Skip converter test when network atomics are available

[benharsh]

The code path for network atomics happens to run into a bug in the converter, which causes the test to fail.

[benharsh]

I was able to reproduce the initial bug on my laptop, and fixed it, assuming that would be the end of this. Alas, there's another bug related to network atomics, so instead of putting more time into this I'm inclined to just skip the test.

[bradcray]

That works for me, thanks! The other fix you mentioned is the one you already merged, is that right? (which resolved this test in other configs?)

The only thing that nags at me a little bit here is what made this only start failing on the 19th. All the PRs that went in on the 18th look pretty innocuous. But I don't think that's probably worth pursuing unless it makes the cause and fix obvious.

In any case, I don't mind skipping it in this config given how "advanced" it is and the converter's focus so far.

[benharsh]

I merged chapel-lang/chapel#28390 a couple of days ago, which fixed the original error message we were seeing. I somewhat foolishly only tested this on my mac, and so I didn't see the failure related to network atomics that we now see in nightly testing.