jekyll-remote-theme 0.5.1 → 0.6.0

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 25fe811cc5335e5f2932bc6db5e74c83b2f7984229d745d13a6ba850396f3e44
-  data.tar.gz: 617b6d962ddea1194a7f0f801124756deada8b6a0960ba9ecabf11bdebb2ef4c
+  metadata.gz: e8b60407212cb023740ee7d4258e9a02c862bed37246db406607af4d3e6228e8
+  data.tar.gz: 448d9d30e4f0e0a2efa300c1ab7bdd8199ff82fcc4afd97264227892d6ab5fe3
 SHA512:
-  metadata.gz: f36daa0ed584f6c5c25f0006b0eab5075c35fc252e7bba19fcbd72a8b225217545c11da7d38dd00c82cbff666cb42fdce3b1d440b4e610cfbc3f74c69f127a80
-  data.tar.gz: a78d981b482904e7f673235a18c5e9a058f33fb0e0714dbeee2788abf64e60552b44b05550634538555daf0fbbd7c95350590f1a2a5d278d5a2630de1f27c234
+  metadata.gz: 36eacae972f56890144e128ec284067fa33f03f528b4f70047b0c60e4f90e80f04426990a6abb6ea87d1d6cbecac1b12effec5dcafbe89c7eb6abd9d1160c357
+  data.tar.gz: '0871deb358244f78aa294b171af43415eb0bed1367c3224554992c1723dbae092d48d0bc321c116fd0297b1e22d7847e030599c748a6fef5b5192a9f1ccdfa93'

data/lib/jekyll-remote-theme/downloader.rb

@@ -6,6 +6,9 @@ module Jekyll
       PROJECT_URL = "https://github.com/benbalter/jekyll-remote-theme"
       USER_AGENT = "Jekyll Remote Theme/#{VERSION} (+#{PROJECT_URL})"
       MAX_FILE_SIZE = 1 * (1024 * 1024 * 1024) # Size in bytes (1 GB)
+      # Limit on the theme's total size once extracted, to stop zip bombs
+      MAX_EXTRACTED_SIZE = 2 * (1024 * 1024 * 1024) # Size in bytes (2 GB)
+      EXTRACT_CHUNK_SIZE = 64 * 1024
       NET_HTTP_ERRORS = [
         Timeout::Error, Errno::EINVAL, Errno::ECONNRESET, EOFError, Net::OpenTimeout,
         Net::HTTPBadResponse, Net::HTTPHeaderSyntaxError, Net::ProtocolError,
@@ -45,23 +48,30 @@ module Jekyll
         @zip_file ||= Tempfile.new([TEMP_PREFIX, ".zip"], :binmode => true)
       end
 
-      # rubocop:disable Metrics/AbcSize
       def download
         Jekyll.logger.debug LOG_KEY, "Downloading #{zip_url} to #{zip_file.path}"
-        http_class.start(zip_url.host, zip_url.port, :use_ssl => true) do |http|
+        HTTP.start(zip_url, :use_ssl => true) do |http|
           http.request(request) do |response|
             raise_unless_success(response)
             enforce_max_file_size(response.content_length)
-            response.read_body do |chunk|
-              zip_file.write chunk
-            end
+            write_body(response)
           end
         end
         @downloaded = true
       rescue *NET_HTTP_ERRORS => e
         raise DownloadError, e.message
       end
-      # rubocop:enable Metrics/AbcSize
+
+      # Content-Length is optional (e.g. chunked responses), so also count the
+      # bytes as they arrive and stop once the limit is exceeded.
+      def write_body(response)
+        bytes = 0
+        response.read_body do |chunk|
+          bytes += chunk.bytesize
+          enforce_max_file_size(bytes)
+          zip_file.write chunk
+        end
+      end
 
       def request
         return @request if defined? @request
@@ -97,67 +107,65 @@ module Jekyll
         # File IO is already open, rewind pointer to start of file to read
         zip_file.rewind
 
