Add ability to repair wheels for other architectures

Author: rhelmot

src/auditwheel/elfutils.py

@@ -5,9 +5,11 @@
 from typing import Iterator
 
 from elftools.common.exceptions import ELFError
+from elftools.elf.dynamic import DynamicSegment
 from elftools.elf.elffile import ELFFile
 
 from .lddtree import parse_ld_paths
+from .libc import Libc
 
 
 def elf_read_dt_needed(fn: str) -> list[str]:
@@ -161,3 +163,37 @@ def filter_undefined_symbols(
         if intersection:
             result[lib] = sorted(intersection)
     return result
+
+
+def elf_get_platform_info(path: str) -> tuple[Libc | None, str | None]:
+    with open(path, "rb") as f:
+        try:
+            elf = ELFFile(f)
+        except ELFError:
+            return (None, None)
+        arch = {
+            "x64": "x86_64",
+            "x86": "i686",
+            "AArch64": "aarch64",
+            "64-bit PowerPC": "ppc64",
+            "IBM S/390": "s390x",
+            "ARM": "armv7l",
+            "RISC-V": "riscv64",
+        }[elf.get_machine_arch()]
+        if arch == "ppc64" and elf.header.e_ident.EI_DATA == "ELFDATA2LSB":
+            arch = "ppc64le"
+
+        libc = None
+        for seg in elf.iter_segments():
+            if not isinstance(seg, DynamicSegment):
+                continue
+            for tag in seg.iter_tags():
+                if tag.entry.d_tag == "DT_NEEDED":
+                    if tag.needed == "libc.so.6":
+                        libc = Libc.GLIBC
+                        break
+                    if tag.needed.startswith("libc.musl-"):
+                        libc = Libc.MUSL
+                        break
+            break
+        return (libc, arch)

src/auditwheel/libc.py

@@ -3,7 +3,6 @@
 import logging
 from enum import IntEnum
 
-from .error import InvalidLibc
 from .musllinux import find_musl_libc
 
 logger = logging.getLogger(__name__)
@@ -15,10 +14,7 @@ class Libc(IntEnum):
 
 
 def get_libc() -> Libc:
-    try:
-        find_musl_libc()
+    if find_musl_libc() is not None:
         logger.debug("Detected musl libc")
         return Libc.MUSL
-    except InvalidLibc:
-        logger.debug("Falling back to GNU libc")
-        return Libc.GLIBC
+    return Libc.GLIBC

src/auditwheel/main_repair.py

@@ -43,13 +43,20 @@ def configure_parser(sub_parsers):
     p.add_argument(
         "--plat",
         action=EnvironmentDefault,
+        required=False,
         metavar="PLATFORM",
         env="AUDITWHEEL_PLAT",
         dest="PLAT",
         help="Desired target platform. See the available platforms under the "
-        f'PLATFORMS section below. (default: "{highest_policy}")',
+        f'PLATFORMS section below. (default on current arch: "{highest_policy}")',
         choices=policy_names,
-        default=highest_policy,
+        default=None,
+    )
+    p.add_argument(
+        "--best-plat",
+        action="store_true",
+        dest="BEST_PLAT",
+        help="Automatically determine the best target platform.",
     )
     p.add_argument(
         "-L",
@@ -115,26 +122,36 @@ def execute(args, p):
     for wheel_file in args.WHEEL_FILE:
         if not isfile(wheel_file):
             p.error("cannot access %s. No such file" % wheel_file)
+        wheel_policy.set_platform_from_wheel(wheel_file)
 
         logger.info("Repairing %s", basename(wheel_file))
 
-        if not exists(args.WHEEL_DIR):
-            os.makedirs(args.WHEEL_DIR)
-
         try:
             wheel_abi = analyze_wheel_abi(wheel_policy, wheel_file, exclude)
         except NonPlatformWheel:
             logger.info(NonPlatformWheel.LOG_MESSAGE)
             return 1
 
