Restrict storing invalid URIs in a URI column

[Anish9901]

To Reproduce

1. Create a empty table
2. Create a URI column called books
3. Insert the following value: The Four Steps to the Epiphany: Successful Strategies for Products that Win
4. Notice that the value is stored successfully

Expected behavior

Any URI with whitespaces should not be stored.

[mathemancer]

But whitespaces are valid in URIs: https://www.rfc-editor.org/rfc/rfc3986#appendix-C . Line breaks are allowed. If we want, we could try to add some "clean up" before storing URIs, e.g., removing or escaping white space, but the RFC doesn't specify semantic meaning (or lack of it) for whitespace in URIs.

I don't think this should be in the Beta Milestone.

[Anish9901]

But whitespaces are valid in URIs: https://www.rfc-editor.org/rfc/rfc3986#appendix-C . Line breaks are allowed.

@mathemancer I read through this part of the spec, as per my understanding whitespaces are allowed while delimiting the URI in a text. If someone has some text that has URI as a part of it, wouldn't it be fair to say that such data should exist in a TEXT column instead?

The spec also recommends adding angle brackets for any URI containing embedded whitespaces. I tried this in gmail & GH, and found out that if you type a URI with an embedded whitespace with and without "<>", only the first part of the URI gets recognized. So, I would say that this is a really uncommon edge case and would be fine if we didn't support it.

In GH:
<http://www.ics.uci.edu/pub/
ietf/uri/historical.html#WARNING>

http://www.ics.uci.edu/pub/
ietf/uri/historical.html#WARNING

If we want, we could try to add some "clean up" before storing URIs

Should we be doing this, given that we never alter user's data directly?

[zackkrida]

Appendix C of the spec does not "allow whitespace in URIs". The section is solely about extracting and reconstructing URIs from surrounding text.

This part of the spec would only be relevant to Mathesar in scenarios where URIs are present in a body of text. If we wanted to, for example, make URIs navigable in a TEXT column, our URI detection code should reconstruct URIs in a manner which ignores whitespace and line breaks.