talum/programming_phoenix.md

## programming_phoenix.md

      
    Raw
  

              programming_phoenix.md
            
          
    Programming Phoenix Ch 3: Controllers & Views

connection the whole universe of things we need to know about the user's request (a struct, map with known set of fields)
connection
|> endpoint
|> router
|> browser_pipeline
|> UsersController

Structs

struct Elixir's main abstraction for working with structured data

built on top of maps
provide protection for bad keys at compile time (instead of runtime like maps)
struct is a map with a __struct__ key

Repository

repository an API for holding things
Controllers

view module containing rendering functions that convert data into a format the end data will consume (HTML/JSON)
template function on that module compiled from a file containing a raw makrup language and embedded Elixir code
Templates


helpers exist
after compilation, templates are functions
Phoenix builds templates using linked lists (no copies of strings)
cache templates at the hardware level
similar to partials, can render another template from a template
Phoenix.View.render("template_name", _assigns, {locals: locals})

Layouts


Controller renders layout view, which renders actual template
each template receives @view_module and @view_template

Questions


Can we talk more about caching at the hardware level?

Programming Phoenix Ch 4: Ecto & Changesets

Ecto


the Elixir framework for persistence
changesets hold all changes you want to perform on the database
migrations used to make the database reflect structure of app

Models


"Just as a controller is a layer to transform requests and responses according to a communication protocol, the model is nothing more than a group of functions to transform data according to our business requirements" (57)
Models can manipulate other structures such as changesets and queries
changesets let Ecto manage record changes, cast parameters, and perform validations
cast on changeset checks for required fields and casts required and optional values to schema types and rejects everything else
Ecto.Changeset defines cast and validate_length
Why changesets? One size does not fit all for update strategies.
Changesets let you decouple update policy from schema
changeset had validation errors because Ecto changeset carries validations and stores info for later use. Changesets track changes.
changeset as bucket to hold everything related to db change, before and after persistence
"If no parameters are specified, we can’t just default to an empty map, because that would be indistinguishable from a blank form submission. Instead, we default params to the :empty atom. By convention, Ecto will produce an invalid changeset, with empty parameters."
pass a changeset into the form helper rather than a model.

Updating Data


able to update data via Repo.insert e.g. Repo.insert(%User{ ...> name: "José", username: "josevalim", password_hash: "<3<3elixir" ...> })
but should prefer to use a changeset e.g. changeset = User.changeset(%User{}, user_params) {:ok, user} = Repo.insert(changeset)
validation errors stored in @changeset.action
example of multiple changesets?