Schemes list is very incomplete

[skalee]

By default, the validator allows http and https URI schemes only. However, that can be overriden by setting option :scheme:

Post.validates :url, uri_format: { scheme: /some|schemes|regexp/ }
Post.validates :url, uri_format: { scheme: :all }

The value :all stands for a predefined list of popular schemes:

uri_format_validator/lib/uri_format_validator/validators/uri_format_validator.rb

Lines 16 to 25 in 605d896

	SCHEMES = %w[
	aaa aaas about acap acct cap cid coap coaps crid data dav dict dns
	example file ftp geo go gopher h323 http https iax icap im imap info ipp
	ipps iris iris.beep iris.lwz iris.xpc iris.xpcs jabber ldap mailto mid
	msrp msrps mtqp mupdate news nfs ni nih nntp opaquelocktoken pkcs11 pop
	pres reload rtsp rtsps rtspu service session shttp sieve sip sips sms
	snmp soap.beep soap.beeps stun stuns tag tel telnet tftp thismessage tip
	tn3270 turn turns tv urn vemmi vnc ws wss xcon xcon-userid xmlrpc.beep
	xmlrpc.beeps xmpp z39.50r z39.50s
	].freeze

However, this list isn't complete. For instance, it lacks ssh URI scheme (ssh://git@github.com:riboseinc/uri_format_validator.git). And I'm worried it's not the only missing one. Any good reason to maintain the schemes list at all, @ronaldtse?

[ronaldtse]

Originally the scheme list came from the authoritative list of schemes at IANA:
https://www.iana.org/assignments/uri-schemes/uri-schemes.xhtml

I agree that scheme checks may be an overkill but non-IANA registered schemes are technically invalid on the Internet. This would also allow us the opportunity to particularly address validity of a single scheme better.

What do you think @skalee? Provide an override to this?

[skalee]

@ronaldtse If it's an authoritative source, then IMHO the option should be named accordingly. Not {scheme: :all}, but maybe {scheme: :iana}, or {scheme: :permanent}, as in example below:

validates :uri, uri_format: {scheme: :iana}

The option {scheme: :all} should allow all schemes, registered or not.

On the other hand, such authoritative list is another thing to maintain. It needs to be updated from time to time. And I see some pretty recent entries (with provisional status) like spotify:// or steam://. I'm bit worried about this burden. But maybe there's some gem which already maintains such list? I'll look for it later.

Also, I see ssh is present on that authoritative list, it has provisional (not permanent) status, that probably explain why it's missing on our short list of schemes with permanent status.