Bug 180360 - Support container.getRegistration() / getRegistrations() inside service workers
Summary: Support container.getRegistration() / getRegistrations() inside service workers
Status: RESOLVED FIXED
Alias: None
Product: WebKit
Classification: Unclassified
Component: WebCore Misc. (show other bugs)
Version: WebKit Nightly Build
Hardware: Unspecified Unspecified
: P2 Normal
Assignee: Chris Dumez
URL:
Keywords: InRadar
Depends on:
Blocks: 180383
  Show dependency treegraph
 
Reported: 2017-12-04 09:15 PST by Chris Dumez
Modified: 2017-12-04 20:00 PST (History)
10 users (show)

See Also:


Attachments
WIP Patch (17.32 KB, patch)
2017-12-04 11:45 PST, Chris Dumez
no flags Details | Formatted Diff | Diff
WIP Patch (20.94 KB, patch)
2017-12-04 12:09 PST, Chris Dumez
no flags Details | Formatted Diff | Diff
WIP Patch (20.78 KB, patch)
2017-12-04 12:18 PST, Chris Dumez
no flags Details | Formatted Diff | Diff
WIP Patch (20.83 KB, patch)
2017-12-04 12:57 PST, Chris Dumez
no flags Details | Formatted Diff | Diff
Archive of layout-test-results from ews105 for mac-elcapitan-wk2 (2.54 MB, application/zip)
2017-12-04 14:16 PST, EWS Watchlist
no flags Details
Archive of layout-test-results from ews126 for ios-simulator-wk2 (2.12 MB, application/zip)
2017-12-04 14:27 PST, EWS Watchlist
no flags Details
Patch (28.81 KB, patch)
2017-12-04 15:11 PST, Chris Dumez
no flags Details | Formatted Diff | Diff
Patch (29.02 KB, patch)
2017-12-04 15:35 PST, Chris Dumez
no flags Details | Formatted Diff | Diff
Patch (31.85 KB, patch)
2017-12-04 15:38 PST, Chris Dumez
no flags Details | Formatted Diff | Diff
Patch (31.05 KB, patch)
2017-12-04 16:40 PST, Chris Dumez
no flags Details | Formatted Diff | Diff
Patch (30.96 KB, patch)
2017-12-04 16:48 PST, Chris Dumez
no flags Details | Formatted Diff | Diff

Note You need to log in before you can comment on or make changes to this bug.
Description Chris Dumez 2017-12-04 09:15:37 PST
Support container.getRegistration() / getRegistrations() inside service workers.
Comment 1 Chris Dumez 2017-12-04 11:45:04 PST
Created attachment 328367 [details]
WIP Patch
Comment 2 Chris Dumez 2017-12-04 12:09:25 PST
Created attachment 328373 [details]
WIP Patch
Comment 3 Chris Dumez 2017-12-04 12:18:29 PST
Created attachment 328374 [details]
WIP Patch
Comment 4 Chris Dumez 2017-12-04 12:57:54 PST
Created attachment 328379 [details]
WIP Patch
Comment 5 EWS Watchlist 2017-12-04 14:16:07 PST
Comment on attachment 328379 [details]
WIP Patch

Attachment 328379 [details] did not pass mac-wk2-ews (mac-wk2):
Output: http://webkit-queues.webkit.org/results/5491816

New failing tests:
imported/w3c/web-platform-tests/service-workers/service-worker/skip-waiting-without-using-registration.https.html
imported/w3c/web-platform-tests/service-workers/service-worker/activation.https.html
Comment 6 EWS Watchlist 2017-12-04 14:16:09 PST
Created attachment 328389 [details]
Archive of layout-test-results from ews105 for mac-elcapitan-wk2

The attached test failures were seen while running run-webkit-tests on the mac-wk2-ews.
Bot: ews105  Port: mac-elcapitan-wk2  Platform: Mac OS X 10.11.6
Comment 7 EWS Watchlist 2017-12-04 14:27:39 PST
Comment on attachment 328379 [details]
WIP Patch

