desmond
My application has many serializers for turning structs into something suitable for clients. These serializers have many serialize/1 functions defined to handle the nil case, list case, and finally the struct itself:
defmodule MyApp.ItemSerializer do
def serialize(nil), do: nil
def serialize(items) when is_list(items) do
Enum.map(items, serialize/1)
end
def serialize(item) do
%{
# ...
}
end
end
I’d like to have a @behaviour MyApp.Serializer directive to guarantee that individual implementations follow the contract of implementing all the serialize/1 functions. The problem is, it looks like callbacks only respect one function implementation, and creating many callbacks with the same function name but different typespecs does not help.
Is this the expected --ahem-- behavior of callbacks?
Where Next?
Trending in Questions
Other Trending Topics
Latest on Elixir Forum
Sponsor Spotlight
Catch errors, track performance, monitor hosts and more.
Categories:
Sub Categories:
Forums
Popular Tags
- #ecto
- #liveview
- #troubleshooting
- #learning-elixir
- #deployment
- #library
- #erlang
- #testing
- #genserver
- #mix
- #absinthe
- #remote-other
- #otp
- #plug
- #how-to-question
- #macros
- #postgres
- #channels
- #elixirconf
- #exunit
- #discussion
- #code-sync
- #javascript
- #podcasts
- #onsite
- #dialyzer
- #docker
- #authentication
- #umbrella
- #full-time-contract
- #podcasts-by-brainlid
- #ecto-query
- #elixir-ls
- #blog-post
- #phoenix_html
- #iex
- #ai
- #graphql
- #genstage
- #elixirconf-us
- #websockets
- #supervisor
- #advent-of-code
- #distillery
- #processes
- #api
- #forms
- #metaprogramming
- #security
- #hex
Our Sponsors
Build Elixir applications with speed and confidence.
Supporting innovation across the BEAM ecosystem.
We build reliable cloud platforms for business-critical systems.
Catch errors, track performance, monitor hosts and more.
Error tracking, logs, uptime, and Just Enough APM. Built for developers.
Practical resources that improve the lives of professional developers.
Develop your skills with books, videos, and courses.
Bringing learning, work and compliance together since 1987.
The team behind Membrane, Popcorn, LiveDebugger. Available for hire.
Courses that'll move you from confusion to "Aha, now I get it!"
Showing Posts 1 to 8- Show Best Posts
- Show All Posts (oldest first)
- Show All Posts (newest first)
tty
This is linked to how functions are differentiated in the BEAM. Functions are differentiated by the module they reside in, their name and arity (number of arguments the function accepts). As such serialize/1 is the same function and the various function clauses a contraction of an if-else.
If having these 3 cases are import the define the 3 different functions in the behaviour individually e.g. serialize_nil/1, serialize_lists/1, serialize_item/1. The behaviour can auto-generate a serialize/1 which calls the relevant callbacks.
If you have many different structs perhaps look into using Protocols.
NobbZ
Why don’t you make it a
protocoll? Then you could implement the list-case once in adefimpl X, for: List, thenilindefimpl X, for: Atom, while having discretedefimpl X, for: Yfor your different structs.desmond
The problem with using a Protocol is I want to be able to support different versions of the serializer for the same object, so
GameSerializer.V1.serialize/1andGameSerializer.V2.serialize/1, with the caller selecting which one to use. So I can’t just call a genericSerializer.serialize/1without knowing what I’m passing in (list, nil, actual object) which I believe rules out the use of a single protocol. Unless I’m missing something?OvermindDL1
How do you determine the version?
desmond
The controller itself calls either the
V1orV2serializer, since the controller/route is namespaced with versions as well. So basically hardcoded.OvermindDL1
Why not encode that into the dispatch argument then?
Or use a protocol per version that can delegate to the lower version for the parts that are the same.
Or use my ProtocolEx to dispatch on both the version and value.
There are even more things you can do.
michalmuskala
Here’s an idea - use a protocol with two callbacks -
serialize_v1andserialize_v2. For implementations where those are close enough or even the same, it’s pretty easy to delegate to a common implementation.OvermindDL1
Yep, precisely my second suggestion. ^.^