@sir the example has two cases of "case 'c':", is this intended?

@sir I find them preferable in scripts because they self-document.

@wno @sir a common misbelief that word-like identifiers, as opposed to single letters, do, in fact, document their own semantics. 😎

@sir *slow claps* the impedance mismatch with go's defaults is cripples my unix-fu every time

@sir Thinking long options are wrong is one thing, but wouldn't it at least have made sense to keep the structure of the flag package? Using the c-ish getops structure is very unlike Go.

@sir Only thing I don't like about POSIX getopt is that it doesn't allow option args after non-option args

@rms I appreciate that tbh, makes all commands consistent

@sir yeh, but usability wise allowing them makes more sense.
Imagine entering ls command and realizing you forgot the -l flag, fixing that would take a few extra key strokes even in vi mode.
Though I agree that we're kind of stuck with whatever is written in POSIX spec.

@rms was reminded of this conversation just now. Consider the following based on POSIX behavior:

xargs -0 rm -f

And based on your suggestion:

xargs -0 -- rm -f

Both of them works on my GNU :^)
Seriously though, I get your point. It does make things consistent, that's the whole point of having a standard.

@sir I was going to look for something like this later, glad to see it pop up on my feed instead o/

Sign in to participate in the conversation

cmpwn.com is a private Mastodon instance for friends of SirCmpwn.