Skip to content

Latest commit

 

History

History
778 lines (616 loc) · 46.7 KB

index.md

File metadata and controls

778 lines (616 loc) · 46.7 KB
Raw
<script src="http://ajax.googleapis.com/ajax/libs/jquery/1.10.2/jquery.min.js"></script> <script src="//netdna.bootstrapcdn.com/bootstrap/3.1.1/js/bootstrap.min.js"></script>

Node.js

  • npm config set prefix '~/.npm-packages' and export PATH="$PATH:$HOME/.npm-packages/bin"

Gulp

  • gulp --tasks to list all gulp tasks

HTML

  • inline: <span>, <a>, <em>, <strong>, <img>,<input>. Inline elements do not cause transitions to a new line, but will be displayed one next to the other horizontally. <br> link The line-break is an odd case, as it’s an inline element that forces a new line. However, as the text carries on on the next line, it’s not a block-level element.
  • block-level: <div>, <p>, <article>,<h1>,<ol><li>,<table>,<form>. Block elements are set like blocks that stack on top of each other and will never display next to one another horizontally
  • inline-block: <label>, <input>, and <textarea>. Inline blocks will act as inline elements (elements are displayed next to each other), but differ in that they can be for instance resized. For example, the <textarea> field can be displayed as a large rectangle, but it can take up space beyond a single line of text.
  • to make <li> horizontal, change its style to display: inline, or float: left
  • onmouseover is not fired when element is disabled, but you can easily wrap inside another elemenent with padding 1px

XPATH

CSS

  • to set with for anchor a tag (or any other inlin element), you need to change it's display to block or inline-block, than you can use max-width or width
  • keep text in one line span { white-space: nowrap; }
  • when you want to have different child of first element, you can with
.target-parent li:first-child .target-child {
  display: none;
}
  • z-index only works on positioned elements (position:absolute, position:relative, or position:fixed).
  • border: 1px needs to be defined before border-style: solid or border-color: red
  • to overwrite !important write another rule with Simply add another CSS rule with !important, and either give the selector a higher specificity (adding an additional tag, id or class to the selector), or add a CSS rule with the same selector at a later point than the existing one (in a tie, the last one defined wins).
  • cascade rules are http://www.w3.org/TR/2011/REC-CSS2-20110607/cascade.html#cascade
  • position element in center margin: auto works for
    but not for , so try this link
.centered {
  position: fixed;
  top: 50%;
  left: 50%;
  /* if you don't know the size... bring your own prefixes */
  transform: translate(-50%, -50%);
  /* if you know the size
  margin-top: -50px;
  margin-left: -100px; */
}

SASS

h3 {
  font-size: 20px;
  div {
    .some-parent-selector & { // this will compile to .some-parent-selector h3 div { font-size: 24px }
      font-size: 24px;
    }
  }
}

Javascript

  • to check if element is visible and to scroll into view

function check_if_in_view($element, options) {
  // https://developer.mozilla.org/en-US/docs/Web/API/Element.scrollIntoView
  // use options parameter to force scrolling even element is visible on current view
  // if options == true/false, the top of the element will be aligned to the top/bottom of the visible area of the scrollable ancestor.
  if (typeof document.getElementsByTagName('body')[0].scrollIntoView == 'undefined')
  {
    console.log("scrollIntoView not defined");
    return true;
  }
  else if (typeof options != 'undefined')
  {
    console.log("ckeckIfInView "+options);
    $element[0].scrollIntoView(options);
  }
  else
  {
    // check if visible
    var offset = $element.offset().top - $(window).scrollTop();
    if(offset < 0){
        console.log("alignWithTop=true");
        $element[0].scrollIntoView();
        return false;
    }
    else if (offset > window.innerHeight){
        console.log("alignWithTop=false");
        $element[0].scrollIntoView(false);
        return false;
    }
    else
    {
      console.log("no scroll, its visible");
      return true;
    }
  }
}

  • if element is hidden, some older browser security restrictions do not allow it to trigger a click with $("#test").click();. So it is better to show element but move it out of visible area style="position:absolute; top:-100px;" link.

