https://github.com/w3c/webappsec/commit/c992d5cd9c93eaa509ee499efd7ef8f5ab9811d8 Seems that origin-when-crossorigin will be replaced by origin-when-cross-origin. However many web platform tests currently (and current public version) are using origin-when-crossorigin. origin-when-crossorigin value should be treated as valid value as well.

Discussed internally and we thought that the wpt should use latest and correct value "origin-when-cross-origin". What we should do is to change this value correctly in referrer-policy web platform test.

Franziskus, just making sure you have seen this bug.

I would like to add some additional information: There's a typo issue in specs (editor's reply in [1], [2], should be Mike). Old version specs [3] and new one [4] should be public soon `origin-when-cross-origin` is correct but Chrome and Firefox already ship `origin-when-crossorigin` and sites already used it In current version, Google supports both of `origin-when-crossorigin` and `origin-when-cross-origin` (I tried with Google Canary and web platform tests are passed with both of them). However, running web platform tests in Firefox still fails in this sort of test (~28 cases failed with setAttribute), because `origin-when-crossorigin` is still considered as invalid value attribute (Bug 1178337 has just fixed recently but only added `origin-when-cross-origin` to the valid enums) [1] https://lists.w3.org/Archives/Public/public-webappsec/2015May/0064.html [2] https://lists.w3.org/Archives/Public/public-webappsec/2015May/0091.html [3] https://www.w3.org/TR/referrer-policy/#referrer-policy-delivery-meta [4] https://w3c.github.io/webappsec-referrer-policy/ Currently referrer policy web platform tests are written based on the old one `origin-when-crossorigin` I discussed with Franziskus and there're 2 ways to go: - Support `origin-when-crossorigin` in Firefox - Fix web platform test to use `origin-when-cross-origin` because `origin-when-crossorigin` will be deprecated soon It makes more sense to go second one. And Franziskus should know about that :) Thanks

Hi Christoph, Could you please take a look into that? This renamed `origin-when-crossorigin` --> `origin-when-cross-origin` in web platform tests and updated meta file as well

Comment on attachment 8750154 [details] [diff] [review] Change `crossorigin` to `cross-origin` Franziskus is the right reviewer for this.

Comment on attachment 8750154 [details] [diff] [review] Change `crossorigin` to `cross-origin` Review of attachment 8750154 [details] [diff] [review]: ----------------------------------------------------------------- I still see some |crossorigin| in spec_json.js. r+ with those fixed as well.

Thanks Franziskus for your reviewing

Try looks good https://treeherder.mozilla.org/#/jobs?repo=try&revision=eb1005956bb3

(In reply to Thomas Nguyen[:tnguyen] from comment #11) > Try looks good > https://treeherder.mozilla.org/#/jobs?repo=try&revision=eb1005956bb3 what's up with all the oranges on Win7?

Something went wrong with windows 7 recently. You could see all of try results have the same result with many failures on Windows 7(click: go next 50)

Hi Josh, could you please review this? This only renamed `origin-when-crossorigin` --> `origin-when-cross-origin` in web platform tests and updated meta file as well.

Comment on attachment 8753752 [details] [diff] [review] Change `crossorigin` to `cross-origin` Review of attachment 8753752 [details] [diff] [review]: ----------------------------------------------------------------- This appears to be racing with https://github.com/w3c/web-platform-tests/pull/3059, so we'll see what happens!

Try result in comment 11

Thanks Josh and Franziskus for your review.

https://hg.mozilla.org/integration/mozilla-inbound/rev/d67e5ce76bf5 https://hg.mozilla.org/integration/mozilla-inbound/rev/54be367a7ed9

for some reasons starting with this push we have perma failures like https://treeherder.mozilla.org/logviewer.html#?job_id=28628738&repo=mozilla-inbound so i had to back this out

Backout: https://hg.mozilla.org/integration/mozilla-inbound/rev/51b499637697 https://hg.mozilla.org/integration/mozilla-inbound/rev/2adb4dd6324d

https://hg.mozilla.org/integration/mozilla-inbound/rev/9a7ac883cb35 https://hg.mozilla.org/integration/mozilla-inbound/rev/07ede99a3cd3

did some investigation and seems your patch is innocent (sorry for the backout) and all the timeouts where more caused by bug 1275281 - sorry for the backout again - re-checked this in

Nice, thanks

https://hg.mozilla.org/mozilla-central/rev/9a7ac883cb35 https://hg.mozilla.org/mozilla-central/rev/07ede99a3cd3