Synchronous XHR bypasses ServiceWorkers.

See  bug 413103 .

SW is skipped for sync loads at  https://code.google.com/p/chromium/codesearch#chromium/src/content/browser/loader/resource_dispatcher_host_impl.cc&sq=package:chromium&l=1598&rcl=1463008567

+kenjibaheux@, jakearchibald@

Tell me if this limitation is only for V1 and you have plan to relax it.

Retitling: "service worker cache" commonly refers to the Cache Storage API, which is not the point of this issue.

I'm not totally sure if this is spec'd or just something we implemented long ago.

I just checked the spec. It seems this behavior is not specced in the Fetch Standard.

Jake, should service worker intercept sync XHR?

It should. Early on in Chrome's implementation, making sync XHR interceptable was really hard for some reason (caused a deadlock if memory serves), so it was intentionally skipped.

Is that still the case?

OK. Tentatively assigning horo@. Setting to P3 since this limitation has been in our impl since inception.

Not sure if the deadlock would still be there... but probably since the basic architecture is still the same.

Yes the deadlock issue is still there.

In the current implementation, the resource loading logic is executed in the main thread of the renderer process even when ServiceWorker fetches resources.
So if the ServiceWorker calls fetch() in the fetch event handler of the sync XHR, the deadlock happens because the main thread is blocked by the sync XHR.

 Issue 726327  has been merged into this issue.

This will be fetch-request-xhr-sync.https.html once https://codereview.chromium.org/2907443002/ lands.

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/7fdac9fc2a776ad2ed4c3df752896e0f33861b69

commit 7fdac9fc2a776ad2ed4c3df752896e0f33861b69
Author: mike <mike@mikepennisi.com>
Date: Thu Jun 01 16:44:45 2017

Upstream service worker "XHR" test to WPT

The original version of this test would only pass if the service worker
did *not* intercept the XHR request. That behavior conflicted with the
Fetch and Service Worker specification and with the test's documentation
itself. Update the test to pass only if the service worker intercepts
the request.

In addition:

- Update URLs to suitable values for the Web Platform Tests project
- Add "use strict" directives to script bodies
- Increase precision of "cleanup" logic scheduling
- Re-name files to adhere to precent set by existing WPT tests
- Catch and report errors within Promise chain rather than relying on
  implicit reporting of uncaught errors.

BUG= 688116 , 602051
R=falken@chromium.org

Review-Url: https://codereview.chromium.org/2907443002
Cr-Commit-Position: refs/heads/master@{#476324}

[modify] https://crrev.com/7fdac9fc2a776ad2ed4c3df752896e0f33861b69/third_party/WebKit/LayoutTests/FlagExpectations/enable-network-service
[add] https://crrev.com/7fdac9fc2a776ad2ed4c3df752896e0f33861b69/third_party/WebKit/LayoutTests/external/wpt/service-workers/service-worker/fetch-request-xhr-sync.https-expected.txt
[add] https://crrev.com/7fdac9fc2a776ad2ed4c3df752896e0f33861b69/third_party/WebKit/LayoutTests/external/wpt/service-workers/service-worker/fetch-request-xhr-sync.https.html
[add] https://crrev.com/7fdac9fc2a776ad2ed4c3df752896e0f33861b69/third_party/WebKit/LayoutTests/external/wpt/service-workers/service-worker/resources/fetch-request-xhr-sync-iframe.html
[add] https://crrev.com/7fdac9fc2a776ad2ed4c3df752896e0f33861b69/third_party/WebKit/LayoutTests/external/wpt/service-workers/service-worker/resources/fetch-request-xhr-sync-worker.js
[delete] https://crrev.com/1704affd92a33a48072fa0dfb0a63af1853a493a/third_party/WebKit/LayoutTests/http/tests/serviceworker/resources/sync-xhr-doesnt-deadlock-iframe.html
[delete] https://crrev.com/1704affd92a33a48072fa0dfb0a63af1853a493a/third_party/WebKit/LayoutTests/http/tests/serviceworker/resources/sync-xhr-doesnt-deadlock.data
[delete] https://crrev.com/1704affd92a33a48072fa0dfb0a63af1853a493a/third_party/WebKit/LayoutTests/http/tests/serviceworker/resources/sync-xhr-doesnt-deadlock.js
[delete] https://crrev.com/1704affd92a33a48072fa0dfb0a63af1853a493a/third_party/WebKit/LayoutTests/http/tests/serviceworker/sync-xhr-doesnt-deadlock-expected.txt
[delete] https://crrev.com/1704affd92a33a48072fa0dfb0a63af1853a493a/third_party/WebKit/LayoutTests/http/tests/serviceworker/sync-xhr-doesnt-deadlock.html

Is this being worked on? Kinda blocked by servicification?

Any context for the ping? No plans to work on this specifically, except as part of the burn down for issue 678905.

Likely blocked on servicification and off-main-thread fetch to avoid deadlocks, might take a lot of care to ensure absolutely no deadlock occurs, probably more trouble than this bug is worth...

 Issue 834802  has been merged into this issue.

The dead lock should be gone now.

Given yhirano@'s comment, and the fact that the tests/expectations still indicate failures ([1] and [2]), are there any plans to get this working, or not really, since the sync XHR is deprecated? (Asking because I'm writing resource load priority layout tests for requests passing through SWs).

[1]: https://wpt.fyi/results/service-workers/service-worker/fetch-request-xhr-sync.https.html?aligned&label=stable
[2]: https://cs.chromium.org/chromium/src/third_party/WebKit/LayoutTests/external/wpt/service-workers/service-worker/fetch-request-xhr-sync.https-expected.txt?q=fetch-request-xhr-s&sq=package:chromium&g=0&l=1

I think we should fix this: sync XHR should go to the service worker. In fact it might just work automatically now if we remove the blocking per yhirano's comment. Would you be interested in taking this?

Will take a look.

Dominic do you have time to look at this? If not I might try to make a patch.
 
Possibly if we just remove this things may work:
https://cs.chromium.org/chromium/src/third_party/blink/renderer/platform/loader/fetch/fetch_parameters.cc?l=110&rcl=7a3fba9b1e3a9b010b84faa8a41ad8a7baea035e

This also affects sync XSLT imports at:
https://cs.chromium.org/chromium/src/third_party/blink/renderer/core/xml/xsl_style_sheet_libxslt.cc?l=232&rcl=7a3fba9b1e3a9b010b84faa8a41ad8a7baea035e

Yep, just finished finals - thanks for the ping looking now.

Just a status update: we found at https://chromium-review.googlesource.com/c/chromium/src/+/1376738 there is probably still a deadlock/livelock somewhere.