Open
Bug 1632837
Opened 5 years ago
Updated 4 years ago
data URL dedicated workers should create their own agent cluster
Categories
(Core :: DOM: Workers, defect, P3)
Core
DOM: Workers
Tracking
()
NEW
People
(Reporter: annevk, Unassigned)
References
(Blocks 1 open bug)
Details
Updated•5 years ago
|
Priority: -- → P3
Comment 1•5 years ago
|
||
Is this going to have disturbing process model implications? That is, it's my hope that:
- The dedicated worker still lives in the same process.
- It just has a different agent cluster id and so the postMessage of SAB will fail.
- There's no way for multiple dedicated workers to be in the same agent cluster because each opaque origin is, by definition, non-equivalent, so the the worker can't collaborate with another data URL dedicated worker to do SPECTRE stuff because none of their SABs can leave their global.
Flags: needinfo?(annevk)
Reporter | ||
Comment 2•5 years ago
|
||
It would allow us to process-isolate data URL / sandboxed (see bug 1632840) dedicated workers, but the specification will never make that a hard requirement. I'm not sure what Fission has decided about such cases.
You can in fact get another dedicated worker into this agent cluster by using blob URLs. Anything else would indeed not work.
Flags: needinfo?(annevk)
Updated•4 years ago
|
Severity: -- → S3
You need to log in
before you can comment on or make changes to this bug.
Description
•