refactor!: Refactor to make easy improvements

[taorepoara]

Has explained in the next comment, this refactor makes this project simpler to use and to improve: #24 (comment)

I've based my work on the serde way of parsing AST into structs to make it more usable.

I've also changed the way I manage attributes to make it easier to add new ones.

[yanganto]

Could we remove attribute? and then the syntax will be simple

#[patch(attribute(derive(Deserialize, Debug), serde(rename = "foo"))]

->

#[patch(derive(Deserialize, Debug), serde(rename = "foo")]

[taorepoara]

Could we remove attribute? and then the syntax will be simple

#[patch(attribute(derive(Deserialize, Debug), serde(rename = "foo"))]

->

#[patch(derive(Deserialize, Debug), serde(rename = "foo")]

That's true it would be so much simple !

We just have to be careful to avoid the possibility between the attributes and the patch configuration attributes, but if there is no way to have conflict I can do that.

I also thought about this syntax (not sure if it's possible).

What do you think?

#[patch(
  #[derive(Deserialize, Debug)],
  #[serde(rename = "foo")]
)]

[yanganto]

#[patch(
  #[derive(Deserialize, Debug)]
  #[serde(rename = "foo")]
)]

It is good and without ,, just parse everything between #[ and ] as inputs.
Such that if the developer forgot to put ,, it will not mater.

[taorepoara]

#[patch(
  #[derive(Deserialize, Debug)]
  #[serde(rename = "foo")]
)]

It is good and without ,, just parse everything between #[ and ] as inputs. Such that if the developer forgot to put ,, it will not mater.

Yes ok, I'll try that. This could allow code injection but it's his own code and I'll see how the formatter manage that

[taorepoara]

I could not implement a good solution for the attribute since the parse_nested_meta try to parse the sub attributes names to Iden which fails on # character.

This could be managed with more low level parsing but could not made it this morning.

I can look at that but only in a few days. I let the attribute for now.

[yanganto]

Hi @taorepoara, it looks good to me.
Is there still anything you want to refactor in the PR?

[taorepoara]

Hi @taorepoara, it looks good to me. Is there still anything you want to refactor in the PR?

No, that all for this one. I'm working on another PR for other features and I'll try to implement the #[patch(#[my-attribute])] syntax later in another PR.