Explicitly check optional arguments if we are NOT loading FASTQ, and warn or throw.

[fnothaft]

LMK your thoughts here. I don't feel strongly one way or the other about adding this code in, but I thought it might be a useful check to have.

[ryan-williams]

I see what you mean now: a consequence of transform being the kitchen-sink that it is (which I am in favor of for now, fwiw; maybe "swiss-army-knife" is the better metaphor :) ) is that there are lots of possible subsets of the available cmdline flags that already don't make sense in the presence of each other.

I think that the cases you're guarding against here are worth checking but are also just a few of the cases that we maybe should check against (from a quick survey of all the available options to transform) if we want transform to have this contract with the user.

So, I'm inclined to think of this as work for another PR: "validate transform args to throw when a user passes inconsistent args".

Let me know if you think I'm mistaken about the existence of this problem predating bigdatagenomics#831 or if you think that the args I've introduced there are pushing us over the edge of needing to do this kind of validation, either just to the flags I've added or to all the flags to transform.

[fnothaft]

Agree with 100% of what you said! Let's push this to another PR; I've filed an issue: bigdatagenomics#841.

[fnothaft]

Closing in favor of bigdatagenomics#841.