Merge pull request #119 from digininja/updatedocker

Fixed meta data and mailto
digininja · Jun 19, 2024 · a76617a · a76617a
2 parents 8a86513 + 4597bf4
commit a76617a
Show file tree

Hide file tree

Showing 4 changed files with 34 additions and 16 deletions.
diff --git a/Gemfile.lock b/Gemfile.lock
@@ -4,24 +4,26 @@ GEM
     mime (0.4.4)
     mime-types (3.5.2)
       mime-types-data (~> 3.2015)
-    mime-types-data (3.2024.0305)
+    mime-types-data (3.2024.0604)
     mini_exiftool (2.11.0)
-    nokogiri (1.16.4-aarch64-linux)
+    nokogiri (1.16.6-aarch64-linux)
       racc (~> 1.4)
-    nokogiri (1.16.4-arm-linux)
+    nokogiri (1.16.6-arm-linux)
       racc (~> 1.4)
-    nokogiri (1.16.4-arm64-darwin)
+    nokogiri (1.16.6-arm64-darwin)
       racc (~> 1.4)
-    nokogiri (1.16.4-x86-linux)
+    nokogiri (1.16.6-x86-linux)
       racc (~> 1.4)
-    nokogiri (1.16.4-x86_64-darwin)
+    nokogiri (1.16.6-x86_64-darwin)
       racc (~> 1.4)
-    nokogiri (1.16.4-x86_64-linux)
+    nokogiri (1.16.6-x86_64-linux)
       racc (~> 1.4)
-    racc (1.7.3)
-    rexml (3.2.6)
+    racc (1.8.0)
+    rexml (3.3.0)
+      strscan
     rubyzip (2.3.2)
     spider (0.5.4)
+    strscan (3.1.0)
 
 PLATFORMS
   aarch64-linux

diff --git a/README.md b/README.md
@@ -107,6 +107,7 @@ Usage: cewl [OPTIONS] ... <url>
 
 ### Running CeWL in a Docker container
 
+
 To quickly use CeWL with Docker, you can use the official `ghcr.io/digininja/cewl` image:
 
 ```sh

diff --git a/cewl.rb b/cewl.rb
@@ -169,7 +169,7 @@ def start! #:nodoc:
 
 				y = []
 				x.select do |a_url, parsed_url|
-                    if (parsed_url.scheme == "http" or parsed_url.scheme == "https") then
+                    if (parsed_url.scheme == "mailto" or parsed_url.scheme == "http" or parsed_url.scheme == "https") then
                       y << [a_url, parsed_url] if allowable_url?(a_url, parsed_url)
                     end
 				end
@@ -440,7 +440,7 @@ def pop
 
 	# Push an item onto the tree
 	def push(value)
-		puts "Pushing #{value}" if @debug
+		puts "Adding #{value} to the tree" if @debug
 		key = value.keys.first
 		value = value.values_at(key).first
 
@@ -454,10 +454,21 @@ def push(value)
 				@children << child
 			else
 				@children.each { |node|
-					if node.data.value == key && node.data.depth<@max_depth
-						child = Tree.new(key, value, node.data.depth + 1, @debug)
-						@children << child
-					end
+                    # Ignore the max depth for mailto links.
+                    # This is not a good way to do this, but it will work for now
+                    # and we all know dirty hacks stay around forever so don't
+                    # expect this to be fixed for a while.
+                    if value =~ /^mailto:/ then
+                      if node.data.value == key then
+                          child = Tree.new(key, value, node.data.depth + 1, @debug)
+                          @children << child
+                      end
+                    else
+                      if node.data.value == key && node.data.depth<@max_depth then
+                          child = Tree.new(key, value, node.data.depth + 1, @debug)
+                          @children << child
+                      end
+                    end
 				}
 			end
 		end
@@ -760,7 +771,7 @@ def usage
 	email_outfile_file = outfile_file
 end
 
-if meta_outfile && email
+if meta_outfile && meta
 	begin
 		meta_outfile_file = File.new(meta_outfile, "w")
 	rescue

diff --git a/compose.yml b/compose.yml
@@ -0,0 +1,4 @@
+services:
+  cewl:
+    build: .
+    image: ghcr.io/digininja/cewl:latest