nanoc-checking 1.0.2 → 1.0.4

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 9d5e37cf90a631812173adcadedb4d8bf9b9d93926b04f9ba25c524ab13bb2dc
-  data.tar.gz: d1e2b8cd49d3b280ec203be153938ba3da0794ea5a57e3cb6ca3ff867bef2e01
+  metadata.gz: b380e10a13a38c13d712399af8308a0cb67b533094287649eb30283fceba9be3
+  data.tar.gz: bd6715348e5527f23bedf1812b56bc49d75e5827b5bb318a5165bf5cf689077f
 SHA512:
-  metadata.gz: b9e144e2d1dc8565072f2d117416b4ba3f89635dfe21b1c1e1e18ad238bcd4f6baf7a285b37c1568c8f514ffb1b018cff6e97c6ea69093e512df2d65ba4932ff
-  data.tar.gz: d196ee9d516d77824de90dc79226cc46f394c0ad32cb897005c684fb2112dbb751eeafae9538f3b9146d8fa89c9fc648aa2909a02f3d9eae2b53fbc4bcf48876
+  metadata.gz: 61e31cc485a14e7a75199c65df33d172dd69caad7d00be40725ec55ca5203f052213846b974dd842be2a33d021d10be184a1266810f9fdcddf89db1a7def572d
+  data.tar.gz: 2472f804af1d3facea1b840d89280070042b27fc03d91475b91db5e41a6783261a84ac62f83d6ea25811112ee85ff045e5a9110bc8ca398f08b4650dcfbc65ae

data/NEWS.md

@@ -1,5 +1,17 @@
 # nanoc-checking news
 
+## 1.0.4 (2024-04-19)
+
+Enhancements:
+
+- Added support for `javascript:` pseudo-URLs (#1698)
+
+## 1.0.3 (2024-03-15)
+
+Fixes:
+
+- Restore compatibility with Nanoc 4.12.20
+
 ## 1.0.2 (2022-01-15)
 
 Fixes:

data/README.md

@@ -1,5 +1,5 @@
 # nanoc-checking
 
-This provides the `check` command and associated functionality for [Nanoc](https://nanoc.ws).
+This provides the `check` command and associated functionality for [Nanoc](https://nanoc.app).
 
-For details, see the [Checking correctness of Nanoc sites](https://nanoc.ws/doc/testing/) chapter of the Nanoc documentation.
+For details, see the [Checking correctness of Nanoc sites](https://nanoc.app/doc/testing/) chapter of the Nanoc documentation.

data/lib/nanoc/checking/checks/external_links.rb

@@ -13,7 +13,7 @@ module Nanoc
           # Find all broken external hrefs
           # TODO: de-duplicate this (duplicated in internal links check)
           filenames = output_html_filenames.reject { |f| excluded_file?(f) }
-          hrefs_with_filenames = ::Nanoc::Extra::LinkCollector.new(filenames, :external).filenames_per_href
+          hrefs_with_filenames = ::Nanoc::Checking::LinkCollector.new(filenames, :external).filenames_per_href
           results = select_invalid(hrefs_with_filenames.keys.shuffle)
 
           # Report them
@@ -21,7 +21,7 @@ module Nanoc
             filenames = hrefs_with_filenames[res.href]
             filenames.each do |filename|
               add_issue(
-                "broken reference to #{res.href}: #{res.explanation}",
+                "broken reference to <#{res.href}>: #{res.explanation}",
                 subject: filename,
               )
             end
@@ -43,6 +43,13 @@ module Nanoc
         end
 
         def validate(href)
+          # Skip javascript: URLs
+          #
+          # This needs to be handled explicitly, because URI.parse does not
+          # like `javascript:` URLs -- presumably because those are not
+          # technically valid URLs.
+          return nil if href.start_with?('javascript:')
+
           # Parse
           url = nil
           begin
@@ -71,7 +78,8 @@ module Nanoc
               next
             end
 
