Use the URL from the response, if it has one #146

annevk · 2015-10-28T05:52:01Z

This mostly affects the way stylesheets work in service workers. See
w3c/ServiceWorker#757 for details.

annevk · 2015-10-28T05:52:16Z

@wanderview, I'd appreciate review.

wanderview · 2015-11-03T14:51:34Z

I don't actually see where Response's url list is set with the URL from a network load. Is it implied in step 4 of HTTP-network fetch?

Let response be the result of making an HTTP request over connection using request, following the relevant requirements from HTTP, and waiting until all the headers are transmitted or fetch is being terminated with reason reason. If fetch is being terminated, set response's termination reason to reason.

Since we're making this logic dependent on response url list being set, we should be explicit about where it is it I think.

annevk · 2015-11-04T10:36:52Z

Okay, so the way to solve this would indeed be to set response's url list to a copy of request's url list in HTTP-network fetch. That way only synthetic responses end up with an empty url list in main fetch.

Instead of always overriding response's url list with request's url list, even after service workers, only do so for synthetic responses. This mostly affects the way stylesheets work in service workers. See w3c/ServiceWorker#757 for details.

annevk · 2015-11-04T11:44:00Z

@wanderview should be fixed now.

I guess to make it even better I could create some sharing of code between service worker responses and network responses. Not sure where to place that though. Ideas?

wanderview · 2015-11-04T15:11:23Z

LGTM, although it would be nice if github provided a way to see the final document as a rendered html page I can view in the browser. Reading HTML diffs makes me a bit cross-eyed.

wanderview · 2015-11-05T14:47:48Z

Gecko bug https://bugzilla.mozilla.org/show_bug.cgi?id=1222008

jakearchibald · 2015-11-09T18:59:55Z

Blink bug https://code.google.com/p/chromium/issues/detail?id=553535

In #146 I tried to let service worker responses that were not synthetic to keep their url list. However, the way I fixed that regressed url lists for a number of responses: HTTP cache responses, about URL responses, etc. This fix instead tries to keep the old setup and only overwrites the url list of a response if it’s empty. Fixes #312.

wanderview · 2017-10-26T19:41:13Z

Blink bug https://code.google.com/p/chromium/issues/detail?id=553535

@mattto @horo-t Has this been implemented in chrome? This issue still says "assigned", but this glitch test suggests the final Response.url is being synthesized across to the intercepted request:

https://sw-passthrough-redirect.glitch.me/

In firefox I currently get:

==> SW fetch() got Response url:https://sw-passthrough-redirect.glitch.me/final.txt redirected:true
==> window fetch() got Response url:https://sw-passthrough-redirect.glitch.me/redirect.txt redirected:false

In chrome I get:

==> SW fetch() got Response url:https://sw-passthrough-redirect.glitch.me/final.txt redirected:true
==> window fetch() got Response url:https://sw-passthrough-redirect.glitch.me/final.txt redirected:true

Note that the window fetch() is seeing "final.txt".

Is this a fetch() only thing right now in chrome, or have you implemented it for all intercepted network requests?

I'm happy if this is implemented in chrome, because I'd like to fix:

https://bugzilla.mozilla.org/show_bug.cgi?id=1222008

wanderview · 2017-10-26T19:48:52Z

I am able to test edge 16 service workers as well, and they also show the final URL.

wanderview · 2017-11-01T14:41:59Z

@jakearchibald pointed me at this which suggests chrome is not using the synthetic Response.url for the CSS case:

https://css-redirect-test.glitch.me/

So it seems chrome is sometimes using synthetic Response.url and sometimes not.

I can't test edge on the CSS case yet because its hitting some error even without a SW.

Edit: link to the correct glitch

jakearchibald · 2017-11-01T15:15:01Z

It seems that Edge doesn't update the base URL on redirected with CSS. URLs remain relative to the request URL, not the response URL.

annevk · 2017-11-01T15:24:25Z

I'm happy for you to explore this further here, but please do escalate eventually to either an open (new) issue or web-platform-tests for better visibility.

wanderview · 2017-11-01T15:47:25Z

I'm happy for you to explore this further here, but please do escalate eventually to either an open (new) issue or web-platform-tests for better visibility.

