I've verified that it fails in dev channel. Haven't tested beta/stable.

Attaching screenshots of it broken in Chrome and working in Firefox for reference.

Deleted: Screen Shot 2016-08-14 at 9.51.45 AM.png 1.2 MB

	Screen Shot 2016-08-14 at 9.51.45 AM.png 1.2 MB View Download

Deleted: Screen Shot 2016-08-14 at 9.52.05 AM.png 1.1 MB

	Screen Shot 2016-08-14 at 9.52.05 AM.png 1.1 MB View Download

You are probably looking for a change made after 390147 (known good), but no later than 390154 (first known bad).
CHANGELOG URL:
  https://chromium.googlesource.com/chromium/src/+log/8ab7647d637da9a6d42178742fcb7a4328d5b6d3..b80c579439ab737f82cddb5c73c91bc1daacd354

So likely the flexbox abspos change:
https://chromium.googlesource.com/chromium/src/+/b80c579439ab737f82cddb5c73c91bc1daacd354

I'm gonna have to look into this but may be wontfix, especially if it is also broken in Edge

If this change isn't web compatible, maybe we should consider reverting both the change and the spec change? What did we gain by changing this?

cc'ing Greg for opinions on web compatibility (Edge has shipped this for a long time)

There have been relatively few reports of broken websites so far. It does gain us a more sensible behavior for positioning abspos children of flexboxes.

Overall it's been pretty light as of late. We had a few heavy hitters early on but we reached out to them (the majority of them were Google properties actually) and they updated. Longtail sites, I haven't seen the same usage of abspos within a flex container.

And just as seen in the other cases, there doesn't seem to be a reason to why they're actually setting this to be absolute positioned.

It looks like this website works now. They're setting right: 0 on the abspos item, I guess they didn't before but I did not take a close look at their code at the time.

Standards discussion ongoing in https://github.com/w3c/csswg-drafts/issues/401

But I basically agree with Greg because we do not get a lot of bugs about this. Either way, closing this bug because the website works now.