Bug 306802 – NavigateEvent#navigationType should be "replace" when navigating to a URL that matches the active document's URL

• Component: DOM
• Status: RESOLVED FIXED

[Dom Christie]

In section 7.4.2.2, the [spec](https://html.spec.whatwg.org/multipage/browsing-the-web.html#navigate-convert-to-replace) states:

> If url equals navigable's active document's URL, and initiatorOriginSnapshot is same origin with navigable's active document's origin, then set historyHandling to "replace".

When tapping a link to a url that matches the current location, the NavigateEvent#navigationType property is "push" rather than "replace". This differs to the Chrome and Firefox implementation, which returns "replace".

[Karl Dubost]

Created attachment 478220 [details]
rendering in safari, firefox, chrome

Confirmed.

[Karl Dubost]

Created attachment 478221 [details]
testcase

PASS Firefox/Chrome
FAIL Safari (STP235)

[Karl Dubost]

The spec says:
https://url.spec.whatwg.org/#concept-url-equals
=============
4.6. URL equivalence

To determine whether a URL A equals URL B, with an optional boolean exclude fragments (default false), run these steps:

1. Let serializedA be the result of serializing A, with exclude fragment set to exclude fragments.
2. Let serializedB be the result of serializing B, with exclude fragment set to exclude fragments.
3. Return true if serializedA is serializedB; otherwise false.
=============


which means that 
    foo.html#blah 
should be equivalent to 
    foo.html
and vice versa.

So when we go back to the spec, it says:
https://html.spec.whatwg.org/multipage/browsing-the-web.html#navigate-convert-to-replace

==============
12. If historyHandling is "auto", then:

1. If url equals navigable's active document's URL, and initiatorOriginSnapshot is same origin with navigable's active document's origin, then set historyHandling to "replace".
2. Otherwise, set historyHandling to "push".
==============


The current code is:

https://searchfox.org/wubkat/rev/efbda44f77fdac3adc639a9427ab93943f2b17b9/Source/WebCore/loader/FrameLoader.cpp#1613-1623
```cpp
bool isSameOrigin = frameLoadRequest.protectedRequesterSecurityOrigin()->isSameOriginDomain(document->protectedSecurityOrigin().get());
if (!isReload(newLoadType)) {
    if (historyHandling == NavigationHistoryBehavior::Auto) {
        if ((document->url() == newURL || document->readyState() != Document::ReadyState::Complete) && isSameOrigin)
            historyHandling = NavigationHistoryBehavior::Replace;
        else
            historyHandling = NavigationHistoryBehavior::Push;
    }
    if (newURL.protocolIsJavaScript() || (documentLoader() && documentLoader()->isInitialAboutBlank()))
        historyHandling = NavigationHistoryBehavior::Replace;
}
```


are document->url() and newURL comparing the same URLs?

For example, this should be equal?
document.url() "https://example.com/page#section1"
newURL "https://example.com/page"

What about the opposite?


This is covered by 
https://wpt.fyi/results/navigation-api/navigate-event/navigate-anchor-same-url.html

[Karl Dubost]

The test http://wpt.live/navigation-api/navigate-event/navigate-anchor-same-url.html
seems to be looping in STP 235.

[Radar WebKit Bug Importer]

<rdar://problem/169999046>

[Chris Dumez]

Pull request: https://github.com/WebKit/WebKit/pull/61221

[EWS]

Committed 309871@main (d56ef888827c): <https://commits.webkit.org/309871@main>

Reviewed commits have been landed. Closing PR #61221 and removing active labels.