Made elf.py more resilient

* try/except for section generator. lief-project/LIEF#1031
* skip symbol name if symbol not present for relocation. This can happen
  for relative relocations FWII

sqlelf/cli.py

@@ -9,6 +9,8 @@
 from sqlelf import elf
 from sqlelf import sql as api_sql
 
+import logging
+
 
 @dataclass
 class ProgramArguments:
@@ -71,11 +73,16 @@ def start(args: list[str] = sys.argv[1:], stdin: TextIO = sys.stdin) -> None:
     )
 
     # Iterate through our arguments and if one of them is a directory explode it out
+    # Assuming program_args.filenames is a list of files and directories
     filenames: list[str] = reduce(
         lambda a, b: a + b,
         map(
             lambda dir: (
-                [os.path.join(dir, f) for f in os.listdir(dir)]
+                [
+                    os.path.join(root, file)
+                    for root, _, files in os.walk(dir)
+                    for file in files
+                ]
                 if os.path.isdir(dir)
                 else [dir]
             ),
@@ -89,6 +96,12 @@ def start(args: list[str] = sys.argv[1:], stdin: TextIO = sys.stdin) -> None:
     if len(filenames) == 0:
         sys.exit("No valid ELF files were provided")
 
+    # Setup the logging config
+    logging.basicConfig(
+        level=logging.INFO,
+        format="%(asctime)s %(levelname)-8s [%(filename)s:%(lineno)d] %(message)s",
+    )
+
     sql_engine = api_sql.make_sql_engine(
         filenames, recursive=program_args.recursive, cache_flags=program_args.cache_flag
     )

sqlelf/elf.py

@@ -21,6 +21,10 @@
 from sqlelf._vendor.elftools.dwarf.die import DIE as DIE_t
 from sqlelf._vendor.elftools.elf.elffile import ELFFile
 
+import logging
+
+LOG = logging.getLogger(__name__)
+
 
 @dataclass
 class Generator:
@@ -244,14 +248,23 @@ def sections_generator() -> Iterator[dict[str, Any]]:
             # as they can be costly
             binary_name = binary.path
             for section in binary.sections:
-                yield {
-                    "path": binary_name,
-                    "name": section.name,
-                    "offset": section.offset,
-                    "size": section.size,
-                    "type": section.type.__name__,
-                    "content": bytes(section.content),
-                }
+                try:
+                    yield {
+                        "path": binary_name,
+                        "name": section.name,
+                        "offset": section.offset,
+                        "size": section.size,
+                        "type": section.type.__name__,
+                        "content": bytes(section.content),
+                    }
+                except RuntimeError:
+                    # TODO(fzakaria): LIEF is failing to parse some section types:
+                    # https://github.com/lief-project/LIEF/issues/1031
+                    # Just skip them for now
+                    LOG.warning(
+                        "Failed to parse section: %s (%s)", section.name, binary_name
+                    )
+                    pass
 
     generator = Generator.make_generator(
         ["path", "name", "offset", "size", "type", "content"],
@@ -443,8 +456,10 @@ def relocations_generator() -> Iterator[dict[str, Any]]:
                     # https://refspecs.linuxbase.org/elf/gabi4+/ch4.reloc.html
                     "is_rela": relocation.is_rela,
                     "purpose": relocation.purpose.__name__,
-                    "section": relocation.section.name if relocation.section else None,
-                    "symbol": relocation.symbol.name,
+                    "section": (
+                        relocation.section.name if relocation.section else None
+                    ),
+                    "symbol": relocation.symbol.name if relocation.symbol else None,
                     "symbol_table": (
                         relocation.symbol_table.name
                         if relocation.symbol_table