Recent searches
Search options
ngl https://url-parts.glitch.me/ is one of the most useful websites i've come across in a long time
, 
; Soni L. @molly0xfff apparently this cannot handle domains .example, .local, localhost, or schemes mailto: or (more generally) non-http(s).
@SoniEx2 or URLs with usernames and passwords
@anthony the auth: URI we proposed in one of the IETF mailing lists is more interesting anyway.
@SoniEx2 I think once you leave http[s] the meaning of everything past the : becomes essentially random, so maybe they just went "lets follow the whatwg url spec" specifically to match browser behaviour. Then it's a matter of just handling all the special case hostnames which I guess they don't (though part of me is "why would eTLDs etc require special handling?"). Does it handle IDN at least?
@ohunt it doesn't even follow the whatwg url spec.
we'd know, we've read it tens of times as part of making fedilinks.
@SoniEx2 @molly0xfff
It also cannot handle a dot after the domain, `https://example.com./index.html`.
@SoniEx2 @molly0xfff or IP addresses.
@molly0xfff pretty cool !
@molly0xfff the URL parser on that site is [broken](https://url-parts.glitch.me/?url=https://example.com/?q=https://example.com/?q=hi)
i'm not sure why they don't just use the browser's native URL parser: `new URL("https://example.com/?q=https://example.com/?q=hi")`
Thanks!
I can see that URL.search works with this URL, though it is quite unusual :).
I do use the URL API, but I also need to create regexes to add spans correctly. I'll have another go.
https://glitch.com/edit/#!/url-parts?path=js%2Fmain.js%3A85%3A22
@molly0xfff it seems to get easily confused:
https://url-parts.glitch.me/?url=https://reyn.ir//
Also https://reyn.ir/reyn.ir/ makes it think there is no path.
@molly0xfff It'd be great if they could stick in some text saying "everything after this particular question mark seperator is just additional tracking stuff that can be deleted". (eg: https://zapier.com/blog/how-to-clean-urls/)
@molly0xfff huh. I wonder what the point of the 'site' component is. Seems strange to pair the scheme with just the registrable domain
@molly0xfff doesn't handle IP addresses
Nice visualization, but not as complete as I'd've expected
@molly0xfff @slightlyoff Cool! It seems to be missing the "authority” part, though. Is that because it doesn't support usernames, perhaps?