+        # Extract each entry by hand rather than calling Zip::Entry#extract,
+        # whose signature changed incompatibly in rubyzip 3.0 (the destination
+        # is now resolved relative to a `destination_directory:` keyword,
+        # mangling the absolute paths we pass). Reading each entry's stream and
+        # writing it ourselves behaves identically across rubyzip 1.x–3.x.
+        @extracted_size = 0
         Zip::File.open(zip_file) do |archive|
-          archive.each { |file| file.extract path_without_name_and_ref(file.name) }
+          archive.each { |entry| extract_entry(entry) }
         end
       ensure
         zip_file.close
         zip_file.unlink
       end
 
-      # Full URL to codeload zip download endpoint for the given theme
-      def zip_url
-        @zip_url ||= Addressable::URI.new(
-          :scheme => theme.scheme,
-          :host   => "codeload.#{theme.host}",
-          :path   => [theme.owner, theme.name, "zip", theme.git_ref].join("/")
-        ).normalize
-      end
+      # Writes a single zip entry to its destination within theme.root.
+      # path_without_name_and_ref sandboxes the destination, guarding against
+      # Zip Slip.
+      def extract_entry(entry)
+        return if entry.name.end_with?("/") # skip directory entries
 
-      # Returns an HTTP class that respects proxy environment variables
-      def http_class
-        @http_class ||= Net::HTTP::Proxy(proxy_host, proxy_port, proxy_user, proxy_pass)
-      end
-
-      # Extracts proxy settings from environment variables
-      def proxy_uri
-        return @proxy_uri if defined?(@proxy_uri)
-
-        proxy_env = find_proxy_env_var
-        @proxy_uri = parse_proxy_uri(proxy_env)
-      end
-
-      def find_proxy_env_var
-        # Check for HTTPS proxy first if the URL uses HTTPS
-        if theme.scheme == "https"
-          ENV["https_proxy"] || ENV["HTTPS_PROXY"] || ENV["http_proxy"] || ENV["HTTP_PROXY"]
-        else
-          ENV["http_proxy"] || ENV["HTTP_PROXY"]
+        enforce_max_extracted_size(@extracted_size + entry.size)
+        dest = path_without_name_and_ref(entry.name)
+        FileUtils.mkdir_p File.dirname(dest)
+        File.open(dest, "wb") do |output|
+          entry.get_input_stream { |input| copy_entry(entry, input, output) }
         end
       end
 
-      def parse_proxy_uri(proxy_env)
-        return nil unless proxy_env
-
-        Addressable::URI.parse(proxy_env)
-      rescue Addressable::URI::InvalidURIError
-        nil
-      end
+      # Copies an entry in chunks, counting the bytes actually inflated rather
+      # than trusting the sizes in the archive. Zip.validate_entry_sizes only
+      # applies to Zip::Entry#extract, so the same check is made here: an entry
+      # may not inflate to more than its declared size.
+      def copy_entry(entry, input, output)
+        written = 0
+        while (chunk = input.read(EXTRACT_CHUNK_SIZE))
+          written += chunk.bytesize
+          if written > entry.size
+            raise DownloadError, "Zip entry #{entry.name} is larger than its declared size"
+          end
 
-      def proxy_host
-        proxy_uri&.host
+          @extracted_size += chunk.bytesize
+          enforce_max_extracted_size(@extracted_size)
+          output.write(chunk)
+        end
       end
 
-      def proxy_port
-        proxy_uri&.port
-      end
+      def enforce_max_extracted_size(size)
+        return unless size > MAX_EXTRACTED_SIZE
 
-      def proxy_user
-        proxy_uri&.user
+        raise DownloadError, "Maximum extracted theme size of #{MAX_EXTRACTED_SIZE} bytes exceeded"
       end
 