We're currently implementing this spec change and will fix any WPT tests that need fixing as part of that effort. I keep asking here in case someone has an objection.

mfalken · 2017-11-02T02:22:03Z

Sorry not ignoring this, I just don't recall what we did off-hand and have been meaning to investigate commits/bugs. @horo-t do you remember what we've changed?

wanderview · 2017-11-02T02:41:12Z

No problem. I was just trying to understand the current situation a bit better in case there were lurking problems. Tom is making good progress on implementing it in firefox.

wanderview · 2017-11-03T13:58:33Z

Note, we may have a run into a thorny spec question with this change here:

https://bugzilla.mozilla.org/show_bug.cgi?id=1222008#c38

Consider:

var xmlDoc = document.implementation.createDocument(null, null, null);
xmlDoc.load('load_cross_origin_xml_document_cors.xml');
// SW intercepts and passes a cross-origin CORS response to respondWith()

In this case the outer network request is set up as same-origin (at least we make that restriction for XMLDocument in gecko). What should we do in the case where the service worker synthesizes a cross-origin CORS Response? Should we expose the cross-origin URL for the DocumentURI? That seems somewhat surprising and potentially breaking.

This problem exists for other same-origin requests as well such as Worker() top level scripts, same-origin fetch(), etc.

How would people feel if we rejected the outer network request if you synthesized a cross-origin Response onto a same-origin Request?

The existing test was asserting the opposite of the standard, which is to use the response URL. See whatwg/fetch#146. This CL does the following: - Tests that a CSS stylesheet fetched via respondWith(fetch(responseUrl) uses responseUrl as its base URL. - Tests that a CSS stylesheet fetched via respondWith(new Response()) uses the response URL (which is the request URL) as its base URL. - Changes the test to not test cross-origin stylesheets. That is more complex than needed for this test, and there is talk of making subresource requests from opaque stylesheets skip the service worker, which would render the test ineffective for testing base URL. - Changes the test to use waitUntil() in the message event to try to ensure the service worker stays alive between the message and fetch events. Bug: 911974 Change-Id: I167dfe86986ec718a50d512f862f1eb49889608b

The existing test was asserting the opposite of the standard, which is to use the response URL. See whatwg/fetch#146. This CL does the following: - Tests that a CSS stylesheet fetched via respondWith(fetch(responseUrl) uses responseUrl as its base URL. - Tests that a CSS stylesheet fetched via respondWith(new Response()) uses the response URL (which is the request URL) as its base URL. - Changes the test to not test cross-origin stylesheets. That is more complex than needed for this test, and there is talk of making subresource requests from opaque stylesheets skip the service worker, which would render the test ineffective for testing base URL. - Changes the test to use waitUntil() in the message event to try to ensure the service worker stays alive between the message and fetch events. Bug: 911974 Change-Id: I167dfe86986ec718a50d512f862f1eb49889608b Reviewed-on: https://chromium-review.googlesource.com/c/1362776 Commit-Queue: Matt Falkenhagen <falken@chromium.org> Reviewed-by: Ben Kelly <wanderview@chromium.org> Cr-Commit-Position: refs/heads/master@{#614277}

This aligns with the standard. See whatwg/fetch#146 Bug: 914135 Change-Id: I229aec6f8473bb6b7cdc88429afa830bc6eb80ed

This aligns with the standard. See whatwg/fetch#146 Chrome Status: https://www.chromestatus.com/feature/4857368017895424 No intent to ship because it's a trivial change. Bug: 914135 Change-Id: I229aec6f8473bb6b7cdc88429afa830bc6eb80ed

This aligns with the standard. See whatwg/fetch#146 Chrome Status: https://www.chromestatus.com/feature/4857368017895424 No intent to ship because it's a trivial change. Bug: 914135 Change-Id: I229aec6f8473bb6b7cdc88429afa830bc6eb80ed Reviewed-on: https://chromium-review.googlesource.com/c/1372109 Commit-Queue: Matt Falkenhagen <falken@chromium.org> Reviewed-by: Kent Tamura <tkent@chromium.org> Cr-Commit-Position: refs/heads/master@{#616248}

