Cleanup API for file and non-special URLs #224
Conversation
This: * Stops the username/password/port APIs from functioning when host is the empty string. * Makes the host/hostname APIs work better with file URLs and adjusts the file host state accordingly. * Make setting host/hostname to the empty string impossible when they have a username/password/port. * Fixes #97.
Also file URL host parsing. See whatwg/url#224 for the corresponding URL Standard change.
|
This is good but the third bullet ( |
|
|
||
| <ol> | ||
| <li><p>If <var>state override</var> is given and <var>url</var> | ||
| <a>includes credentials</a>, <a>syntax violation</a>, return. |
There was a problem hiding this comment.
Is "return" the same as "terminate this algorithm" elsewhere?
There was a problem hiding this comment.
Yeah. I'm slowly switching to Infra terminology.
There was a problem hiding this comment.
I think the inconsistency is pretty confusing; when I implemented in whatwg-url I accidentally translated "return" as "bail out of the entire URL parsing algorithm".
There was a problem hiding this comment.
Isn't that what's happening?
There was a problem hiding this comment.
No; it's bailing out of this particular state and back into the main parser loop. I.e. it goes back to step 11 ("Keep running the following state machine by switching on state. If after a run pointer points to EOF code point, go to the next step. Otherwise, increase pointer by one and continue with the state machine.")
There was a problem hiding this comment.
Err, nevermind, that is what's happening. I guess what I meant is that I translated this as "bail out of this step and go back to step 11". Basically I saw two different types of bail-outs, and since I knew that "abort these steps" meant the one, I assumed that "return" meant the other one. I think it's best to change them all at once or not at all.
|
When updating the setters, unfortunately we need to update HTML's location and HTMLHyperlinkElementUtils as well, usually... :-/ |
|
@jasnell yeah, it's mostly an edge case I wanted to take care of to avoid messing up serialization. In particular creating a serialization that could not be parsed into the same components. |
Also file URL host parsing. See whatwg/url#224 for the corresponding URL Standard change.
This change depends on whatwg/url#224 landing first. Tests: web-platform-tests/wpt#4696. Location API bits are not tested since we generally treat loading from file URLs as out-of-scope of the web platform, but accounting for it standards-wise seemed trivial enough.
|
It occurs to me this might be cleaner and require fewer changes in the dependent APIs if you still went to the host/hostname state, but as the first step it just switched to the file host state if the scheme is file and state override is given. |
|
Would you prefer that? I have to say I have very mixed feelings about the whole state override setup. |
|
I think I would. I found the state override setup confusing at first but now that I understand how it works it seems fine. And given that, preserving the idea that we can just do the API stuff by directly moving into the state override makes sense to me. |
|
Done. I also verified that the tests still pass with whatwg-url. |
|
LGTM but still would prefer not starting to use "return" without switching over everything all at once. |
Also file URL host parsing. See whatwg/url#224 for the corresponding URL Standard change.
This change builds on whatwg/url#224. Tests: web-platform-tests/wpt#4696. Location API bits are not tested since we generally treat loading from file URLs as out-of-scope for the web platform, but accounting for it standards-wise seemed trivial enough.
- Update to spec
- Add opaque hosts
- File state did not correctly deal with lack of base URL
- Cleanup API for file and non-special URLs
- Allow % and IPv6 addresses in non-special URL hosts
- Use specific names for percent-encode sets
- Add empty host concept for file and non-special URLs
- Clarify IPv6 serializer
- Fix existing mistakes
- Add missing ':' to forbidden host code point list.
- Correct IPv4 parser empty label behavior
- Maintain type equivalence in URLContext with spec
- scheme, username, and password should always be strings
- host, port, query, and fragment may be strings or null
- Align scheme state more closely with the spec
- Make sure the `special` variable is always synced with
URL_FLAG_SPECIAL.
PR-URL: nodejs#12523
Fixes: nodejs#10608
Fixes: nodejs#10634
Refs: whatwg/url#185
Refs: whatwg/url#225
Refs: whatwg/url#224
Refs: whatwg/url#218
Refs: whatwg/url#243
Refs: whatwg/url#260
Refs: whatwg/url#268
Reviewed-By: James M Snell <jasnell@gmail.com>
Reviewed-By: Daijiro Wachi <daijiro.wachi@gmail.com>
Reviewed-By: Joyee Cheung <joyeec9h3@gmail.com>
- Update to spec
- Add opaque hosts
- File state did not correctly deal with lack of base URL
- Cleanup API for file and non-special URLs
- Allow % and IPv6 addresses in non-special URL hosts
- Use specific names for percent-encode sets
- Add empty host concept for file and non-special URLs
- Clarify IPv6 serializer
- Fix existing mistakes
- Add missing ':' to forbidden host code point list.
- Correct IPv4 parser empty label behavior
- Maintain type equivalence in URLContext with spec
- scheme, username, and password should always be strings
- host, port, query, and fragment may be strings or null
- Align scheme state more closely with the spec
- Make sure the `special` variable is always synced with
URL_FLAG_SPECIAL.
