Just tidying up bits of indexing in #CE8SSJV3N github.com/nf-core/rnaseq/pull/1958

Sorry, I thought I could review it, but in the end I don't know enough the pipeline to review it. Even though on the code aspect, I don't see anything wrong.

It's just a case of preventing redundant indexing. There was already a Salmon indexing step at the top, but we were still allowing the inner workflow to do it sometimes, and for sortMeRNA it was actually indexing twice sometimes. So just making sure those don't happen any more

This does make sense. I don't quite get however, why do you need specifically salmon index only if you have auto-strandness ?

So, Salmon indexing is activated automatically when it's used as the pseudoaligner, which should make sense (you'll see that in the code but using the variable 'pseudoaligner' rather than the literal salmon string. But Salmon is also used (whatever the pseudoaligner setting) within the preprocessing subworkflow to 'sniff' the strandedness of data when users supply 'auto' as strandedness. When both are the case (salmon alignment and auto strand) 'salmon' will actually get added twice to the list of indexing tools, but we don't care since we just check for presence of the string