Three possible uses for IRIs

[Sean Conner]

It was thus said that the Great colecmac at protonmail.com once stated:
> 
> I'm unsure whether that would require an IRI parser or not, but I'd feel more
> confident with one. However, there is already a client torture test that *sort of*
> covers this. It's not designed as an IRI test, but it includes invalid
> characters in a link line.
> 
> gemini://gemini.conman.org/test/torture/0031
> 
> That page contains a link line that looks like this:
> 
> => <0032> "Beware the bad link"
> 
> And the Go stdlib will actually correct this link and output a correct
> absolute one. So in Amfora, it will go to the correct URL, which is
> gemini://gemini.conman.org/test/torture/%3C0032%3E

  What you failed to quote from that test is:

	I'm not entirely sure what the proper response should be ...

  And it was a last minute thing to add the link to %3C0032%3E---I was
thinking it was more of an Easter Egg type of thing than what the actual
result should be.  

> I've set up my own test that contains a more complex Unicode character: 蛸.
> It tests the path, as well as Unicode in the query strings.
> You can access it at: gemini://makeworld.gq/test/iri-link.gmi

  I tried both the Gemini Client Torture Test 31, and your link with the
Gemini portal at portal.mozz.us.  The results were interesting.  If failed
the Gemini Client Torture Test, but loaded the page with the Unicode
character on your site.  So at least it supports percent encoding of
characters outside the ASCII range.

  -spc (So that's one more data point ... )