60720 2011-05-12 13:22:52 -0700 IndexedDB index does not enforce uniqueness 2011-11-23 14:52:19 -0800 1 1 1 Unclassified WebKit New Bugs 528+ (Nightly build) Unspecified Unspecified RESOLVED INVALID P2 Normal --- 0 pilgrim webkit-unassigned dgrogan fishd hans jsbell pilgrim tony oldest_to_newest 403102 0 pilgrim 2011-05-12 13:22:52 -0700 Original test: http://mxr.mozilla.org/mozilla2.0/source/dom/indexedDB/test/test_indexes.html?force=1 This is an adaptation of part of a test in Mozilla's IndexedDB test suite. It creates an object store with one record, then creates a unique index on one of the keys, then attempts to add another record with a duplicate key (which should violate that unique index). Mozilla correctly throws a CONSTAINT_ERR, but WebKit does not throw. 403103 1 93328 pilgrim 2011-05-12 13:23:13 -0700 Created attachment 93328 test case 508429 2 jsbell 2011-11-23 14:52:19 -0800 The attached test is flawed - it's assuming that the index constraints are enforced synchronously during the add() operation and result in an exception being thrown. Per the spec, with the async API (which is what the test is using) the constraint is enforced asynchronously and the uniqueness constraint error is reported via the request object (i.e. the onerror handler will fire). The test could be corrected with: 73d72 < evalAndExpectException("objectStore.add({ name: 'Bob', height: 62, weight: 170 }, '237-23-7738');", "IDBDatabaseException.CONSTRAINT_ERR"); 75c74,76 < done(); --- > req = evalAndLog("req = objectStore.add({ name: 'Bob', height: 62, weight: 170 }, '237-23-7738');", "IDBDatabaseException.CONSTRAINT_ERR"); > req.onsuccess = unexpectedSuccessCallback; > req.onerror = done; ... but WebKit already has a test for this: LayoutTests/storage/indexeddb/index-unique.html 93328 2011-05-12 13:23:13 -0700 2011-05-12 13:23:13 -0700 test case index-unique-constraint.html text/html 3043 pilgrim PCFET0NUWVBFIGh0bWw+CjwhLS0KICBvcmlnaW5hbCB0ZXN0OiBodHRwOi8vbXhyLm1vemlsbGEu