SC2015 inconsistent behavior

[rozcietrzewiacz]

SC2015 inconsistent behavior

[koalaman]

This inconsistency was intentionally introduced in d603ee1 to reduce the noise of A && B || C when C is a single, allowlisted command (echo, printf, exit, return), since these generally aren't a problem. It does not aim to generally determine whether C is a harmless or side effect free command or structure, as this is an infinite rabbit hole.

Somewhat unrelated, the suggestion shouldn't trigger when B is a test, and this is added in 76ff702.

[rozcietrzewiacz]

@koalaman Oh 76ff702 seems definitely related - it actually solves my main issue 🤩 Thanks!

So a little background, in case someone else stumbles upon the same: I got here in the first place because of simple initial variable checks in a script, such as:

#!/bin/sh

#Check env vars
[ "$NECESSARY_ENV_VAR1" ] && [ "$NECESSARY_ENV_VAR2" ] || {
  echo "Missing correct set of environment vars (NECESSARY_ENV_VAR1 and NECESSARY_ENV_VAR2)"
  exit 1
}

...

This used to trigger a SC2015, so I dug into it and noticed that it doesn't trigger with a simpler [ "$NECESSARY_ENV_VAR1" ] && [ "$NECESSARY_ENV_VAR2" ] || exit 1... And this was inconsistent, but I agree it should rather be fixed by focusing on the nature of B in the A && B || C problem.