exclude docs from wheel; harden bundle item handling; single-source version

build/release/version.py

@@ -143,6 +143,97 @@ def released_tags_for(version: str, tags: list[str]) -> list[str]:
     return sorted({version, f"v{version}"} & set(tags))
 
 
+# Version literals that are deliberately not the package version. Keyed by
+# repo-relative path; each entry documents *why* the literal may stay.
+VERSION_STRING_ALLOWLIST: dict[str, frozenset[str]] = {
+    # Historical release records: rewriting these would destroy the audit trail.
+    "docs/publishing.md": frozenset({"0.4.2", "0.4.3", "0.4.4", "0.5.0", "0.5.1", "0.5.2"}),
+    # Unrelated third-party/CUDA versions quoted in prose.
+    "docs/get_started.md": frozenset({"12.9.0"}),
+}
+
+# Paths never scanned: tests deliberately use synthetic versions, reports are
+# generated output, and this module's own allowlist/sentinel literals would
+# otherwise flag themselves.
+VERSION_SCAN_EXCLUDED_PREFIXES = (
+    "tests/",
+    "build/release/reports/",
+    "build/nuitka/",
+    "build/lib/",
+    "dist/",
+    "SDK/",
+    "data/",
+)
+VERSION_SCAN_EXCLUDED_FILES = ("build/release/version.py",)
+
+_VERSION_LITERAL_RE = re.compile(
+    r"(?<![<>=!~\w.])v?(\d+\.\d+\.\d+(?:[-+][0-9A-Za-z.]+)?)(?![0-9A-Za-z.])"
+)
+_SCANNED_SUFFIXES = (".md", ".py", ".txt", ".toml", ".cfg", ".ini", ".rst")
+
+
+def tracked_text_files(repo_root: Path) -> list[Path]:
+    """Tracked files worth scanning for version literals."""
+    repo_root = Path(repo_root)
+    try:
+        completed = subprocess.run(
+            ["git", "-C", str(repo_root), "ls-files", "-z"],
+            capture_output=True,
+            text=True,
+            check=False,
+        )
+        names = [n for n in (completed.stdout or "").split("\0") if n]
+    except OSError:
+        names = []
+
+    if not names:
+        names = [
+            str(path.relative_to(repo_root)).replace("\\", "/")
+            for path in repo_root.rglob("*")
+            if path.is_file()
+        ]
+
+    selected = []
+    for name in names:
+        if name.startswith(VERSION_SCAN_EXCLUDED_PREFIXES):
+            continue
+        if name in VERSION_SCAN_EXCLUDED_FILES:
+            continue
+        if not name.endswith(_SCANNED_SUFFIXES):
+            continue
+        selected.append(Path(name))
+    return selected
+
+
+def find_version_literals(
+    repo_root: Path, version: str
+) -> list[tuple[str, int, str]]:
+    """``(path, lineno, literal)`` for every version-looking string != ``version``.
+
+    Dependency constraints are skipped by the regex lookbehind (``numpy>=1.24.4``
+    is not a version reference), and historical/third-party literals require an
+    explicit entry in :data:`VERSION_STRING_ALLOWLIST`.
+    """
+    repo_root = Path(repo_root)
+    findings: list[tuple[str, int, str]] = []
+    for relative in tracked_text_files(repo_root):
+        absolute = repo_root / relative
+        try:
+            text = absolute.read_text(encoding="utf-8", errors="ignore")
+        except OSError:
+            continue
+        if "\0" in text:
+            continue
+        allowed = VERSION_STRING_ALLOWLIST.get(relative.as_posix(), frozenset())
+        for lineno, line in enumerate(text.splitlines(), start=1):
+            for match in _VERSION_LITERAL_RE.finditer(line):
+                literal = match.group(1)
+                if literal == version or literal in allowed:
+                    continue
+                findings.append((relative.as_posix(), lineno, literal))
+    return findings
+
+
 def conda_envs_root() -> Path | None:
     """Locate the conda envs directory the same way build/wheel/win.bat does."""
     override = os.getenv("CONDA_ENVS_ROOT")