gh-112844: Add SBOM for external dependencies (#115789)

3 files changed, 266 insertions, 17 deletions

diff --git a/.github/CODEOWNERS b/.github/CODEOWNERS
index 1d0bce1..e8eed40 100644
--- a/.github/CODEOWNERS
+++ b/.github/CODEOWNERS
@@ -247,6 +247,7 @@ Lib/test/test_interpreters/   @ericsnowcurrently
 /Tools/wasm/                  @brettcannon
 
 # SBOM
+/Misc/externals.spdx.json     @sethmlarson
 /Misc/sbom.spdx.json          @sethmlarson
 /Tools/build/generate_sbom.py @sethmlarson
 
diff --git a/Misc/externals.spdx.json b/Misc/externals.spdx.json
new file mode 100644
index 0000000..2acfccb
--- /dev/null
+++ b/Misc/externals.spdx.json
@@ -0,0 +1,174 @@
+{
+  "SPDXID": "SPDXRef-DOCUMENT",
+  "packages": [
+    {
+      "SPDXID": "SPDXRef-PACKAGE-bzip2",
+      "checksums": [
+        {
+          "algorithm": "SHA256",
+          "checksumValue": "ab8d1b0cc087c20d4c32c0e4fcf7d0c733a95da12cedc6d63b3f0a9af07427e2"
+        }
+      ],
+      "downloadLocation": "https://github.com/python/cpython-source-deps/archive/refs/tags/bzip2-1.0.8.tar.gz",
+      "externalRefs": [
+        {
+          "referenceCategory": "SECURITY",
+          "referenceLocator": "cpe:2.3:a:bzip:bzip2:1.0.8:*:*:*:*:*:*:*",
+          "referenceType": "cpe23Type"
+        }
+      ],
+      "licenseConcluded": "NOASSERTION",
+      "name": "bzip2",
+      "primaryPackagePurpose": "SOURCE",
+      "versionInfo": "1.0.8"
+    },
+    {
+      "SPDXID": "SPDXRef-PACKAGE-libffi",
+      "checksums": [
+        {
+          "algorithm": "SHA256",
+          "checksumValue": "9d802681adfea27d84cae0487a785fb9caa925bdad44c401b364c59ab2b8edda"
+        }
+      ],
+      "downloadLocation": "https://github.com/python/cpython-source-deps/archive/refs/tags/libffi-3.4.4.tar.gz",
+      "externalRefs": [
+        {
+          "referenceCategory": "SECURITY",
+          "referenceLocator": "cpe:2.3:a:libffi_project:libffi:3.4.4:*:*:*:*:*:*:*",
+          "referenceType": "cpe23Type"
+        }
+      ],
+      "licenseConcluded": "NOASSERTION",
+      "name": "libffi",
+      "primaryPackagePurpose": "SOURCE",
+      "versionInfo": "3.4.4"
+    },
+    {
+      "SPDXID": "SPDXRef-PACKAGE-openssl",
+      "checksums": [
+        {
+          "algorithm": "SHA256",
+          "checksumValue": "e6a77c273ebb284fedd8ea19b081fce74a9455936ffd47215f7c24713e2614b2"
+        }
+      ],
+      "downloadLocation": "https://github.com/python/cpython-source-deps/archive/refs/tags/openssl-3.0.13.tar.gz",
+      "externalRefs": [
+        {
+          "referenceCategory": "SECURITY",
+          "referenceLocator": "cpe:2.3:a:openssl:openssl:3.0.13:*:*:*:*:*:*:*",
+          "referenceType": "cpe23Type"
+        }
+      ],
+      "licenseConcluded": "NOASSERTION",
+      "name": "openssl",
+      "primaryPackagePurpose": "SOURCE",
+      "versionInfo": "3.0.13"
+    },
+    {
+      "SPDXID": "SPDXRef-PACKAGE-sqlite",
+      "checksums": [
+        {
+          "algorithm": "SHA256",
+          "checksumValue": "6f0364a27375435a34137b138ca4fedef8d23eec6493ca1dfff33bfc0c34fda4"
+        }
+      ],
+      "downloadLocation": "https://github.com/python/cpython-source-deps/archive/refs/tags/sqlite-3.45.1.0.tar.gz",
+      "externalRefs": [
+        {
+          "referenceCategory": "SECURITY",
+          "referenceLocator": "cpe:2.3:a:sqlite:sqlite:3.45.1.0:*:*:*:*:*:*:*",
+          "referenceType": "cpe23Type"
