We should use smart pointers everywhere in ReplaceSelectionCommand based on the new rule we came up with.
Created attachment 426071 [details] Patch
Comment on attachment 426071 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=426071&action=review LGTM. r=me. > Source/WebCore/editing/ReplaceSelectionCommand.cpp:562 > + ScriptDisallowedScope::InMainThread scriptDisallowedScope; Why do we need this here?
(In reply to Jiewen Tan from comment #2) > Comment on attachment 426071 [details] > Patch > > View in context: > https://bugs.webkit.org/attachment.cgi?id=426071&action=review > > LGTM. r=me. > > > Source/WebCore/editing/ReplaceSelectionCommand.cpp:562 > > + ScriptDisallowedScope::InMainThread scriptDisallowedScope; > > Why do we need this here? Because we're accessing the render objects. We don't want any script execution to happen there.
Huh, weird. Not sure why we didn't hit that compilation error locally.
Created attachment 426072 [details] Fixed builds
Comment on attachment 426072 [details] Fixed builds This has already been reviewed by Jiewen.
Committed r276006 (236558@main): <https://commits.webkit.org/236558@main>
<rdar://problem/76689476>