Move most IDB object encoding/decoding from WebKit2 to WebCore
Created attachment 261175 [details] Patch v1 For an EWS run - Not including ChangeLogs because I have other changes locally that would cause this to not apply on EWS. Definitely reviewable, though.
Created attachment 261193 [details] Patch v2
Comment on attachment 261193 [details] Patch v2 View in context: https://bugs.webkit.org/attachment.cgi?id=261193&action=review r=me if ews passes > Source/WebKit2/Shared/WebCoreArgumentCoders.h:405 > template<> struct ArgumentCoder<WebCore::IDBGetResult> { IDBGetResult feels left out :(
(In reply to comment #3) > Comment on attachment 261193 [details] > Patch v2 > > View in context: > https://bugs.webkit.org/attachment.cgi?id=261193&action=review > > r=me if ews passes > > > Source/WebKit2/Shared/WebCoreArgumentCoders.h:405 > > template<> struct ArgumentCoder<WebCore::IDBGetResult> { > > IDBGetResult feels left out :( It can't be moved right now because it uses DataReference which is WK2-only. But good news - While resolving https://bugs.webkit.org/show_bug.cgi?id=149117, IDBGetResult will be going away entirely. :)
The EFL failure: ../../Source/WebCore/Modules/indexeddb/IDBKeyPath.h: In static member function 'static bool WebCore::IDBKeyPath::decode(Decoder&, WebCore::IDBKeyPath&) [with Decoder = IPC::ArgumentDecoder]': ../../Source/WebCore/Modules/indexeddb/IDBKeyPath.h:140:1: error: control reaches end of non-void function [-Werror=return-type] ...Even though all switch cases are accounted for. :( Stupid compiler. Will fix.
https://trac.webkit.org/changeset/189810