @@
+.DEFAULT_GOAL := default
+
+default:
+	@for dir in $(shell find . -mindepth 1 -maxdepth 1 -type d); do \
+		if [ -f "$$dir/Makefile" ]; then \
+			$(MAKE) -C $$dir; \
+		fi; \
+	done
+
+%:
+	@for dir in $(shell find . -mindepth 1 -maxdepth 1 -type d); do \
+		if [ -f "$$dir/Makefile" ]; then \
+			$(MAKE) -C $$dir $@; \
+		fi; \
+	done

syft/pkg/cataloger/java/test-fixtures/jar-metadata/Makefile

@@ -1,5 +1,10 @@
 CACHE_DIR = cache
 CACHE_PATH = $(shell pwd)/cache
+FINGERPRINT_FILE=$(CACHE_DIR).fingerprint
+
+ifndef CACHE_DIR
+    $(error CACHE_DIR is not set)
+endif
 
 JACKSON_CORE = jackson-core-2.15.2
 SBT_JACKSON_CORE = com.fasterxml.jackson.core.jackson-core-2.15.2
@@ -8,28 +13,53 @@ API_ALL_SOURCES = api-all-2.0.0-sources
 SPRING_INSTRUMENTATION = spring-instrumentation-4.3.0-1.0
 MULTIPLE_MATCHING = multiple-matching-2.11.5
 
-$(CACHE_DIR):
-	mkdir -p $(CACHE_DIR)
 
-$(CACHE_DIR)/$(JACKSON_CORE).jar: $(CACHE_DIR)
+.DEFAULT_GOAL := fixtures
+
+# requirement 1: 'fixtures' goal to generate any and all test fixtures
+fixtures: $(CACHE_DIR)
+
+# requirement 2: 'fingerprint' goal to determine if the fixture input that indicates any existing cache should be busted
+fingerprint: $(FINGERPRINT_FILE)
+
+$(CACHE_DIR): $(CACHE_DIR)/$(JACKSON_CORE).jar $(CACHE_DIR)/$(SBT_JACKSON_CORE).jar $(CACHE_DIR)/$(OPENSAML_CORE).jar  $(CACHE_DIR)/$(API_ALL_SOURCES).jar $(CACHE_DIR)/$(SPRING_INSTRUMENTATION).jar $(CACHE_DIR)/$(MULTIPLE_MATCHING).jar
+
+$(CACHE_DIR)/$(JACKSON_CORE).jar:
+	mkdir -p $(CACHE_DIR)
 	cd $(JACKSON_CORE) && zip -r $(CACHE_PATH)/$(JACKSON_CORE).jar .
 
-$(CACHE_DIR)/$(SBT_JACKSON_CORE).jar: $(CACHE_DIR)
+$(CACHE_DIR)/$(SBT_JACKSON_CORE).jar:
+	mkdir -p $(CACHE_DIR)
 	cd $(SBT_JACKSON_CORE) && zip -r $(CACHE_PATH)/$(SBT_JACKSON_CORE).jar .
 
-$(CACHE_DIR)/$(OPENSAML_CORE).jar: $(CACHE_DIR)
+$(CACHE_DIR)/$(OPENSAML_CORE).jar:
+	mkdir -p $(CACHE_DIR)
 	cd $(OPENSAML_CORE) && zip -r $(CACHE_PATH)/$(OPENSAML_CORE).jar .
 
-$(CACHE_DIR)/$(API_ALL_SOURCES).jar: $(CACHE_DIR)
+$(CACHE_DIR)/$(API_ALL_SOURCES).jar:
+	mkdir -p $(CACHE_DIR)
 	cd $(API_ALL_SOURCES) && zip -r $(CACHE_PATH)/$(API_ALL_SOURCES).jar .
 
-$(CACHE_DIR)/$(SPRING_INSTRUMENTATION).jar: $(CACHE_DIR)
+$(CACHE_DIR)/$(SPRING_INSTRUMENTATION).jar:
+	mkdir -p $(CACHE_DIR)
 	cd $(SPRING_INSTRUMENTATION) && zip -r $(CACHE_PATH)/$(SPRING_INSTRUMENTATION).jar .
 