-            if /^3..$/.match?(res.code)
+            case res.code
+            when /^3..$/
               if i == 4
                 return Result.new(href, 'too many redirects')
               end
@@ -87,7 +95,7 @@ module Nanoc
               end
 
               url = URI.parse(location)
-            elsif res.code == '200'
+            when '200'
               return nil
             else
               return Result.new(href, res.code)

data/lib/nanoc/checking/checks/internal_links.rb

@@ -18,14 +18,14 @@ module Nanoc
         def run
           # TODO: de-duplicate this (duplicated in external links check)
           filenames = output_html_filenames
-          uris = ::Nanoc::Extra::LinkCollector.new(filenames, :internal).filenames_per_href
+          uris = ::Nanoc::Checking::LinkCollector.new(filenames, :internal).filenames_per_href
 
           uris.each_pair do |href, fns|
             fns.each do |filename|
               next if valid?(href, filename)
 
               add_issue(
-                "broken reference to #{href}",
+                "broken reference to <#{href}>",
                 subject: filename,
               )
             end
@@ -89,7 +89,7 @@ module Nanoc
           # FIXME: do not depend on current working directory
           origin = File.absolute_path(origin)
 
-          relative_origin = origin[@config.output_dir.size..-1]
+          relative_origin = origin[@config.output_dir.size..]
           excludes = config.fetch(:exclude_origins, [])
           excludes.any? { |pattern| Regexp.new(pattern).match(relative_origin) }
         end

data/lib/nanoc/checking/checks/mixed_content.rb

@@ -14,7 +14,7 @@ module Nanoc
 
         def run
           filenames = output_html_filenames
-          resource_uris_with_filenames = ::Nanoc::Extra::LinkCollector.new(filenames).filenames_per_resource_uri
+          resource_uris_with_filenames = ::Nanoc::Checking::LinkCollector.new(filenames).filenames_per_resource_uri
 
           resource_uris_with_filenames.each_pair do |uri, fns|
             next unless guaranteed_insecure?(uri)

data/lib/nanoc/checking/link_collector.rb

@@ -0,0 +1,130 @@
+# frozen_string_literal: true
+
+module ::Nanoc
+  module Checking
+    class LinkCollector
+      # HTML5 element attributes
+      URI_ATTRS = {
+        'a' => %i[href ping],
+        'area' => %i[href ping],
+        'audio' => %i[src],
+        'base' => %i[href],
+        'blockquote' => %i[cite],
+        'form' => %i[action],
+        'iframe' => %i[src],
+        'img' => %i[src srcset],
+        'link' => %i[href],
+        'object' => %i[data],
+        'script' => %i[src],
+        'source' => %i[src srcset],
+        'video' => %i[poster src],
+      }.freeze
+      # HTML+RDFa global URI attributes
+      GLOBAL_ATTRS = %i[about resource].freeze
+
+      def initialize(filenames, mode = nil)
+        @filenames = filenames
+        @filter =
+          case mode
+          when nil
+            ->(_h) { true }
+          when :external
+            ->(h) { external_href?(h) }
+          when :internal
+            ->(h) { internal_href?(h) }
+          else
+            raise ArgumentError, 'Expected mode argument to be :internal, :external or nil'
+          end
+      end
+
+      def filenames_per_href
+        grouped_filenames { |filename| hrefs_in_file(filename) }
+      end
+
+      def filenames_per_resource_uri
+        grouped_filenames { |filename| resource_uris_in_file(filename) }
+      end
+
+      def external_href?(href)
+        return false if internal_href?(href)
+
+        href =~ %r{^(//|[a-z-]+:)}
+      end
+
+      def internal_href?(href)
+        return false if href.nil?
+
+        href.start_with?('file:/')
+      end
+
+      # all links
+      def hrefs_in_file(filename)
+        uris_in_file filename, nil
+      end
+
+      # embedded resources, used by the mixed-content checker
+      def resource_uris_in_file(filename)
+        uris_in_file filename, %w[audio base form iframe img link object script source video]
+      end
+
+      private
+
+      def grouped_filenames
+        require 'nokogiri'
+        grouped_filenames = {}
+        @filenames.each do |filename|
+          yield(filename).each do |resouce_uri|
+            grouped_filenames[resouce_uri] ||= Set.new
+            grouped_filenames[resouce_uri] << filename
+          end
+        end
+        grouped_filenames
+      end
+
+      def uris_in_file(filename, tag_names)
+        uris = Set.new
+        base_uri = URI("file://#{filename}")
+        doc = Nokogiri::HTML(::File.read(filename))
+        doc.traverse do |tag|
+          next unless tag_names.nil? || tag_names.include?(tag.name)
+
+          attrs = []
+          attrs += URI_ATTRS[tag.name] unless URI_ATTRS[tag.name].nil?
+          attrs += GLOBAL_ATTRS if tag_names.nil?
+          next if attrs.nil?
+
+          attrs.each do |attr_name|
+            next if tag[attr_name].nil?
+
+            if attr_name == :srcset
+              uris = uris.merge(tag[attr_name].split(',').map { |v| v.strip.split[0].strip }.compact)
+            elsif %i[about ping resource].include?(attr_name)
+              uris = uris.merge(tag[attr_name].split.map(&:strip).compact)
+            else
+              uris << tag[attr_name.to_s]
+            end
+          end
+        end
+
+        # Strip fragment
+        uris.map! { |uri| uri.gsub(/#.*$/, '') }
+
+        # Resolve paths relative to the filename, return invalid URIs as-is
+        uris.map! do |uri|
+          if uri.start_with?('//')
+            # Don’t modify protocol-relative URLs. They’re absolute!
+            uri
+          else
+            begin
+              URI.join(base_uri, uri).to_s
+            rescue
+              uri
+            end
+          end
+        end
+
+        uris.select(&@filter)
+      end
+    end
+  end
+end

