I built a quick and dirty tool to check the broken links on our website.

This is the output:

Visiting http://hanamirb.org/guides/1.0/upgrade-notes/v060  Not Found
Visiting http://hanamirb.org/guides/1.0/upgrade-notes/v070  Not Found
Visiting http://hanamirb.org/guides/1.0/upgrade-notes/v080  Not Found
Visiting http://hanamirb.org/guides/1.0/upgrade-notes/v090  Not Found
Visiting http://hanamirb.org/guides/1.0/upgrade-notes/v100  Not Found
Visiting http://hanamirb.org/guides/head/validations/boolean-logic  Not Found
Visiting http://hanamirb.org/guides/head/validations/custom-predicates  Not Found
Visiting http://hanamirb.org/guides/head/validations/advanced-usage  Not Found
Visiting http://hanamirb.org/guides/upgrade-notes/v070  Not Found
Visiting http://hanamirb.org/guides/upgrade-notes/v100  Not Found
Visiting http://hanamirb.org/guides/upgrade-notes/v100beta1  Not Found
Visiting http://hanamirb.org/guides/upgrade-notes/v090  Not Found
Visiting http://hanamirb.org/guides/upgrade-notes/v080  Not Found
Visiting http://hanamirb.org/guides/upgrade-notes/v060  Not Found
Visiting http://hanamirb.org/guides/architectures/container  Not Found
Visiting http://hanamirb.org/guides/architectures/application  Not Found

We should add a new page with repository query examples.

TODO

• joins;
• groub by;
• where with block (ILIKE, IN, etc);
• order;
• some SQL functions (type casting for example);
• #exist?

If you want to see something special please, add comment with it

Right before the section in the GS guide called "Building Forms To Create Records", there's some test output shown:

9 runs, 16 assertions, 0 failures, 0 errors, 0 skips

This test output has changed in v1.2 to:

12 runs, 20 assertions, 0 failures, 0 errors, 0 skips

I guess there have been some new tests added in the interim?

I am submitting this as an issue and not a PR because I couldn't find a 1.2-specific getting started guide at the moment.

I wrote this code my application and I run the bundle exec hanami s
so I got message uninitialized constant Web .
I think Web is not initialized when accept_json.rb is loaded.

Below codes work well.

# apps/web/views/accept_json.rb
module Web
  module Views
    module AcceptJson
      def self.included(view)
        view.class_eval do
          format :json
        end
      end
    end
  end
end

or

# apps/web/application.rb
module Web
  class Application < Hanami::Application
    configure do
      # ...
      view.prepare do
        require_relative './views/accept_json'
        include Web::Views::AcceptJson
      end
    end
  end
end

Some time ago sqlite was set to be default adapter. After that deploy on heroku doesn't work(since heroku doesn't support sqlite). Instruction should be updated to don't confuse newcomers which would see that it doesn't work

@hanami/core WDYT?

Hi folks:

I open this issue because when you navigate the guides from a mobile it's a little irritating when you click a link and then you have to scroll down again.

You can see a screenshot to see what is the first thing you see when enter in the guides:

I think we could have a select instead the list, what do you think?

The Getting Started guide has a confusing inconsistency here:

With our validations in place, we can limit our entity creation and redirection to cases where the incoming params are valid:

# apps/web/controllers/books/create.rb
module Web::Controllers::Books
  class Create
    include Web::Action

    expose :book

    params do
      required(:book).schema do
        required(:title).filled(:str?)
        required(:author).filled(:str?)
      end
    end

    def call(params)
      if params.valid?
        @book = BookRepository.create(Book.new(params[:book]))

        redirect_to '/books'
      else
        self.status = 422
      end
    end
  end
end

You may have noticed that there isn't an else condition to that if statement.
What happens when params aren't valid?

The paragraph after the code says that there isn’t an else condition, but there is one – it says self.status = 422. There is no other nearby code block that the paragraph could be referring to.

When I look at the history of getting-started.md, I see that commit 4f5f071 “Guides: explain the differences between #halt and #status=” created this inconsistency. That commit mainly edited the HTTP Status section of the Action Control Flow page, but edited two code blocks in Getting Started too. The else condition was added without the following paragraph being updated.

I don’t know enough to suggest whether the else condition should be removed or the paragraph after the code should be revised.

I've been going through the guide and I've hit several places where it seems to be out of date or otherwise wrong.

