In most of the tests shown, there are “missteps” in the repeating linear gradients. These missteps are places where a pattern is on where it should be off, or vice versa. They’re really obvious in the first three test cases on the referenced page, though they show up in all the test cases eventually. I can’t describe exactly what you’ll see at page load since these missteps in repetition show up in different places for different element background area widths. This is why, when you resize the browser window, the missteps dance around. It also seems pretty clear that the rate of missteps decreases as the pattern width widens. For example, test #10 only shows them intermittently as the window size changes, so the pattern is correct most of the time but wrong every now and again. For test #1 it’s the inverse: most window widths will show one or more missteps, but every now and again there’s a width where no missteps occur within the visible background area. And even better, the missteps land in different spots depending on how you declare your gradient, even when two declarations should be equivalent. That’s why each test has two different syntax patterns: to illustrate that the missteps are sensitive to syntax. Checked in both Safari 5.1.7 and WebKit Nightly Version 5.1.7 (6534.57.2, r119050).
I can repro this bug on Win7 Chrome28 so platform should be All
Safari, Chrome, and Firefox all agree on rendering for this test case. I don't believe there is any remaining compatibility issue. If I am missing a subtlety in the results, could you please REOPEN this bug with some additional details to help me understand?
Yep, looks like this was resolved at some point in the last decade. No re-opening from me. Thanks, Brent!
(In reply to Eric Meyer from comment #3) > Yep, looks like this was resolved at some point in the last decade. No > re-opening from me. Thanks, Brent! Thank you for confirming -- and sorry this sat unreviewed for so long :-(.