Custom node hot reload + gui search install

It would be really great if it were possible to reload self-written nodes without restarting the entire server.
Additionally, a search (search for n8n-node-{searchquery} in the npm store) and an installation (or update) of these nodes directly in the Ui would be great.

About the hot reloading. I guess you are doing what we wrote in the tutorial. There is a better way. Run two terminals, one doing lerna run watch --parallel and the other one doing npm run start. This is much faster than npm run dev, since the project it’s not built each time. Sadly, if you change something in the node’s UI (the description property) you still have to kill the process running npm run start and run it again. But again, this is much faster. The reason for that it’s that n8n loads all the nodes’ UI in memory on start. So, even when the changes are being transpiled when you save the file, n8n does not about these changes until it starts again.

I have those 2 commands as two aliases in my command line. So I do not have to type all that each time.

alias n.watch="lerna run watch --parallel"
alias n.start="npm run start"

It’s on my TODO to add that to the tutorial.

Let me know if that helps.

No, it is about adjustments to the productive system. If I have tested a node completely on the test server and then want to update it on the productive system, there is currently nothing else I can do to restart the complete productive system.

This is also where the idea of these community nodes comes in.
It would simply be inconvenient to restart the server every time just because you install or update a community node.

Not only the function of the node should be updated, but also the description property.

So in general it has not much to do with the development of own nodes but more with the installation and update of community nodes without having to restart the server.

Ahh I see. I misunderstood your feature request. Now I get it. Thanks for the explanation.

I think this can moved to done! :smile:

2 Likes