b3JnL21vemlsbGEyLjAvc291cmNlL2RvbS9pbmRleGVkREIvdGVzdC90ZXN0X2luZGV4ZXMuaHRt bD9mb3JjZT0xCiAgbGljZW5zZSBvZiBvcmlnaW5hbCB0ZXN0OgogICAgIiBBbnkgY29weXJpZ2h0 IGlzIGRlZGljYXRlZCB0byB0aGUgUHVibGljIERvbWFpbi4KICAgICAgaHR0cDovL2NyZWF0aXZl Y29tbW9ucy5vcmcvcHVibGljZG9tYWluL3plcm8vMS4wLyAiCi0tPgo8aHRtbD4KPGhlYWQ+Cjxs aW5rIHJlbD0ic3R5bGVzaGVldCIgaHJlZj0iLi4vLi4vLi4vZmFzdC9qcy9yZXNvdXJjZXMvanMt dGVzdC1zdHlsZS5jc3MiPgo8c2NyaXB0IHNyYz0iLi4vLi4vLi4vZmFzdC9qcy9yZXNvdXJjZXMv anMtdGVzdC1wcmUuanMiPjwvc2NyaXB0Pgo8c2NyaXB0IHNyYz0iLi4vLi4vLi4vZmFzdC9qcy9y ZXNvdXJjZXMvanMtdGVzdC1wb3N0LWZ1bmN0aW9uLmpzIj48L3NjcmlwdD4KPHNjcmlwdCBzcmM9 Ii4uL3Jlc291cmNlcy9zaGFyZWQuanMiPjwvc2NyaXB0Pgo8L2hlYWQ+Cjxib2R5Pgo8cCBpZD0i ZGVzY3JpcHRpb24iPjwvcD4KPGRpdiBpZD0iY29uc29sZSI+PC9kaXY+CjxzY3JpcHQ+CgpkZXNj cmlwdGlvbigiVGVzdCBJbmRleGVkREI6IGluZGV4IHdpdGggdW5pcXVlIGNvbnN0cmFpbnQiKTsK aWYgKHdpbmRvdy5sYXlvdXRUZXN0Q29udHJvbGxlcikKICAgIGxheW91dFRlc3RDb250cm9sbGVy LndhaXRVbnRpbERvbmUoKTsKCmZ1bmN0aW9uIHRlc3QoKQp7CiAgICBpbmRleGVkREIgPSBldmFs QW5kTG9nKCJpbmRleGVkREIgPSB3aW5kb3cuaW5kZXhlZERCIHx8IHdpbmRvdy53ZWJraXRJbmRl eGVkREIgfHwgd2luZG93Lm1vekluZGV4ZWREQjsiKTsKICAgIHNob3VsZEJlRmFsc2UoImluZGV4 ZWREQiA9PSBudWxsIik7CiAgICBJREJEYXRhYmFzZUV4Y2VwdGlvbiA9IGV2YWxBbmRMb2coIklE QkRhdGFiYXNlRXhjZXB0aW9uID0gd2luZG93LklEQkRhdGFiYXNlRXhjZXB0aW9uIHx8IHdpbmRv dy53ZWJraXRJREJEYXRhYmFzZUV4Y2VwdGlvbjsiKTsKICAgIHNob3VsZEJlRmFsc2UoIklEQkRh dGFiYXNlRXhjZXB0aW9uID09IG51bGwiKTsKICAgIElEQlRyYW5zYWN0aW9uID0gZXZhbEFuZExv ZygiSURCVHJhbnNhY3Rpb24gPSB3aW5kb3cuSURCVHJhbnNhY3Rpb24gfHwgd2luZG93LndlYmtp dElEQlRyYW5zYWN0aW9uOyIpOwogICAgc2hvdWxkQmVGYWxzZSgiSURCVHJhbnNhY3Rpb24gPT0g bnVsbCIpOwoKICAgIG5hbWUgPSB3aW5kb3cubG9jYXRpb24ucGF0aG5hbWU7CiAgICBkZXNjcmlw dGlvbiA9ICJNeSBUZXN0IERhdGFiYXNlIjsKICAgIHJlcXVlc3QgPSBldmFsQW5kTG9nKCJpbmRl eGVkREIub3BlbihuYW1lLCBkZXNjcmlwdGlvbikiKTsKICAgIHJlcXVlc3Qub25zdWNjZXNzID0g b3BlblN1Y2Nlc3M7CiAgICByZXF1ZXN0Lm9uZXJyb3IgPSB1bmV4cGVjdGVkRXJyb3JDYWxsYmFj azsKfQoKZnVuY3Rpb24gb3BlblN1Y2Nlc3MoKQp7CiAgICBkZWJ1Zygib3BlblN1Y2Nlc3MoKToi