-$(CACHE_DIR)/$(MULTIPLE_MATCHING).jar: $(CACHE_DIR)
+$(CACHE_DIR)/$(MULTIPLE_MATCHING).jar:
+	mkdir -p $(CACHE_DIR)
 	cd $(MULTIPLE_MATCHING) && zip -r $(CACHE_PATH)/$(MULTIPLE_MATCHING).jar .
 
 # Jenkins plugins typically do not have the version included in the archive name, 
 # so it is important to not include it in the generated test fixture
-$(CACHE_DIR)/gradle.hpi: $(CACHE_DIR)
-	cd jenkins-plugins/gradle/2.11 && zip -r $(CACHE_PATH)/gradle.hpi .
+$(CACHE_DIR)/gradle.hpi:
+	mkdir -p $(CACHE_DIR)
+	cd jenkins-plugins/gradle/2.11 && zip -r $(CACHE_PATH)/gradle.hpi .
+
+# requirement 3: we always need to recalculate the fingerprint based on source regardless of any existing fingerprint
+.PHONY: $(FINGERPRINT_FILE)
+$(FINGERPRINT_FILE):
+	@find . ! -path '*/cache*'  -type f -exec sha256sum {} \; | sort -k2 > $(FINGERPRINT_FILE)
+	@#cat $(FINGERPRINT_FILE) | sha256sum | awk '{print $$1}'
+
+# requirement 4: 'clean' goal to remove all generated test fixtures
+clean:
+	rm -rf $(CACHE_DIR)/* $(FINGERPRINT_FILE)

syft/pkg/cataloger/java/test-fixtures/java-builds/Makefile

@@ -1,17 +1,18 @@
 PKGSDIR=packages
+FINGERPRINT_FILE=$(PKGSDIR).fingerprint
 
 ifndef PKGSDIR
     $(error PKGSDIR is not set)
 endif
 
-all: jars archives native-image
 
-clean: clean-examples
-	rm -f $(PKGSDIR)/*
+.DEFAULT_GOAL := fixtures
 
-clean-examples: clean-gradle clean-maven clean-jenkins clean-nestedjar
+# requirement 1: 'fixtures' goal to generate any and all test fixtures
+fixtures: jars archives native-image
 
-.PHONY: maven gradle clean clean-gradle clean-maven clean-jenkins clean-examples clean-nestedjar jars archives
+# requirement 2: 'fingerprint' goal to determine if the fixture input that indicates any existing cache should be busted
+fingerprint: $(FINGERPRINT_FILE)
 
 jars: $(PKGSDIR)/example-java-app-maven-0.1.0.jar $(PKGSDIR)/example-java-app-gradle-0.1.0.jar $(PKGSDIR)/example-jenkins-plugin.hpi $(PKGSDIR)/spring-boot-0.0.1-SNAPSHOT.jar
 
@@ -71,8 +72,16 @@ $(PKGSDIR)/example-java-app: $(PKGSDIR)/example-java-app-maven-0.1.0.jar
 $(PKGSDIR)/gcc-amd64-darwin-exec-debug:
 	./build-example-macho-binary.sh $(PKGSDIR)
 
-# we need a way to determine if CI should bust the test cache based on the source material
-.PHONY: cache.fingerprint
-cache.fingerprint:
-	find example* build* gradle* Makefile -type f -exec sha256sum {} \; | sort | tee /dev/stderr | tee cache.fingerprint
-	sha256sum cache.fingerprint
+# requirement 3: we always need to recalculate the fingerprint based on source regardless of any existing fingerprint
+.PHONY: $(FINGERPRINT_FILE)
+$(FINGERPRINT_FILE):
+	@find example-* build-* Makefile -type f -exec sha256sum {} \; | sort -k2 > $(FINGERPRINT_FILE)
+	@#cat $(FINGERPRINT_FILE) | sha256sum | awk '{print $$1}'
+
+# requirement 4: 'clean' goal to remove all generated test fixtures
+clean: clean-examples
+	rm -rf $(PKGSDIR) $(FINGERPRINT_FILE)
+
+clean-examples: clean-gradle clean-maven clean-jenkins clean-nestedjar
+
+.PHONY: maven gradle clean clean-gradle clean-maven clean-jenkins clean-examples clean-nestedjar jars archives

