[worklets] Introduce an "owner document" for each WorkletGlobalScope. #389
Conversation
|
@annevk does this work? |
|
I guess I was not specific enough. The problem is mostly with WorkletGlobalScope. I see that you defer to host specifications for them lifetime, but I don't think that's good enough nor necessary. We know that all worklet globals will be owned by a Window/document so we might as well write that down. |
bfgeek
changed the title
|
@annevk PTAL, I think this is enough? Each WorkletGlobalScope has an "owner document". This also introduces a "destroy a WorkletGlobalScope" algorithm - may want to change the event loop in html spec for this instead... |
|
Do you have tests for this? In particular, if an initial about:blank document gets replaced, does the WorkletGlobalScope go away or stay? I'm asking since @wanderview wants to change ownership of workers to be global object based and maybe we want worklets to work like that too? |
|
We don't currently have tests for this; (@nhiroki we should add tests for this :) ) But we currently do kill the globals when the document get's destroyed, @nhiroki please correct me if I'm wrong: |
|
@bfgeek You're right. I'm not sure how we can confirm it in web-platform-tests. For workers we could test it by creating a worker in a frame, closing the frame and postMessage+setTimeout like [1], but worklets don't have such a communication path. [1] https://github.com/w3c/web-platform-tests/blob/master/workers/WorkerGlobalScope_close.htm |
|
Filed a chromium bug: https://crbug.com/712133 |
|
@nhiroki with audio worklets it should be easy to test, but they are still rather speculative at this point. Can a paint worklet draw anything that can then be imported by canvas or some such? Or does it intentionally not leak anything from its insides? (Sorry, I used to know this, but I forgot.) |
|
@annevk Thank you for the suggestion. I'll try the approach. |
|
It seems weird btw for the owner to not be the global object as you do expose the accessor to the worklet on the global object. |
Closes w3c/css-houdini-drafts#1000. This provides a baseline by porting over all the existing text from https://drafts.css-houdini.org/worklets/, modernizing and restructuring it along the way. It does not yet fix many of the open logged issues (although it does fix some; see below). Notable changes from that document: * Rearranged sections to better match workers, and my sense of flow. * Moved worklet script fetching to be siblings with all the other script fetching algorithms. * Improved clarity and guidance on what specifications that define worklets should do, including fleshing out the fake worklet example. * Changed "create a WorkletGlobalScope", which took one set of arguments, to "create a worklet global scope", which just takes a Worklet instance. This appears to match better how the algorithm is used, e.g. in https://drafts.css-houdini.org/css-paint-api/#draw-a-paint-image step 10. * Updated "report an error" to bail out for non-EventTarget globals, like WorkletGlobalScope. Closes #2611. * Updated worklets to only be exposed in secure contexts. Closes w3c/css-houdini-drafts#505. * Makes the lifetime of creating and terminating WorkletGlobalScopes more explicit. Closes w3c/css-houdini-drafts#224. Closes w3c/css-houdini-drafts#389. * Explicitly start and stop the event loop for a given WorkletGlobalScope upon creation/termination. Closes w3c/css-houdini-drafts#843. Closes w3c/css-houdini-drafts#318 for real. * Fixes creation of new worklet global scopes to only run the top-level module scripts added via addModule(), which will automatically run their dependencies. Previously it would run all module scripts loaded into the worklet, so dependencies would be run in the order they were fetched, not as part of the top-down module evaluation process. Closes w3c/css-houdini-drafts#264.
|
With whatwg/html#6056, we can close this. The lifetime of WorkletGlobalScopes is now very explicit. |
Fixes #385.