+        if args.BEST_PLAT:
+            if args.PLAT:
+                p.error("Cannot specify both --best-plat and --plat")
+            args.PLAT = wheel_abi.overall_tag
+
+        if not exists(args.WHEEL_DIR):
+            os.makedirs(args.WHEEL_DIR)
+
+        highest_policy = wheel_policy.get_policy_name(wheel_policy.priority_highest)
+        if args.PLAT is None:
+            args.PLAT = highest_policy
         policy = wheel_policy.get_policy_by_name(args.PLAT)
         reqd_tag = policy["priority"]
 
         if reqd_tag > wheel_policy.get_priority_by_name(wheel_abi.sym_tag):
             msg = (
                 'cannot repair "%s" to "%s" ABI because of the presence '
                 "of too-recent versioned symbols. You'll need to compile "
-                "the wheel on an older toolchain." % (wheel_file, args.PLAT)
+                "the wheel on an older toolchain or pick a newer platform."
+                % (wheel_file, args.PLAT)
             )
             p.error(msg)
 
@@ -184,3 +201,6 @@ def execute(args, p):
 
         if out_wheel is not None:
             logger.info("\nFixed-up wheel written to %s", out_wheel)
+
+        if args.BEST_PLAT:
+            args.PLAT = None

src/auditwheel/main_show.py

@@ -34,6 +34,8 @@ def execute(args, p):
     if not isfile(args.WHEEL_FILE):
         p.error("cannot access %s. No such file" % args.WHEEL_FILE)
 
+    wheel_policy.set_platform_from_wheel(args.WHEEL_FILE)
+
     try:
         winfo = analyze_wheel_abi(wheel_policy, args.WHEEL_FILE, frozenset())
     except NonPlatformWheel:

src/auditwheel/musllinux.py

@@ -3,12 +3,10 @@
 import logging
 import pathlib
 import re
-import subprocess
 from typing import NamedTuple
 
-from auditwheel.error import InvalidLibc
-
 LOG = logging.getLogger(__name__)
+VERSION_RE = re.compile(b"[^.](?P<major>\\d+)\\.(?P<minor>\\d+)\\.(?P<patch>\\d+)\0")
 
 
 class MuslVersion(NamedTuple):
@@ -17,31 +15,28 @@ class MuslVersion(NamedTuple):
     patch: int
 
 
-def find_musl_libc() -> pathlib.Path:
+def find_musl_libc(library_path: str | None = None) -> pathlib.Path | None:
     try:
-        (dl_path,) = list(pathlib.Path("/lib").glob("libc.musl-*.so.1"))
+        (dl_path,) = list(pathlib.Path(library_path or "/lib").glob("libc.musl-*.so.1"))
     except ValueError:
-        LOG.debug("musl libc not detected")
-        raise InvalidLibc
+        return None
 
     return dl_path
 
 
-def get_musl_version(ld_path: pathlib.Path) -> MuslVersion:
+def get_musl_version(ld_path: pathlib.Path) -> MuslVersion | None:
     try:
-        ld = subprocess.run(
-            [ld_path], check=False, errors="strict", stderr=subprocess.PIPE
-        ).stderr
+        with open(ld_path, "rb") as fp:
+            text = fp.read()
     except FileNotFoundError:
-        LOG.error("Failed to determine musl version", exc_info=True)
-        raise InvalidLibc
-
-    match = re.search(
-        r"Version " r"(?P<major>\d+)." r"(?P<minor>\d+)." r"(?P<patch>\d+)", ld
-    )
-    if not match:
-        raise InvalidLibc
-
-    return MuslVersion(
-        int(match.group("major")), int(match.group("minor")), int(match.group("patch"))
-    )
+        return None
+
+    for match in VERSION_RE.finditer(text):
+        return MuslVersion(
+            int(match.group("major")),
+            int(match.group("minor")),
+            int(match.group("patch")),
+        )
+
+    LOG.error("Failed to determine musl version", exc_info=True)
+    return None