Cleanup hexdump implementation

pwndbg/color/hexdump.py

@@ -26,7 +26,7 @@ config_separator = theme.add_color_param(
 config_highlight_group_lsb = theme.add_param(
     "hexdump-highlight-group-lsb",
     "underline",
-    "highlight LSB of each group. Applies only if hexdump-adjust-group-endianess"
+    "highlight LSB of each group. Applies only if hexdump-use-big-endian"
     " actually changes byte order.",
 )
 

pwndbg/commands/hexdump.py

@@ -13,7 +13,7 @@ pwndbg.gdblib.config.add_param("hexdump-width", 16, "line width of hexdump comma
 pwndbg.gdblib.config.add_param("hexdump-bytes", 64, "number of bytes printed by hexdump command")
 pwndbg.gdblib.config.add_param(
     "hexdump-group-width",
-    4,
+    -1,
     "number of bytes grouped in hexdump command (If -1, the architecture's pointer size is used)",
 )
 pwndbg.gdblib.config.add_param(
@@ -57,7 +57,7 @@ parser.add_argument(
 
 @pwndbg.commands.ArgparsedCommand(parser)
 @pwndbg.commands.OnlyWhenRunning
-def hexdump(address=None, count=pwndbg.gdblib.config.hexdump_bytes):
+def hexdump(address, count=pwndbg.gdblib.config.hexdump_bytes):
     if hexdump.repeat:
         address = hexdump.last_address
         hexdump.offset += 1
@@ -65,16 +65,31 @@ def hexdump(address=None, count=pwndbg.gdblib.config.hexdump_bytes):
         hexdump.offset = 0
 
     address = int(address)
-    address &= pwndbg.gdblib.arch.ptrmask
+    if address > pwndbg.gdblib.arch.ptrmask:
+        new_address = address & pwndbg.gdblib.arch.ptrmask
+        print(
+            message.warn("0x%x is larger than the maximum address, truncating to 0x%x instead"),
+            address,
+            new_address,
+        )
+        address = new_address
+
     count = max(int(count), 0)
     width = int(pwndbg.gdblib.config.hexdump_width)
+
     group_width = int(pwndbg.gdblib.config.hexdump_group_width)
     group_width = pwndbg.gdblib.typeinfo.ptrsize if group_width == -1 else group_width
+
+    # TODO: What if arch endian is big, and use_big_endian is false?
     flip_group_endianess = (
         pwndbg.gdblib.config.hexdump_group_use_big_endian and pwndbg.gdblib.arch.endian == "little"
     )
 
-    if count > address > 0x10000:
+    # The user may have input the start and end range to dump instead of the
+    # starting address and the number of bytes to dump. If the address is above
+    # some minimum address, and the count is larger than that address, we assume
+    # this is the case and correct it
+    if address > 0x10000 and count > address:
         count -= address
 
     try:
@@ -84,17 +99,19 @@ def hexdump(address=None, count=pwndbg.gdblib.config.hexdump_bytes):
         print(e)
         return
 
-    for i, line in enumerate(
-        pwndbg.hexdump.hexdump(
-            data,
-            address=address,
-            width=width,
-            group_width=group_width,
-            flip_group_endianess=flip_group_endianess,
-            offset=hexdump.offset,
-        )
-    ):
+    result = pwndbg.hexdump.hexdump(
+        data,
+        address=address,
+        width=width,
+        group_width=group_width,
+        flip_group_endianess=flip_group_endianess,
+        offset=hexdump.offset,
+    )
+    for i, line in enumerate(result):
         print(line)
+
+    # If this command is entered again with no arguments, remember where we left off printing
+    # TODO: This is broken if the user inputs a count less than the width
     hexdump.offset += i
 
 

pwndbg/hexdump.py

@@ -2,9 +2,10 @@
 Hexdump implementation, ~= stolen from pwntools.
 """
 
-import copy
 import string
 
+import pwnlib.util.lists
+
 import pwndbg.color.hexdump as H
 import pwndbg.color.theme as theme
 import pwndbg.gdblib.config
@@ -13,12 +14,8 @@ color_scheme = None
 printable = None
 
 
-def groupby(array, count, fill=None):
-    array = copy.copy(array)
-    while fill and len(array) % count:
-        array.append(fill)
-    for i in range(0, len(array), count):
-        yield array[i : i + count]
+def groupby(width, array, fill=None):
+    return pwnlib.util.lists.group(width, array, underfull_action="fill", fill_value=fill)
 
 
 config_colorize_ascii = theme.add_param(
@@ -41,7 +38,7 @@ def load_color_scheme():
     global color_scheme, printable
     #
     # We want to colorize the hex characters and only print out
-    # printable values on the righ hand side.
+    # printable values on the right hand side.
     #
     color_scheme = {i: H.normal("%02x" % i) for i in range(256)}
     printable = {i: H.normal(".") for i in range(256)}
@@ -73,32 +70,37 @@ def hexdump(
 ):
     if not color_scheme or not printable:
         load_color_scheme()
+
+    # If there's nothing to print, just print the offset and address and return
+    if len(data) == 0:
+        yield H.offset("+%04x " % len(data)) + H.address("%#08x  " % (address + len(data)))
+
+        # Don't allow iterating over this generator again
+        return
+
     data = list(bytearray(data))
-    base = address
     last_line = None
     skipping = False
-    for i, line in enumerate(groupby(data, width, -1)):
+    for i, line in enumerate(groupby(width, data, fill=-1)):
         if skip and line == last_line:
-            if not skipping:
-                skipping = True
-                yield "..."
-            continue
+            if skipping:
+                continue
+
+            skipping = True
+            yield "..."
         else:
             skipping = False
             last_line = line
 
         hexline = []
 
-        if address:
-            hexline.append(H.offset("+%04x " % ((i + offset) * width)))
-
-        hexline.append(H.address("%#08x  " % (base + (i * width))))
+        hexline.append(H.offset("+%04x " % ((i + offset) * width)))
+        hexline.append(H.address("%#08x  " % (address + (i * width))))
 
-        for group in groupby(line, group_width):
-            group_length = len(group)
+        for group in groupby(group_width, line):
             group = reversed(group) if flip_group_endianess else group
             for idx, char in enumerate(group):
-                if flip_group_endianess and idx == group_length - 1:
+                if flip_group_endianess and idx == group_width - 1:
                     hexline.append(H.highlight_group_lsb(color_scheme[char]))
                 else:
                     hexline.append(color_scheme[char])
@@ -106,22 +108,9 @@ def hexdump(
             hexline.append(" ")
 
         hexline.append(H.separator("%s" % config_separator))
-        for group in groupby(line, group_width):
+        for group in groupby(group_width, line):
             for char in group:
                 hexline.append(printable[char])
             hexline.append(H.separator("%s" % config_separator))
 
-        yield ("".join(hexline))
-
-    # skip empty footer if we printed something
-    if last_line:
-        return
-
-    hexline = []
-
-    if address:
-        hexline.append(H.offset("+%04x " % len(data)))
-
-    hexline.append(H.address("%#08x  " % (base + len(data))))
-
-    yield "".join(hexline)
+        yield "".join(hexline)

tests/test_hexdump.py

@@ -18,7 +18,7 @@ def run_tests(stack, use_big_endian, expected):
 
     # Test empty hexdump
     result = gdb.execute("hexdump 0", to_string=True)
-    assert result == "0x000000  \n"
+    assert result == "+0000 0x000000  \n"
 
     results = []
     # TODO: Repetition is not working in tests