Musings, JS and React

• Formspree is awesome!
• HTML5 form validation is okay as a first declarative validation layer.

Musings, Ruby

• undef_method: Only in class definitions, you can probably use this in a file.
• Getting method names:
  • __method__ to get the name of the method, good for passing to another abstraction thing.
  • __callee__ looks up the name dynamically and refers to the name under which the method was called.
• Updating with bang methods:
  • ActiveRecord::RecordInvalid inherits ActiveRecord::ActiveRecordError, which inherits from StandardError.
  • When updating via update_attributes, we get ActiveModel::Errors, not ActiveRecord::RecordInvalid.
• Is the adapter design pattern an Anti-corruption layer?
• Single vs Multi-tenant SaaS:
  • Create a database, provision some servers, add a load balancer, add some caching, and call it a day.
  • Any-scale system: a system that’s built to remove scale from that equation.
  • Multi-tenant: Load balancer, gateway, subnets, kinesis, lambdas, Redshift…
  • Multi-tenant: Teams share resources. Higher cost up front but each new cx doesn’t really cost much.
  • Single-tenant: More expensive except for the human aspect.
  • At Kumu, every project is backed by a CouchDB database.
• Rubocop tells us to put mixins in separate attachments.
• Ruby and Amazon Lambda: Currently, it doesn’t support Ruby, but you can use it as a go-between for your S3 to Rails use case.
• lambda or proc?
• Flaky Ruby tests:
  • Capybara: Avoid dangerous methods that do not wait: visit(path) and current_path, all(selector), accessors (text, value, title). Use safe methods that do wait: find(selector), find_field, has_selector?.
  • DB pollution: database_cleaner: transaction vs truncation. Truncation is required when running request specs with Capybara. The thread running your Rails server uses a separate database connection from the one running your test.
  • Non-deterministic attributes: Don’t use faker, just hardcode.
  • ID-dependent things: When tests run in a transaction, the auto-increment value still runs. Truncation = auto-increment will be reset.
  • To reproduce: run RSpec using the same seed, run the entire list of files, use the --bisect flag.
• Sentinel value: “A special value in the context of an algorithm which uses its presence as a condition of termination.”

Result Objects: Errors without Exceptions

Reference

• Resonad gem:
  • Errors are part of the return value, not an exception.
  • Expected errors are available through result.error, and unexpected bugs are exceptions.
  • All expected error cases are automatically “caught”, without having to guess what they are.

result = register_new_user(params)
if result.success?
  handle_success(result.value)
else
  handle_failure(result.error)
end

• When not to use:
  • When the method should always succeed.
  • When there is only a single failure case, just return nil or whatever.
  • If you can recover Locally, and you can just return a Null Object.
• Result objects are appropriate:
  • When an operation can fail in multiple different ways.
  • When operations need to be chained together, and each individual operation can fail.
  • When it’s important for callers to know that the method can fail.
  • Works with service/interactor/command objects.

• Implementation: Resonad, dry-monads, Github::Result, monadic, result-monad.

• Refactoring to data: “This is an example of what I call “data-driven design” or just “thinking in data”. All the specific details are described in data: numbers, strings, arrays, hashes, structs, etc. All the behaviour (i.e. the implementation) is generic, and controlled by the data.”
• When thinking of inheritance, ask yourself if it is possible to think of the inheritance chain in terms of just data (array, hash, and do a concretion based on that).