Multi wildcard support finalization

[lrecknagel]

First - thanks for all the efforts already put in this project and also the PR with I`am picking up!

2nd - I hope I have picked up the prior work correctly, please tell me if I take a wrong path here.
(Followed the suggested manual: https://gist.github.com/jsumners/461ef7a64545108635cc437fde112721)

I´ve worked on finishing the multi-wildcard feature #36 with the missing test-coverage.
Also added handling for a undefined / null case which I`ve found.

Following checks have been made:

• npm run cov-ui
• npm run test
• npm run bench (see attached report)

fast-redact-bench-wc_finalization.txt

Please give me ping if I forgot something - hope we can bring this feature to life!

Best Lucas

[mcollina]

Is it possible that target is not an object? It's not covered by tests anyway

[lrecknagel]

I think its possible, applies for the example below:

const redact = fastRedact({
  paths: ['a[*].x.d[*].i.*.i']
})
const o = {
  a: [
    { x: { d: [ { i: undefined, j: 'NR' } ] } }
  ]
}

Only applies when the given path expects a deeper nested object but is undefined somewhere.
I`am not sure if this should be a case to be considered.

[mcollina]

what if subtarget is also an object?

[lrecknagel]

Could you pls explain your point here - I dont get it right now
subTarget can be an object, and it works imho

For the example, which is also in the tests (https://github.com/lrecknagel/fast-redact/blob/multi-wildcard-support-finalization/test/index.js#L1228)

const redactOut = fastRedact({
  paths: ['a[*].x.d[*].p.*.i'],
})
const obj = {
  a: [
    { x: { d: [ { i: undefined, j: 'NR' } ] }},
  ]
}
redactOut(obj)

we get the following state for subtarget:

[davidmarkclements]

exceptional work, thank you!

couple of points, also would you mind posting benchmarks of main branch vs your branch, and also adding benchmarks that cover this new code (if it's not covered already that is)

[mcollina]

Can you also rebase on top of main so we get some CI running? Thanks!

[mcollina]

lgtm on passing CI

[lrecknagel]

Thanks for the hints and comments! I´ve working on the benchmark part and will push it asap.

[lrecknagel]

• Two simple multi-wc benchmarks added.
• See bench of main and the current one:

001_fast-redact-bench-main.txt
002_fast-redact-bench-wc_finalization.txt

Depending on getObj which gives a relatively flat object it might be interesting to add another multi-wc bench on deeper nested objects. Would be great if you could express your opinion on that shortly.

[mcollina]

I'm good with the benchmarks I'm seeing

[davidmarkclements]

thanks @lrecknagel - it all seems good to me, I think let's get this merged, any extra benchmarks that could be PR'd after are always welcome but not demanded