Degree symbol breaks URI detector

Question

show that the URI detector breaks when it encounters a degree symbol (°) or U+00B9:

The link was copied and pasted from the browser's address field.

I'm not sure if RFC 3986 has been superseded or not, but by that standard, technically, URLs are still supposed to be US-ASCII. Support for any other characters is a feature of the client. This URL can be represented as https://www.google.com.pe/maps/place/3%C2%B008%2706.5%2522S%2B59%C2%B054%2708.7%2522W/%40-3.1351316%2C-59 — choster, Commented Mar 6, 2016 at 15:45
That should be https://www.google.com.pe/maps/place/3%C2%B008%2706.5%2522S%2B59%C2%B054%2708.7%2522W/%40-3.1351316%2C-59.9374329%2C14z rather. — choster, Commented Mar 6, 2016 at 15:52
With maps.google creating that in links, and many browsers supporting it, I think we should code to the "real world." It's not a "standard" if the majority aren't following it! — WGroleau, Commented Mar 6, 2016 at 21:12
7320 updates 3986, and suggests supporting anything allowed by HTML 4.01 (which still isn't clear on this). datatracker.ietf.org/doc/rfc7320/?include_text=1 — WGroleau, Commented Mar 6, 2016 at 21:27
I can't quite agree with the idea that we should intentionally create a link that doesn't work with the rationale that the one which DOES work (and is human-readable) is non-standard. — WGroleau, Commented Mar 13, 2016 at 4:47

Community · Accepted Answer · 2017-03-20 10:29:48Z

4

The gist of the highest voted answer is:

The URL parser is correctly not including the non-encoded diamond character in the URL, because valid URLs simply cannot include them.

There might be a way of arguing for RFC 7320 updating 3986.

CommunityBot

1

answered Mar 11, 2016 at 11:21

JoErNanOMod

52.8k19 silver badges45 bronze badges

1

Engineers: the bureaucrats of making stuff.
– choster
Commented Oct 18, 2016 at 17:54

Add a comment |

1 Answer 1