has_subblock() should check for end of block

See #17 (comment)

README.md

@@ -18,6 +18,8 @@ New features:
 Other changes:
 - The `value` attribute of scene item blocks, which was not being used, has been
   removed.
+- Check more carefully for sub-blocks
+  ([#17](https://github.com/ricklupton/rmscene/issues/17#issuecomment-1701071477)).
 
 ### v0.4.0
 

src/rmscene/tagged_block_reader.py

@@ -175,6 +175,12 @@ def read_subblock(self, index: int) -> Iterator[SubBlockInfo]:
 
     def has_subblock(self, index: int) -> bool:
         """Check if a subblock with the given index is next."""
+        # It's possible that if we are at the end of the block, the next bytes
+        # (the size of the next block) could happen to match the tag and index
+        # of what we are looking for -- so check explicitly for end of block.
+        if self.current_block:
+            if self.bytes_remaining_in_block() <= 0:
+                return False
         return self.data.check_tag(index, TagType.Length4)
 
     def _check_position(self, block_info: BlockInfo):

tests/test_tagged_block_reader.py

@@ -159,3 +159,21 @@ def test_read_string_utf():
     s = stream("1c05000000" "030161c397")
     result = s.read_string(1)
     assert result == "a×"
+
+
+def test_has_subblock_checks_for_end_of_block():
+    # See https://github.com/ricklupton/rmscene/issues/17#issuecomment-1701071477
+    #
+    # Construct some potentially confusing data -- the 0x2c is the start of the
+    # next block, but if we don't take care, `has_subblock(2)` could see it as a
+    # subblock instead.
+    data_hex = """
+    03000000 00010103
+    1f 0219
+    2c000000 00010100
+    """
+
+    s = stream(data_hex)
+    with s.read_block():
+        assert s.read_id(1)
+        assert s.has_subblock(2) == False