Attachment 328379 [details] did not pass ios-sim-ews (ios-simulator-wk2):
Output: http://webkit-queues.webkit.org/results/5491833

New failing tests:
imported/w3c/web-platform-tests/service-workers/service-worker/skip-waiting-without-using-registration.https.html
imported/w3c/web-platform-tests/service-workers/service-worker/activation.https.html
Comment 8 EWS Watchlist 2017-12-04 14:27:40 PST
Created attachment 328391 [details]
Archive of layout-test-results from ews126 for ios-simulator-wk2

The attached test failures were seen while running run-webkit-tests on the ios-sim-ews.
Bot: ews126  Port: ios-simulator-wk2  Platform: Mac OS X 10.12.6
Comment 9 Chris Dumez 2017-12-04 15:11:46 PST
Created attachment 328395 [details]
Patch
Comment 10 Chris Dumez 2017-12-04 15:35:51 PST
Created attachment 328396 [details]
Patch
Comment 11 Chris Dumez 2017-12-04 15:38:20 PST
Created attachment 328397 [details]
Patch
Comment 12 Chris Dumez 2017-12-04 15:45:52 PST
Comment on attachment 328397 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=328397&action=review

> LayoutTests/imported/w3c/ChangeLog:16
> +        a cleanup step to unregister. The test now passes all checks.

PR: https://github.com/w3c/web-platform-tests/pull/8559
Comment 13 youenn fablet 2017-12-04 16:16:08 PST
Comment on attachment 328397 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=328397&action=review

> Source/WebCore/dom/ScriptExecutionContext.cpp:588
> +void ScriptExecutionContext::postTaskTo(const DocumentOrWorkerIdentifier& contextIdentifier, WTF::Function<void(ScriptExecutionContext&)>&& task)

I am always wondering whether to use Identifier or const Identifier& in such cases.
I have difficulties keeping it consistent. Since it is a Variant, it is better to pass it by ref?

> Source/WebCore/dom/ScriptExecutionContext.cpp:592
> +    switchOn(contextIdentifier, [&](DocumentIdentifier identifier) {

space between ] and ( is apparently our style.

> Source/WebCore/dom/ScriptExecutionContext.cpp:596
> +        document->postTask([task = WTFMove(task)](ScriptExecutionContext& scope) {

auto& here and in other various places below ?

> Source/WebCore/workers/service/ServiceWorkerContainer.cpp:232
> +    Ref<SecurityOrigin> topOrigin = context->topOrigin();

This may end up topOrigin to be destroyed on the main thread. Shouldn't we isolate it?
Ditto below probably.

> Source/WebCore/workers/service/ServiceWorkerContainer.cpp:320
> +    }

You can probably write it as:
auto registrations = WTF::map(WTFMove(registrations), [this] (auto&& registrationData) {
    return ServiceWorkerRegistration::getOrCreate(*scriptExecutionContext(), *this, WTFMove(registrationData));
});

> Source/WebCore/workers/service/ServiceWorkerContainer.cpp:545
> +    m_pendingGetRegistrationsRequests.clear();

I would try to use one map that keeps DeferredPromise instead of typed DOMPromiseDeferred.

> Source/WebKit/WebProcess/Storage/WebSWClientConnection.cpp:140
> +    ASSERT(isMainThread());

We do not do that often when receiving IPC messages.
Should we try to do that more aggressively? Or should these assertions be moved to the lambdas instead?
Comment 14 Chris Dumez 2017-12-04 16:27:25 PST
Comment on attachment 328397 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=328397&action=review

>> Source/WebCore/dom/ScriptExecutionContext.cpp:588
>> +void ScriptExecutionContext::postTaskTo(const DocumentOrWorkerIdentifier& contextIdentifier, WTF::Function<void(ScriptExecutionContext&)>&& task)
> 
> I am always wondering whether to use Identifier or const Identifier& in such cases.
> I have difficulties keeping it consistent. Since it is a Variant, it is better to pass it by ref?

