[Feature Request] CLI / Entity generator: Optionally derive serde::{Serialize/Deserialize}

[elbart]

Dear SeaORM team,

another question / request from my side.

I just generated Sea ORM entities from my existing database structure using the sea-orm-cli tool. This worked like a charm. however it does not have an option to also add the the derive attributes for serde::{Serialize,Deserialize} traits. I know there is JSON support using into_json() however a serde_json::Value is nothing to handle nicely after the query executed for post query modifications. Now there is two questions:

1. I saw in all your examples for actix and rocket, that the Serialize/Desrialize attributes are there (e.g. https://github.com/SeaQL/sea-orm/blob/master/examples/actix4_example/src/post.rs#L4). Are there cases where Serialize/Deserialize would not work on the generated Entity struct?
2. Would it be ok to start a PR to propose such a change to the sea-orm-cli (add another clap option) and the sea-orm-codegen or was there a general decision to not go this way at all?

Best regards!

[tyt2y3]

Hi Tim, thanks a lot.

This worked like a charm.

I am glad to hear that! (It still needs some polishing)

Yes. I just don't want to force every attribute to impl serde for now, as it will be some extra burden for custom types. It seems that right now it is perfectly doable (expect for the extra compile time).

So it's a good idea to add it as an optional argument, and we might turn it on by default at some point.

[baoyachi]

Hi Tim, thanks a lot.

This worked like a charm.

I am glad to hear that! (It still needs some polishing)

Yes. I just don't want to force every attribute to impl serde for now, as it will be some extra burden for custom types. It seems that right now it is perfectly doable (expect for the extra compile time).

So it's a good idea to add it as an optional argument, and we might turn it on by default at some point.

I gree

[elbart]

Hi Tim, thanks a lot.

This worked like a charm.

I am glad to hear that! (It still needs some polishing)

Yes. I just don't want to force every attribute to impl serde for now, as it will be some extra burden for custom types. It seems that right now it is perfectly doable (expect for the extra compile time).

So it's a good idea to add it as an optional argument, and we might turn it on by default at some point.

Thanks for your feedback! I hope I prepared the PR in #237 as expected...