Marking this as blocked on  bug 706118  due to csswg-test being currently in-flight merging into wpt.

 Issue 712950  has been merged into this issue.

From that other issue, https://codereview.chromium.org/2809593002/ has some discussion about this.

Vitally:

> Matching up those from css2.1/20110323 to the current testsuite should
be easy as the filenames are unlikely to have changed; those directly in
css2.1 will be far harder.

To add: those in LayoutTests/css2.1/t[[:digit:]]*-c[[:digit:]]*-.* almost certainly all come from what's now in css/CSS2/css1 in WPT (and some of them are now reftests), which would allow us to get rid of more than half of them.

Note, https://github.com/w3c/web-platform-tests/tree/master/css/CSS2/css1 is currently not imported.

Does anything think it would be a good idea to import that directory from wpt and delete all of LayoutTests/css2.1?

Certainly replacing LayoutTests with the WPT version is almost certainly a clear win.  I'd be happy just doing that if meade@ (style TL) is OK with it.

I don't have context on what exactly the 'css1' directory covers though to say what the value is in having it at all, perhaps it also reflects an old version of specs and is superceded by another directory that covers the modern versions?
  +geoffers for his comment.

I'd be fine with replacing LayoutTests with WPT versions, but I also don't have historical context on what's in the css1 directory...

> Does anything think it would be a good idea to import
> https://github.com/w3c/web-platform-tests/tree/master/css/CSS2/css1 from 
> wpt and delete all of LayoutTests/css2.1?

Not quite. The suites do not match. 

LayoutTests/css2.1 contains 493 tests. These have became multiple subdirectories in w-p-t. Searching by filenames:

shand-border -> CSS2/borders
shand-font -> CSS2/fonts
case|escapes|ident|atkeyw|atrule|import|comments| -> CSS2/syntax
fwd-parsing -> CSS2/css1
syntax-01 -> ?????
ex-len -> CSS2/css1
etc......

LayoutTests/css2.1/20110323 contains 263 tests. Again, these map to multiple subdirectories in w-p-t. 

abspos-non-replaced-width-margin -> CSS2/csswg-issues/submitted/css2.1/
at-import -> CSS2/cascade

What I'd do is:

1) survey existing tests: what w-p-t directories our old imports map to? Perfect day task for an intern?

2) Figure out what additional directories should we import, and import them.

3) Blow away CSS2 stuff.

If we have filename -> dir mapping, steps 2&3 can be done piecemeal, one w-p-t directory import + partial test removal.

I'd love to see our ancient test suites replaced. The only gotcha is that our old test suites have been fully debugged, there are no failing expectations in css2.1/. There are lots of failing expectations in external/wpt/css/CSS2

> 1) survey existing tests: what w-p-t directories our old imports map to? Perfect day task for an intern?

Looking at the more recent import, a curious glance over would suggest they're from all over the place, and given the build system it's entirely plausible they're not so nicely arranged in the (unbuilt) directories in wpt.

The older import I think was of the entire testsuite as then existed, so will probably be even more all over the place.

Another thing that occurs to me is that given we only run reference tests from wpt and some of the tests we have in LayoutTests/css2.1 are pixel tests there's a risk that we'll lose coverage as a result of that.

> I'd love to see our ancient test suites replaced. The only gotcha is that our old test suites have been fully debugged, there are no failing expectations in css2.1/. There are lots of failing expectations in external/wpt/css/CSS2

Glancing over https://wpt.fyi/css/CSS2 doesn't lead to me to find too many tests that fail in three browsers (minus Chrome, because the Chrome results are currently broken because of a ChromeDriver bug, awaiting a new ChromeDriver release), which means some may well just be showing up Chrome bugs.

It also doesn't look like there's *that* much in way of changes to the tests themselves, so it's probably doable for someone to just look over all the changes made and check whether the issues they're resolving need done upstream as well.

This issue has been Available for over a year. If it's no longer important or seems unlikely to be fixed, please consider closing it out. If it is important, please re-triage the issue.

Sorry for the inconvenience if the bug really should have been left as Available.

For more details visit https://www.chromium.org/issue-tracking/autotriage - Your friendly Sheriffbot

Not sure whether this is worth doing but keeping open for now.