  • to detect jQuery and load if needed follow this (you can put check onload window.onload = function(e) { it (type jQuery == 'undefined')...}) so you can wait for eventual later on page jQuery include

  • select elements by data attribute $('a[data-color=true]') or $('a:data(collor)');

  • to check if element(s) exists if( $('#selector').length )

  • document ready can be detected in $(document).ready(function(){}) or because ready can be used only on current document selector is not needed $(function(){});

  • when <a></a> does not have href that it will not follow thelink so you do not need to preventDefault. Use this for toggling some part

  • hover on mobile devices does not work what is expected. It stays in hover state until the user press (click) on the screen next time.

  • for popups with buttons, if you want to close popup clicking anywhere outside button, button should not be included in background div to clearly separate when popup should be closed. if it is inside we can workaround with targeting more specifically and stop propagation $('#letsdoit').click( function(e) { e.stopPropagation();});

  • in rails-ujs click on links with remote:true is not efected with e.preventDefault()

  • get the user agent online http://ifconfig.me/all

  • detect browsers version

    function is_uiwebview(){ return /(iPhone|iPod|iPad).*AppleWebKit(?!.*Safari)/i.test(navigator.userAgent); } function is_android(){ return /Android/i.test(navigator.userAgent); } function is_idevice(){ return /(iPhone|iPod|iPad)/i.test(navigator.userAgent); }

  • detect browser type in rails

    def is_android?
    request.user_agent =~ /Android/i
    end

    def is_iphone_app?
    request.user_agent =~ /(iPhone|iPod|iPad).*AppleWebKit(?!.*Safari)/i
    end

    def is_idevice?
    request.user_agent =~ /(iPhone|iPod|iPad)/i
    end

    def is_iphone?
    request.user_agent =~ /(iPhone|iPod)/i
    end

SQL

  • PG::InvalidColumnReference: ERROR: SELECT DISTINCT ON expressions must match initial ORDER BY expressions is a problem when you do not include column a for which you want to order, so it should be SELECT DINSTINCT ON (dinstinct,a)...

REGEX

  • to regex email in text stream
r = Regexp.new(/\b[a-zA-Z0-9._%+-]+@[a-zA-Z0-9.-]+\.[a-zA-Z]{2,4}\b/)     
emails = content.scan(r).uniq
  • in Rails you can use default devise validates_format_of :recipient_email_address, with: User.email_regexp, allow_blank: true
  • (a|b) when you want to match a OR b
  • .*? when you want to force minimum rather than maximum match
  • great source for regular expression explanation http://www.regexr.com/ http://regex101.com/
  • grep lines not containig word, you need to use param -P. Vim example find non disable buttons: :grep -P "^.*button(?!.*disable.*)" app/views/* -R
  • to extract url params (subject,body) from mailto href you can use URI to un-escape html characters: subject_match = mailto.match(/[?&]subject=([^&]+)[&$]/)[1] if subject_match subject_from_mailto = URI.unescape subject_match end

iSO android and other mobile phone

  • fadein and fadeout does not work well on mobilephone iOS android, its better to use display: block, display: none
  • click event does not work on iOS when element is not clicable. simple solution is to add style: "cursor: pointer".
  • on android when file field input is display: none, you can not trigger click on it. Better is to use style="position:absolute; left:-1000px"

Vim

Heroku

Heroku problems that we shold not care

  • Could not create resource with vendor, please try again later
  • Launching... Push failed: could not store slugis a heroku

To prevent heroku from sleep, add scheduler on every hour: rake dyno_ping --trace, and in lib/tasks/dyno_ping.rake

desc "Ping"
task :dyno_ping do
#  require "net/http"  # uncomment on ruby 2.0
  uri = URI("http://time-tracklng.herokuapp.com/");
  Net::HTTP.get_response(uri)
end

Vagrant

  • if you get a problem with SSL certificate
An error occurred while downloading the remote file. The error
message, if any, is reproduced below. Please fix this error and try
again.

