@xyhhx @RustyBertrand these day (for some fields) you can likely find the per-print on arXiv and it likely has better formatting.

But if the paper is not on arXiv, scihub is definitely the moral (if not legal) choice.