- From: Karen Anderson (IE) <Karen.Anderson@microsoft.com>
- Date: Tue, 1 May 2012 23:46:58 +0000
- To: "public-web-perf@w3.org" <public-web-perf@w3.org>
- Message-ID: <748A9FD1BD28E74F971C8D69F63494720B2875@BL2PRD0310MB350.namprd03.prod.outlook.co>
Hi All, I have made changes to the tests identified by Action 99 concerning the timing attributes order. We identified in previous calls that we do not necessarily need to distinguish the time spent in the onbeforeunload and onunload event handlers and only need to ensure that the timing order is correct. Further having sleeps within these event handlers is inappropriate for some UAs. Accordingly, I have: * Removed the sleeps from blank_page_unload.html and replaced with a single variable assignment to ensure that the handler executes, but does not necessarily need to consume time. * Removed the tests in test_timing_attributes_order.htm that verify the timing deltas in these event handlers. In addition, I found an interoperable solution such that all UAs are passing the document readiness test. The highlight of the test update is that the test contains a frame which will traverse the ready states. As ready states are discovered it adds properties to its document object for review by the parent page. The parent page adds this frame dynamically, ensuring the earliest attachment to these states. Test verification checks for the properties added to the frame document object and appropriately passes/fails the test. The changes are live and can be reviewed at: http://w3c-test.org/webperf/tests/submission/Microsoft/NavigationTiming/test_timing_attributes_order.htm http://w3c-test.org/webperf/tests/submission/Microsoft/NavigationTiming/test_document_readiness_exist.html Please let me know if you have any questions. Thanks, Karen
Received on Tuesday, 1 May 2012 23:48:19 UTC