SSL certificate problem: unable to get local issuer certificate
More details here: http://curl.haxx.se/docs/sslcerts.html

curl performs SSL certificate verification by default, using a "bundle"
 of Certificate Authority (CA) public keys (CA certs). If the default
 bundle file isn't adequate, you can specify an alternate file
 using the --cacert option.
If this HTTPS server uses a certificate signed by a CA represented in
 the bundle, the certificate verification probably failed due to a
 problem with the certificate (it might be expired, or the name might
 not match the domain name in the URL).
If you'd like to turn off curl's verification of the certificate, use
 the -k (or --insecure) option.

solution is to sudo cp /etc/ssl/certs/ca-certificates.crt /opt/vagrant/embedded/cacert.pem link or to add config.vm.box_download_insecure = true to Vagrantfile link

Cucumber

  • @temp is shared between blocks

  • pause can be implemented with: STDIN.getc

    require 'capybara/dsl' => true include Capybara::DSL => Object Capybara.default_driver = :selenium => :selenium visit "http://google.co.uk" => ""

  • selenium-webdriver scripts can be debuged with

    driver = ""
    eval File.open('./scraping_selenium.rb').read

or something like

require 'rubygems'
require 'ruby-debug'
x = 23
puts "welcome"
debugger
puts "end"

Rspec

  • double('output').as_null_object ignore other messages (unexpected messages are recorded) so this will work:

    output.should_receive(:puts).with('Second guess:')

  • let will create memoized local variable so you do not need @var

  • https://www.relishapp.com/rspec/

  • https://github.com/jnicklas/capybara

  • https://github.com/cucumber/cucumber

  • http://rubydoc.info/github/thoughtbot/shoulda-matchers/master/index

  • https://github.com/thoughtbot/factory_girl

  • let(:form) { mock_model(Form, :success_message => "Success message") }

  • gem 'spork-rails'

    spork --bootstrap in spec/spec_helper.rb move everthing inside Spork.prefork add to .rspec --drb

  • gem 'guard-rspec'

    guard init spork in Guard file on guard 'spork', :test_unit => false

  • Do not depend on implementation https://www.youtube.com/watch?v=m3nMrWpIRKo use stub and mock, and check with integration test whole functionality

  • should_receive can be before action, but response.should and flash.should have to be after

  • testing static pages:

    subject { page } shared_examples_for "all static pages" do it { should have_selector('h1', text: heading) } end describe "GET Home page" do before { visit root_path } let(:heading) { 'Home page' }

    it_should_behave_like "all static pages"

    end

  • testing model validation:

    before do @user = User.new(name: "Example User", email: "[email protected]") end describe "when email is not present" do before { @user.email = " " } it { should_not be_valid } end

Rails

  • form_for method param can be ignored since for new object is POST, updating is PATCH... but if you are using form objects where nested elements sometimes exists, sometimes not, this method should be fixed (also in routing)

  • when using username: first_name.last_name instead of id resources :users, :constraints => { :id => /[^\/]+(?=\.html\z|\.json\z)|[^\/]+/ }, except: :new do, users with first_name in ['new','edit','show'] will not be able to work properly

  • keys in locale files for i18n can be relative to current file: lazy lookup <%= t '.title' %> in app/views/books/index.html.erb is the same as <%= t 'books.index.title' %>. Also works for partials (skip leading _). Template could use locale extension sr.html for example: app/views/books/index.sr.html.erb. List of locale files on i18n gem.

  • pry in rails with pry -r /home/orlovic/rails/produceruncatarse/config/environment.rb or in user pry-rails gem

  • if you need asset_path in some class (for example uploader) you can use this ActionController::Base.helpers.asset_path 'image.png'

  • access url route helpers in model Rails.application.routes.url_helpers.menu_url(link) and define host with Rails.application.routes.default_url_options[:host] = Rails.application.secrets.default_url_options. You can delegate :url_helpers, to: 'Rails.application.routes' in class definition.

  • rails use cache for activerecord, and if you need to send query again and fetch fresh data, you can use .reload

  • truncate_html is when you want to truncate text inside html. When you are using raw than it is advisable to use sanitize

  • target with data- attributes when you put some javascript on elements. When new markup come, you just need to copy data attributes. Also they can contain some real data: id or array of some id's. In rails do not use data: { total_value: 0} because its hard to search. It's better to use "data-total-value" => 0

  • when form is submitted and when there are errors, scaffold rails render again :new action, but the browser url is changed to the requested url so if than user refresh the page, it is another template (usually it's a index action of the same controller, but could be another controller)

  • when you want to create nested models on the same form, its nice to generate template on server and change id rails cast stack. If you are only creating its easier to just use hidden template with name that is the same as for nested model, for example email_message[proposal_attributes][proposal_item_attributes][][price]. Form need to be configured to send array instead of hash for nested parameters parameters, that is child_index: ''. Edit is problematic without id, so this is usefull only for creating.

