fix: setting up node with modified config

[gabrielmer]

Description

We currently modify a copy of our configurations when initializing a node in

nwaku/waku/factory/waku.nim

Lines 104 to 106 in 1713f56

	proc init*(T: type Waku, conf: WakuNodeConf): Result[Waku, string] =
	var confCopy = conf
	let rng = crypto.newRng()

However, we don't use this modified configuration when setting up the protocols and initializing the node, which causes mismatches between them.

Added a fix to use the same version of the configuration at thewakunode2 application layer.

Changes

• using modified config copy in wakunode2

[github-actions]

You can find the image built from this PR at

quay.io/wakuorg/nwaku-pr:3036

Built from f081254

[NagyZoltanPeter]

Hm, nice catch! Thank you for it!

[Ivansete-status]

Thanks for it! Just added a question

[Ivansete-status]

I actually think these are the only needed changes.
Did you try to see what happens with just that?

[gabrielmer]

Yes, so the need of each change is the following:

1. In this file, apart from this line, changed the name to refer that we are using a copy of the original configuration and that we're not modifying the user's input

2. In apps/wakunode2/wakunode2.nim, we create a copy of the original configuration and pass it to all the procs (so we don't modify the original configuration passed by the user)

3. In tests/wakunode2/test_app.nim, had to change from let to var for the configurations passed to init() now that we modify the object in the proc

Let me know if there's one of these points you think we could spare

[NagyZoltanPeter]

Do we use/re-use the original user configuration afterward? Just curious...

[gabrielmer]

Do we use/re-use the original user configuration afterward? Just curious...

After that point we actually don't use the original configuration, but did it just to not override the original user configuration which generally isn't a good pattern. In principle we can avoid doing the copy and change the original but idk, it doesn't feel right haha