-      def proxy_pass
-        proxy_uri&.password
+      # Full URL to codeload zip download endpoint for the given theme
+      def zip_url
+        @zip_url ||= Addressable::URI.new(
+          :scheme => theme.scheme,
+          :host   => "codeload.#{theme.host}",
+          :path   => [theme.owner, theme.name, "zip", theme.git_ref].join("/")
+        ).normalize
       end
 
       def theme_dir_exists?
@@ -165,7 +173,7 @@ module Jekyll
       end
 
       def theme_dir_empty?
-        Dir["#{theme.root}/*"].empty?
+        Dir.empty?(theme.root)
       end
 
       # Codeload generated zip files contain a top level folder in the form of

data/lib/jekyll-remote-theme/http.rb

@@ -0,0 +1,48 @@
+# frozen_string_literal: true
+
+module Jekyll
+  module RemoteTheme
+    # Net::HTTP setup shared by the theme download and the GitHub API lookup
+    # for @latest, so both use the same proxy settings and timeouts.
+    module HTTP
+      OPEN_TIMEOUT = 10 # seconds
+      READ_TIMEOUT = 60 # seconds
+      # Proxy environment variables to try, by request scheme
+      PROXY_SCHEMES = {
+        "https" => %w(https http).freeze,
+        "http"  => %w(http).freeze,
+      }.freeze
+
+      module_function
+
+      # Opens a connection to the host of `uri` (an Addressable::URI) and
+      # yields the Net::HTTP session.
+      def start(uri, use_ssl: uri.scheme == "https", env: ENV, &block)
+        proxy = proxy_uri(uri, env)
+        Net::HTTP.start(
+          uri.host, uri.port,
+          proxy&.host, proxy&.port, proxy&.user, proxy&.password,
+          :use_ssl      => use_ssl,
+          :open_timeout => OPEN_TIMEOUT,
+          :read_timeout => READ_TIMEOUT,
+          &block
+        )
+      end
+
+      # The proxy to use for `uri`, from the environment, or nil. HTTPS
+      # requests use https_proxy, falling back to http_proxy. Hosts listed in
+      # no_proxy/NO_PROXY, and loopback addresses, bypass the proxy.
+      def proxy_uri(uri, env = ENV)
+        PROXY_SCHEMES.fetch(uri.scheme, PROXY_SCHEMES["http"]).each do |scheme|
+          target = URI::Generic.build(:scheme => scheme, :host => uri.host,
+                                      :port   => uri.inferred_port)
+          proxy = target.find_proxy(env)
+          return proxy if proxy
+        end
+        nil
+      rescue URI::Error
+        nil
+      end
+    end
+  end
+end

data/lib/jekyll-remote-theme/munger.rb

@@ -20,11 +20,13 @@ module Jekyll
         end
 
         Jekyll.logger.info LOG_KEY, "Using theme #{theme.name_with_owner}"
+        # after_reset fires on every rebuild (e.g. `jekyll serve`). Only the
+        # first run downloads the theme, so only it registers the cleanup.
         unless munged?
           downloader.run
           configure_theme
+          enqueue_theme_cleanup
         end
-        enqueue_theme_cleanup
 
         theme
       end
@@ -35,8 +37,10 @@ module Jekyll
         site.theme&.is_a?(Jekyll::RemoteTheme::Theme)
       end
 
+      # On a rebuild, reuse the theme configured by the first run rather than
+      # building a new one, which would create another temp directory.
       def theme
-        @theme ||= Theme.new(raw_theme)
+        @theme ||= munged? ? site.theme : Theme.new(raw_theme)
       end
 
       def raw_theme
@@ -63,13 +67,22 @@ module Jekyll
       end
 
       # Initialize GitHub metadata munger if it was loaded after :after_init hook
+      # or if the existing munger needs to be updated for the current site
       def initialize_github_metadata
         return unless defined?(Jekyll::GitHubMetadata::SiteGitHubMunger)
-        return if Jekyll::GitHubMetadata::SiteGitHubMunger.global_munger
 
