It's true, that adding separate script on /etc/event.d/ is "cleaner", in a way that it have less chances or messing something up (not that I've heard about ereswap breaking anyone's rcS-late, ever, but there are always chances), or is just easier for manual managing. It's generally perceived, that touching default, boot-critical scripts - like rcS-late - is less "clean" way of doing things, and I absolutely agree with this.

Hoever, without modifying rcS_late, we have situation, where one script enables regular swap on eMMC, and another one disables it, just a moment later. It makes booting a little longer due to unnecessary operation being performed (although, in practice, it's hardly-to-non noticeable), but most importantly, just don't match my aesthetic, and can create problems, if one decide to go without any eMMC swap, at all.

That's why I've decided to go with the way of sed-modifying rcS-late, which was paincfully more hard to write, than just adding another script to /etc/event.d/, especially, that they way sed command was written, tries to be compatible with modified rcS-late's, too (unless the very swap lines vere touched, but in that case, one wouldn't need ereswap, probably).

After all, we have major respected packages touching boot-critical script - including things from CSSU - so why should we restrain ourselves

/Estel