Mechanize ne soumet pas le formulaire

Question

J'essaie de me connecter à un site via mechanize, et je remplis tous les formulaires, puis je les soumets, mais à chaque fois que j'essaie, je me retrouve sur la même page, donc soit je suis redirigé (sans message d'erreur), soit le formulaire ne se soumet pas. Comment cela se fait-il ?

Le code

require 'mechanize'
class Scraper

  def initialize
    @a = Mechanize.new { |agent|
      agent.follow_meta_refresh = true
    }
  end

  def login

    @a.get("https://login.salesforce.com/") do |page|
      red = page.forms.first do |form|
        form.field_with(:type => "email").value = username
        form.field_with(:type => "password").value  =  password
      end.click_button
      pp red
      #puts main.title
    end
  end
end
s = Scraper.new
s.login

La page initiale :

   #<Mechanize::Page
   {url #<URI::HTTPS:0x007f8f39d6fb30 URL:https://login.salesforce.com/>}
   {meta_refresh}
   {title "salesforce.com - Customer Secure Login Page"}
   {iframes
    #<Mechanize::Page::Frame
     "marketing"
     "https://www.salesforce.com/login-messages/messages.html?noroundedcorner">}
   {frames}
   {links
    #<Mechanize::Page::Link "Salesforce" "http://www.salesforce.com">
    #<Mechanize::Page::Link
     "Forgot your password?"
     "/secur/forgotpassword.jsp?locale=us">
    #<Mechanize::Page::Link
     "Sign up for free."
     "https://www.salesforce.com/form/trial/freetrial.jsp?d=70130000000Enus">}
   {forms
    #<Mechanize::Form
     {name "login"}
     {method "POST"}
     {action "https://login.salesforce.com/"}
     {fields
      [hidden:0x3fc79cec63ac type: hidden name: un value: ]
      [hidden:0x3fc79cec6244 type: hidden name: width value: ]
      [hidden:0x3fc79cec60c8 type: hidden name: height value: ]
      [hidden:0x3fc79cec5efc type: hidden name: hasRememberUn value: true]
      [hidden:0x3fc79cec5d58 type: hidden name: startURL value: ]
      [hidden:0x3fc79cec5bc8 type: hidden name: loginURL value: ]
      [hidden:0x3fc79cec5a38 type: hidden name: loginType value: ]
      [hidden:0x3fc79cec987c type: hidden name: useSecure value: true]
      [hidden:0x3fc79cec969c type: hidden name: local value: ]
      [hidden:0x3fc79cec9520 type: hidden name: lt value: standard]
      [hidden:0x3fc79cec9340 type: hidden name: qs value: ]
      [hidden:0x3fc79cec9174 type: hidden name: locale value: ]
      [hidden:0x3fc79cec8f80 type: hidden name: oauth_token value: ]
      [hidden:0x3fc79cec8db4 type: hidden name: oauth_callback value: ]
      [hidden:0x3fc79cec8be8 type: hidden name: login value: ]
      [hidden:0x3fc79cec89cc type: hidden name: serverid value: ]
      [hidden:0x3fc79cec8814 type: hidden name: display value: page]
      [field:0x3fc79cec8670 type: email name: username value: ]
      [field:0x3fc79cec84e0 type: password name: pw value: ]}
     {radiobuttons}
     {checkboxes
      [checkbox:0x3fc79cec833c type: checkbox name: rememberUn value: ]}
     {file_uploads}
     {buttons [button:0x3fc79cecac2c type:  name: Login value: ]}>}>

La dernière page :

#<Mechanize::Page
{url #<URI::HTTPS:0x007f9d1d250960 URL:https://login.salesforce.com/>}

{meta_refresh}
   {title "salesforce.com - Customer Secure Login Page"}
   {iframes
    #<Mechanize::Page::Frame
     "marketing"
     "https://www.salesforce.com/login-messages/messages.html?    r=https%3A%2F%2Flogin.salesforce.com%2F&noroundedcorner">}
   {frames}
   {links
    #<Mechanize::Page::Link "Salesforce" "http://www.salesforce.com">
    #<Mechanize::Page::Link
     "Forgot your password?"
     "/secur/forgotpassword.jsp?locale=us">
    #<Mechanize::Page::Link
     "Sign up for free."
     "https://www.salesforce.com/form/trial/freetrial.jsp?d=70130000000Enus">}
   {forms
    #<Mechanize::Form
     {name "login"}
     {method "POST"}
     {action "https://login.salesforce.com/"}
     {fields
      [hidden:0x3fce8e93aaac type: hidden name: un value: ]
      [hidden:0x3fce8e93a8a4 type: hidden name: width value: ]
      [hidden:0x3fce8e93a638 type: hidden name: height value: ]
      [hidden:0x3fce8e93a390 type: hidden name: hasRememberUn value: true]
      [hidden:0x3fce8e93a19c type: hidden name: startURL value: null]
      [hidden:0x3fce8e939f58 type: hidden name: loginURL value: null]
      [hidden:0x3fce8e939cc4 type: hidden name: loginType value: ]
      [hidden:0x3fce8e9399a4 type: hidden name: useSecure value: true]
      [hidden:0x3fce8e93979c type: hidden name: local value: ]
      [hidden:0x3fce8e939648 type: hidden name: lt value: standard]
      [hidden:0x3fce8e93d414 type: hidden name: qs value:r=https%3A%2F%2Flogin.salesforce.com%2F]
      [hidden:0x3fce8e93d284 type: hidden name: locale value: ]
      [hidden:0x3fce8e93d0cc type: hidden name: oauth_token value: ]
      [hidden:0x3fce8e93cf50 type: hidden name: oauth_callback value: ]
      [hidden:0x3fce8e93cd98 type: hidden name: login value: ]
      [hidden:0x3fce8e93cc44 type: hidden name: serverid value: ]
      [hidden:0x3fce8e93cab4 type: hidden name: display value: page]
      [field:0x3fce8e93c780 type: email name: username value: ]
      [field:0x3fce8e93c4c4 type: password name: pw value: ]}
     {radiobuttons}
     {checkboxes
      [checkbox:0x3fce8e93c334 type: checkbox name: rememberUn value: ]}
     {file_uploads}
     {buttons [button:0x3fce8e93b9d4 type:  name: Login value: ]}>}>

Quel est le problème avec mon code ?

Answer 1

Le site web utilise Javascript pour traiter la connexion, ce que mechanize ne peut pas gérer. Vous pourriez utiliser quelque chose comme Sélénium pour accéder au site.

Answer 2

Sur le formulaire, il y a des champs cachés un , width , height qui doivent également contenir le nom d'utilisateur et quelques chiffres, en plus de username champ contenant le nom d'utilisateur.

  [hidden:0x3fce8e93aaac type: hidden name: un value: ]
  [hidden:0x3fce8e93a8a4 type: hidden name: width value: ]
  [hidden:0x3fce8e93a638 type: hidden name: height value: ]

Vous pouvez utiliser l'inspecteur de Chrome sous l'onglet Réseau pour surveiller le contenu de la requête qui est effectivement envoyée au serveur (activez l'option "préserver le journal"), après avoir été modifiée par le javascript.