jekyll: use git to fetch remote resources

Signed-off-by: CrazyMax <crazy-max@users.noreply.github.com>

Dockerfile

@@ -47,6 +47,7 @@ ARG JEKYLL_ENV
 ARG DOCS_URL
 ENV TARGET=/out
 RUN --mount=type=bind,target=.,rw \
+    --mount=type=cache,target=/tmp/docker-docs-clone \
     --mount=type=cache,target=/src/.jekyll-cache <<EOT
   set -eu
   CONFIG_FILES=_config.yml$([ "$JEKYLL_ENV" = "production" ] && echo ",_config_production.yml" || true)

Gemfile

@@ -16,8 +16,8 @@ end
 #  more info: https://github.com/docker/docs/issues/14788
 gem 'rouge', '3.27.0'
 
-gem 'archive-zip', '0.12.0'
 gem 'front_matter_parser', '1.0.1'
+gem 'git', '1.12.0'
 gem 'html-proofer', '3.19.4'
 gem 'mdl', '0.11.0'
 gem 'octopress-hooks', '2.6.2'

Gemfile.lock

@@ -1,11 +1,9 @@
 GEM
   remote: https://rubygems.org/
   specs:
-    addressable (2.8.0)
-      public_suffix (>= 2.0.2, < 5.0)
-    archive-zip (0.12.0)
-      io-like (~> 0.3.0)
-    chef-utils (17.10.0)
+    addressable (2.8.1)
+      public_suffix (>= 2.0.2, < 6.0)
+    chef-utils (18.0.161)
       concurrent-ruby
     colorator (1.1.0)
     concurrent-ruby (1.1.10)
@@ -18,6 +16,9 @@ GEM
     ffi (1.15.5)
     forwardable-extended (2.6.0)
     front_matter_parser (1.0.1)
+    git (1.12.0)
+      addressable (~> 2.8)
+      rchardet (~> 1.8)
     html-proofer (3.19.4)
       addressable (~> 2.3)
       mercenary (~> 0.3)
@@ -29,7 +30,6 @@ GEM
     http_parser.rb (0.8.0)
     i18n (1.12.0)
       concurrent-ruby (~> 1.0)
-    io-like (0.3.1)
     jekyll (4.2.2)
       addressable (~> 2.4)
       colorator (~> 1.0)
@@ -92,13 +92,14 @@ GEM
     pathutil (0.16.2)
       forwardable-extended (~> 2.6)
     posix-spawn (0.3.15)
-    public_suffix (4.0.7)
+    public_suffix (5.0.0)
     racc (1.6.0)
     rainbow (3.1.1)
     rake (13.0.6)
-    rb-fsevent (0.11.1)
+    rb-fsevent (0.11.2)
     rb-inotify (0.10.1)
       ffi (~> 1.0)
+    rchardet (1.8.0)
     rexml (3.2.5)
     rouge (3.27.0)
     safe_yaml (1.0.5)
@@ -118,8 +119,8 @@ PLATFORMS
   x86_64-linux
 
 DEPENDENCIES
-  archive-zip (= 0.12.0)
   front_matter_parser (= 1.0.1)
+  git (= 1.12.0)
   html-proofer (= 3.19.4)
   jekyll (= 4.2.2)
   jekyll-last-modified-at

_plugins/fetch_remote.rb

@@ -1,9 +1,8 @@
-require 'archive/zip'
 require 'front_matter_parser'
+require 'git'
 require 'jekyll'
 require 'json'
 require 'octopress-hooks'
-require 'open-uri'
 require 'rake'
 
 require_relative 'util.rb'
@@ -12,14 +11,6 @@ module Jekyll
   class FetchRemote < Octopress::Hooks::Site
     priority :highest
 
-    def self.download(url, dest)
-      uri = URI.parse(url)
-      result = File.join(dest, File.basename(uri.path))
-      puts "    Downloading #{url}"
-      IO.copy_stream(URI.open(url), result)
-      return result
-    end
-
     def self.copy(src, dest)
       if (tmp = Array.try_convert(src))
         tmp.each do |s|
@@ -50,16 +41,29 @@ module Jekyll
       beginning_time = Time.now
       puts "Starting plugin fetch_remote.rb..."
       site.config['fetch-remote'].each do |entry|
-        puts "  Repo #{entry['repo']} (#{entry['ref']})"
-        Dir.mktmpdir do |tmpdir|
-          tmpfile = FetchRemote.download("#{entry['repo']}/archive/#{entry['ref']}.zip", tmpdir)
-          Dir.mktmpdir do |ztmpdir|
-            puts "    Extracting #{tmpfile}"
-            Archive::Zip.extract(
-              tmpfile,
-              ztmpdir,
-              :create => true
-            )
+        puts "  Repo #{entry['repo']}"
+
+        gituri = Git::URL.parse(entry['repo'])
+        clonedir = "#{Dir.tmpdir}/docker-docs-clone#{gituri.path}"
+        if Dir.exist?(clonedir)
+          puts "    Opening #{clonedir}"
+          begin
+            git = Git.open(clonedir)
+            git.chdir do
+              puts "    Fetching #{entry['ref']}"
+              git.checkout(entry['ref'])
+              git.fetch
+            end
+          rescue => e
+            FileUtils.rm_rf(clonedir)
+            puts "    Cloning repository into #{clonedir}"
+            Git.clone("#{entry['repo']}.git", Pathname.new(clonedir), branch: entry['ref'])
+          end
+        else
+          puts "    Cloning repository into #{clonedir}"
+          Git.clone("#{entry['repo']}.git", Pathname.new(clonedir), branch: entry['ref'])
+        end
+
         entry['paths'].each do |path|
           if File.extname(path['dest']) != ""
             if path['src'].size > 1
@@ -76,9 +80,9 @@ module Jekyll
           files = FileList[]
           path['src'].each do |src|
             if "#{src}".start_with?("!")
-                  files.exclude(File.join(ztmpdir, "*/"+"#{src}".delete_prefix("!")))
+              files.exclude(File.join(clonedir, "/"+"#{src}".delete_prefix("!")))
             else
-                  files.include(File.join(ztmpdir, "*/#{src}"))
+              files.include(File.join(clonedir, "/#{src}"))
             end
           end
 
@@ -87,7 +91,7 @@ module Jekyll
               s = File.realpath(s)
               # traverse source directory
               FileUtils::Entry_.new(s, nil, false).wrap_traverse(proc do |ent|
-                    file_clean = ent.path.delete_prefix(ztmpdir).split("/").drop(2).join("/")
+                file_clean = ent.path.delete_prefix(clonedir).split("/").drop(1).join("/")
                 destent = FileUtils::Entry_.new(d, ent.rel, false)
                 puts "      #{file_clean} => #{destent.path}"
 
@@ -121,8 +125,6 @@ module Jekyll
           end
         end
       end
-        end
-      end
 
       end_time = Time.now
       puts "done in #{(end_time - beginning_time)} seconds"