• hanami new bookshelf --database=postgres will generate a project which uses minitest, not rspec, but all the code examples use rspec
• the rspec examples use the old foo.should.match bar syntax, which nowadays is not recommended. Some of the test examples just don't run at all (those that use must_equal and co; there is also one random assert, which of course also fails in an rspec context)

I can try to fix some of these issues, although I would need some guidance as to whether to change the test examples to minitest (since that is the out-of-the-box default), or change them to expect syntax, or whatever.

In the v1.1, we'll change the good practices for interactors:

• no more params in initializers (well, in fact, you can actually do what you want in initializer, but it's best if you keep it for dependency injections)
• params go through #call
• #call can now take anything as parameters, except no parameters (so the minimum declaration is def call(*); end and not def call; end

Old format (̀#call without parameters) will still be working (I guess we'll deprecate it in v1.2?)

Hello,
I think that it'll be very helpful. Because it'll allow automatically update the site.

Related link:
https://iamstarkov.com/deploy-gh-pages-from-travis/

/cc @hanami/core-team

Examples:
Without bundle exec
hanami generate action web dashboard#index
http://hanamirb.org/guides/actions/overview/

With and without lead percent
http://hanamirb.org/guides/projects/rake/

I suppose bundle exec is excess. With rbenv and one version of hanami it works properly without bundle exec. OhMyZsh with some simple settings adds bundle exec to rake commands automatically.

Perhaps these guides are not responsible with bundler and it needs to be moved to Command Line section.

The page header contains reference to Atom feed http://feeds.feedburner.com/hanamirb , but it does not exist.

Ref: hanami/model#351 (comment)

hanami new bookshelf --database=postgres

should just be

hanami new bookshelf

We don't use any advanced features of Postgres, and it makes the Getting Started guide more complicated. Some users may have to spend time installing Postgres (which can be a pain!) or dealing with authentication issues... all just to learn the basics of Hanami!

I think we should include a note telling users that we're using SQLite and that in a production app they should use Postgres or some other production-ready database (and show them it's easy to use postgres instead: hanami new bookshelf --database=postgres)

Hi there,

I have an idea for restructuring the guides into their own respective repos.

For example, rather than having the assets guides in this repo (hanami/hanami.github.io) at source/guides/assets/, we could put the guides inside the hanami/assets repo, at guides/

Then the build tool here could just clone each of the repos and build the site based on that. Travis could listen to each repo for commits and rebuild the site for every one, to ensure it's up-to-date.

Here's the current sources/guides folder, and where they'd be relocated to (/guides in each repo unless specified):

actions/              => hanami/controller
architectures/        => hanami/hanami (guides/architectures)
command-line/         => hanami/hanami (guides/command-line)
helpers/              => hanami/helpers
mailers/              => hanami/mailer
models/               => hanami/model  (guides/)
upgrade-notes/        => hanami/hanami (guides/upgrade-notes) - or stays here
applications/         => hanami/hanami (guides/applications)
assets/               => hanami/assets
getting-started.md    => hanami/hanami (or stays here)
index.md              => stays here
migrations/           => hanami/model (guides/migrations)
routing/              => hanami/router
views/                => hanami/view

This would fix the issue of having to make two separate PR's for changes that affect the guides. Though you'd still have to do that for changes that affect the getting-started guide. The blog would stay in this repo.

This would also help people searching for something in, let's say, hanami/controller that is answered in the guide for that section.

Thoughts?
@hanami/core-team @hanami/contributors

Ref hanami/hanami#658

We can both understand if a specific presentational logic behaves correctly and/or assert if the rendered markup.

I feel like a) there's something missing at the end ("assert if the rendered markup is correct") or b) the sentence should be slightly different ("assert the contents of the rendered markup").

Rewrite the mission statement in the main guides page:
http://hanamirb.org/guides/

Ref: hanami/hanami#658

Given example maybe should be replaced by another one, not bound to form, because passing wrong params cause error

Add a book#test_0002_displays list of errors when params contains errors:
NameError: undefined local variable or method `form' for #<Hanami::View::Rendering::Scope:0x007fe0fa942e00>
Did you mean?  fork

Error happens because template is rendering from Web::Views::Books::Create instead Web::Views::Books::New and Create view don't have form method

And this may be confusing, especially when passing tutorial in first time.

Ref: hanami/hanami#649

In certain part of documentation follow this link , .one is used and the reason is that it throws error if nothing found using limit, while making use of .one it will return nil.I think it must be a standard to use .one for all the queries where we use .limit.

http://hanamirb.org/guides/1.0/models/sql-queries/
In the above link under section limit, this can be added

The guides simply say:

If we want to mount an application, we should use mount.