+        }
+      ],
+      "licenseConcluded": "NOASSERTION",
+      "name": "sqlite",
+      "primaryPackagePurpose": "SOURCE",
+      "versionInfo": "3.45.1.0"
+    },
+    {
+      "SPDXID": "SPDXRef-PACKAGE-tcl-core",
+      "checksums": [
+        {
+          "algorithm": "SHA256",
+          "checksumValue": "1d3f2015e49e269cf681373d433cd54d88d5ef7443fe87f5f50f5fcfe9003e73"
+        }
+      ],
+      "downloadLocation": "https://github.com/python/cpython-source-deps/archive/refs/tags/tcl-core-8.6.13.1.tar.gz",
+      "externalRefs": [
+        {
+          "referenceCategory": "SECURITY",
+          "referenceLocator": "cpe:2.3:a:tcl_tk:tcl_tk:8.6.13.1:*:*:*:*:*:*:*",
+          "referenceType": "cpe23Type"
+        }
+      ],
+      "licenseConcluded": "NOASSERTION",
+      "name": "tcl-core",
+      "primaryPackagePurpose": "SOURCE",
+      "versionInfo": "8.6.13.1"
+    },
+    {
+      "SPDXID": "SPDXRef-PACKAGE-tk",
+      "checksums": [
+        {
+          "algorithm": "SHA256",
+          "checksumValue": "6056203b8a6aaf6ea89d90a7b55dc7f407e55c093f731a98fd830a712a3c81d3"
+        }
+      ],
+      "downloadLocation": "https://github.com/python/cpython-source-deps/archive/refs/tags/tk-8.6.13.1.tar.gz",
+      "externalRefs": [
+        {
+          "referenceCategory": "SECURITY",
+          "referenceLocator": "cpe:2.3:a:tcl_tk:tcl_tk:8.6.13.1:*:*:*:*:*:*:*",
+          "referenceType": "cpe23Type"
+        }
+      ],
+      "licenseConcluded": "NOASSERTION",
+      "name": "tk",
+      "primaryPackagePurpose": "SOURCE",
+      "versionInfo": "8.6.13.1"
+    },
+    {
+      "SPDXID": "SPDXRef-PACKAGE-xz",
+      "checksums": [
+        {
+          "algorithm": "SHA256",
+          "checksumValue": "a15c168e39e87d750c3dc766edc7f19bdda57dacf01e509678467eace91ad282"
+        }
+      ],
+      "downloadLocation": "https://github.com/python/cpython-source-deps/archive/refs/tags/xz-5.2.5.tar.gz",
+      "externalRefs": [
+        {
+          "referenceCategory": "SECURITY",
+          "referenceLocator": "cpe:2.3:a:xz_project:xz:5.2.5:*:*:*:*:*:*:*",
+          "referenceType": "cpe23Type"
+        }
+      ],
+      "licenseConcluded": "NOASSERTION",
+      "name": "xz",
+      "primaryPackagePurpose": "SOURCE",
+      "versionInfo": "5.2.5"
+    },
+    {
+      "SPDXID": "SPDXRef-PACKAGE-zlib",
+      "checksums": [
+        {
+          "algorithm": "SHA256",
+          "checksumValue": "e3f3fb32564952006eb18b091ca8464740e5eca29d328cfb0b2da22768e0b638"
+        }
+      ],
+      "downloadLocation": "https://github.com/python/cpython-source-deps/archive/refs/tags/zlib-1.3.1.tar.gz",
+      "externalRefs": [
+        {
+          "referenceCategory": "SECURITY",
+          "referenceLocator": "cpe:2.3:a:zlib:zlib:1.3.1:*:*:*:*:*:*:*",
+          "referenceType": "cpe23Type"
+        }
+      ],
+      "licenseConcluded": "NOASSERTION",
+      "name": "zlib",
+      "primaryPackagePurpose": "SOURCE",
+      "versionInfo": "1.3.1"
+    }
+  ],
+  "spdxVersion": "SPDX-2.3"
+}
\ No newline at end of file
diff --git a/Tools/build/generate_sbom.py b/Tools/build/generate_sbom.py
index 201c81c..6aa4946 100644
--- a/Tools/build/generate_sbom.py
+++ b/Tools/build/generate_sbom.py
@@ -7,9 +7,8 @@ import glob
 import pathlib
 import subprocess
 import sys
+import urllib.request
 import typing
-import zipfile
-from urllib.request import urlopen
 
