Variable Pool Falsy Value Handling#

Two related bugs caused valid falsy values (False, 0, empty strings) to be silently dropped during workflow variable resolution. Both share the same root cause — implicit Python truthiness checks used in place of explicit None checks — and were fixed in separate PRs.

Bug 1: VariablePool._get_nested_attribute() drops falsy nested values#

Affected component: api/core/workflow/entities/variable_pool.py (now in graphon external package)

When resolving a nested attribute on an ObjectSegment (e.g., some_object.count where count = 0), the original _get_nested_attribute() returned None for any falsy value because it relied on obj.get(attr) whose falsy return was indistinguishable from a missing key 1.

Fix (PR #26155): Replace the single isinstance guard with a two-condition check:

# Before
if not isinstance(obj, dict):
    return None
return obj.get(attr)

# After
if not isinstance(obj, dict) or attr not in obj:
    return None
return variable_factory.build_segment(obj.get(attr))

The attr not in obj membership test explicitly distinguishes "key is absent" from "key exists with a falsy value." Additionally, variable_factory.build_segment() was updated to short-circuit when the input is already a Segment instance 1, preventing double-wrapping during extraction.

Tests covering None, "", 0, and False were added at api/tests/unit_tests/core/workflow/entities/test_variable_pool.py 1.

Bug 2: WorkflowEntry.mapping_user_inputs_to_variable_pool() drops falsy user inputs#

Affected component: api/core/workflow/workflow_entry.py — used during single-step (debugger) node execution.

The input resolution falls back from a full key (node_id.variable_key) to a short key (variable_key). The fallback was gated on a truthiness check 2:

input_value = user_inputs.get(node_variable)
if not input_value: # BUG: also triggers for False, 0, ""
    input_value = user_inputs.get(node_variable_key)
if input_value is None:
    continue

When user_inputs["node.enabled"] = False, not input_value is True, so the code falls through to the short-key lookup (user_inputs.get("enabled")). If that key is also absent, input_value becomes None and the continue skips writing the value to the pool — even though the caller explicitly provided False 2.

Fix (PR #25908): Change the fallback guard to an explicit None check:

input_value = user_inputs.get(node_variable)
if input_value is None: # FIXED: only falls back when key is truly absent
    input_value = user_inputs.get(node_variable_key)

The current mapping_user_inputs_to_variable_pool() implementation uses the is None fallback pattern alongside an is None guard before continue 3. Note that a separate variable_pool.get() truthiness guard at line 563 still uses not variable_pool.get(variable_selector), which may silently pass for pools containing falsy values — a remaining risk area 4.

The Recurring Pattern#

Both bugs recur easily because Python developers habitually use truthiness checks as "absent" guards. If you encounter a workflow where 0, False, or "" values disappear mid-execution, audit any if not value: or if value: guard that gates variable pool writes or reads.

⚠️ Regression risk: Issue #38963 confirmed that as of commit 40df83de, line 579 of workflow_entry.py still contained if not input_value: — either a regression or a missed code path from PR #25908 2.

Key Source References#