From 448fe8fddbbb030603a8e7ac87be4d3454c6800b Mon Sep 17 00:00:00 2001
From: Chris <cawatson1993@gmail.com>
Date: Wed, 10 Jul 2019 17:19:04 -0700
Subject: [PATCH] Update session cache and remove dnt options

---
 README.md                     |  2 --
 src/arachnid.cr               | 28 +++++++++++++++++++++++++---
 src/arachnid/agent.cr         |  6 +++---
 src/arachnid/session_cache.cr | 17 +++++------------
 4 files changed, 33 insertions(+), 20 deletions(-)

diff --git a/README.md b/README.md
index d80ba45..b602ecd 100644
--- a/README.md
+++ b/README.md
@@ -184,7 +184,6 @@ Arachnid has a ton of configration options which can be passed to the mehthods l
 - **read_timeout** - Read timeout
 - **connect_timeout** - Connect timeout
 - **max_redirects** - Maximum amount of redirects to follow
-- **do_not_track** - Sets the DNT header
 - **default_headers** - Default HTTP headers to use for all hosts
 - **host_header** - HTTP host header to use
 - **host_headers** - HTTP headers to use for specific hosts
@@ -199,7 +198,6 @@ Arachnid has a ton of configration options which can be passed to the mehthods l
 
 There are also a few class properties on `Arachnid` itself which are used as the defaults, unless overrided.
 
-- **do_not_track**
 - **max_redirects**
 - **connect_timeout**
 - **read_timeout**
diff --git a/src/arachnid.cr b/src/arachnid.cr
index 69bf7e9..6789c59 100644
--- a/src/arachnid.cr
+++ b/src/arachnid.cr
@@ -13,9 +13,6 @@ module Arachnid
   # Specifies whether robots.txt should be honored globally
   class_property? robots : Bool = false
 
-  # Should we set the DNT (Do Not Track) header?
-  class_property? do_not_track : Bool = false
-
   # Maximum amount of redirects to follow
   class_property max_redirects : Int32 = 5
 
@@ -62,3 +59,28 @@ module Arachnid
     end
   end
 end
+
+require "json"
+
+# Let's build a sitemap of crystal-lang.org
+# Links will be a hash of url to resource title
+links = {} of String => String
+
+# Visit a particular host, in this case `crystal-lang.org`. This will
+# not match on subdomains.
+Arachnid.host("https://crystal-lang.org") do |spider|
+  # Ignore the API secion. It's a little big.
+  spider.ignore_urls_like(/\/(api)\//)
+
+  spider.every_html_page do |page|
+    puts "Visiting #{page.url.to_s}"
+
+    # Ignore redirects for our sitemap
+    unless page.redirect?
+      # Add the url of every visited page to our sitemap
+      links[page.url.to_s] = page.title.to_s.strip
+    end
+  end
+end
+
+File.write("crystal-lang.org-sitemap.json", links.to_pretty_json)
diff --git a/src/arachnid/agent.cr b/src/arachnid/agent.cr
index b2d4e71..761a9e7 100644
--- a/src/arachnid/agent.cr
+++ b/src/arachnid/agent.cr
@@ -64,11 +64,11 @@ module Arachnid
 
     # Creates a new `Agent` object.
     def initialize(
+      http_client = nil,
       host : String? = nil,
       read_timeout : Int32? = nil,
       connect_timeout : Int32? = nil,
       max_redirects : Int32? = nil,
-      do_not_track : Bool? = nil,
       default_headers : Hash(String, String)? = nil,
       host_header : String? = nil,
       host_headers : Hash(String | Regex, String)? = nil,
@@ -111,10 +111,10 @@ module Arachnid
       @max_depth = max_depth
 
       @sessions = SessionCache.new(
+        http_client || HTTPClient::Default,
         read_timeout,
         connect_timeout,
-        max_redirects,
-        do_not_track
+        max_redirects
       )
 
       @cookies = CookieJar.new
diff --git a/src/arachnid/session_cache.cr b/src/arachnid/session_cache.cr
index ff2aac1..e81fe23 100644
--- a/src/arachnid/session_cache.cr
+++ b/src/arachnid/session_cache.cr
@@ -5,6 +5,9 @@ module Arachnid
   # Stores active HTTP Sessions organized by scheme, host-name and port.
   class SessionCache
 
+    # The HTTPClient class to use for requests
+    property client : HTTPClient.class
+
     # Optional read timeout.
     property read_timeout : Int32
 
@@ -14,24 +17,19 @@ module Arachnid
     # Max redirects to follow.
     property max_redirects : Int32?
 
-    # Should we set a DNT (Do Not Track) header?
-    property? do_not_track : Bool
-
     @sessions = {} of Tuple(String?, String?, Int32?) => HTTPClient
 
     # Create a new session cache
     def initialize(
-      client,
+      client : HTTPClient?.class = nil,
       read_timeout : Int32? = nil,
       connect_timeout : Int32? = nil,
-      max_redirects : Int32? = nil,
-      do_not_track : Bool? = nil
+      max_redirects : Int32? = nil
     )
       @client = client || HTTPClient::Default
       @read_timeout = read_timeout || Arachnid.read_timeout
       @connect_timeout = connect_timeout || Arachnid.connect_timeout
       @max_redirects = max_redirects || Arachnid.max_redirects
-      @do_not_track = do_not_track || Arachnid.do_not_track?
     end
 
     # Determines if there is an active session for the given URL
@@ -60,11 +58,6 @@ module Arachnid
       endpoint.fragment = nil
       endpoint.path = ""
 
-      # Set headers
-      headers = {
-        "DNT" => @do_not_track ? "1" : "0"
-      }
-
       unless @sessions.has_key?(key)
         session = @client.new(
           endpoint: endpoint,