It would be helpful to specify that the mount instruction should go to config/environment.rb.

I was wondering what the license of your source is, more specifically the design and it's assets.

Following the example in Custom Helpers Guide (http://lotusrb.org/guides/helpers/custom-helpers/), I had an application loading error:
uninitialized constant Web::Helpers (NameError)
with Lotus v0.5.0 and Ruby 2.2.3.

It seems to be related to the loading order defined in application.rb file:

# apps/web/application.rb
      # ...

      load_paths << [
        'controllers',
        'views',
        'helpers'
      ]
# ...

      view.prepare do
        include Lotus::Helpers
        include Web::Helpers::Shuffler
      end

I guess helpers must be loaded before views before being included into them.
Inserting 'helpers' on top of load_paths fixed the issue:

# apps/web/application.rb
      # ...

      load_paths << [
        'helpers',
        'controllers',
        'views'
      ]
# ...

I this is right, could you fix the guide?
Thanks.

Steps to reproduce

Reduce the width of your browser until you see the dropdown (or browse to the getting started guide on your phone).

Then select a topic from the dropdown.

Result

Correct page shows but dropdown still shows 'Getting started'

Hanami::Interactor, in hanami/utils is a very useful component.

We should consider adding a Guide to the website about how to use it (and the benefits of it).

Perhaps it could be an extension of the Getting Started guide, working from the bookshelf example.

There's no Hanami::Utils section on the site, so not sure if that's worth adding.

Happy to help write/review this guide, if we choose to move forward with it.

I'm doing the getting started tutorial:

http://lotusrb.org/guides/getting-started/

But I'm stuck here:

It seems that if there's no trailing slash on /books, it says "File not found":

By adding the trailing slash, it displays correctly:

So I had to change visit to use trailing slash in order to make the tests pass.

Any ideas why this is happening?

Why

https://www.searchenginejournal.com/importance-of-404-error-page-top-404-designed-pages/64899/

Links

https://help.github.com/articles/creating-a-custom-404-page-for-your-github-pages-site/
https://isfeargoidte.me.uk/2016/03/22/404-error-github-pages-middleman/

I may have found mistake in http://hanamirb.org/guides/1.1/architecture/interactors/.

That page contain code above:

# spec/bookshelf/interactors/add_book_spec.rb:
require 'spec_helper'

describe AddBook do
  let(:interactor) { AddBook.new }

  it "succeeds" do
    expect(interactor.call).to be_a_success
  end
end

# lib/bookshelf/interactors/add_book.rb
require 'hanami/interactor'

class AddBook
  include Hanami::Interactor

  def initialize
    # set up the object
  end

  def call(book_attributes)
    # get it done
  end
end

Copy and paste these codes, run bundle exec rake test, I saw an error like this:

  1) Error:
AddBook#test_0001_succeeds:
ArgumentError: wrong number of arguments (given 0, expected 1)
    /Users/hoge/work/myruby/hanami_practice/bookshelf/lib/bookshelf/interactors/add_book.rb:10:in `call'
    /Users/hoge/.rbenv/versions/2.4.1/lib/ruby/gems/2.4.0/gems/hanami-utils-1.1.1/lib/hanami/interactor.rb:356:in `block in call'

I suggest discussing what we want to add to guides. I created a simple list with my ideas. If you have other ideas - let's discuss and update list ❤️

List

Validations

• Copy validation docs from validation repository
• Say that hanami-v based on dry-v

Controllers

• Add information about flash (now we have only information about testing)
• Rendering XML and JSON data in controllers
• Streaming and File Downloads

View

• Improve example with forms (https://github.com/hanami/helpers/blob/0cf7e87eb3ae5abb52f1636af3da6b1855ada19e/lib/hanami/helpers/form_helper.rb#L149-L181)
• Write about DELETE and POST links. In rails, you can find some magic about this. I think we have to describe that developer have to make a form for this.

Router

• Router params. Something like this: http://guides.rubyonrails.org/routing.html#customizing-resourceful-routes
• How to parse JSON

Introduction

• Build simple JSON API project with hanami (something like "Getting Started")

Model

• check this PR: #343

Other guides

• How to build custom CLI
• Add information about hanami-utils
• Security Guide
• Debugging hanami app
• Profiling hanami app
• How to contribute to hanami

cc @hanami/core

There should be a guide on database configuration explaining where to put the database link and what it should look like with majorly used databases (sqlite, mysql, postgresql)

We have a little problem with travis now. For fixing this, I see two ways:

1. Ask @joneslee85 to help us and set new password to secret;
2. Update build config in travis for a new person (travis guide how to set secret value).

