Fix cirular import in policy submodule

The `external_versioned_symbols` and `versioned_symbols_policy`
functions used to live in two separate files, importing code from the
`__init__.py` file in the same module. As this file itself imports both
of these symbols, this creates a cirtular import. We can easily break
this by moving the two functions into the same `__init__.py` file.

Author: lkollar

src/auditwheel/policy/__init__.py

@@ -3,17 +3,20 @@
 import json
 import logging
 import platform as _platform_module
+import re
 import sys
 from collections import defaultdict
 from os.path import abspath, dirname, join
 from pathlib import Path
+from typing import Any, Generator
+
+from auditwheel.elfutils import filter_undefined_symbols, is_subdir
 
 from ..libc import Libc, get_libc
 from ..musllinux import find_musl_libc, get_musl_version
-from .external_references import lddtree_external_references
-from .versioned_symbols import versioned_symbols_policy
 
 _HERE = Path(__file__).parent
+LIBPYTHON_RE = re.compile(r"^libpython\d+\.\d+m?.so(.\d)*$")
 
 logger = logging.getLogger(__name__)
 
@@ -207,6 +210,116 @@ def _load_policy_schema():
     return schema
 
 
+def versioned_symbols_policy(
+    wheel_policy: WheelPolicies, versioned_symbols: dict[str, set[str]]
+) -> int:
+    def policy_is_satisfied(
+        policy_name: str, policy_sym_vers: dict[str, set[str]]
+    ) -> bool:
+        policy_satisfied = True
+        for name in set(required_vers) & set(policy_sym_vers):
+            if not required_vers[name].issubset(policy_sym_vers[name]):
+                for symbol in required_vers[name] - policy_sym_vers[name]:
+                    logger.debug(
+                        "Package requires %s, incompatible with "
+                        "policy %s which requires %s",
+                        symbol,
+                        policy_name,
+                        policy_sym_vers[name],
+                    )
+                policy_satisfied = False
+        return policy_satisfied
+
+    required_vers: dict[str, set[str]] = {}
+    for symbols in versioned_symbols.values():
+        for symbol in symbols:
+            sym_name, _, _ = symbol.partition("_")
+            required_vers.setdefault(sym_name, set()).add(symbol)
+    matching_policies: list[int] = []
+    for p in wheel_policy.policies:
+        policy_sym_vers = {
+            sym_name: {sym_name + "_" + version for version in versions}
+            for sym_name, versions in p["symbol_versions"].items()
+        }
+        if policy_is_satisfied(p["name"], policy_sym_vers):
+            matching_policies.append(p["priority"])
+
+    if len(matching_policies) == 0:
+        # the base policy (generic linux) should always match
+        raise RuntimeError("Internal error")
+
+    return max(matching_policies)
+
+
+def lddtree_external_references(
+    wheel_policies: list, lddtree: dict, wheel_path: str
+) -> dict:
+    # XXX: Document the lddtree structure, or put it in something
+    # more stable than a big nested dict
+    def filter_libs(libs: set[str], whitelist: set[str]) -> Generator[str, None, None]:
+        for lib in libs:
+            if "ld-linux" in lib or lib in ["ld64.so.2", "ld64.so.1"]:
+                # always exclude ELF dynamic linker/loader
+                # 'ld64.so.2' on s390x
+                # 'ld64.so.1' on ppc64le
+                # 'ld-linux*' on other platforms
+                continue
+            if LIBPYTHON_RE.match(lib):
+                # always exclude libpythonXY
+                continue
+            if lib in whitelist:
+                # exclude any libs in the whitelist
+                continue
+            yield lib
+
+    def get_req_external(libs: set[str], whitelist: set[str]) -> set[str]:
+        # get all the required external libraries
+        libs = libs.copy()
+        reqs = set()
+        while libs:
+            lib = libs.pop()
+            reqs.add(lib)
+            for dep in filter_libs(lddtree["libs"][lib]["needed"], whitelist):
+                if dep not in reqs:
+                    libs.add(dep)
+        return reqs
+
+    ret: dict[str, dict[str, Any]] = {}
+    for p in wheel_policies:
+        needed_external_libs: set[str] = set()
+        blacklist = {}
+
+        if not (p["name"] == "linux" and p["priority"] == 0):
+            # special-case the generic linux platform here, because it
+            # doesn't have a whitelist. or, you could say its
+            # whitelist is the complete set of all libraries. so nothing
+            # is considered "external" that needs to be copied in.
+            whitelist = set(p["lib_whitelist"])
+            blacklist_libs = set(p["blacklist"].keys()) & set(lddtree["needed"])
+            blacklist = {k: p["blacklist"][k] for k in blacklist_libs}
+            blacklist = filter_undefined_symbols(lddtree["realpath"], blacklist)
+            needed_external_libs = get_req_external(
+                set(filter_libs(lddtree["needed"], whitelist)), whitelist
+            )
+
+        pol_ext_deps = {}
+        for lib in needed_external_libs:
+            if is_subdir(lddtree["libs"][lib]["realpath"], wheel_path):
+                # we didn't filter libs that resolved via RPATH out
+                # earlier because we wanted to make sure to pick up
+                # our elf's indirect dependencies. But now we want to
+                # filter these ones out, since they're not "external".
+                logger.debug("RPATH FTW: %s", lib)
+                continue
+            pol_ext_deps[lib] = lddtree["libs"][lib]["realpath"]
+        ret[p["name"]] = {
+            "libs": pol_ext_deps,
+            "priority": p["priority"],
+            "blacklist": blacklist,
+        }
+    return ret
+
+
 __all__ = [
     "lddtree_external_references",
     "versioned_symbols_policy",

src/auditwheel/policy/external_references.py

@@ -19,11 +19,7 @@
 )
 from .genericpkgctx import InGenericPkgCtx
 from .lddtree import lddtree
-from .policy import (
-    WheelPolicies,
-    lddtree_external_references,
-    versioned_symbols_policy,
-)
+from .policy import WheelPolicies, lddtree_external_references, versioned_symbols_policy
 
 log = logging.getLogger(__name__)
 WheelAbIInfo = namedtuple(
@@ -201,7 +197,9 @@ def get_versioned_symbols(libs):
     return result
 
 
-def get_symbol_policies(wheel_policy, versioned_symbols, external_versioned_symbols, external_refs):
+def get_symbol_policies(
+    wheel_policy, versioned_symbols, external_versioned_symbols, external_refs
+):
     """Get symbol policies
     Since white-list is different per policy, this function inspects
     versioned_symbol per policy when including external refs
@@ -223,7 +221,9 @@ def get_symbol_policies(wheel_policy, versioned_symbols, external_versioned_symb
             ext_symbols = external_versioned_symbols[soname]
             for k in iter(ext_symbols):
                 policy_symbols[k].update(ext_symbols[k])
-        result.append((versioned_symbols_policy(wheel_policy, policy_symbols), policy_symbols))
+        result.append(
+            (versioned_symbols_policy(wheel_policy, policy_symbols), policy_symbols)
+        )
     return result