Calling schema_dsl() with a field that has no name before the colon crashes with IndexError: list index out of range.
Repro:
from llm.utils import schema_dsl
schema_dsl(':just a description') # IndexError
schema_dsl('name, :description') # IndexErrorRoot cause (llm/utils.py line 394-395): after field.split(":", 1), field_info can be an empty string. field_info.strip().split() then returns [], so field_parts[0] raises IndexError.
The fix is to continue when field_parts is empty, silently skipping the nameless field — consistent with how blank field strings are already filtered out during field splitting.
Calling
schema_dsl()with a field that has no name before the colon crashes withIndexError: list index out of range.Repro:
Root cause (
llm/utils.pyline 394-395): afterfield.split(":", 1),field_infocan be an empty string.field_info.strip().split()then returns[], sofield_parts[0]raisesIndexError.The fix is to
continuewhenfield_partsis empty, silently skipping the nameless field — consistent with how blank field strings are already filtered out during field splitting.