KTsKICAgIGRiID0gZXZhbEFuZExvZygiZGIgPSBldmVudC50YXJnZXQucmVzdWx0Iik7CgogICAg b2JqZWN0U3RvcmVOYW1lID0gIlBlb3BsZSI7CgogICAgcmVxdWVzdCA9IGV2YWxBbmRMb2coInJl cXVlc3QgPSBkYi5zZXRWZXJzaW9uKCcxJykiKTsKICAgIHJlcXVlc3Qub25zdWNjZXNzID0gY3Jl YXRlQW5kUG9wdWxhdGVPYmplY3RTdG9yZTsKICAgIHJlcXVlc3Qub25lcnJvciA9IHVuZXhwZWN0 ZWRFcnJvckNhbGxiYWNrOwp9CgpmdW5jdGlvbiBjcmVhdGVBbmRQb3B1bGF0ZU9iamVjdFN0b3Jl KCkKewogICAgZGVsZXRlQWxsT2JqZWN0U3RvcmVzKGRiKTsKCiAgICBvYmplY3RTdG9yZSA9IGV2 YWxBbmRMb2coIm9iamVjdFN0b3JlID0gZGIuY3JlYXRlT2JqZWN0U3RvcmUob2JqZWN0U3RvcmVO YW1lKTsiKTsKCiAgICBkZWJ1ZygiRmlyc3QsIGFkZCBhbGwgb3VyIGRhdGEgdG8gdGhlIG9iamVj dCBzdG9yZS4iKTsKICAgIG9iamVjdFN0b3JlRGF0YSA9IGV2YWxBbmRMb2coIm9iamVjdFN0b3Jl RGF0YSA9IHsga2V5OiAnMjM3LTIzLTc3MzInLCB2YWx1ZTogeyBuYW1lOiAnQm9iJywgaGVpZ2h0 OiA2MCwgd2VpZ2h0OiAxMjAgfSB9OyIpOwogICAgcmVxdWVzdCA9IGV2YWxBbmRMb2coInJlcXVl c3QgPSBvYmplY3RTdG9yZS5hZGQob2JqZWN0U3RvcmVEYXRhLnZhbHVlLCBvYmplY3RTdG9yZURh dGEua2V5KTsiKTsKICAgIHJlcXVlc3Qub25lcnJvciA9IHVuZXhwZWN0ZWRFcnJvckNhbGxiYWNr OwogICAgcmVxdWVzdC5vbnN1Y2Nlc3MgPSBjcmVhdGVBbmRUZXN0SW5kZXhlczsKfQoKZnVuY3Rp b24gY3JlYXRlQW5kVGVzdEluZGV4ZXMoKQp7CiAgICBkZWJ1ZygiTm93IGNyZWF0ZSB0aGUgaW5k ZXhlcy4iKTsKICAgIGluZGV4RGF0YSA9IGV2YWxBbmRMb2coImluZGV4RGF0YSA9IHsgbmFtZTog J25hbWUnLCBrZXlQYXRoOiAnbmFtZScsIG9wdGlvbnM6IHsgdW5pcXVlOiB0cnVlIH0gfTsiKTsK ICAgIGV2YWxBbmRMb2coIm9iamVjdFN0b3JlLmNyZWF0ZUluZGV4KGluZGV4RGF0YS5uYW1lLCBp bmRleERhdGEua2V5UGF0aCwgaW5kZXhEYXRhLm9wdGlvbnMpOyIpOwoKICAgIGRlYnVnKCJDaGVj ayB0aGF0IHRoZSBuYW1lIGluZGV4IGVuZm9yY2VzIGl0cyB1bmlxdWUgY29uc3RyYWludC4iKTsK ICAgIG9iamVjdFN0b3JlID0gZXZhbEFuZExvZygib2JqZWN0U3RvcmUgPSBkYi50cmFuc2FjdGlv bihvYmplY3RTdG9yZU5hbWUsIElEQlRyYW5zYWN0aW9uLlJFQURfV1JJVEUpLm9iamVjdFN0b3Jl KG9iamVjdFN0b3JlTmFtZSk7Iik7CiAgICBldmFsQW5kRXhwZWN0RXhjZXB0aW9uKCJvYmplY3RT dG9yZS5hZGQoeyBuYW1lOiAnQm9iJywgaGVpZ2h0OiA2Miwgd2VpZ2h0OiAxNzAgfSwgJzIzNy0y My03NzM4Jyk7IiwgIklEQkRhdGFiYXNlRXhjZXB0aW9uLkNPTlNUUkFJTlRfRVJSIik7CgogICAg ZG9uZSgpOwp9Cgp2YXIgc3VjY2Vzc2Z1bGx5UGFyc2VkID0gdHJ1ZTsKCnRlc3QoKTsKCjwvc2Ny aXB0Pgo8L2JvZHk+CjwvaHRtbD4KCg==