wlott

This turned out to be quite a long post, so I'll summarize first: - complications listed above are just the tip of the iceberg. - Supreme Commander addressed this issue...

> I'd be happy to see this, but we should add a regression test that compares this behavior to the web standard. That will both ensure that we're actually observing...

It looks like there are additional bugs that cause the test fixture stuff to produce what I suspect are false positives. I add a new test fixture with: ```html Test...

I've been working on fixing this all day, and well, it has snowballed way beyond what I first thought it would entail. Part of the problem is that I read...

Okay then, I'll start working out the details. (I retired a couple weeks ago and was thinking that it would be fun to contribute to some open source projects--this can...

Bumped into something I'm not sure how to handle. Near the end of [3.1. Managing Overflow: the overflow-x, overflow-y, and overflow properties](https://www.w3.org/TR/css-overflow-3/#overflow-control), it says: > The [visible](https://www.w3.org/TR/css-overflow-3/#valdef-overflow-visible)/[clip](https://www.w3.org/TR/css-overflow-3/#valdef-overflow-clip) values of [overflow](https://www.w3.org/TR/css-overflow-3/#propdef-overflow)...

I am close. Oh so close. I even turned on the `scroll_{width,height}` checking in the generated tests for visible and clip nodes. But a small number of tests are failing....

I had it backwards in my write-up. When I use `

Ah, then the problem looks to be with the test extraction code. Or other assumptions. Because that inner div has no children, gentest produces a leaf node: ``` let node0...

Stranger and stranger. When given this test case: ``` ``` Chrome produces different scrollable overflow with zero-width children for in-flow and out-of-flow. `id="inflow"` gets a 200px scroll region but `id="out-of-flow"`...