test: fix pummel/test-net-connect-memleak

[Trott]

A loop that generates a long array is resulting in a RangeError. Moving
to Array.prototype.fill() along with the ** operator instead of using a
loop fixes the issue.

Checklist

• make -j4 test (UNIX), or vcbuild test (Windows) passes
• commit message follows commit guidelines

[nodejs-github-bot]

@Trott build started: https://ci.nodejs.org/blue/organizations/jenkins/node-test-pull-request-lite-pipeline/detail/node-test-pull-request-lite-pipeline/210/pipeline

[Trott]

Since pummel tests are not run in CI, a lite CI is likely sufficient for this. (It will lint the test at least.)

[addaleax]

LGTM

The test could use a larger makeover and be moved to parallel, I think. With async_hooks, we now have a public API to make GC observable from userland, so we can avoid using a huge amount of memory just for that :)

[Trott]

@addaleax I pondered seeing what would happen if I moved it to sequential as it only took four seconds to run on my laptop, but I worried that assert(reclaimed > 128 * 1024); // It's around 256 MB on x64. would be unreliable across platforms.

[addaleax]

@Trott Yeah, sorry for being unclear – I’d only suggest moving out of pummel if we get rid of the huge array that we use here.

(If you want, it’s totally fine to leave a TODO here and/or ping me when this lands 🙂 )

[Trott]

@Trott Yeah, sorry for being unclear

@addaleax Oh no, you were clear. Sorry. I was just mentioning my own considerations for moving it, but yeah, it's all irrelevant if we move to async_hooks like you suggested. TBH, in retrospect, I probably shouldn't have even left the comment. :-D

[mhdawson]

LGTM

[Trott]

Landed in ff958ad

[addaleax]

The promised follow-up is in #21794 :)