- From: Jonas Sicking <notifications@github.com>
- Date: Fri, 31 Jul 2015 00:01:47 -0700
- To: whatwg/fetch <fetch@noreply.github.com>
- Message-ID: <whatwg/fetch/issues/93/126587623@github.com>
Don't we at this point all agree that we shouldn't expose the `Initiator` to script? So discussing which `Initiator` types exist seems pretty pointless, no? A worker essentially has exactly the same processing model as a script. The only difference is that right before evaluating the script a new global is created. But after that global has been created, the worker is processed exactly the same way that a script is. So the security implications of processing a worker is the same as a script. The caching policies that you'd want to have for a worker matches what you'd want to use for a script. The templating or transpiling that you'd want to do for a worker matches what you'd want to do for a script. The URL policies that you'd want to use for a worker is the same as you'd want to use for a scripts and other subresources. The only other difference that I could see between a worker and a script is that we allow cross-origin loads of scripts but not of workers. Though to be honest that's sort of silly that we have that limitation and something that would be good to fix (possibly with exception of ServiceWorkers). Documents on the other hand handle entirely different content types. Any time that a document is loaded a navigation is done, which means that redirects have a very different effect. The cache policy you'd want to use is likely different. You're much more likely to want to use human-readable URLs for a document than for a worker (again, possibly ServiceWorkers excepted). Etc. --- Reply to this email directly or view it on GitHub: https://github.com/whatwg/fetch/issues/93#issuecomment-126587623
Received on Friday, 31 July 2015 07:02:17 UTC