Use ResourceResponse::ResponseUrl() to set the base URL, and use ResponseResponse::GetType() to determine whether the resonse is CORS-same-origin. This CL has three web-exposed changes. 1. Use the response URL rather than the last request URL as the base URL of the stylesheet. This aligns with the standard. See whatwg/fetch#146 and WPT results indicate Firefox, Edge, and Safari use the response URL. This only matters if the response came from a service worker, as the URLs only differ when the service worker intercepts the request and responds with a different URL via respondWith(fetch(other_url)). This is covered by the WPT: service-workers/service-worker/fetch-request-css-base-url.https.html The test doesn't completely pass yet because the search query part of the URL gets chopped off for FetchEvent.request.referrer, but the base URL is correct. Chrome Status: https://www.chromestatus.com/feature/5642183499579392 2. Consider A->B->A redirects to be cross-origin rather than same-origin. Previously, this was considered same-origin. See the discussion in whatwg/fetch#737 and change whatwg/fetch#834. This change makes the following WPT test pass: css/cssom/stylesheet-same-origin.sub.html It also affects the web test: http/tests/security/cannot-read-cssrules-redirect.html This test is updated to match the behavior change. It can be removed later since it is redundant with the WPT test. 3. Consider load failures to be cross-origin rather than same-origin. That is, accessing |styleSheet.rules| throws a SecurityError if the load failed. This aligns with the specification: - cssRules checks the `origin-clean` flag: https://drafts.csswg.org/cssom/#dom-cssstylesheet-cssrules - This is set to true iff CORS-same-origin: https://html.spec.whatwg.org/multipage/links.html#link-type-stylesheet - CORS-same-origin is false on kError: https://html.spec.whatwg.org/multipage/urls-and-fetching.html#cors-same-origin This change makes the following WPT test pass: css/cssom/stylesheet-same-origin.sub.html It also affects the web tests: register-bypassing-scheme-partial.html require-sri-for-style-blocked.php These tests are updated to match the behavior change. Chrome Status: https://www.chromestatus.com/feature/5642183499579392 Intent to Ship: https://groups.google.com/a/chromium.org/d/msg/blink-dev/7OSy00oxVpk/siufiQVBBwAJ Bug: 911974 Change-Id: I9add3162596963eee66f60f339cfd9911bc151cd Reviewed-on: https://chromium-review.googlesource.com/c/1367331 Commit-Queue: Joel Hockey <joelhockey@chromium.org> Reviewed-by: Joel Hockey <joelhockey@chromium.org> Reviewed-by: Rune Lillesveen <futhark@chromium.org> Reviewed-by: Mike West <mkwst@chromium.org> Cr-Commit-Position: refs/heads/master@{#616580}

…se URL., a=testonly Automatic update from web-platform-tests WPT: service worker: fix test for CSS base URL. The existing test was asserting the opposite of the standard, which is to use the response URL. See whatwg/fetch#146. This CL does the following: - Tests that a CSS stylesheet fetched via respondWith(fetch(responseUrl) uses responseUrl as its base URL. - Tests that a CSS stylesheet fetched via respondWith(new Response()) uses the response URL (which is the request URL) as its base URL. - Changes the test to not test cross-origin stylesheets. That is more complex than needed for this test, and there is talk of making subresource requests from opaque stylesheets skip the service worker, which would render the test ineffective for testing base URL. - Changes the test to use waitUntil() in the message event to try to ensure the service worker stays alive between the message and fetch events. Bug: 911974 Change-Id: I167dfe86986ec718a50d512f862f1eb49889608b Reviewed-on: https://chromium-review.googlesource.com/c/1362776 Commit-Queue: Matt Falkenhagen <falken@chromium.org> Reviewed-by: Ben Kelly <wanderview@chromium.org> Cr-Commit-Position: refs/heads/master@{#614277} -- wpt-commits: 3e60f8f48ebed7895b6d58a88d35ccb5a80ed3c6 wpt-pr: 14373

…or the base URL., a=testonly Automatic update from web-platform-tests service worker: XSLT: Use response URL for the base URL. This aligns with the standard. See whatwg/fetch#146 Chrome Status: https://www.chromestatus.com/feature/4857368017895424 No intent to ship because it's a trivial change. Bug: 914135 Change-Id: I229aec6f8473bb6b7cdc88429afa830bc6eb80ed Reviewed-on: https://chromium-review.googlesource.com/c/1372109 Commit-Queue: Matt Falkenhagen <falken@chromium.org> Reviewed-by: Kent Tamura <tkent@chromium.org> Cr-Commit-Position: refs/heads/master@{#616248} -- wpt-commits: bc5122da94555fdf1b82b520599d860e722d8c05 wpt-pr: 14474

…se URL., a=testonly Automatic update from web-platform-tests WPT: service worker: fix test for CSS base URL. The existing test was asserting the opposite of the standard, which is to use the response URL. See whatwg/fetch#146. This CL does the following: - Tests that a CSS stylesheet fetched via respondWith(fetch(responseUrl) uses responseUrl as its base URL. - Tests that a CSS stylesheet fetched via respondWith(new Response()) uses the response URL (which is the request URL) as its base URL. - Changes the test to not test cross-origin stylesheets. That is more complex than needed for this test, and there is talk of making subresource requests from opaque stylesheets skip the service worker, which would render the test ineffective for testing base URL. - Changes the test to use waitUntil() in the message event to try to ensure the service worker stays alive between the message and fetch events. Bug: 911974 Change-Id: I167dfe86986ec718a50d512f862f1eb49889608b Reviewed-on: https://chromium-review.googlesource.com/c/1362776 Commit-Queue: Matt Falkenhagen <falkenchromium.org> Reviewed-by: Ben Kelly <wanderviewchromium.org> Cr-Commit-Position: refs/heads/master{#614277} -- wpt-commits: 3e60f8f48ebed7895b6d58a88d35ccb5a80ed3c6 wpt-pr: 14373 UltraBlame original commit: 6853099d97952c5f90ad15cb41bf48ef979c4e13