I am not sure in the case of a Variant so I figured better safe than sorry.

>> Source/WebCore/dom/ScriptExecutionContext.cpp:592
>> +    switchOn(contextIdentifier, [&](DocumentIdentifier identifier) {
> 
> space between ] and ( is apparently our style.

Uh? ok.

>> Source/WebCore/dom/ScriptExecutionContext.cpp:596
>> +        document->postTask([task = WTFMove(task)](ScriptExecutionContext& scope) {
> 
> auto& here and in other various places below ?

ok.

>> Source/WebCore/workers/service/ServiceWorkerContainer.cpp:232
>> +    Ref<SecurityOrigin> topOrigin = context->topOrigin();
> 
> This may end up topOrigin to be destroyed on the main thread. Shouldn't we isolate it?
> Ditto below probably.

Will isolateCopy().

>> Source/WebCore/workers/service/ServiceWorkerContainer.cpp:320
>> +    }
> 
> You can probably write it as:
> auto registrations = WTF::map(WTFMove(registrations), [this] (auto&& registrationData) {
>     return ServiceWorkerRegistration::getOrCreate(*scriptExecutionContext(), *this, WTFMove(registrationData));
> });

Ok, wasn't familiar with WTF::map()

>> Source/WebCore/workers/service/ServiceWorkerContainer.cpp:545
>> +    m_pendingGetRegistrationsRequests.clear();
> 
> I would try to use one map that keeps DeferredPromise instead of typed DOMPromiseDeferred.

Will look into this.
Comment 15 Chris Dumez 2017-12-04 16:40:43 PST
Created attachment 328409 [details]
Patch
Comment 16 youenn fablet 2017-12-04 16:43:59 PST
Comment on attachment 328409 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=328409&action=review

> Source/WebCore/workers/service/ServiceWorkerContainer.cpp:292
> +    Ref<SecurityOrigin> topOrigin = context->topOrigin();

No need for this variable and refing topOrigin probably as well.
Comment 17 Chris Dumez 2017-12-04 16:48:24 PST
Created attachment 328411 [details]
Patch
Comment 18 WebKit Commit Bot 2017-12-04 17:12:05 PST
Comment on attachment 328411 [details]
Patch

Rejecting attachment 328411 [details] from commit-queue.

Failed to run "['/Volumes/Data/EWS/WebKit/Tools/Scripts/webkit-patch', '--status-host=webkit-queues.webkit.org', '--bot-id=webkit-cq-03', 'land-attachment', '--force-clean', '--non-interactive', '--parent-command=commit-queue', 328411, '--port=mac']" exit_code: 2 cwd: /Volumes/Data/EWS/WebKit

Last 500 characters of output:
..99b2274  master     -> origin/master
Partial-rebuilding .git/svn/refs/remotes/origin/master/.rev_map.268f45cc-cd09-0410-ab3c-d52691b4dbfc ...
Currently at 225506 = 6dd634539ab447fd1507bbe274df2c73431ab7b7
r225507 = 99b227420d9a3b88cc44f866efee9ec3cb1c6da1
Done rebuilding .git/svn/refs/remotes/origin/master/.rev_map.268f45cc-cd09-0410-ab3c-d52691b4dbfc
First, rewinding head to replay your work on top of it...
Fast-forwarded master to refs/remotes/origin/master.
Total errors found: 0 in 0 files

Full output: http://webkit-queues.webkit.org/results/5494450
Comment 19 Chris Dumez 2017-12-04 18:06:32 PST
Comment on attachment 328411 [details]
Patch

Clearing flags on attachment: 328411

Committed r225513: <https://trac.webkit.org/changeset/225513>
Comment 20 Chris Dumez 2017-12-04 18:06:34 PST
All reviewed patches have been landed.  Closing bug.
Comment 21 Radar WebKit Bug Importer 2017-12-04 20:00:51 PST
<rdar://problem/35846638>