syft/pkg/cataloger/kernel/test-fixtures/Makefile

@@ -1,7 +1,21 @@
-all:
+FINGERPRINT_FILE=cache.fingerprint
 
-# we need a way to determine if CI should bust the test cache based on the source material
-.PHONY: cache.fingerprint
-cache.fingerprint:
-	find Makefile **/Dockerfile -type f -exec sha256sum {} \; | sort | tee /dev/stderr | tee cache.fingerprint
-	sha256sum cache.fingerprint
+
+.DEFAULT_GOAL := fixtures
+
+# requirement 1: 'fixtures' goal to generate any and all test fixtures
+fixtures:
+	@echo "nothing to do"
+
+# requirement 2: 'fingerprint' goal to determine if the fixture input that indicates any existing cache should be busted
+fingerprint: $(FINGERPRINT_FILE)
+
+# requirement 3: we always need to recalculate the fingerprint based on source regardless of any existing fingerprint
+.PHONY: $(FINGERPRINT_FILE)
+$(FINGERPRINT_FILE):
+	@find Makefile **/Dockerfile -type f -exec sha256sum {} \; | sort -k2 > $(FINGERPRINT_FILE)
+	@#cat $(FINGERPRINT_FILE) | sha256sum | awk '{print $$1}'
+
+# requirement 4: 'clean' goal to remove all generated test fixtures
+clean:
+	rm -f $(FINGERPRINT_FILE)

syft/pkg/cataloger/python/test-fixtures/image-multi-site-package/Dockerfile

@@ -1,9 +1,8 @@
 # digest is for linux/amd64
-FROM ubuntu:20.04@sha256:cc9cc8169c9517ae035cf293b15f06922cb8c6c864d625a72b7b18667f264b70
+FROM ubuntu:20.04@sha256:cc9cc8169c9517ae035cf293b15f06922cb8c6c864d625a72b7b18667f264b70 AS base
 
 # install Python 3.8 and Python 3.9
-ENV DEBIAN_FRONTEND=noninteractive
-RUN apt-get update && apt-get install -y python3.8 python3.9 python3-pip python3-venv python3.9-venv python3.8-venv
+RUN apt-get update && DEBIAN_FRONTEND=noninteractive apt-get install -y python3.8 python3.9 python3-pip python3-venv python3.9-venv python3.8-venv
 
 # install pip and virtualenv for both Python versions
 RUN python3.8 -m pip install --upgrade pip virtualenv
@@ -35,3 +34,17 @@ RUN /app/project2/venv/bin/pip install click==8.0.3 pyyaml==6.0
 RUN /app/project2/venv/bin/pip install inquirer==3.2.4 runs==1.2.2 xmod==1.8.1 six==1.16.0 wcwidth==0.2.13 blessed==1.20.0 editor==1.6.6 readchar==4.1.0
 
 WORKDIR /app
+
+# let's not waste disk space... we only need the above state we've setup, not all of the os-level packages
+RUN rm -rf /app/project1/venv/share
+RUN rm -rf /app/project2/venv/share
+RUN find /app/project1/venv/lib/python3.9/site-packages/* -type d ! -name '*.dist-info' -exec rm -rf {} +
+RUN find /app/project2/venv/lib/python3.8/site-packages/* -type d ! -name '*.dist-info' -exec rm -rf {} +
+RUN find /usr/local/lib/python3.8/dist-packages/* -type d ! -name '*.dist-info' -exec rm -rf {} +
+RUN find /usr/local/lib/python3.9/dist-packages/* -type d ! -name '*.dist-info' -exec rm -rf {} +
+
+FROM scratch
+
+COPY --from=base /app/ /app/
+COPY --from=base /usr/local/lib/python3.8/ /usr/local/lib/python3.8/
+COPY --from=base /usr/local/lib/python3.9/ /usr/local/lib/python3.9/