…or the base URL., a=testonly Automatic update from web-platform-tests service worker: XSLT: Use response URL for the base URL. This aligns with the standard. See whatwg/fetch#146 Chrome Status: https://www.chromestatus.com/feature/4857368017895424 No intent to ship because it's a trivial change. Bug: 914135 Change-Id: I229aec6f8473bb6b7cdc88429afa830bc6eb80ed Reviewed-on: https://chromium-review.googlesource.com/c/1372109 Commit-Queue: Matt Falkenhagen <falkenchromium.org> Reviewed-by: Kent Tamura <tkentchromium.org> Cr-Commit-Position: refs/heads/master{#616248} -- wpt-commits: bc5122da94555fdf1b82b520599d860e722d8c05 wpt-pr: 14474 UltraBlame original commit: 93f7c970ddcf711010d0a29451229eee220771f0

annevk force-pushed the networkurlwins branch from c0897b0 to 9effdfc Compare November 2, 2015 07:57

Use the URL from the response, if it has one

ed37f5e

Instead of always overriding response's url list with request's url list, even after service workers, only do so for synthetic responses. This mostly affects the way stylesheets work in service workers. See w3c/ServiceWorker#757 for details.

annevk force-pushed the networkurlwins branch from 9effdfc to ed37f5e Compare November 4, 2015 11:42

annevk merged commit ed37f5e into master Nov 4, 2015

annevk deleted the networkurlwins branch November 4, 2015 15:30

jungkees mentioned this pull request Jun 9, 2016

what should the document base URI for an intercepted navigation w3c/ServiceWorker#787

Closed

annevk mentioned this pull request Jun 9, 2016

Make sure all responses have a URL list #322

Merged

This was referenced Nov 3, 2017

Create F2F agenda - 7 November 2017 w3c/ServiceWorker#1206

Open

consider failing same-origin fetch requests that get a cross-origin cors Response synthesized by a service worker #629

Closed

chromium-wpt-export-bot mentioned this pull request Dec 5, 2018

WPT: service worker: fix test for CSS base URL. web-platform-tests/wpt#14373

Merged

chromium-wpt-export-bot mentioned this pull request Dec 12, 2018

service worker: XSLT: Use response URL for the base URL. web-platform-tests/wpt#14474

Merged

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Use the URL from the response, if it has one #146

Use the URL from the response, if it has one #146

annevk commented Oct 28, 2015

annevk commented Oct 28, 2015

wanderview commented Nov 3, 2015

annevk commented Nov 4, 2015

annevk commented Nov 4, 2015

wanderview commented Nov 4, 2015

wanderview commented Nov 5, 2015

jakearchibald commented Nov 9, 2015

wanderview commented Oct 26, 2017

wanderview commented Oct 26, 2017

wanderview commented Nov 1, 2017 •

edited

Loading

jakearchibald commented Nov 1, 2017

annevk commented Nov 1, 2017

wanderview commented Nov 1, 2017

mfalken commented Nov 2, 2017

wanderview commented Nov 2, 2017

wanderview commented Nov 3, 2017

Use the URL from the response, if it has one #146

Use the URL from the response, if it has one #146

Conversation

annevk commented Oct 28, 2015

annevk commented Oct 28, 2015

wanderview commented Nov 3, 2015

annevk commented Nov 4, 2015

annevk commented Nov 4, 2015

wanderview commented Nov 4, 2015

wanderview commented Nov 5, 2015

jakearchibald commented Nov 9, 2015

wanderview commented Oct 26, 2017

wanderview commented Oct 26, 2017

wanderview commented Nov 1, 2017 • edited Loading

jakearchibald commented Nov 1, 2017

annevk commented Nov 1, 2017

wanderview commented Nov 1, 2017

mfalken commented Nov 2, 2017

wanderview commented Nov 2, 2017

wanderview commented Nov 3, 2017

wanderview commented Nov 1, 2017 •

edited

Loading