Joseph Anthony Pasquale Holsten

Great, I'll see how they're executing. I was a little afraid of entering n levels of encoding hell, so I just pass arguments along to `execvp`, see https://github.com/discoteq/flock/blob/master/src/flock.c#L304 I'm not...

Ah, they're being "smart" and passing along to a subshell: https://github.com/karelzak/util-linux/blob/master/sys-utils/flock.c#L239 Can you find one other person to agree this is the appropriate behaviour? I'm torn.

Looks like it was added back in https://github.com/karelzak/util-linux/commit/baf39af15b1fe8570e6430788ec7cd4959fbc5d9#diff-1219b3b65dcec8078370f245769671b1R193 And the only explanation in [its NEWS](https://github.com/karelzak/util-linux/blob/baf39af15b1fe8570e6430788ec7cd4959fbc5d9/NEWS#L3) is: > * flock: replaced with flock-2.0.2 by H. Peter Anvin

maybe if it's a single string, pass to a subshell; if it's multiple, exec directly? I feel like something has similar behaviour but I can't remember what.

ah, well that use case convinces me. I'd usually [this pattern](https://github.com/discoteq/flock/blob/master/man/flock.1.ronn#L78) for that use case, but that's obnoxious for one-offs. ok, I'll look into implementing this.

so sorry, I dropped this right on the floor. I know I can't get to this till at least this evening, so here are my notes: At the moment, `-c`...

If this is interesting to you, please review the implementation in #23

A brief search shows me this is at least correct for Linux. Need to confirm this is correct for illumos and bsd system types. @mohammad-akhlaghi want to draft the changeset...

Well, I don’t love putting this into the build script, but fixed in an ugly way is better than broken! I’m sorry I’m not really able to draft this in...

Ugh. I obviously want the spirit of this PR, but not this PR itself. Perhaps GitHub actions?