  • folders in app/assets/* are hidden when you are accessing files, for example: localhost:3000/assets/file.txt should be here app/assets/javascript/file.txt or app/assets/img/file.txt (last one overrides prev). In production files that are listed in application.css.erb would be served as part of application.css, but all other files are served as is on their locations

  • to undo rails generate model survey you can call rails destroy model survey

  • All files from assets are provided but not all are included in all layouts (default is app/views/layouts/application.html.erb)

  • comments like // *= require_tree . does not exclude it

  • rake stats shows number of code lines, tests ...

  • debugging tools: finish (up backtrace function)

  • rendering views instead of writing js partials http://stackoverflow.com/questions/16328472/rails-render-a-view-not-a-partial-from-within-a-view

  • in views, render "something" will render partial _something, render "controller/something" will render template, render :partial => 'contr/something' will render partial

  • render @products, will render partial _product with local variable product

  • to run production version follow this

    RAILS_ENV=production rake db:setup RAILS_ENV=production rake assets:precompile rails s -e production rails c production

  • you can use raw or .html_safe but it is advised to use sanitize with it to strip script tags

  • for ajax actions (links with remote:true) it should be implemented both, JS and HTML version, since user can use right click on in new tab for that links. It can be respond to format.html, or in before filter before_action :reload_on_html_request, only: [ :active, :pause] and

  def reload_on_html_request
    # we have ajax link (not buttons) that can be opened in new window which causes an error... we redirect them to back (referrer)
    # http://api.rubyonrails.org/classes/ActionController/Redirecting.html
    # http://stackoverflow.com/questions/771656/correctly-doing-redirect-to-back-in-ruby-on-rails-when-referrer-is-not-availabl
    if ! request.xhr?
      redirect_to :back # that is the same as request.env["HTTP_REFERER"]
    end
  rescue ActionController::RedirectBackError
    redirect_to root_path
  end
  • if you have ajax links, it is advisable to implement notification in case of error since ajax error tends to be silent:
    // ajax error handling, currently only for debugging                                        
    LOG && $(document).on('ajax:error', '[data-remote]', function(e, xhr, status, error) {      
      LOG && console.log("ajax:error [data-remote] e.target="+e.target+" status="+status+" error="+error);
      //disable eventual popups so user can see the message                                     
      $('.active').removeClass('active');
      flash_alert("Please refresh the page. Server responds with: \"" +status+ " " + error + "\".");
    });
  • data-disable-with is provided with rails-ujs and if you want to manually send ajax requests than you need to manually fire up ajax:complete events. For example $(this).trigger("ajax:complete"); in:
  $(document).on('click', '[data-check-if-in-editing-mode]', checkIfInEditingMode);
  function checkIfInEditingMode(e) {
    LOG && console.log("checkIfInEditingMode");
    e.preventDefault();
    // do not submit automatically
    // if there are open questions title or description, warn the user
    if ($(this).data('check-if-in-editing-mode') != "ignore" && $('[data-in-editing-mode]').length )
    {
      $('[data-in-editing-mode]').first().each( function() {
        checkIfInView($(this));
        alert($(this).data('in-editing-mode'));
      });
      $(this).trigger("ajax:complete");
    }
    else
    {
      $.ajax({
        url: $(e.target).attr('href'),
        type: 'put',
        dataType: 'script',
        error: function(jqXHR, textStatus, errorThrown) {
          flash_alert("Please refresh the page. Server responds with: \"" +textStatus+ " " + errorThrown + "\".");
          $('[data-check-if-in-editing-mode]').trigger("ajax:complete");
        },
        success: function( data, textStatus, jqXHR ) {
          $('[data-check-if-in-editing-mode]').trigger("ajax:complete");
        },
      });
    }
  }

  • turbolinks loads new page, replaces body and title, and use pushState. Any javascript is run after replacement so you can reference all elements in <script> tags on the beggining of the page (if you really want to). Execution order of non dymanically (defer or async) script are the same as they appear on the page (inline script is halted until external is loaded). But with turbolinks, as with other dynamic loading where you can not know the order, execution is done after replacing, but before external script is loaded, so you probably can not have javascript_include_tag inside a view, because the code it provides will be available after all script on a page is done, and if you have several includes, the order is not preserved since the shortest one can be executed before all others. So the solution is to render all javascript inside the template with <%= render 'app/assets/javascript/custom_script', formats: [:js] %>

