Re: [whatwg/url] Allow protocol setter to switch between special and non-special schemes (Issue #674)

The current algorithm we use is as follows ( See URL::setProtocol at https://trac.webkit.org/browser/webkit/trunk/Source/WTF/wtf/URL.cpp#L390 ):

1. Take the value and remove any colon and anything after it
2. If the value is empty, does not start with an ASCII alpha character, or contains anything but ASCII alpha characters, '+', '-', or '.' (currently implemented in URLParser::maybeCanonicalizeScheme), return the unchanged URL.
3. If the "URL" is an invalid URL (an HTML anchor tag) then try parsing value then ':' then the invalid URL string (this step is strange and seems worth removing and not standardizing)
4. If value is "file" and the URL has credentials or a port then return the unchanged URL. (This is an implementation of https://github.com/whatwg/url/pull/269 in a different place)
5. If the URL's protocol is already "file" and has a host, return the unchanged URL. (This seems related to step 4, but I'm not sure what the history is.)
6. Serialize the current URL to a string, remove all characters before the first colon, then parse value with a colon and that string.

It seems like there are restrictions around "file", reparsing the whole URL, and removal of stuff after the colon that are common to all browsers.

-- 
You are receiving this because you are subscribed to this thread.
Reply to this email directly or view it on GitHub:
https://github.com/whatwg/url/issues/674#issuecomment-977544481