IndexedDB: Support NO_DUPLICATE cursors on LevelDB back-end
Created attachment 94958 [details] Patch
Comment on attachment 94958 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=94958&action=review LGTM > Source/WebCore/ChangeLog:8 > + This is tested by storage/indexeddb/mozilla/indexes.html The layout tests aren't run with the leveldb backend on the bots, right? You have a local change that forces leveldb?
(In reply to comment #2) > (From update of attachment 94958 [details]) > View in context: https://bugs.webkit.org/attachment.cgi?id=94958&action=review > > LGTM > > > Source/WebCore/ChangeLog:8 > > + This is tested by storage/indexeddb/mozilla/indexes.html > > The layout tests aren't run with the leveldb backend on the bots, right? You have a local change that forces leveldb? Exactly.
Comment on attachment 94958 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=94958&action=review > Source/WebCore/ChangeLog:5 > + IndexedDB: Support NO_DUPLICATE cursors on LevelDB back-end If possible, a link to this part of the spec in the ChangeLog would be nice. > Source/WebCore/storage/IDBLevelDBBackingStore.cpp:839 > + CursorImplCommon(LevelDBTransaction* transaction, const Vector<char>& lowKey, bool lowOpen, const Vector<char>& highKey, bool highOpen, bool forward, bool unique) The bools are out of hand. I don't think you should hold up this change, but a future cleanup might be nice to use enums, an options struct or something else to clean up these signatures.
(In reply to comment #4) > (From update of attachment 94958 [details]) Thanks for the review! > View in context: https://bugs.webkit.org/attachment.cgi?id=94958&action=review > > > Source/WebCore/ChangeLog:5 > > + IndexedDB: Support NO_DUPLICATE cursors on LevelDB back-end > > If possible, a link to this part of the spec in the ChangeLog would be nice. Done. > > > Source/WebCore/storage/IDBLevelDBBackingStore.cpp:839 > > + CursorImplCommon(LevelDBTransaction* transaction, const Vector<char>& lowKey, bool lowOpen, const Vector<char>& highKey, bool highOpen, bool forward, bool unique) > > The bools are out of hand. I don't think you should hold up this change, but a future cleanup might be nice to use enums, an options struct or something else to clean up these signatures. Will do.
Committed r87491: <http://trac.webkit.org/changeset/87491>