-        munger = Jekyll::GitHubMetadata::SiteGitHubMunger.new(site)
-        munger.munge!
-        Jekyll::GitHubMetadata::SiteGitHubMunger.global_munger = munger
+        # Create a new munger if one doesn't exist or if it's for a different site
+        current_munger = Jekyll::GitHubMetadata::SiteGitHubMunger.global_munger
+        needs_init = current_munger.nil? ||
+          (Jekyll::GitHubMetadata.respond_to?(:site) &&
+            Jekyll::GitHubMetadata.site != site)
+
+        if needs_init
+          Jekyll.logger.debug LOG_KEY, "Initializing GitHub metadata munger"
+          munger = Jekyll::GitHubMetadata::SiteGitHubMunger.new(site)
+          munger.munge!
+          Jekyll::GitHubMetadata::SiteGitHubMunger.global_munger = munger
+        end
       end
 
       def enqueue_theme_cleanup

data/lib/jekyll-remote-theme/theme.rb

@@ -171,11 +171,7 @@ module Jekyll
       end
 
       def make_api_request(api_url)
-        Net::HTTP.start(
-          api_url.host,
-          api_url.port,
-          :use_ssl => api_url.scheme == "https"
-        ) do |http|
+        HTTP.start(api_url) do |http|
           request = Net::HTTP::Get.new(api_url.request_uri)
           request["Accept"] = "application/vnd.github.v3+json"
           request["User-Agent"] = Downloader::USER_AGENT

data/lib/jekyll-remote-theme/version.rb

@@ -2,6 +2,6 @@
 
 module Jekyll
   module RemoteTheme
-    VERSION = "0.5.1"
+    VERSION = "0.6.0"
   end
 end

data/lib/jekyll-remote-theme.rb

@@ -16,6 +16,7 @@ module Jekyll
     class DownloadError < StandardError; end
 
     autoload :Downloader,  "jekyll-remote-theme/downloader"
+    autoload :HTTP,        "jekyll-remote-theme/http"
     autoload :MockGemspec, "jekyll-remote-theme/mock_gemspec"
     autoload :Munger,      "jekyll-remote-theme/munger"
     autoload :Theme,       "jekyll-remote-theme/theme"

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: jekyll-remote-theme
 version: !ruby/object:Gem::Version
-  version: 0.5.1
+  version: 0.6.0
 platform: ruby
 authors:
 - Ben Balter
@@ -92,7 +92,7 @@ dependencies:
         version: 1.3.0
     - - "<"
       - !ruby/object:Gem::Version
-        version: '3.0'
+        version: '4.0'
   type: :runtime
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
@@ -102,7 +102,7 @@ dependencies:
         version: 1.3.0
     - - "<"
       - !ruby/object:Gem::Version
-        version: '3.0'
+        version: '4.0'
 - !ruby/object:Gem::Dependency
   name: jekyll_test_plugin_malicious
   requirement: !ruby/object:Gem::Requirement
@@ -279,6 +279,7 @@ extra_rdoc_files: []
 files:
 - lib/jekyll-remote-theme.rb
 - lib/jekyll-remote-theme/downloader.rb
+- lib/jekyll-remote-theme/http.rb
 - lib/jekyll-remote-theme/mock_gemspec.rb
 - lib/jekyll-remote-theme/munger.rb
 - lib/jekyll-remote-theme/theme.rb
@@ -294,14 +295,14 @@ required_ruby_version: !ruby/object:Gem::Requirement
   requirements:
   - - ">="
     - !ruby/object:Gem::Version
-      version: 2.7.0
+      version: '3.0'
 required_rubygems_version: !ruby/object:Gem::Requirement
   requirements:
   - - ">="
     - !ruby/object:Gem::Version
       version: '0'
 requirements: []
-rubygems_version: 4.0.10
+rubygems_version: 3.6.9
 specification_version: 4
 summary: Jekyll plugin for building Jekyll sites with any GitHub-hosted theme
 test_files: []