Hexdump command (#3337)

* Add option to dump memory as code

* Change for loops

* Change return

* Add tests.

* Fix bug with backslash

docs/commands/memory/hexdump.md

@@ -2,7 +2,7 @@
 # hexdump
 
 ```text
-usage: hexdump [-h] [address] [count]
+usage: hexdump [-h] [-C [{py,c}]] [address] [count]
 
 ```
 
@@ -19,6 +19,7 @@ Hexdumps data at the specified address or module name.
 |Short|Long|Help|
 | :--- | :--- | :--- |
 |-h|--help|show this help message and exit|
+|-C|--code|Output as Python or C code data definition (default: py)|
 
 <!-- END OF AUTOGENERATED PART. Do not modify this line or the line below, they mark the end of the auto-generated part of the file. If you want to extend the documentation in a way which cannot easily be done by adding to the command help description, write below the following line. -->
 <!-- ------------\>8---- ----\>8---- ----\>8------------ -->

pwndbg/commands/hexdump.py

@@ -53,6 +53,24 @@ def address_or_module_name(s) -> int:
         raise argparse.ArgumentTypeError("Unknown hexdump argument type.")
 
 
+def format_c(data: bytes) -> str:
+    toks = [f"{b:#02x}" for b in data]
+    lines = []
+    for i in range(0, len(toks), 16):
+        elem = ", ".join(toks[i : i + 16])
+        lines.append(f"    {elem},")
+    body = "\n".join(lines)
+    return f"static const unsigned char data[] = {{\n{body}\n}};\n"
+
+
+def format_py(data: bytes) -> str:
+    lines = []
+    for i in range(0, len(data), 16):
+        seg = "".join(f"\\x{b:02x}" for b in data[i : i + 16])
+        lines.append(f'    b"{seg}"')
+    return "data = (\n{}\n)\n".format("\n".join(lines))
+
+
 parser = argparse.ArgumentParser(
     description="Hexdumps data at the specified address or module name."
 )
@@ -66,11 +84,20 @@ parser.add_argument(
 parser.add_argument(
     "count", nargs="?", default=pwndbg.config.hexdump_bytes, help="Number of bytes to dump"
 )
+parser.add_argument(
+    "-C",
+    "--code",
+    type=str,
+    nargs="?",
+    const="py",
+    choices=("py", "c"),
+    help="Output as Python or C code data definition (default: py)",
+)
 
 
 @pwndbg.commands.Command(parser, category=CommandCategory.MEMORY)
 @pwndbg.commands.OnlyWhenRunning
-def hexdump(address, count=pwndbg.config.hexdump_bytes) -> None:
+def hexdump(address, count=pwndbg.config.hexdump_bytes, code: str | None = None) -> None:
     if hexdump.repeat:
         address = hexdump.last_address
     else:
@@ -129,19 +156,23 @@ def hexdump(address, count=pwndbg.config.hexdump_bytes) -> None:
         print(e)
         return
 
-    result = pwndbg.hexdump.hexdump(
-        data,
-        address=address,
-        width=width,
-        group_width=group_width,
-        flip_group_endianness=flip_group_endianness,
-        offset=hexdump.offset,
-    )
-
-    for line in result:
-        print(line)
+    if code:
+        source = format_py(data) if code == "py" else format_c(data)
+        print(source)
+        hexdump.offset += len(data)
+    else:
+        result = pwndbg.hexdump.hexdump(
+            data,
+            address=address,
+            width=width,
+            group_width=group_width,
+            flip_group_endianness=flip_group_endianness,
+            offset=hexdump.offset,
+        )
+        for line in result:
+            print(line)
 
-    hexdump.offset += count
+    hexdump.offset += len(data)
 
 
 hexdump.last_address = 0

tests/library/gdb/tests/test_hexdump.py

@@ -158,3 +158,39 @@ def test_hexdump_limit_check(start_binary):
 
     # Reset to default for subsequent tests if any
     gdb.execute(f"set hexdump-limit-mb {default_limit_mb}")
+
+
+def test_hexdump_code_py_format(start_binary):
+    start_binary(BINARY)
+    sp = pwndbg.aglib.regs.rsp
+
+    pwndbg.aglib.memory.write(sp, b"abcdefgh\x01\x02\x03\x04\x05\x06\x07\x08" * 16)
+
+    SIZE = 21
+    out = gdb.execute(f"hexdump -C py $rsp {SIZE}", to_string=True)
+
+    expected = (
+        "data = (\n"
+        '    b"\\x61\\x62\\x63\\x64\\x65\\x66\\x67\\x68\\x01\\x02\\x03\\x04\\x05\\x06\\x07\\x08"\n'
+        '    b"\\x61\\x62\\x63\\x64\\x65"\n'
+        ")\n"
+    )
+    assert out.rstrip("\n") == expected.rstrip("\n")
+
+
+def test_hexdump_code_c_format(start_binary):
+    start_binary(BINARY)
+    sp = pwndbg.aglib.regs.rsp
+
+    pwndbg.aglib.memory.write(sp, b"abcdefgh\x01\x02\x03\x04\x05\x06\x07\x08" * 16)
+
+    SIZE = 21
+    out = gdb.execute(f"hexdump -C c $rsp {SIZE}", to_string=True)
+
+    expected = (
+        "static const unsigned char data[] = {\n"
+        "    0x61, 0x62, 0x63, 0x64, 0x65, 0x66, 0x67, 0x68, 0x1, 0x2, 0x3, 0x4, 0x5, 0x6, 0x7, 0x8,\n"
+        "    0x61, 0x62, 0x63, 0x64, 0x65,\n"
+        "};\n"
+    )
+    assert out.rstrip("\n") == expected.rstrip("\n")