I seem to have found a combinatorial bug with the following three options using STAR_2.4.0j_modified (Linux_x86_64_static binary).
If I specify all three of the following options I get an immediate segfault upon getting to "Started mapping" in the Log (all aligned output files are empty).
--outSAMtype BAM SortedByCoordinate
--outSAMattributes All
--quantMode TranscriptomeSAM
Any two of the above three run fine. The segfault appears only with all three together. (Or I am just going crazy at this point).
Thank you,
Alexander Stoddard