dry-rb/chat

WE MOVED TO ZULIP => https://dry-rb.zulipchat.com <=

Tim Cooper

@coop

I totally get what you mean, I’m just linking to how bundler is doing it. Their docs site is generated from https://github.com/bundler/bundler-site. The dry-rb.org docs site is generated from https://github.com/dry-rb/dry-rb.org/.

Fran Worley

@fran-worley

Sorry, my brain is not engaging... I know understand what you mean!

I thought you meant that bundler does some magic with rubygem docs for versions... must get a :coffee: before I do anything else!

Tim Cooper

@coop

Tim Cooper

@coop

@timriley I’ve been reading over your article again and I have a feeling that you meant to say that an Article is an entity not a value object. From your article you quote the definition of a value object as:

A value object is one whose notion of equality is based on the values of the attributes it contains, not its object identity.

I agree with this but your article class includes an id field:

class Article < Dry::Types::Struct
  attribute :id, Types::Strict::Int
  attribute :title, Types::Strict::String
  attribute :body, Types::Strict::String
end

I think we both will agree that Article.new(id: 1, title: “title”, body: “body”) != Article.new(id: 1, title: “amended title”, body: “body”) in code land but in the real world the objects are the same, they’re both referring to the exact same article (the second article is modified and therefore its title is different).

IMO a value object is something that is only identifiable by its attributes, things like money, name, address and credit cards.

I did enjoy the article though, just my 2 cents.

Simon Schmid

@sled

@coop :+1:

Tim Riley

@timriley

Good point, @coop. id being there is kind of a smell that it’s not a true value object.

Simon Schmid

@sled

but it leaves me wondering how to handle objects which could be in an invalid state :)

Tim Riley

@timriley

Even though the the equality methods on Dry::Types::Struct still take into account all the attributes.

@sled we just treat invalid data as just that - a hash of invalid params/attributes

pass it around until like that the user can make it valid

Simon Schmid

@sled

@timriley what do you think about an anonymous class which represents the hash as simple struct with attr_readers ?

kind of a "representer" of the invalid data

Tim Riley

@timriley

What does that get you over a hash?

Simon Schmid

@sled

@timriley compatibility with existing libraries which can't handle hashes directly

Tim Riley

@timriley

heh. Well I guess that’d be helpful then :)

Simon Schmid

@sled

mh or maybe a simple decorator which just delegates to the fetch method

Tim Cooper

@coop

http://www.amazon.com/Domain-Driven-Design-Tackling-Complexity-Software/dp/0321125215 - if you haven’t read this I would strongly recommend reading through it. A lot of the dry / rom stuff talks about building domain orientated systems, this book was the original instigator (I believe).

Tim Riley

@timriley

Cool, thanks for the recommendation @coop

Simon Schmid

@sled

thanks! :)

Tim Cooper

@coop

I actually have a copy you could borrow next time you’re in Brisbane :P

Tim Riley

@timriley

hehe

Tim Riley

@timriley

@nepalez Congrats on the first dry-initializer release :) Would you like to pen a little blog post for the dry-rb site announcing it?

Tim Riley

@timriley

@coop thinking I could maybe add a little parenthetical aside explaining the difference between entities and value objects – do you reckon that’d help?

Tim Cooper

@coop

I think it would help.

Tim Riley

@timriley

@coop How’s this read?

(Technically, these examples are entities rather than strict value types, because we include the record IDs in their attributes, which you'd expect to be the key indicator of equality. However, in practice, Dry::Types::Struct instances still check for equality across all attributes, and in our apps we only ever instantiate these objects from known-good database data, which means they still behave effectively as value objects. So let's continue!)

Tim Cooper

@coop

That clears it up.

Tim Riley

@timriley

Cool, and thanks for the discussion!

Tim Cooper

@coop

@timriley do you ever mutate these and write them back to the database? And when I say “mutate” I’m talking about returning a new value object with the changed attributes.

Tim Riley

@timriley

Nope, we just work with hashes when we’re persisting changes. Get a params hash from a form, adjust as necessary in an “operation” object, run it through validation schema, pass it to a ROM command.

Tim Cooper

@coop

From your example I could imagine something: changed_article = article.with_title(“my new title”) and then pass changed_article to the DB.

Ah ok.

Tim Riley

@timriley

Yeah. I think a lot of this will be clearer once I can start to piece a few things together.

But I’m trying to avoid 6,000 word articles ;)

Tim Cooper

@coop

The idea of hashes is fine, we do something similar for event sourcing, each aggregate keeps list of events to persist to the database which are effectively an array of hashes.

We wrote so much code that is similar to dry-types.

Tim Riley

@timriley

heh

Tim Cooper

@coop

All of our events are typed… We enforce types when constructing events. I’ve tried substituting for dry-t and dry-v but it’s quite a lot of work. It will be worth it because dry-t and dry-v integrate so nicely, we’ll be able to delete a lot of code at the end.

Tim Riley

@timriley

Cool, that’ll be nice to get in place :)

Andy Holland

@AMHOL

hanami/validations#98

Fran Worley

@fran-worley

@AMHOL Some sound arguments, it's just a shame that there can't be some kind of consensus amongst rails alternative ruby 'frameworks'.

Andy Holland

@AMHOL

@fran-worley Yeah, the point about not being able to wrap the API was the only real valid argument IMO, but each to their own I guess

Fran Worley

@fran-worley

@AMHOL I can kind of see the argument for having lots of dependencies but if it keeps code DRYer then why wouldn't you...

Andy Holland

@AMHOL

Yeah, I think the adding more gems argument is a non-issue

Fran Worley

@fran-worley

I'd rather have 30 small dependencies that use each other than 5 large ones that have to re-implement common functionality. I just appreciate that others don't agree.

Andy Holland

@AMHOL

Unless those gems make changes to the runtime environment

:+1: * :100:

Fran Worley

@fran-worley

Yeah. Also the point about the wrapable API, isn't this being addressed to some degree by the discussions on custom input processors and Error handlers

and surely there is nothing to stop you writing your own macros to reintroduce the validates :title, :presence type syntax ?

Where communities thrive

People

Repo info

Activity