  • accessing helper functions in controller can be done using view_context (including ApplicationHelper is not generally a good idea)

  • mail attachments.inline['logo.png'] = File.read('app/assets/images/logo.png') will be attached with disposition="attachments" if there is no <%= image_tag attachments['logo.png'].url %> in a view url. Rails look for its reference and if it does not find it will not include it as "inline". One way to include all inline attachments but not to show them (that is the case when you already have a inline img tag inside the view, for example forwarding emails)

    <% attachments.each do |a| %> <% if a.content_disposition[0..5] == "inline" %> <% image_tag a.url, size: "0x0" %> <% end %> <% end %>

This way they are using its own content_id. To include inline attachments for existing mails, you have to use the same content_id

 @email_message.temp_images.each do |mandrill_image_cid, mandrill_image_hash|
    # mandrill_image_hash.keys = ["name", "type", "content"] and they are always base64
    attachments.inline[mandrill_image_cid] = { mime_type: mandrill_image_hash["type"], content: Base64.decode64(mandrill_image_hash["content"]), content_id: mandrill_image_hash["name"]} # we have to set mime type and content_id because we are using the original email with refers this content_id
  end

And as we have refenece for those images in email body, we do not need to use img_tag with size "0".

  • sorting could be done with this helper that is used like <th><%= sortable "id" %></th>:

    def sortable(column, title = nil) title ||= column.titleize css_class = (column == sort_column) ? "current #{sort_direction}" : nil direction = (column == sort_column && sort_direction == "asc") ? "desc" : "asc" link_to title, { :sort => column, :direction => direction }, { :class => css_class } end

and this two methods:

def sort_column
  User.column_names.include?(params[:sort]) ? params[:sort] : "id"
end

def sort_direction
  %w[asc desc].include?(params[:direction]) ? params[:direction] : "desc"
end

so in controller we can write `@users = User.order(sort_column+" "+sort_direction)

Rails console

  • puts Readline::HISTORY.entries.split("exit").last[0..-2].join("\n") show last commands so you can copy and paste to some code

Active record

  • in User class usually we have before_filter :generate_referral_token
  def generate_referral_token
    # http://stackoverflow.com/questions/6021372/best-way-to-create-unique-token-in-rails
    self.referral_token = loop do
      # http://ruby-doc.org/stdlib-2.1.0/libdoc/securerandom/rdoc/SecureRandom.html
      random_token = SecureRandom.urlsafe_base64(nil, false)[0..REFERRAL_TOKEN_LENGTH-1].upcase
      break random_token unless self.class.exists?(referral_token: random_token)
    end
  end
  • if question mark is used in activerecord query scope :customers, -> { where('role = ?', ROLE_EMPLOYER)} than you should use table name because in joins some other table could have role column scope :customers, -> { where('users.role = ?', ROLE_EMPLOYER)}, but the best is if you use hash scope :customers, -> { where( role: ROLE_EMPLOYER)} (you can even build new customer with User.customers.build
  • has_and_belongs_to_many needs a table model1_model2 (where model1 name is less than model2 in alphabet)
  • for HABTM you can use f.collection_check_boxes :min_qualification_ids, MinQualification.active.all, :id, :name. Trick here is to use min_qualification_ids. That _ids is provided for all has_many association. In controller you need just to update parent record and all has_many_and_belongs associated records mapping will be updated (here is table min_qualifications_users)
min_qualifications_params = params.require(:user).permit( min_qualification_ids: [])
non_empty_min_qualifications_params = min_qualifications_params.each_with_object({}) { |(k,v),o| o[k] = v.reject &:blank? }
@user.update!(non_empty_min_qualifications_params)
  • you can join several tables,just use their name: Job <-> MinQualitication <-> User
class Job
  def suggested_jobseekers
    User.active_job_seekers.joins(min_qualifications: [:jobs]).where(jobs_min_qualifications: { id: id} )
  end
end

  • validate with if contition validates_presence_of :user_id, if: Proc.new{ |u| u.active? }

  • when you add validation validate_presence_of :username which you generate, you should also write migration for existing users and before_create hook to populate that field, update seed file

