fix: add another heuristic for libc discovery (#564)

* fix: add another heuristic for libc discovery

* [pre-commit.ci] auto fixes from pre-commit.com hooks

for more information, see https://pre-commit.ci

* fix heuristic

---------

Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com>

Author: mayeut

scripts/create-arch-wheels.sh

@@ -7,11 +7,11 @@ set -eux
 
 SCRIPT_DIR="$(CDPATH='' cd -- "$(dirname -- "$0")" && pwd -P)"
 INTEGRATION_TEST_DIR="${SCRIPT_DIR}/../tests/integration"
-mkdir -p "${INTEGRATION_TEST_DIR}/arch-wheels"
+mkdir -p "${INTEGRATION_TEST_DIR}/arch-wheels/glibc"
+mkdir -p "${INTEGRATION_TEST_DIR}/arch-wheels/musllinux_1_2"
 
 # "mips64le" built with buildpack-deps:bookworm and renamed cp313-cp313
-# "386" "amd64" "arm/v5" "arm/v7" "arm64/v8"
-for ARCH in  "ppc64le" "riscv64" "s390x"; do
+for ARCH in  "386" "amd64" "arm/v5" "arm/v7" "arm64/v8" "ppc64le" "riscv64" "s390x"; do
   docker run --platform linux/${ARCH} -i --rm -v "${INTEGRATION_TEST_DIR}:/tests" debian:trixie-20250203 << "EOF"
 # for, "arm/v5" QEMU will report armv7l, running on aarch64 will report aarch64, force armv5l/armv7l
 case "$(dpkg --print-architecture)" in
@@ -21,7 +21,19 @@ case "$(dpkg --print-architecture)" in
 esac
 DEBIAN_FRONTEND=noninteractive apt-get update
 DEBIAN_FRONTEND=noninteractive apt-get install -y --no-install-recommends gcc python3-pip python3-dev
-python3 -m pip wheel --no-deps -w /tests/arch-wheels /tests/testsimple
+python3 -m pip wheel --no-deps -w /tests/arch-wheels/glibc /tests/testsimple
 EOF
+done
 
+for ARCH in "386" "amd64" "arm/v6" "arm/v7" "arm64/v8" "ppc64le" "riscv64" "s390x"; do
+  docker run --platform linux/${ARCH} -i --rm -v "${INTEGRATION_TEST_DIR}:/tests" alpine:3.21 << "EOF"
+# for, "arm/v5" QEMU will report armv7l, running on aarch64 will report aarch64, force armv5l/armv7l
+case "$(cat /etc/apk/arch)" in
+  armhf) export _PYTHON_HOST_PLATFORM="linux-armv6l";;
+  armv7) export _PYTHON_HOST_PLATFORM="linux-armv7l";;
+  *) ;;
+esac
+apk add gcc binutils musl-dev python3-dev py3-pip
+python3 -m pip wheel --no-deps -w /tests/arch-wheels/musllinux_1_2 /tests/testsimple
+EOF
 done

src/auditwheel/lddtree.py

@@ -522,6 +522,17 @@ def ldd(
                 if libc != Libc.GLIBC:
                     msg = f"found a dependency on GLIBC but the libc is already set to {libc}"
                     raise InvalidLibc(msg)
+        if libc is None:
+            # try the filename as a last resort
+            if path.name.endswith(("-arm-linux-musleabihf.so", "-linux-musl.so")):
+                libc = Libc.MUSL
+            elif path.name.endswith(("-arm-linux-gnueabihf.so", "-linux-gnu.so")):
+                # before python 3.11, musl was also using gnu
+                soabi = path.stem.split(".")[-1].split("-")
+                valid_python = tuple(f"3{minor}" for minor in range(11, 100))
+                if soabi[0] == "cpython" and soabi[1].startswith(valid_python):
+                    libc = Libc.GLIBC
+
         if ldpaths is None:
             ldpaths = load_ld_paths(libc).copy()
         # Propagate the rpaths used by the main ELF since those will be

src/auditwheel/libc.py

@@ -5,7 +5,7 @@
 import re
 import subprocess
 from dataclasses import dataclass
-from enum import IntEnum
+from enum import Enum
 from pathlib import Path
 
 from .error import InvalidLibc
@@ -19,9 +19,14 @@ class LibcVersion:
     minor: int
 
 
-class Libc(IntEnum):
-    GLIBC = (1,)
-    MUSL = (2,)
+class Libc(Enum):
+    value: str
+
+    GLIBC = "glibc"
+    MUSL = "musl"
+
+    def __str__(self) -> str:
+        return self.value
 
     def get_current_version(self) -> LibcVersion:
         if self == Libc.MUSL:

src/auditwheel/policy/__init__.py

@@ -11,6 +11,7 @@
 
 from ..architecture import Architecture
 from ..elfutils import filter_undefined_symbols
+from ..error import InvalidLibc
 from ..lddtree import DynamicExecutable
 from ..libc import Libc
 from ..tools import is_subdir
@@ -62,8 +63,13 @@ def __init__(
             raise ValueError(msg)
         if libc == Libc.MUSL:
             if musl_policy is None:
-                musl_version = libc.get_current_version()
-                musl_policy = f"musllinux_{musl_version.major}_{musl_version.minor}"
+                try:
+                    musl_version = libc.get_current_version()
+                    musl_policy = f"musllinux_{musl_version.major}_{musl_version.minor}"
+                except InvalidLibc:
+                    logger.warning(
+                        "can't determine musl libc version, latest known version will be used."
+                    )
             elif _MUSL_POLICY_RE.match(musl_policy) is None:
                 msg = f"Invalid 'musl_policy': '{musl_policy}'"
                 raise ValueError(msg)
@@ -106,6 +112,9 @@ def __init__(
         self._policies.sort()
 
         if self._libc_variant == Libc.MUSL:
+            if musl_policy is None:
+                self._musl_policy = "_".join(self._policies[1].name.split("_")[0:3])
+                self._policies = [self._policies[0], self._policies[1]]
             assert len(self._policies) == 2, self._policies
 
     def __iter__(self) -> Generator[Policy]: