diff --git a/gdb-port/parser.py b/gdb-port/parser.py
index 766210dd62d76115cb51c59bef8e957786b9ebf8..34d8e29c97cb06c28453473f22c0a674a9790a7c 100644
--- a/gdb-port/parser.py
+++ b/gdb-port/parser.py
@@ -95,7 +95,6 @@ class ParserStack:
 		self.bytes_at_enter = []
 		self.cumulative_byte_differences = []
 		self.total_byte_differences = []
-		self.bytes_since_prev_h_do_parse = []
 		self.stack_events = []
 		# Represents the index of a "waterline" on the stack_events for the purpose of counting allocations
 		# This is the index of the last event (stack push or pop) where the relevant allocations have been committed to TopLevelParse's memory stats.
@@ -153,7 +152,6 @@ class ParserStack:
 			# But the user doesn't really care, so going this route requires keeping track of allocations that would be committed,
 			# and presenting it a unified manner at the UI layer
 			# Additionally, it would probably make more sense to store it at the ParserStack, storing (frame, bytes) pairs
-			self.bytes_since_prev_h_do_parse.append(thusfar)
 			# TODO: where should commit_at_push() be called? HDoParseBreakpoint determines whether to stop, which ParserStack doesn't know about. could be a bool parameter to push()
 			if should_commit:
 				self.commit_at_push()