On the reproducibility of experiments of indexing repetitive document collections