Shorten the text in /css/css-view-transitions/new-content-is-inline.html to not linewrap

[dholbert]

Test List

https://wpt.fyi/results/css/css-view-transitions/new-content-is-inline.html

Rationale

Prior to web-platform-tests/wpt#45809 , the text-strings in this WPT did not linewrap at all in Chrome or Safari -- so that demonstrates that linewrapping & multi-line text strings are not part of what it's trying to test.[1]

But when it was migrated to use 20px-sized Ahem (in web-platform-tests/wpt#45809 ), the text started linewrapping. That happens to trigger Firefox bug 2012343 (which is also independently and more-directly tested by another view transitions test, multiline-span-with-overflowing-text-and-box-decorations.html.

Hence, I propose that we adjust this test to shorten the text so that it doesn't linewrap anymore, in the same way that I recently did for the neighbor old-content-is-inline.html in #1250.

Benefits of this proposal:

• Avoids triggering the unrelated Firefox bug (which as noted is already tested by a separate WPT) so that Firefox can pass this test.
• Lets us remove the fuzzy annotation (I believe all browsers pass with no fuzz if we don't linewrap).
• Makes this test much more consistent with the neighboring analogous test old-content-is-inline.html

[1] Moreover, if I modify a historical pre-Ahem version of the testcase slightly to make the text barely wrap in Chrome, it fails in Chrome; this isn't a Chrome bug, but is rather an indication that the test itself isn't robust in the face of linewrapping.

[dholbert]

Elaborating on this slightly:

Benefits of this proposal:
[...]

• Makes the test much more consistent with old-content-is-inline.html

With the PR merged, the two tests are back to being basically-identical in their markup (after they diverged for a while, after #1250 added Ahem to one but not the other). Diffing them shows that they're nearly the same now, aside from the thing they're testing (which is whether the "new" or "old" state is hidden), and I think that's the intent.

Having them be nicely diffable is worthwhile from understanding how the tests relate to each other.