3330 2005-06-07 20:42:30 -0700 REGRESSION (125-300): JavaScript "children" array includes all descendants, not just immediate children 2005-07-03 07:36:12 -0700 1 1 1 Unclassified WebKit DOM 412 Mac OS X 10.4 VERIFIED FIXED http://colloquy.info/test/children.html P2 Critical --- 1 timothy mjs hyatt oldest_to_newest 10842 0 timothy 2005-06-07 20:42:30 -0700 Radar: 4063841 Summary: The "children" array on nodes returns children's children in on flattened array. This is a regression from Safari 1.2 and from the DOM specs. A test case is setup here. It "children.length" should return 5 for the number of div elements. http://colloquy.info/test/children.html Steps to Reproduce: 1) Access the children array on a DOM node with JavaScript. Expected Results: Only immediate children should be in the array, not the children's children, etc. Actual Results: The array has more elements than expected if the immediate children have children elements also. They all get flattened out into the main array (a recursive behavior). Regression: This is a regression from Safari 1.2 builds. Notes: Test case at: http://colloquy.info/test/children.html 10846 1 hyatt 2005-06-07 20:45:35 -0700 This is actually a bug with collections in general. traverseNextItem begins by calling traverseNextNode, but traverseNextNode is a deep traversal. This means that shallow traversals like children are deep when they should not be when you make repeated calls to traverseNextItem to iterate (as calcLength does). The same is true of some of the other collections that should not be deep, and so this bug could be the cause of other HTML collection DOM bugs. 11130 2 2168 timothy 2005-06-08 22:35:51 -0700 Created attachment 2168 Fix 11131 3 mjs 2005-06-08 23:33:21 -0700 I sent review comment to Timothy by email. 12685 4 timothy 2005-06-20 07:55:59 -0700 This was fixed with the attached patch, and landed by Maciej. Thought I would close this and help out. 2168 2005-06-08 22:35:51 -0700 2005-06-08 22:35:51 -0700 Fix 3330.patch text/plain 1385 timothy SW5kZXg6IENoYW5nZUxvZwo9PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09 PT09PT09PT09PT09PT09PT09PT09PT09PT09ClJDUyBmaWxlOiAvY3ZzL3Jvb3QvV2ViQ29yZS9D aGFuZ2VMb2csdgpyZXRyaWV2aW5nIHJldmlzaW9uIDEuNDI1MApkaWZmIC11IC1wIC1yMS40MjUw IENoYW5nZUxvZwotLS0gQ2hhbmdlTG9nCTIwMDUvMDYvMDggMDg6Mzk6MTYJMS40MjUwCisrKyBD aGFuZ2VMb2cJMjAwNS8wNi8wOSAwNDo1MDoyNQpAQCAtMSwzICsxLDExIEBACisyMDA1LTA2LTA4 ICBUaW1vdGh5IEhhdGNoZXIgIDx0aW1vdGh5QGNvbGxvcXV5LmluZm8+CisKKyAgICAgICAgUmV2 aWV3ZWQgYW5kIGNvbW1pdHRlZCBieSB4eHguCisKKyAgICAgICAgLSBmaXhlZCBodHRwOi8vYnVn emlsbGEub3BlbmRhcndpbi5vcmcvc2hvd19idWcuY2dpP2lkPTMzMzAKKyAgICAgICAgYW5kIDxy ZGFyOi8vcHJvYmxlbS80MDYzODQxPgorICAgICAgICBSRUdSRVNTSU9OICgxMjUtMzAwKTogSmF2 YVNjcmlwdCAiY2hpbGRyZW4iIGFycmF5IGluY2x1ZGVzIGFsbCBkZXNjZW5kYW50cywgbm90IGp1 c3QgaW1tZWRpYXRlIGNoaWxkcmVuCisKIDIwMDUtMDYtMDggIEFuZGVycyBDYXJsc3NvbiAgPGFu ZGVyc2NhQG1hYy5jb20+CiAKIAlSZXZpZXdlZCBhbmQgY29tbWl0dGVkIGJ5IE1hY2llai4KSW5k ZXg6IGtodG1sL2h0bWwvaHRtbF9taXNjaW1wbC5jcHAKPT09PT09PT09PT09PT09PT09PT09PT09 PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PQpSQ1MgZmlsZTogL2N2 cy9yb290L1dlYkNvcmUva2h0bWwvaHRtbC9odG1sX21pc2NpbXBsLmNwcCx2CnJldHJpZXZpbmcg cmV2aXNpb24gMS4yNQpkaWZmIC11IC1wIC1yMS4yNSBodG1sX21pc2NpbXBsLmNwcAotLS0gaHRt bF9taXNjaW1wbC5jcHAJMjAwNS8wNS8xMSAwMDo1ODoyOAkxLjI1CisrKyBodG1sX21pc2NpbXBs LmNwcAkyMDA1LzA2LzA5IDA0OjUwOjI2CkBAIC0xMjksNyArMTI5LDkgQEAgTm9kZUltcGwgKkhU TUxDb2xsZWN0aW9uSW1wbDo6dHJhdmVyc2VOZQogewogICAgIGFzc2VydChjdXJyZW50KTsKIAot ICAgIGN1cnJlbnQgPSBjdXJyZW50LT50cmF2ZXJzZU5leHROb2RlKG1fYmFzZS5nZXQoKSk7Cisg ICAgaWYgKHR5cGUgPT0gTk9ERV9DSElMRFJFTiAmJiBtX2Jhc2UuZ2V0KCkgIT0gY3VycmVudCkK KyAgICAgICAgY3VycmVudCA9IGN1cnJlbnQgLT4gbmV4dFNpYmxpbmcoKTsKKyAgICBlbHNlIGN1 cnJlbnQgPSBjdXJyZW50LT50cmF2ZXJzZU5leHROb2RlKG1fYmFzZS5nZXQoKSk7CiAKICAgICB3 aGlsZSAoY3VycmVudCkgewogICAgICAgICBpZihjdXJyZW50LT5ub2RlVHlwZSgpID09IE5vZGU6 OkVMRU1FTlRfTk9ERSkgewo=