However, those spec sections were removed by @fantasai in w3c/csswg-drafts@829a074, and the remaining spec text is a lot shorter.

That spec-edit was for w3c/csswg-drafts#1432, which had another quite-relevant spec edit that landed as well:
w3c/csswg-drafts@a595144

This second spec-edit was in response to a resolution in w3c/csswg-drafts#1432 (comment) which removed the influence of align-content/justify-content on abspos descendants, I think.

That may mean that all of the tests with align-content or justify-content in their names probably are wrong and need amended reference cases... Right now the tests are asserting that these properties do influence the positioning of the abspos children, as the spec used to require; but now the spec says they do not have any influence, I think. (Checking my understanding in w3c/csswg-drafts#1432 (comment) )

RE your main question here: I don't know offhand what the correct spec text to link to would be, at this point, but I can try to find out (and see if any of the other tests are also obsolete).

@dholbert is there a change in Gecko needed to align with the spec changes that happened here? Would you expect that after that, the Gecko results would be largely the same as Chrome and Firefox? If so, then most of these tests may not have value in finding interop problems, and we might be able to leave just the few that do have different cross-browser results, and a few negative tests to ensure that align-content/justify-content have no effect on abspos descendants.

Regardless of the details shake out to be, your feedback and help is much appreciated!

I need to page this spec area back into my head (and review the edits that have happened since I last looked at it), but it looks to me like a change in Gecko is indeed needed, at least for the align-content/justify-content behavior. I filed https://bugzilla.mozilla.org/show_bug.cgi?id=1682641 on that. For those tests, I think you're right that we don't need a zillion negative tests -- just a few would probably do.

That's less than half of the tests in this directory, though. I'll aim to get back to you with thoughts about the other tests in this directory (the *-self tests) within the next few days.

Thanks for your work on migrating/merging these tests with the rest of the suite!

Thanks @dholbert, much appreciated! I'll soon be out for the holidays but I'll check back in eventually :)

Sorry, the holidays interrupted my "the next few days" plans here. :) Back from holidays now, and this is on my list of things-to-look-at-first.

I think I misunderstood w3c/csswg-drafts@829a074 and I forgot that another section of spec text still existed. As a result, I think these tests are still valid (particularly the justify-content and align-content ones that I was initially suspicious of).

For flexbox, these tests are essentially anchored in Flexbox section 4.1, https://drafts.csswg.org/css-flexbox-1/#abspos-items , which says:

The static position of an absolutely-positioned child of a flex container is determined such that the child is positioned as if it were the sole flex item in the flex container, assuming both the child and the flex container were fixed-size boxes of their used size

(I had mistakenly thought this text had been removed, but it's still there, and I think still authoritative.)

This spec-text implies that the static position should be influenced by the flex container's justify-content (which Chrome and Firefox agree on FWIW as shown in this testcase, modulo a few small pieces failing in Chrome due to https://bugs.chromium.org/p/chromium/issues/detail?id=1011718#c12 ).

It also implies that the abspos children's static position should be influenced by the flex container's align-content, if the flex container is multiline (i.e. has flex-wrap:wrap). Chrome does not agree with Firefox on that point -- Chrome seems to just disregard align-content entirely when determining static position of abspos children in a multi-line flex container. That seems probably like a Chrome bug which I should file if I haven't already done so. That discrepancy can be seen in this testcase where Chrome renders all of the children the same in each section, despite the varying align-content values on their flex container.

So: bottom-line, the flex- tests should probably all link to https://www.w3.org/TR/css-flexbox-1/#abspos-items , and similarly the grid tests should probably all link to https://www.w3.org/TR/css-grid-1/#abspos-items , and they should probably move to those respective subdirectories.

( @foolip I think the last paragraph of my comment above is the answer to your original question here. It sounded like you might be up for making those changes/migrations -- if so, I applaud & support you doing so. :) Thanks!)

Thank you for digging into that, @dholbert! I take that to mean we should move the tests into css/flexbox/ and css/grid/ and would be happy to send a PR for that.

Yes. Thank you!

I've sent #27174 to do the move.