 CPYTHON_ROOT_DIR = pathlib.Path(__file__).parent.parent.parent
 
@@ -125,30 +124,41 @@ def filter_gitignored_paths(paths: list[str]) -> list[str]:
     return sorted([line.split()[-1] for line in git_check_ignore_lines if line.startswith("::")])
 
 
-def main() -> None:
-    sbom_path = CPYTHON_ROOT_DIR / "Misc/sbom.spdx.json"
-    sbom_data = json.loads(sbom_path.read_bytes())
+def get_externals() -> list[str]:
+    """
+    Parses 'PCbuild/get_externals.bat' for external libraries.
+    Returns a list of (git tag, name, version) tuples.
+    """
+    get_externals_bat_path = CPYTHON_ROOT_DIR / "PCbuild/get_externals.bat"
+    externals = re.findall(
+        r"set\s+libraries\s*=\s*%libraries%\s+([a-zA-Z0-9.-]+)\s",
+        get_externals_bat_path.read_text()
+    )
+    return externals
 
-    # We regenerate all of this information. Package information
-    # should be preserved though since that is edited by humans.
-    sbom_data["files"] = []
-    sbom_data["relationships"] = []
 
-    # Ensure all packages in this tool are represented also in the SBOM file.
-    actual_names = {package["name"] for package in sbom_data["packages"]}
-    expected_names = set(PACKAGE_TO_FILES)
-    error_if(
-        actual_names != expected_names,
-        f"Packages defined in SBOM tool don't match those defined in SBOM file: {actual_names}, {expected_names}",
-    )
+def check_sbom_packages(sbom_data: dict[str, typing.Any]) -> None:
+    """Make a bunch of assertions about the SBOM package data to ensure it's consistent."""
 
-    # Make a bunch of assertions about the SBOM data to ensure it's consistent.
     for package in sbom_data["packages"]:
         # Properties and ID must be properly formed.
         error_if(
             "name" not in package,
             "Package is missing the 'name' field"
         )
+
+        # Verify that the checksum matches the expected value
+        # and that the download URL is valid.
+        if "checksums" not in package or "CI" in os.environ:
+            download_location = package["downloadLocation"]
+            resp = urllib.request.urlopen(download_location)
+            error_if(resp.status != 200, f"Couldn't access URL: {download_location}'")
+
+            package["checksums"] = [{
+                "algorithm": "SHA256",
+                "checksumValue": hashlib.sha256(resp.read()).hexdigest()
+            }]
+
         missing_required_keys = REQUIRED_PROPERTIES_PACKAGE - set(package.keys())
         error_if(
             bool(missing_required_keys),
@@ -180,6 +190,26 @@ def main() -> None:
             f"License identifier must be 'NOASSERTION'"
         )
 
+
+def create_source_sbom() -> None:
+    sbom_path = CPYTHON_ROOT_DIR / "Misc/sbom.spdx.json"
+    sbom_data = json.loads(sbom_path.read_bytes())
+
+    # We regenerate all of this information. Package information
+    # should be preserved though since that is edited by humans.
+    sbom_data["files"] = []
+    sbom_data["relationships"] = []
+
+    # Ensure all packages in this tool are represented also in the SBOM file.
+    actual_names = {package["name"] for package in sbom_data["packages"]}
+    expected_names = set(PACKAGE_TO_FILES)
+    error_if(
+        actual_names != expected_names,
+        f"Packages defined in SBOM tool don't match those defined in SBOM file: {actual_names}, {expected_names}",
+    )
+
+    check_sbom_packages(sbom_data)
+
     # We call 'sorted()' here a lot to avoid filesystem scan order issues.
     for name, files in sorted(PACKAGE_TO_FILES.items()):
         package_spdx_id = spdx_id(f"SPDXRef-PACKAGE-{name}")
@@ -224,5 +254,49 @@ def main() -> None:
     sbom_path.write_text(json.dumps(sbom_data, indent=2, sort_keys=True))
 
 
+def create_externals_sbom() -> None:
+    sbom_path = CPYTHON_ROOT_DIR / "Misc/externals.spdx.json"
+    sbom_data = json.loads(sbom_path.read_bytes())
+
+    externals = get_externals()
+    externals_name_to_version = {}
+    externals_name_to_git_tag = {}
+    for git_tag in externals:
+        name, _, version = git_tag.rpartition("-")
+        externals_name_to_version[name] = version
+        externals_name_to_git_tag[name] = git_tag
+
+    # Ensure all packages in this tool are represented also in the SBOM file.
+    actual_names = {package["name"] for package in sbom_data["packages"]}
+    expected_names = set(externals_name_to_version)
+    error_if(
+        actual_names != expected_names,
+        f"Packages defined in SBOM tool don't match those defined in SBOM file: {actual_names}, {expected_names}",
+    )
+
+    # Set the versionInfo and downloadLocation fields for all packages.
+    for package in sbom_data["packages"]:
+        package["versionInfo"] = externals_name_to_version[package["name"]]
+        download_location = (
+            f"https://github.com/python/cpython-source-deps/archive/refs/tags/{externals_name_to_git_tag[package['name']]}.tar.gz"
+        )
+        download_location_changed = download_location != package["downloadLocation"]
+        package["downloadLocation"] = download_location
+
+        # If the download URL has changed we want one to get recalulated.
+        if download_location_changed:
+            package.pop("checksums", None)
+
+    check_sbom_packages(sbom_data)
+
+    # Update the SBOM on disk
+    sbom_path.write_text(json.dumps(sbom_data, indent=2, sort_keys=True))
+
+
+def main() -> None:
+    create_source_sbom()
+    create_externals_sbom()
+
+
 if __name__ == "__main__":
     main()