data/lib/nanoc/checking/version.rb

@@ -2,6 +2,6 @@
 
 module Nanoc
   module Checking
-    VERSION = '1.0.2'
+    VERSION = '1.0.4'
   end
 end

data/lib/nanoc/checking.rb

@@ -13,6 +13,7 @@ require_relative 'checking/check'
 require_relative 'checking/checks'
 require_relative 'checking/command_runners'
 require_relative 'checking/dsl'
+require_relative 'checking/link_collector'
 require_relative 'checking/runner'
 require_relative 'checking/loader'
 require_relative 'checking/issue'

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: nanoc-checking
 version: !ruby/object:Gem::Version
-  version: 1.0.2
+  version: 1.0.4
 platform: ruby
 authors:
 - Denis Defreyne
 autorequire:
 bindir: bin
 cert_chain: []
-date: 2022-01-15 00:00:00.000000000 Z
+date: 2024-04-19 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: nanoc-cli
@@ -19,7 +19,7 @@ dependencies:
         version: '4.12'
     - - ">="
       - !ruby/object:Gem::Version
-        version: 4.12.4
+        version: 4.12.5
   type: :runtime
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
@@ -29,7 +29,7 @@ dependencies:
         version: '4.12'
     - - ">="
       - !ruby/object:Gem::Version
-        version: 4.12.4
+        version: 4.12.5
 - !ruby/object:Gem::Dependency
   name: nanoc-core
   requirement: !ruby/object:Gem::Requirement
@@ -39,7 +39,7 @@ dependencies:
         version: '4.12'
     - - ">="
       - !ruby/object:Gem::Version
-        version: 4.12.4
+        version: 4.12.5
   type: :runtime
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
@@ -49,7 +49,7 @@ dependencies:
         version: '4.12'
     - - ">="
       - !ruby/object:Gem::Version
-        version: 4.12.4
+        version: 4.12.5
 description: Provides checking functionality for Nanoc
 email: denis+rubygems@denis.ws
 executables: []
@@ -74,14 +74,16 @@ files:
 - lib/nanoc/checking/commands/check.rb
 - lib/nanoc/checking/dsl.rb
 - lib/nanoc/checking/issue.rb
+- lib/nanoc/checking/link_collector.rb
 - lib/nanoc/checking/loader.rb
 - lib/nanoc/checking/runner.rb
 - lib/nanoc/checking/version.rb
-homepage: https://nanoc.ws/
+homepage: https://nanoc.app/
 licenses:
 - MIT
 metadata:
   rubygems_mfa_required: 'true'
+  source_code_uri: https://github.com/nanoc/nanoc/tree/nanoc-checking-v1.0.4/nanoc-checking
 post_install_message:
 rdoc_options: []
 require_paths:
@@ -90,14 +92,14 @@ required_ruby_version: !ruby/object:Gem::Requirement
   requirements:
   - - ">="
     - !ruby/object:Gem::Version
-      version: '2.6'
+      version: '2.7'
 required_rubygems_version: !ruby/object:Gem::Requirement
   requirements:
   - - ">="
     - !ruby/object:Gem::Version
       version: '0'
 requirements: []
-rubygems_version: 3.3.4
+rubygems_version: 3.5.9
 signing_key:
 specification_version: 4
 summary: Checking support for Nanoc