Nicolas Goy

It seems the reproduction is not 100% of the time. I am trying to find the conditions and I will give you an update.

I updated the repo with some path modifications and now it does it 100% for me.

Yes, but it should at least watch itself (the css file) even with `source(none)`. Also the weird thing is that it is dependent on the folders it is located. As...

Maybe we don't need to add `--paid` option as it would be specific to a a `paid` flag, but I would suggest batch edition of tag and comment. The idea...

I agree with @jcalabro in what we need to have separate `truncate` or `append`. Maybe we can take some inspiration from the rust approach? Basically, the rust fs open options...

While speaking of this, there is also an append permission in NFSv4 ACL for directories that give permission to only create sub-directories and not files. This is not directly related...

I think this is a good feature, but I think it could be supported with a more general syntax. This PR started to implement a new set of syntax: https://github.com/jazzband/Watson/pull/484...

Ok, I got bitten by this, so I tried implementing a PR, but it is more subtle that I anticipated. In [`before_send_inject_reloader/3`](https://github.com/phoenixframework/phoenix_live_reload/blob/e8a8d800e63c34fabb217af480b742af78a35872/lib/phoenix_live_reload/live_reloader.ex#L152C7-L152C7) we get the full response, with a ``...

Yes `has_body?` always returns `true`, because it is called on the fully rendered page, so the html actually have a body, even if this body is not within the root...

> We can't guarantee that the user has a root layout, or a root layout file, so that unfortunately is not going to work. We could just add a comment...