What do you prefer @hanami/core?

The blog page only links to the 5 most recent posts: http://hanamirb.org/blog/

I couldn't find any other way to see older blog posts. I just searched google for what I was looking for).

It makes sense to only show 5 most recent by default, but others should be reachable.
We can either add pagination, or make an 'archive' index page that links to all the blog posts.

I checked new crystal site design and found that they have excellent community page:
https://crystal-lang.org/community/

I think we can create something similar for hanami too. WDYT?
/cc @hanami/core

Now we have only has_many association example. I think it will be good to add has_one example too.
WDYT?

/cc @hanami/documentation @jodosha @cllns

As from the section of the getting started page working-with-entities

For now, we want it to carry title and author information. Let's add these attributes.

# lib/bookshelf/entities/book.rb
class Book < Hanami::Entity
end

The text implies that readers should add attributes which I suppose is not true.

Hello,
Anchor icon which is used on the blog posts is differs from the site styleguide (IMHO)

May I submit PR for it, or this works as designed ?

The bundle directory is quite large and takes a long time to clone down here in Kangaland. I would suggest removing this directory and instead relying on the Gemfile.lock.

My concern with it being so huge is that it is a barrier-to-entry for new OSS contributors, who may also come from countries with slow internet connections.

WDYT?

Hi,

This is really a small issue, and maybe it's by even by design, but I find it really annoying that whenever I open a new page in /guides, the focus goes to the search box on top, because I can't scroll with PageUp/PageDown right away. And if I scroll with the mouse, when I click on a link (typically on the right-hand side panel), it does not activate the link but removes the focus from the search box. Then I have to click again. And at least on Firefox, the search box, when it receives the focus, extends too much and pushes the top-right-hand links div down, pushing the whole page contents down, so that means that the link will move up after the first click.

I hope that makes sense.

Those are links to:

• /guides/1.0/actions/rack-integration/
• /guides/1.1/actions/rack-integration/
• /guides/actions/rack-integration/ (which is same as 1.1 right now)

I think we should just link to the last one. What do others think? @davydovanton @hanami/core

Since 1.0 there's #fields_for_collection

It's basically the same as #fields_for for has_many relationships or can be used for bulk updates/creations

Right now we use dotenv in development and test environments, but not production. This issue has effected several people in the chat.

We should consider adding a guide for Environment variables. Specifically, explain how they're set in development and test, as well as offer some suggestions for how to set them in production (heroku config:set,.env.production combined with an entry in .gitignore, or even .bashrc).

We can talk about SERVE_STATIC_ASSETS and APPNAME_SESSIONS_SECRET there too.

We already have a mini-guide for deploying to Heroku on the homepage, in a tab, but we should put this same information (with more explanation) into a Guide. Perhaps we should make it generic, but focus on Heroku, since it's a standard enough that we can help people.

The cause for this is that @sovetnik asked about a deploying issue in the Gitter chat. We should make it so people don't have to ask in the chat.

Here https://github.com/hanami/hanami.github.io/blob/build/source/guides/head/actions/sessions.md it says that

Custom storage technologies are autoloaded via require "rack/session/#{ adapter_name }". 

However, when I create file with class Rack::Session::Foo loadable by require 'rack/session/foo', it's not autoloaded and server complains about uninitialized constant. At the moment I just require it manualy at top of application.rb, but the guide should probably be reworded somehow (not sure how).

See more informations here: https://discourse.hanamirb.org/t/select-clause-in-a-repository/396/2

Just started at https://github.com/translation-gang/ru.hanami.github.io
Primary maintainer is @sovetnik.

Noticed a bug on the helpers/escape guide today:

@davydovanton any ideas?

I think it's an important theme for guides. That's why I suggest to add this. We can use this SO question for example:
https://stackoverflow.com/questions/43080678/join-query-in-hanami-model

Reference to #343

As from the Architectures page, there is a sentence said that

Indeed, by opening lib/bookshelf.rb, we find the Bookshelf module, which is the main namespace of our project.

From my workspace, I don't see Module Bookshelf on such file. My lib/bookshelf.rb content is as follow
Hanami::Utils.require!("#{__dir__}/bookshelf")

Not sure if such text is still valid.