PR-URL: #12523
Fixes: #10608
Fixes: #10634
Refs: whatwg/url#185
Refs: whatwg/url#225
Refs: whatwg/url#224
Refs: whatwg/url#218
Refs: whatwg/url#243
Refs: whatwg/url#260
Refs: whatwg/url#268
Reviewed-By: James M Snell <jasnell@gmail.com>
Reviewed-By: Daijiro Wachi <daijiro.wachi@gmail.com>
Reviewed-By: Joyee Cheung <joyeec9h3@gmail.com>
- Update to spec
- Add opaque hosts
- File state did not correctly deal with lack of base URL
- Cleanup API for file and non-special URLs
- Allow % and IPv6 addresses in non-special URL hosts
- Use specific names for percent-encode sets
- Add empty host concept for file and non-special URLs
- Clarify IPv6 serializer
- Fix existing mistakes
- Add missing ':' to forbidden host code point list.
- Correct IPv4 parser empty label behavior
- Maintain type equivalence in URLContext with spec
- scheme, username, and password should always be strings
- host, port, query, and fragment may be strings or null
- Align scheme state more closely with the spec
- Make sure the `special` variable is always synced with
URL_FLAG_SPECIAL.
PR-URL: #12523
Fixes: #10608
Fixes: #10634
Refs: whatwg/url#185
Refs: whatwg/url#225
Refs: whatwg/url#224
Refs: whatwg/url#218
Refs: whatwg/url#243
Refs: whatwg/url#260
Refs: whatwg/url#268
Reviewed-By: James M Snell <jasnell@gmail.com>
Reviewed-By: Daijiro Wachi <daijiro.wachi@gmail.com>
Reviewed-By: Joyee Cheung <joyeec9h3@gmail.com>
- Update to spec
- Add opaque hosts
- File state did not correctly deal with lack of base URL
- Cleanup API for file and non-special URLs
- Allow % and IPv6 addresses in non-special URL hosts
- Use specific names for percent-encode sets
- Add empty host concept for file and non-special URLs
- Clarify IPv6 serializer
- Fix existing mistakes
- Add missing ':' to forbidden host code point list.
- Correct IPv4 parser empty label behavior
- Maintain type equivalence in URLContext with spec
- scheme, username, and password should always be strings
- host, port, query, and fragment may be strings or null
- Align scheme state more closely with the spec
- Make sure the `special` variable is always synced with
URL_FLAG_SPECIAL.
PR-URL: #12523
Fixes: #10608
Fixes: #10634
Refs: whatwg/url#185
Refs: whatwg/url#225
Refs: whatwg/url#224
Refs: whatwg/url#218
Refs: whatwg/url#243
Refs: whatwg/url#260
Refs: whatwg/url#268
Reviewed-By: James M Snell <jasnell@gmail.com>
Reviewed-By: Daijiro Wachi <daijiro.wachi@gmail.com>
Reviewed-By: Joyee Cheung <joyeec9h3@gmail.com>
- Update to spec
- Add opaque hosts
- File state did not correctly deal with lack of base URL
- Cleanup API for file and non-special URLs
- Allow % and IPv6 addresses in non-special URL hosts
- Use specific names for percent-encode sets
- Add empty host concept for file and non-special URLs
- Clarify IPv6 serializer
- Fix existing mistakes
- Add missing ':' to forbidden host code point list.
- Correct IPv4 parser empty label behavior
- Maintain type equivalence in URLContext with spec
- scheme, username, and password should always be strings
- host, port, query, and fragment may be strings or null
- Align scheme state more closely with the spec
- Make sure the `special` variable is always synced with
URL_FLAG_SPECIAL.
PR-URL: #12523
Fixes: #10608
Fixes: #10634
Refs: whatwg/url#185
Refs: whatwg/url#225
Refs: whatwg/url#224
Refs: whatwg/url#218
Refs: whatwg/url#243
Refs: whatwg/url#260
Refs: whatwg/url#268
Reviewed-By: James M Snell <jasnell@gmail.com>
Reviewed-By: Daijiro Wachi <daijiro.wachi@gmail.com>
Reviewed-By: Joyee Cheung <joyeec9h3@gmail.com>
Do not allow username,password,port for URLs without a host or file URLs Imports tests from web-platform-tests/wpt@0e6a90f and tracks URL spec change at whatwg/url#224 <!-- Reviewable:start --> --- This change is [<img src="https://reviewable.io/review_button.svg" height="34" align="absmiddle" alt="Reviewable"/>](https://reviewable.io/reviews/servo/rust-url/411) <!-- Reviewable:end -->
This change builds on whatwg/url#224. Tests: web-platform-tests/wpt#4696. Location API bits are not tested since we generally treat loading from file URLs as out-of-scope for the web platform, but accounting for it standards-wise seemed trivial enough.
This:
the empty string.
the file host state accordingly.
have a username/password/port.