Sergey Sharybin
d3b6117068
Callback handle was assigning to wrong storage in RE_draw_lock_cb. So far it was completely harmless because all the callbacks are using the same handle, so test_break_callback was using correct handle, and since draw_lock_callback didn't use handle at all nobody noticed this. But this typo lead to draw_lock_callback using NULL instead of real RenderJob, which is bad for the feature we're working in GSoC branch.