Introduce IDBServerConnection (and start moving things to it) IDBServerConnection will be a purely asynchronous interface for the database connection, transactions, and cursors. It's interface will be 100% asynchronous and callback based. Eventually none of the IDB*Backend classes will need IDBBackingStore* classes at all, and will use IDBServerConnection instead
Created attachment 216644 [details] Patch v1
Comment on attachment 216644 [details] Patch v1 Attachment 216644 [details] did not pass efl-ews (efl): Output: http://webkit-queues.appspot.com/results/22749454
Created attachment 216645 [details] Patch v2 - Implement dtor
Comment on attachment 216645 [details] Patch v2 - Implement dtor Attachment 216645 [details] did not pass efl-ews (efl): Output: http://webkit-queues.appspot.com/results/22699675
Comment on attachment 216645 [details] Patch v2 - Implement dtor Attachment 216645 [details] did not pass gtk-ews (gtk): Output: http://webkit-queues.appspot.com/results/22908597
Created attachment 216648 [details] Patch v3
Comment on attachment 216648 [details] Patch v3 View in context: https://bugs.webkit.org/attachment.cgi?id=216648&action=review > Source/WebCore/Modules/indexeddb/IDBServerConnection.h:40 > +// This interface is meant to provide a single asynchronous layer between the web-facing frontend Does it provide what it means to provide? The comment should either be more assertive, or a specific FIXME. > Source/WebCore/Modules/indexeddb/IDBServerConnection.h:42 > +// If the completion of an operation is important monitoring it must be through the use of a callback function. Important? > Source/WebCore/Modules/indexeddb/leveldb/IDBServerConnectionLevelDB.cpp:29 > +#include "IDBBackingStoreInterface.h" Is there a reason why this is outside ifdefs? We usually put headers inside ifdefs in cpp files, and outside ifdefs in headers.
http://trac.webkit.org/changeset/159144