Created attachment 375532 [details] test case Open the attached test case. Expected: A green circle drawn at (0, 0). Result: A green circle drawn at (100, 100). The 'cx' and 'cy' attributes are changed through changing the style of the circle element. Because the strings are incorrect css values, the expectation is no change in the values of these attributes should happen.
Seems like the test case is working now.
I am unable to reproduce this bug in Safari 15.5 on macOS 12.4 and "green" circle is drawn on (100,100) and it is confirmed by Comment 01 as well. From other browser perspective, Chrome Canary 105 matches with Safari behavior and draw it at (100,100) while Firefox Nightly 104 draws it at (0,0). Thanks!
Based on Ahmad's testing, closing as CONFIG CHANGED.
Aren't we saying that this is not to be fixed, and if so, why? Per the report, (0, 0) is expected, but we are getting (100, 100), same as in 2019. This looks similar to bug 94352.
REOPEN based on Alexey's comment. I didn't realize Said filed this against a spec -- I assumed matching Chrome was sufficient for compatibility.
<rdar://problem/97224363>
NOTE - we are passing all tests here: https://wpt.fyi/results/svg/geometry/parsing?label=master&label=experimental&aligned&q=invalid and matching with Firefox Nightly, which is same the case with this test case. Safari & Firefox are matching, do we expect to match Chrome here. If yes, then I think we might need to update tests as well.