Improve sample hex output

Closes #47

cobrafuzz/fuzzer.py

@@ -366,12 +366,8 @@ def _write_sample(self, buf: bytes, prefix: str = "crash-") -> None:
             logging.info("Crash dir created (%s)", self._crash_dir)
 
         crash_path = self._crash_dir / (prefix + m.hexdigest())
-
-        with crash_path.open("wb") as f:
-            f.write(buf)
-        logging.info("sample was written to %s", crash_path)
-        if len(buf) < 200:
-            logging.info("sample = %s", buf.hex())
+        crash_path.write_bytes(buf)
+        logging.info(util.hexdump(title=f"Sample written to {crash_path.name}:", data=buf))
 
     def _initialize_process(self, wid: int) -> tuple[MPProcess, mp.Queue[Update]]:
         queue: mp.Queue[Update] = self._mp_ctx.Queue()

cobrafuzz/util.py

@@ -236,3 +236,14 @@ def disable_logging() -> Iterator[None]:
         yield
     finally:
         logging.disable(previous_level)
+
+
+def hexdump(title: str, data: bytes) -> str:
+    length = 16
+    result = [title]
+    for i in range(0, len(data), length):
+        chunk = data[i : i + length]
+        hex_chunk = " ".join(f"{b:02x}" for b in chunk)
+        ascii_chunk = "".join(chr(b) if 32 <= b < 127 else "." for b in chunk)
+        result.append(f"{i:08x}: {hex_chunk:<{length*3}} {ascii_chunk}")
+    return "\n".join(result)

tests/unit/test_fuzzer.py

@@ -13,7 +13,7 @@
 import dill  # type: ignore[import-untyped]
 import pytest
 
-from cobrafuzz import fuzzer, simplifier, state as st
+from cobrafuzz import fuzzer, simplifier, state as st, util
 from tests import utils
 
 
@@ -102,7 +102,7 @@ def test_write_sample(
     assert artifact.is_file()
     with artifact.open("rb") as af:
         assert af.read() == sample
-    assert length >= 200 or f"sample = {sample.hex()}" in caplog.text
+    assert util.hexdump(title="", data=sample) in caplog.text
 
 
 def test_regression_valid(caplog: pytest.LogCaptureFixture, tmp_path: Path) -> None:
@@ -464,8 +464,7 @@ def test_start_error(
     assert caplog.record_tuples == [
         ("root", logging.INFO, "START units: 1, workers: 1, seeds: 0"),
         ("root", logging.INFO, "Test error message"),
-        ("root", logging.INFO, f"sample was written to {tmp_path / filename}"),
-        ("root", logging.INFO, "sample = 6465616462656566"),
+        ("root", logging.INFO, util.hexdump(title=f"Sample written to {filename}:", data=data)),
         ("root", logging.INFO, "Found 1 crashes, stopping."),
     ]
 
@@ -522,8 +521,7 @@ def simplify(self) -> None:
         ("root", logging.INFO, "START units: 1, workers: 1, seeds: 0"),
         ("root", logging.INFO, "Test error message"),
         ("root", logging.INFO, f"Crash dir created ({crash_path})"),
-        ("root", logging.INFO, f"sample was written to {crash_path / filename}"),
-        ("root", logging.INFO, "sample = 6465616462656566"),
+        ("root", logging.INFO, util.hexdump(title=f"Sample written to {filename}:", data=data)),
         ("root", logging.INFO, "Found 1 crashes, stopping."),
     ]
 

tests/unit/test_util.py

@@ -230,3 +230,35 @@ def test_non_adaptive_choice() -> None:
     c.append(2)
     assert c._population == [1, 2]
     assert c._distribution is None
+
+
+@pytest.mark.parametrize(
+    ("title", "data", "expected"),
+    [
+        ("", b"", ""),
+        (
+            "some title",
+            b"x",
+            "some title\n00000000: 78                                               x",
+        ),
+        (
+            "some other title",
+            b"xy",
+            "some other title\n00000000: 78 79                                            xy",
+        ),
+        (
+            "title:",
+            b"0123456780123456",
+            "title:\n00000000: 30 31 32 33 34 35 36 37 38 30 31 32 33 34 35 36  0123456780123456",
+        ),
+        (
+            "title:",
+            b"012345678012345678",
+            "title:\n"
+            "00000000: 30 31 32 33 34 35 36 37 38 30 31 32 33 34 35 36  0123456780123456\n"
+            "00000010: 37 38                                            78",
+        ),
+    ],
+)
+def test_hexdump(title: str, data: bytes, expected: str) -> None:
+    assert util.hexdump(title, data) == expected