  • accept_nested_attributes :jobs is needed if you want to use f.fields_for :jobs do |fjob| for user.jobs.new nested form. Usually you add validations in jobs model, for example validates :user, presence: true, but than you need to insert funky inverse_of in user model has_many :jobs, :dependent => :destroy, inverse_of: :user if you want this validation to pass for nested attributes

  • to make some fields private you can use link

    class YourModel < ActiveRecord::Base private def my_private_attribute self[:my_private_attribute] end def my_private_attribute=(val) write_attribute :my_private_attribute, val end end

  • its better to use destroy instead of delete since it will trigger destroying associated models but its better not to remove anything since it is blocking operation

  • when you are adding new validations (for example title needs to be present) write also a migration file to make sure all items in database are valid. To check if some unvalid object exists in database you can run Rails.application.eager_load! and e=ActiveRecord::Base.descendants.select { |c| c.all.select {|i| !i.save}.present? }.map {|c| c.all.select {|i| !i.save}.map { |i| i.save;puts c;i.errors} }.

  • [default_scope is appended to all queries][http://rails-bestpractices.com/posts/806-default_scope-is-evil], only way to remove it is uncope

  • Assigning an object to a belongs_to association does not automatically save the object. It does not save the associated object either.

  • you can specify order for association: has_many :educations, -> {order 'graduation_year DESC'}, :dependent => :destroy

  • you can attach multiple polymorphic association with conditions:

    has_many :attachments, as: :attachmentable, conditions: { is_image: false } has_many :images, as: :attachmentable, class_name: 'Attachments', conditions: { is_image: true }

  • naming conventions is to have foreign_table_name_id for associations (belongs_to :foreign_table_name), type for single table inheritance (class Firm < Company), imageable_id and imageable_type for polymorphic associations (belongs_to :imageable, polymorphic: true)

  • polymorphic items belogs to one object so you should not use someting like new_user.location = Location.where(:address => address).first_or_create because it will steal location object from old user. Better is to always build new new_user.build_location address: address

  • polymorphic example user.location = Location.new address: address will create (!) new location object, but user.build_location address: address won't

  • instead of has_many :applied_jobs, through: :applications, source: :job it is better to use user.applications.map(&:job) link

  • new_record? does not work with associated count. It is beter to use length ie survey.questions.length

Gems

  • paperclip:
    • default image path in assets pipeline is using helper in controller ActionController::Base.helpers.asset_path('missing_:style.png')
    • new = menu_item.dup will copy attributes. to copy image new.photo = menu_item.photo it will download image and upload images
  • performance:
    • miniprofiler for analysing load time, disable with http://mysite.com?pp=disable
  • when using pagination on page where you have some activating/archiving Kaminary gem with ajax does not know which params should be black listed so use something like paginate @jobs, params: { id: nil, action: 'index' } in response index.js response file like this ajax example
  • get the country based on ip address:
  • browser-timezone-rails is best gem to display local time withouht any calculation (using jsTimezoneDetect)
  • pdf generator wickedpdf

HTTPARTY

CANCAN

  • if you set can :manage, :all in ability file, then after that you should define only what cannot. Only way that you need to define can after 'can :manage, :all' is for index, create or update actions, if you want to populate with hash, for example can :index, Produc, { user: current_user }, no block definitions for abilitity for managers since they can magane all!
  • for load_and_authorize_resource for :index, if you define ability in a block instead of hash, then load_resource will not populate @products since it does not know how to do it. For :show, :edit, :update and :destroy it will fetch by params[:id], for :new and :create it will create new one if you define hash, and it will be overwritten with params[:class] attributes

CARRIERWAVE

If you're using ActiveRecord, CarrierWave will indicate invalid URLs and download failures automatically with attribute validation errors. If you aren't, or you disable CarrierWave's validate_download option, you'll need to handle those errors yourself.

