Undocumented :translate-nl option to Str.encode

[zoffixznet]

Added to Rakudo in rakudo/rakudo@0a2eef8 but isn't documented.

[jnthn]

For what it's worth, it's not in roast either yet - which in turn reflects that I'm still pondering whether it's the right name and the right peg to hang it on.

[zoffixznet]

It's kinda long to type :)

[jnthn]

I don't expect it to be typed often. :-) I was wondering if translate-crlf would be more clear what it's about.

[jnthn]

By the way, the reason I don't expect it to be typed often is because most encoding happens at I/O boundaries, and we can make good defaults for 99% of cases (we already do so). So this would mostly be relevant to folks who are implementing their own I/O-like things. Not entirely common.

[JJ]

Still not in roast. Would it be worth the while to document it now, indicating it's Rakudo-specific?

[jnthn]

I think, having had time to get used to the name, we can roast it. We still need to do an audit of the various IO types where it might be used and make sure they expose it and pass it along to the decoders and encoders. I don't know if on handles or similar where you can read and write we should expose an in and out form. But for the decoder case of it as being discussed here, then it's pretty clear there's only one thing it's talking about.

[JJ]

OK. I can do the roast test, and then come back here to add the documentation. Anyone else wants to do it, feel free.

[JJ]

It's now in roast after above PR was accepted, so I'll proceed to document it.