syft/pkg/cataloger/redhat/test-fixtures/Makefile

@@ -1,21 +1,38 @@
 RPMSDIR=rpms
+FINGERPRINT_FILE=$(RPMSDIR).fingerprint
 
 ifndef RPMSDIR
     $(error RPMSDIR is not set)
 endif
 
-all: rpms
 
-clean:
-	rm -rf $(RPMSDIR)
+.DEFAULT_GOAL := fixtures
+
+# requirement 1: 'fixtures' goal to generate any and all test fixtures
+fixtures: rpms
+
+# requirement 2: 'fingerprint' goal to determine if the fixture input that indicates any existing cache should be busted
+fingerprint: $(FINGERPRINT_FILE)
 
 rpms:
 	mkdir -p $(RPMSDIR)
-	cd $(RPMSDIR) && curl https://dl.fedoraproject.org/pub/epel/7/x86_64/Packages/a/abc-1.01-9.hg20160905.el7.x86_64.rpm -O
-	cd $(RPMSDIR) && curl https://dl.fedoraproject.org/pub/epel/7/x86_64/Packages/z/zork-1.0.3-1.el7.x86_64.rpm -O
+	@# see note from https://dl.fedoraproject.org/pub/epel/7/README
+	@#  ATTENTION
+	@# ======================================
+	@# The contents of this directory have been moved to our archives available at:
+	@#
+	@# http://archives.fedoraproject.org/pub/archive/epel/
 
-# we need a way to determine if CI should bust the test cache based on the source material
-.PHONY: $(RPMSDIR).fingerprint
-$(RPMSDIR).fingerprint:
-	find Makefile -type f -exec sha256sum {} \; | sort | tee /dev/stderr | tee $(RPMSDIR).fingerprint
-	sha256sum $(RPMSDIR).fingerprint
+	cd $(RPMSDIR) && curl -LO https://archives.fedoraproject.org/pub/archive/epel/7/x86_64/Packages/a/abc-1.01-9.hg20160905.el7.x86_64.rpm
+	cd $(RPMSDIR) && curl -LO https://archives.fedoraproject.org/pub/archive/epel/7/x86_64/Packages/z/zork-1.0.3-1.el7.x86_64.rpm
+
+# requirement 3: we always need to recalculate the fingerprint based on source regardless of any existing fingerprint
+.PHONY: $(FINGERPRINT_FILE)
+$(FINGERPRINT_FILE):
+	@find Makefile -type f -exec sha256sum {} \; | sort -k2 > $(FINGERPRINT_FILE)
+	@#cat $(FINGERPRINT_FILE) | sha256sum | awk '{print $$1}'
+
+# requirement 4: 'clean' goal to remove all generated test fixtures
+.PHONY: clean
+clean:
+	rm -rf $(RPMSDIR) $(FINGERPRINT_FILE)

syft/pkg/cataloger/redhat/test-fixtures/image-minimal/Dockerfile

@@ -2,4 +2,8 @@
 FROM rockylinux:9.3.20231119@sha256:45cc42828cc5ceeffa3a9b4f6363fb582fac3ab91f77bf403daa067f8f049f96
 
 ADD remove.sh /remove.sh
-RUN /remove.sh
+RUN /remove.sh
+
+# let's only keep what we need for testing (not the intermediate layers)
+FROM scratch
+COPY --from=0 / /

test/cli/.gitignore

@@ -0,0 +1 @@
+results

test/cli/cyclonedx_valid_test.go

@@ -33,7 +33,7 @@ func TestValidCycloneDX(t *testing.T) {
 		{
 			name:       "validate cyclonedx output",
 			subcommand: "scan",
-			args:       []string{"-o", "cyclonedx-json"},
+			args:       []string{"-o", "cyclonedx-json", "-o", "cyclonedx-json=results/sbom.cdx.json"},
 			fixture:    imageFixture,
 			assertions: []traitAssertion{
 				assertSuccessfulReturnCode,

