RESOLVED FIXED99619
IndexedDB: Explicitly send null/undefined/integers to frontend IDBCallbacks 
https://bugs.webkit.org/show_bug.cgi?id=99619
Summary IndexedDB: Explicitly send null/undefined/integers to frontend IDBCallbacks
Alec Flett
Reported 2012-10-17 11:16:48 PDT
IndexedDB: Explicitly send null/undefined/integers to frontend IDBCallbacks
Attachments
Patch (27.58 KB, patch)
2012-10-17 11:23 PDT, Alec Flett 		no flags
DetailsFormatted DiffDiff
View All Add attachment proposed patch, testcase, etc.
Alec Flett
Comment 1 2012-10-17 11:23:18 PDT
Created attachment 169221 [details] Patch
Alec Flett
Comment 2 2012-10-17 11:24:11 PDT
jsbell/dgrogan: and here's the webkit side. can you take a look?
David Grogan
Comment 3 2012-10-17 14:51:50 PDT
Comment on attachment 169221 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=169221&action=review LGTM > Source/WebCore/Modules/indexeddb/IDBRequest.cpp:382 > + return onSuccess(SerializedScriptValue::numberValue(value)); These are deserialized right away but it looks there's no way to go from an integer to a ScriptValue without going through SerializedScriptValue?
Joshua Bell
Comment 4 2012-10-17 14:59:51 PDT
(In reply to comment #3) > These are deserialized right away but it looks there's no way to go from an integer to a ScriptValue without going through SerializedScriptValue? Maybe add a FIXME to add numberValue / undefinedValue to ScriptValue?
Alec Flett
Comment 5 2012-10-17 15:12:55 PDT
I'll add a FIXME - the static numbervalue/undefinedValue is pretty darn quick, relatively speaking and this code will be ripe for refactoring soon.
Alec Flett
Comment 6 2012-10-17 15:13:44 PDT
abarth@ - r? this is fairly mechanical
Adam Barth
Comment 7 2012-10-17 15:18:41 PDT
Comment on attachment 169221 [details] Patch Ok, but this is kind of dangerous in that someone passing a 0 literal might know realize which overload they're calling... Can we use different names, or do they all need to be called onSuccess?
Alec Flett
Comment 8 2012-10-17 15:21:56 PDT
abarth@ - FWIW the compiler should catch that ambiguity but yes, the fact that they're ALL named onSuccess is a bit problematic. This may also change as a part of our refactoring...
WebKit Review Bot
Comment 9 2012-10-17 15:57:35 PDT
Comment on attachment 169221 [details] Patch Clearing flags on attachment: 169221 Committed r131661: <http://trac.webkit.org/changeset/131661>
WebKit Review Bot
Comment 10 2012-10-17 15:57:39 PDT
All reviewed patches have been landed. Closing bug.
Note You need to log in before you can comment on or make changes to this bug.