fix: cache family when calling libc()

[styfle]

We can assume that the same process will return the same report and thus we can cache the resulting family in memory to make it faster for subsequent invocations.

For example, this function is called in a loop here:
https://github.com/npm/cli/blob/780afc50e3a345feb1871a28e33fa48235bc3bd5/workspaces/arborist/lib/arborist/build-ideal-tree.js#L212

References

Fixes #116 (comment)

[Tofandel]

I just tried this approach in #120 but there also seems to be an issue with this when getReport is called later in the program it takes very very long to complete (think 3minutes which is forever in computer time, that might be a bug in node because cpu is at 0% during that time, or not, after all this method is meant for debugging only), caching it is still not enough, it needs to be ran before anything in the program runs because then it's instant (2ms on my machine)

Here is a log of console.time('report') and console.timeEnd('report') before and after the getReport call
when run in the end of the upgrade command:
⠼report: 3:10.573 (m:ss.mmm)
when run in the beginning of the process at the top level
report: 1.943ms

[Tofandel]

For anyone wondering why it's freezing and it's taking so much time to get a report, it has to do with this line which does a synchronus reverse DNS lookup in getReport https://github.com/nodejs/node/blob/5633c62219e199baac833e8862d60333d85dc3d3/src/node_report_utils.cc#L29

A simple test case

console.time("report");
process.report.getReport();
console.timeEnd("report");
await fetch('https://example.com/');
console.time("report");
process.report.getReport();
console.timeEnd("report");
await fetch('https://example.com/');
await fetch('https://example.com/');
console.time("report");
process.report.getReport();
console.timeEnd("report");

report: 1.909ms
report: 20.188s
report: 40.220s

This does look like a bug and I reported it to the node repo
nodejs/node#55576