test/cli/scan_cmd_test.go

@@ -143,8 +143,22 @@ func TestPackagesCmdFlags(t *testing.T) {
 			name: "squashed-scope-flag-hidden-packages",
 			args: []string{"scan", "-o", "json", "-s", "squashed", hiddenPackagesImage},
 			assertions: []traitAssertion{
-				assertPackageCount(162),
-				assertNotInOutput("vsftpd"), // hidden package
+				assertPackageCount(14),
+				// package 1: alpine-baselayout-data@3.6.5-r0 (apk)
+				// package 2: alpine-baselayout@3.6.5-r0 (apk)
+				// package 3: alpine-keys@2.4-r1 (apk)
+				// package 4: apk-tools@2.14.4-r0 (apk)
+				// package 5: busybox-binsh@1.36.1-r29 (apk)
+				// package 6: busybox@1.36.1-r29 (apk)
+				// package 7: ca-certificates-bundle@20240705-r0 (apk)
+				// package 8: libcrypto3@3.3.1-r3 (apk)
+				// package 9: libssl3@3.3.1-r3 (apk)
+				// package 10: musl-utils@1.2.5-r0 (apk)
+				// package 11: musl@1.2.5-r0 (apk)
+				// package 12: scanelf@1.3.7-r2 (apk)
+				// package 13: ssl_client@1.36.1-r29 (apk)
+				// package 14: zlib@1.3.1-r1 (apk)
+				assertNotInOutput(`"name":"curl"`), // hidden package
 				assertSuccessfulReturnCode,
 			},
 		},
@@ -152,9 +166,33 @@ func TestPackagesCmdFlags(t *testing.T) {
 			name: "all-layers-scope-flag",
 			args: []string{"scan", "-o", "json", "-s", "all-layers", hiddenPackagesImage},
 			assertions: []traitAssertion{
-				assertPackageCount(163), // packages are now deduplicated for this case
+				assertPackageCount(24),
+				// package 1: alpine-baselayout-data@3.6.5-r0 (apk)
+				// package 2: alpine-baselayout@3.6.5-r0 (apk)
+				// package 3: alpine-keys@2.4-r1 (apk)
+				// package 4: apk-tools@2.14.4-r0 (apk)
+				// package 5: brotli-libs@1.1.0-r2 (apk)
+				// package 6: busybox-binsh@1.36.1-r29 (apk)
+				// package 7: busybox@1.36.1-r29 (apk)
+				// package 8: c-ares@1.28.1-r0 (apk)
+				// package 9: ca-certificates-bundle@20240705-r0 (apk)
+				// package 10: ca-certificates@20240705-r0 (apk)
+				// package 11: curl@8.9.1-r1 (apk)
+				// package 12: libcrypto3@3.3.1-r3 (apk)
+				// package 13: libcurl@8.9.1-r1 (apk)
+				// package 14: libidn2@2.3.7-r0 (apk)
+				// package 15: libpsl@0.21.5-r1 (apk)
+				// package 16: libssl3@3.3.1-r3 (apk)
+				// package 17: libunistring@1.2-r0 (apk)
+				// package 18: musl-utils@1.2.5-r0 (apk)
+				// package 19: musl@1.2.5-r0 (apk)
+				// package 20: nghttp2-libs@1.62.1-r0 (apk)
+				// package 21: scanelf@1.3.7-r2 (apk)
+				// package 22: ssl_client@1.36.1-r29 (apk)
+				// package 23: zlib@1.3.1-r1 (apk)
+				// package 24: zstd-libs@1.5.6-r0 (apk)
 				assertInOutput("all-layers"),
-				assertInOutput("vsftpd"), // hidden package
+				assertInOutput(`"name":"curl"`), // hidden package
 				assertSuccessfulReturnCode,
 			},
 		},
@@ -165,9 +203,9 @@ func TestPackagesCmdFlags(t *testing.T) {
 				"SYFT_SCOPE": "all-layers",
 			},
 			assertions: []traitAssertion{
-				assertPackageCount(163), // packages are now deduplicated for this case
+				assertPackageCount(24), // packages are now deduplicated for this case
 				assertInOutput("all-layers"),
-				assertInOutput("vsftpd"), // hidden package
+				assertInOutput(`"name":"curl"`), // hidden package
 				assertSuccessfulReturnCode,
 			},
 		},

