Hi,
I really do not think this is production ready.
1. Simple async code.
Simple example:
// m1.js
const label = 'Hi world';
setInterval(() => console.log('From m1:', label), 1000);
exports.label = label;
// m2.js
const dload = require('../index');
mo.fs = require('fs');
mo.path = require('path');
mo.m1 = require('./m1.js');
mo.co(function * (co_next) { let label = mo.m1.label;
console.log('Old label:', label);
setInterval(() => {
dload.reload(mo.path.join(__dirname, './m1.js'));
label = mo.m1.label;
console.log('New label:', label);
}, 1200);
})();
Run m2, then change the label manually a few times and save. All sorts of stuff in there:
From m1: Hello world
From m1: Hi worlds
From m1: Hi world
From m1: Hello world
New label: Bye bye
From m1: Hi worlds
From m1: Hi world
From m1: Hello world
From m1: Hi world
From m1: Hi world
From m1: Hello worlds
From m1: Hello world
From m1: Hello world
From m1: Hi world
From m1: Hi world
From m1: Hello world
From m1: Hi world
From m1: Hi worlds
From m1: Hello world
From m1: Hello world
From m1: Hi worlds
From m1: Hi world
From m1: Hello world
From m1: Bye bye
I know there's the _release handler, but what if my interval is not in my code, but in some library my original module required? I would be totally scared to run this in production and try reloading things when all kinds of resources are attached out there. E.g. I sometimes keep anything db-related in a separate repository. And that stuff touches say Redis, MongoDB and RabbitMQ. I know I will have to write manual _releasse code. But if I miss just one thing, it's not good. And if I let my colleagues use that? They don't even know what to release! OTOH if the server is killed, I _know_ everything is shut down. And further, even with that, let's say there's a long-running task out there. Even if you do _release, you might not be able to interrupt it properly so your resource-consuming task will keep running, _and_ you'll load a new copy of the same thing.
In this state, I'm not sure if I would even use it to hot-reload configuration variables (static, sync code), let alone async modules.
2. I guess it's early, but the docs and the API should be better. On the first read it was not clear to me what exactly I need to do. Especially since you rewrite modules in the same code where you try to explain dload. I would kick out that example and write something like this:
# Usage
const dload = require('dload');
mo.myHotModule = require('./my-hot-module');
// ... do some change to myHotModule, save the file
dload.reload(<path>);
// mo.myHotModule is reloaded.
Even your module name itself (dload) is confusing and not reflecting what it is for.
3. Use cases
It's not clear to me what is the usecase of this. You want to be able to load a module, use it, then _rewrite_ it from your own code, and then reload the changes?
If I were to think of some use cases, I would say something like auto-reload is needed. Kind of like nodemon or supervisor, but reloading just those files. You'd give a list of files to watch, and then your server would not restart, only those modules would.
Or maybe you have some sandbox code, like for a live, web ide, but in that case I would never let that sandbox code in my main thread - because a simple setInterval(() => console.log()) will kill my main thread and not even let me reload the module. I'd rather spawn the sandbox code in a new process in such cases.
Such as it is, I would not use this module. If I need code reload, a simple restart is usually tollerant. And if my code is critical and I need online redeploy, I would use simple container tech for this, because that was it's purpose in the first place.
Please take the above as constructive critique. I love the idea, I just think you're not there yet (for me).