By analyzing the blame information on this pull request, we identified @markelog, @jeresig and @guybedford to be potential reviewers

Ref jquery-migrate issue.

I wonder if we should add tests not only for the jQuery constructor for these selectors, but specifically and isolated for these functions

If we'll start warning about this soon (as in potentially Migrate 3.0) is there a UI ticket for this?

@markelog I'm not sure what you're suggesting. We have tests for the added selectors.

@timmywil

we have stuff like -

assert.equal( jQuery( "#table td:visible" ).length, 1, "hidden cell is not perceived as visible (#4512). Works on table elements" );

but should we have stuff like -

  assert.equal(jQuery.expr.pseudos.visible(td), false);

instead?
Should we have direct tests? Cause tests through jQuery constructor or jQuery#(find | is) are proxies, if they work, doesn't mean the jQuery.expr.pseudos.visible methods works as expected, maybe jQuery constructor as unindent decorator did something along the way and positive result is just a side-effect.

In other words, i'm talking about perfectionism :)

if they work, doesn't mean the jQuery.expr.pseudos.visible methods works as expected

I think it does with the tests we have.

I think it does with the tests we have.

Some coverage tools would also agree with me in that assessment, but yeah, i guess we can take care of that when we implement it.

Some coverage tools would also agree with me in that assessment

How so? It's still hitting the codepaths.

For example, the 2 relevant lines in "src/css/hiddenVisibleSelectors.js" are 7 and 10, and they are executed, so coverage would be good, right?