Proxy new objectstore/index methods through IPC

Proxy new objectstore/index methods through IPC

8 new messages corresponding to 8 new methods in IDBDatabaseBackendInterface.

BUG=161958
Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=175066

[alecflett, 2012-12-13 23:37:52]

Here are the new messages for the big index/objectstore refactor. 

These are 8 new messages, replacing 13 existing messages:

get - replaces ObjectStoreGet/IndexGet/IndexGetKey
put - replaces ObjectStorePut
setIndexKeys - replaces ObjectStoreSetIndexKeys
setIndexesReady - replaces ObjectStoreSetIndexesReady
openCursor - replaces ObjectStoreOpenCursor, IndexOpenCursor, IndexOpenKeyCursor
count - replaces ObjectStoreCount, IndexCount
deleteRange - replaces ObjectStoreDelete
clear - replaces ObjectStoreClear

[Tom Sepez, 2012-12-14 01:34:43]

lgtm with nits.

https://codereview.chromium.org/11567029/diff/1/content/browser/in_process_we...
File content/browser/in_process_webkit/indexed_db_dispatcher_host.cc (right):

https://codereview.chromium.org/11567029/diff/1/content/browser/in_process_we...
content/browser/in_process_webkit/indexed_db_dispatcher_host.cc:541:
(*map)[params.transaction_id] += params.value.size();
Safe from overflow because a corrupt renderer can't just set value.size()
without supplying data in the IPC to deserialize? What if renderer repeatedly
lies about transaction_id, do we leak slots?

https://codereview.chromium.org/11567029/diff/1/content/browser/in_process_we...
File content/browser/in_process_webkit/indexed_db_dispatcher_host.h (right):

https://codereview.chromium.org/11567029/diff/1/content/browser/in_process_we...
content/browser/in_process_webkit/indexed_db_dispatcher_host.h:115: typedef
std::map<int64, int64> WebIDBTransactionIDToSizeMap;
Sizes generally should be unsigned.

[jsbell, 2012-12-14 23:04:34]

I want to read through this again but initial comments...

https://codereview.chromium.org/11567029/diff/2002/content/browser/in_process...
File content/browser/in_process_webkit/indexed_db_dispatcher_host.cc (right):

https://codereview.chromium.org/11567029/diff/2002/content/browser/in_process...
content/browser/in_process_webkit/indexed_db_dispatcher_host.cc:504: // XXX
alecf
Don't forget to remove this.

https://codereview.chromium.org/11567029/diff/2002/content/browser/in_process...
content/browser/in_process_webkit/indexed_db_dispatcher_host.cc:644: // xxx end
alecf
Don't forget to remove.

https://codereview.chromium.org/11567029/diff/2002/content/browser/in_process...
File content/browser/in_process_webkit/indexed_db_dispatcher_host.h (right):

https://codereview.chromium.org/11567029/diff/2002/content/browser/in_process...
content/browser/in_process_webkit/indexed_db_dispatcher_host.h:20: struct
IndexedDBHostMsg_DatabaseDeleteRange_Params;
This was only called "DeleteRange" when we had Delete(key) vs.
DeleteRange(keyRange). Could be simplified now.

https://codereview.chromium.org/11567029/diff/2002/content/common/indexed_db/...
File content/common/indexed_db/indexed_db_dispatcher.h (right):

https://codereview.chromium.org/11567029/diff/2002/content/common/indexed_db/...
content/common/indexed_db/indexed_db_dispatcher.h:160: void
RequestIDBDatabaseDeleteRange(
s/DeleteRange/Delete/

[alecflett, 2013-01-03 00:09:32]

ok, everything updated, with minor fixes found during webkit side migration.

[jsbell, 2013-01-03 01:17:35]

lgtm with nits

https://codereview.chromium.org/11567029/diff/8001/content/browser/in_process...
File content/browser/in_process_webkit/indexed_db_dispatcher_host.cc (right):

https://codereview.chromium.org/11567029/diff/8001/content/browser/in_process...
content/browser/in_process_webkit/indexed_db_dispatcher_host.cc:534: // XXX
can't swap here because params has a const
This comment still looks like a note-to-self?

https://codereview.chromium.org/11567029/diff/8001/content/common/indexed_db/...
File content/common/indexed_db/indexed_db_dispatcher.cc (right):

https://codereview.chromium.org/11567029/diff/8001/content/common/indexed_db/...
content/common/indexed_db/indexed_db_dispatcher.cc:348: for (size_t i = 0; i <
value->size(); ++i) {
Do this using memcpy instead, with appropriate compile time assertions that
sizeof value[0] == sizeof params.value[0] ?

[commit-bot: I haz the power, 2013-01-03 21:47:29]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/alecflett@chromium.org/11567029/11004

[commit-bot: I haz the power, 2013-01-04 00:48:37]

Change committed as 175066