Change URL parsing sequence for window open steps

[domenic]

Previously, we would only check at the last minute. This would be buggy in the case where we create a new navigable, and then throw an exception. Instead we should be sure to throw before creating a new navigable.

In addition to fixing this somewhat-obviously-wrong bug, this also causes URL parsing to happen even in cases where a named window is targeted, but no window with that name exists. Previously, the window open steps would do nothing in such cases; now, they throw an exception when given an unparseable URL.

Closes #10681.

@shannonbooth, your review would be appreciated. Also, are you interested in helping to write web platform tests for the normative change described above, about adding validation? I checked by source inspection that it matches how Chromium behaves, but depending on the results we might need to do more work to match the browser majority here.

• At least two implementers are interested (and none opposed):
  • This matches Chromium and Gecko from code inspection (but not WebKit)
• Tests are written and can be reviewed and commented upon at:
  • Untestable, per Add ability to test with popup blocker enabled web-platform-tests/wpt#3867
• Implementation bugs are filed:
  • Chromium: N/A
  • Gecko: N/A
  • WebKit: https://bugs.webkit.org/show_bug.cgi?id=283153
  • Deno (only for timers, structured clone, base64 utils, channel messaging, module resolution, web workers, and web storage): N/A
  • Node.js (only for timers, structured clone, base64 utils, channel messaging, and module resolution): N/A
• MDN issue is filed: N/A, too minor
• The top of this comment includes a clear commit message to use.

(See WHATWG Working Mode: Changes for more details.)

---

/nav-history-apis.html ( diff )

[annevk]

I hope this works out testing-wise as this seems like a very good change.

[shannonbooth]

The changes look good to me, and I also hope that testing on other implementations works out as the normative change is intuitive to me.

Also, are you interested in helping to write web platform tests for the normative change described above, about adding validation? I checked by source inspection that it matches how Chromium behaves, but depending on the results we might need to do more work to match the browser majority here.

I will give it a shot when I get some time! Not fully familiar with the process, but will figure it out :^)

[domenic]

I looked into this and I think this is actually not testable.

The only case where we would previously avoid URL parsing and do nothing is when the "rules for choosing a navigable" return null. This would occur "If the user agent has been configured such that in this instance it will not find a navigable" or "If currentNavigable's active window does not have transient activation and the user agent has been configured to not show popups (i.e., the user agent has a "popup blocker" enabled)".

However, neither of this configurations is possible in web platform tests. The former is basically nonexistant in modern browsers (maybe it's for like kiosk-mode browsers?) and the latter cannot be tested because the WPT infra hard-codes the popup blocker to off (web-platform-tests/wpt#3867).

So, I will go ahead and merge this.