forked from mirrors/gecko-dev
5c9f8df90d
This check is unnecessary. We have a central place to compute whether scriptability is allowed that already accounts for the docshell flag, see Scriptability::SetDocShellAllowsScript, from bug 840488: https://searchfox.org/mozilla-central/rev/a4d62e09a4c46aef918667fa759bf9ae898dc258/docshell/base/nsDocShell.cpp#2532 It is also incorrect, for two reasons: * It should really be GetCanExecuteScripts(), so that it works properly in subframes. * We have a whitelist of principals that should always be allowed to execute script (see xpc::PrincipalImmuneToScriptPolicy), which is not accounted for here. So if we load an add-on document (or pdf.js after bug 866634) in a docshell with disabled script execution, we still won't load the out-of-line scripts. The latter is what's happening in bug 810815, as TB uses the docshell flag to disable script execution on emails. This is still a bit of a behavior change in the sense that after this, IIUC, scripts should download but not execute. I think that's fine and we have no test that depends on the current behavior. An alternative to this patch would be to change this check to get the docshell's document's node principal, and check that against PrincipalImmuneToScriptPolicy (and while at it fix the check to use GetCanExecuteScripts()). But this seems simpler. Differential Revision: https://phabricator.services.mozilla.com/D71489 |
||
|---|---|---|
| .. | ||
| build | ||
| moz.build | ||
| nsEmbedCID.h | ||
| nsIEmbeddingSiteWindow.idl | ||
| nsIPrintingPromptService.idl | ||
| nsIPrintPreviewNavigation.idl | ||
| nsIWebBrowser.idl | ||
| nsIWebBrowserChrome.idl | ||
| nsIWebBrowserChrome3.idl | ||
| nsIWebBrowserChromeFocus.idl | ||
| nsIWebBrowserPrint.idl | ||
| nsWebBrowser.cpp | ||
| nsWebBrowser.h | ||
| nsWebBrowserContentPolicy.cpp | ||
| nsWebBrowserContentPolicy.h | ||