[Webkit-unassigned] [Bug 197759] New: Support wasm-unsafe-eval
bugzilla-daemon at webkit.org
bugzilla-daemon at webkit.org
Thu May 9 16:48:53 PDT 2019
https://bugs.webkit.org/show_bug.cgi?id=197759
Bug ID: 197759
Summary: Support wasm-unsafe-eval
Product: WebKit
Version: Safari Technology Preview
Hardware: Macintosh
OS: macOS 10.14
Status: NEW
Severity: Normal
Priority: P2
Component: WebAssembly
Assignee: webkit-unassigned at lists.webkit.org
Reporter: rob at agilebits.com
The following issue mentions Chrome, but the same applies to Safari: https://github.com/WebAssembly/content-security-policy/issues/7
It's impossible to use wasm in Safari without the `unsafe-eval` directive. This proposal describes a new directive, `wasm-unsafe-eval`, that would allow executing arbitrary wasm code without having to also allow JS eval(): https://github.com/WebAssembly/content-security-policy/blob/master/proposals/CSP.md
Chrome has the following issue open to support that directive: https://bugs.chromium.org/p/chromium/issues/detail?id=948834
We would prefer to use the stream APIs with proper CSP support, but WK does not yet even have the stream functions: https://bugs.webkit.org/show_bug.cgi?id=173105
Given that and the fact that Chrome has the stream functions but not proper CSP support (https://bugs.chromium.org/p/chromium/issues/detail?id=961485), an acceptable alternative for now would be to support `wasm-unsafe-eval` in the `script-src` directive.
Note that Firefox and Edge allow wasm compilation and execution without any change to the CSP whatsoever. That behavior is effectively the same as adding `wasm-unsafe-eval` automatically. We aren't advocating that position either and would prefer to see those browsers disable wasm unless `wasm-unsafe-eval` is present or an acceptable SRI hash or origin is present.
--
You are receiving this mail because:
You are the assignee for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.webkit.org/pipermail/webkit-unassigned/attachments/20190509/c5288ef8/attachment.html>
More information about the webkit-unassigned
mailing list