Prepare to switch from WTF::Variant to std::variant
Created attachment 440232 [details] Patch
Created attachment 440236 [details] Patch
Comment on attachment 440236 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=440236&action=review > Source/JavaScriptCore/runtime/CacheUpdate.cpp:41 > CacheUpdate::CacheUpdate(CacheUpdate&& other) Just had another idea for this. Can we try this? CacheUpdate::CacheUpdate(CacheUpdate&&) = default;
Comment on attachment 440236 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=440236&action=review >> Source/JavaScriptCore/runtime/CacheUpdate.cpp:41 >> CacheUpdate::CacheUpdate(CacheUpdate&& other) > > Just had another idea for this. Can we try this? > > CacheUpdate::CacheUpdate(CacheUpdate&&) = default; We could also experiment with removing the explicit move constructor and assignment operator from the class entirely. I am pretty sure the compiler will generate the correct code if we just don’t say anything.
Created attachment 440243 [details] Patch
Removing it causes compiler errors, but =default works fine.
Created attachment 440279 [details] Patch
There are several appearances in for example CDMProxy.h and related files that are not being changed. In the GTK/WPE ports, these things are built when experimental features are enabled and I don't recall if we have them enabled in EWS so they won't fail. Anyway, I don't know if you're taking a step by step approach but those would need changing as well or things will break on our side.
GTK/EWS developer builds have experimental features enabled.
Committed r283906 (242780@main): <https://commits.webkit.org/242780@main> All reviewed patches have been landed. Closing bug and clearing flags on attachment 440279 [details].
<rdar://problem/84105079>