wpt computing-row-measure-0.html fix

2 tests in this test suite seem inconsistent:

test#2 asserts that

tbody.height=10px > tr.height=1px > td.height=1px
implies td.offsetHeight = 1px

test#4 asserts that

tbody.height=10px > tr > td.height=1px
implies td.offsetHeight = 10px

Edge 17 is the only browser that agrees with #2 and #4
FF agrees with #2, but not #4
Chrome agrees with #4, but not #2
Safari agrees with #4, but not #2

To me, #2 and #4 seem to be in conflict.
Either tbody height propagates to rows, or it does not.

The problem is that #2 is overconstrained.

My suggestion is that tbody height always propagates to tr.

Bug: 958381
Change-Id: I28bfd108c67968d31d0372b536c316c997d2d958
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2586097
Reviewed-by: Ian Kilpatrick <[email protected]>
Commit-Queue: Ian Kilpatrick <[email protected]>
Cr-Commit-Position: refs/heads/master@{#845515}

css/css-tables/height-distribution/computing-row-measure-0.html

@@ -54,7 +54,7 @@ <h4>"Computing row measures"</h4>
         [
             "Checking intermediate min-content height for span 1 (2)",
             document.getElementById('two').offsetHeight,
-            1
+            10
         ],
         [
             "Checking intermediate min-content height for span 1 (3)",