test/cli/test-fixtures/Makefile

@@ -1,6 +1,22 @@
-# change these if you want CI to not use previous stored cache
-CLI_CACHE_BUSTER := "e5cdfd8"
+FINGERPRINT_FILE=cache.fingerprint
+
+.DEFAULT_GOAL := fixtures
+
+# requirement 1: 'fixtures' goal to generate any and all test fixtures
+fixtures:
+	@echo "nothing to do"
+
+# requirement 2: 'fingerprint' goal to determine if the fixture input that indicates any existing cache should be busted
+fingerprint: $(FINGERPRINT_FILE)
+
+# requirement 3: we always need to recalculate the fingerprint based on source regardless of any existing fingerprint
+.PHONY: $(FINGERPRINT_FILE)
+$(FINGERPRINT_FILE):
+	@find image-* -type f -exec sha256sum {} \; | sort -k2 > $(FINGERPRINT_FILE)
+	@#cat $(FINGERPRINT_FILE) | sha256sum | awk '{print $$1}'
+
+# requirement 4: 'clean' goal to remove all generated test fixtures
+.PHONY: clean
+clean:
+	rm -f $(FINGERPRINT_FILE)
 
-.PHONY: cache.fingerprint
-cache.fingerprint:
-	find image-* -type f -exec md5sum {} + | awk '{print $1}' | sort | md5sum | tee cache.fingerprint && echo "$(CLI_CACHE_BUSTER)" >> cache.fingerprint

test/cli/test-fixtures/image-hidden-packages/Dockerfile

@@ -1,4 +1,4 @@
-FROM centos:7.9.2009@sha256:dead07b4d8ed7e29e98de0f4504d87e8880d4347859d839686a31da35a3b532f
-# all-layers scope should pickup on vsftpd
-RUN yum install -y vsftpd
-RUN yum remove -y vsftpd
+FROM --platform=linux/amd64 alpine:3.20.2@sha256:eddacbc7e24bf8799a4ed3cdcfa50d4b88a323695ad80f317b6629883b2c2a78
+
+RUN apk add --no-cache curl
+RUN apk del curl

test/install/Makefile

@@ -1,5 +1,7 @@
 NAME=syft
 
+FINGERPRINT_FILE := cache.fingerprint
+
 # for local testing (not testing within containers) use the binny-managed version of cosign.
 # this also means that the user does not need to install cosign on their system to run tests.
 COSIGN_BINARY=../../.tool/cosign
@@ -21,8 +23,6 @@ ACCEPTANCE_CMD=sh -c '../../install.sh -v -b /usr/local/bin && syft version && r
 PREVIOUS_RELEASE=v0.33.0
 ACCEPTANCE_PREVIOUS_RELEASE_CMD=sh -c "../../install.sh -b /usr/local/bin $(PREVIOUS_RELEASE) && syft version"
 
-# CI cache busting values; change these if you want CI to not use previous stored cache
-INSTALL_TEST_CACHE_BUSTER=894d8ca
 
 define title
     @printf '\n≡≡≡[ $(1) ]≡≡≡≡≡≡≡≡≡≡≡≡≡≡≡≡≡≡≡≡≡≡≡≡≡≡≡≡≡≡≡≡≡≡≡≡≡≡≡≡≡≡≡≡≡≡≡≡≡≡≡≡≡≡\n'
@@ -130,7 +130,8 @@ busybox-1.36:
 
 ## For CI ########################################################
 
-.PHONY: cache.fingerprint
-cache.fingerprint:
-	$(call title,Install test fixture fingerprint)
-	@find ./environments/* -type f -exec md5sum {} + | awk '{print $1}' | sort | tee /dev/stderr | md5sum | tee cache.fingerprint && echo "$(INSTALL_TEST_CACHE_BUSTER)" >> cache.fingerprint
+# requirement 3: we always need to recalculate the fingerprint based on source regardless of any existing fingerprint
+.PHONY: $(FINGERPRINT_FILE)
+$(FINGERPRINT_FILE):
+	@find ./environments/* -type f -exec sha256sum {} \; | sort -k2 > $(FINGERPRINT_FILE)
+	@#cat $(FINGERPRINT_FILE) | sha256sum | awk '{print $$1}'