hapax: PeterL: tried playing with your v implementation, but before could even begin i noticed litmus.sh couldn't process your seals, gpg --verify did just fine

hapax: my first thought was something broke when making it posix-compatible

hapax: but then i realized your signature headers were longer (guess you used some newer gpg?)

hapax: the culprit was the busybox xxd(1) implementation, lines 279 and 291, -p essentialy truncates to 30 cols unless set manually with -c

hapax: re v itself: is there a particular reason you used unix patch(1) instead of phf's vpatch from his vtools? i believe the risk is the same as with the xxd in litmus.sh

hapax: asciilifeform: you may find the xxd case interesting, guess it's another reason why shellscripting is a bad idea in general