  • another way is to use key attribute, for example: user.key or user.avatar.key = 'http://asd.asd'

DEVISE

  • in facebook omniauth callbacks params do not see actuall param my_param in user_omniauth_authorize_path(:facebook, my_param: 1). You can access it in request.env['omniauth.params']['my_param']

  • you should not write after_create :my_func hook in User model, since devise resource.save will return true, but resource.error will contain some fanky error Email can't be blank link

  • if you do not allow users do destroy accounts, in your class RegistrationsController < Devise::RegistrationsController you should overide destroy with something like (notification)

    def destroy ExceptionNotifier.notify_exception(Exception.new('destroy user'), env: request.env, data: { current_user: current_user }); end

BACKGROUND WORKER

  • is your server need to use third party api, it is advisable to put it in background, so in case of traffic congestion (for example you and target api) user do not receive 'not responding page' since browsers have timeout 30 sec.
  • SuckerPunch is done in the same thread and can be run on heroku on single dyno (drawback is that it is cleaned on each restart, deploy, even error with exception)
  • DelayedJob is peristent since it stores serialized object in db, and when it comes to run it, it retrive object and run method (drawback is number of ActiveRecord connections to db, since each worker need two, one to retreive the object, another to use with ActiveRecord. Watch out when you backup database, delayed jobs should not be restored).
  • Sidekiq is identical to Socker Punch, fast and run on single thread, better than delayed_job since it can run 20 processes in same time, threadsafe, works on one heroky dyno with unicorn webserver set up heroku config:set REDIS_PROVIDER=REDISTOGO_URL (drawback is that you need to use Redis server for in memory database storage) If there is an error invalid byte sequence in US-ASCII try to export LANG=en_US.UTF-8 before installing a gem

For rvm put source ~/.bash_profile in your ~/.bashrc file and you do not need to enable terminal settings to "Run command as login shell"

Ruby

  • dummy ruby objects OpenStruct

  • condition if or ? : in parameters lists to check if function argument should be passed can be used in two ways for hash argumests: merge or conditional. Since hash argument need key/value pair you can not pass a(b:2,{c:3} since {c:3} is not pair. note that we usually need extra () around conditions

    • merge example form_for @user, {}.merge( @contition ? {class: 'main'} : {} ) (key 'class' is not always passed in form), note that we have to pass hash to merge and we can not use {class: 'main'} if @condition since it will try to merge nil and raise exception if @condition == false
    • conditional form_for @user, class: ('main' if @condition)(key 'class' always exists as param, but value is nil unless @contidion).

  • if you want to see all methods for some object, you can a.methods.grep /what/. If you want to list all instance methods, but not methods of superclass a.class.instance_methods(false). To see where it is defined a.method(:what).source_location

  • if you want to return hash as result on some collection (map always return array), you can use { "x" => 1, "y" => 2, "z" => 3 }.each_with_object({}) {|(k,v),o| o[k.to_sym]=v }

  • remove duplicate repettitive spaces from string str.split.join(" ")

  • default params for function could be set like def f(a=1);end, and default values for variables if not defined var = true unless defined? var (do not use var||=true since it will override var=false)

  • if you want to save output of ruby script ruby a.rb |tee output and want to see output in realtime you need to call flush after each puts $stdout.flush or you can set that for whole script STDOUT.sync = true

  • rescue_from are searched from bootom to top.

  • inline rescue can use last return value rescue $! tapas

  • break two nested loops (two levels)

    bank.branches do |branch| break unless branch.employees.each do |employee| break if employee.name == "John Doe" end end

  • number of methods for some object can be large, to search method use grep, for example request.methods.grep /url/

  • self in ruby is implicitly assumed, but you can use it for class methods or if you have attr_accessor :name, and have name="asd", name= is local variable assignment

  • p ["a", "b"].map(&:upcase) is equivalent to p ["a", "b"].map{|string| string.upcase}

  • regular expression http://www.rubular.com/

  • put -w for warning log https://github.com/bbatsov/ruby-style-guide

  • lambda literal syntax is: method = ->(a, b) { a + b }

  • lambda method for multi line blocks:

method = lambda do |a, b|
  tmp = a * 7
  tmp * b / 50
end
  • if dont know if path is array use: [*paths].each { |path| do_something(path) }
  • %w{word} is non onterpolated, W{#{some variable} word} is interpolated array of words separated by space
  • ! has greater priority then == so use () to determine precedence, for example if ! a==3 is not the same as if a!=3

Jekyll

http://cheat.markdunkley.com/

Difference betweet page and post:

  • Post file name contains the title and the date of the post. invalid date in the filename cause an error.
  • Posts are Comparable objects, which means two posts can be compared. The comparison is made by the post date and the post slugs.
  • The generated default relative url of Post and Page are different (eg. /2000/01/01/my-post.html and /about.html).
  • Page can be placed anywhere but Post can only be placed under "_posts" folder.
  • Post has more data out of the box for use in Liquid templates (title, url, date, id, categories, next, previous, tags, content).
  • Posts has unique id.

