@foone I bet they had a debug build that did something fun inside that branch. Maybe they put some sentinel value there to track which caller of malloc was responsible.
Not sure why their compiler didn't optimize the branch away in the release build, though