Interesting plugins:

BASH

  • map local domain names blabla.dev to 127.0.0.1 with:
sudo apt-get install dnsmashq
sudo sh -c "echo 'local=/dev/\n\
address=/dev/127.0.0.1\
' > /etc/dnsmasq.d/dev-tld"
sudo service dnsmasq restart
  • you can run command as another user in two ways
    • su deployer -c 'whoami can add login -l option su -l deployer -c 'rvm list' so PATH is extended with ~/.rvm/bin/ and rvm script works
    • sudo -u deployer whoami simpler, no need for quotes. Add option -i or --login to run as login shell. Wrap command inside bash to properly load XDG_RUNTIME_DIR for rails c or other ruby commands: sudo -i -u deployer /bin/bash -c "cd /vagrant && rails c". Also when you use pipe | or &&, you should wrap inside bash.
  • top command for memory and cpu usage on linux. To sort by memory press f and move up/down and than exist q
  • start windows, in gnome-terminal if negative is used, than it is bottom alligned. try also with wmctrl
alias s="wmctrl -e 1,340,100,-1,-1 -r orlovic;google-chrome http://localhost:3000\&;gnome-terminal --geometry=300x24+0-0 -e 'sh -c \"vi;exec bash\"';gnome-terminal --geometry=80x24-0+0 -e 'sh -c \"git pull;rake db:migrate;rails s;exec bash\"';"
  • xwininfo to click on window to find it's 0x440003d which we can use to create shortcut xdotool windowactivate 0x440003d. Go to Settings -> Keyboard -> Shortcuts -> Custom Schortcuts -> + . Name: main editor command xdotool windowsactivate 0x123123 shortcut Alt+M.
  • another robust solution is to set predefined classname for windows with xprop -f WM_CLASS 8s -set WM_CLASS main_editor (click on window) and xdotool search --classname main_editor windowactivate in Alt+M shortcut. I'm using ALT+HJKL; with browser_preview, main_editor, run_commands, server_log, blog_editor. Command to select all is
for name in browser_preview main_editor run_commands server_log blog_editor; do
  echo "Select $name"
  xprop -f WM_CLASS 8s -set WM_CLASS $name
  echo "thanks"
done
  • to scroll in terminal window, instead shift+page_up we can bind to key simulate keyup, just run xbindkeys_show and xbindkeys -k and create file . I usually map vim shortcuts CTRL+ , so for terminal winodws I use ALT+. There are already two ALT bash shortcut Alt+F Alt+B to move cursor one word (use capitalize F to not open File menu) (CTRL version move by one character) .
# ~/.xbindkeysrc
"xdotool keyup bracketright && xdotool key --clearmodifiers shift+Page_Up"
    Alt+Mod2 + bracketright
"xdotool keyup bracketleft && xdotool key --delay 0 shift+Page_Down"
    Alt+Mod2 + bracketleft
# "xte 'keydown Shift_R' 'key Page_Up' 'keyup Shift_R'"
#    m:0x18 + c:35
#    Alt+Mod2 + bracketright
  • find and remove files find . -type f -name "FILE-TO-FIND" -exec rm -f {} \;
  • test the speed on remote server:
    • download: curl -o /dev/null http://speedtest.qsc.de/1GB.qsc
    • upload: generate large file fallocate -l 1G gentoo_root.img and use scp to test upload link
  • you can put any shell (vim, rails s) to suspend state with Control + z. Than you can put it in background bg if needed. You can use that shell for inspection other things. When you are finished, you can switch back to vim, or rails s, with foreground fg
  • escape single quote ' in linux scripts with $'Hello I\'m here' link
  • escape space in variables with FILENAME=$(printf %q "$FILENAME") link
  • run command in variables with eval $VAR

HACK

  • html escaping is not easy . for example
<IMG SRC=/ onerror="console.log('cao. bolje da filtrirate sadrzaj za js code');location.assign('http://www.iznajmljivanjeprojektoranovisad.in.rs')"></img>

GIT

http://blog.trk.in.rs/2016/05/05/git-tips/

Other