jekyll-sanity-checker 1.0.0 → 1.0.2

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA1:
-  metadata.gz: c87617b0a35196dbf0935ed56672d5125d96ae6c
-  data.tar.gz: cfc830023fefb829707cd486de73e6bf64fdb5df
+  metadata.gz: 51965e489751bb5c1ec0187035e7a81d34ac1f88
+  data.tar.gz: bc471f5d84895af28fa695de6bc1d705e0663d8e
 SHA512:
-  metadata.gz: 0ceee07753ade7b5da1ebec8c12c82510a3fade2b670867ce7f2ec13b47072d2ae221f46990e48e0ef3149757e02d7de2377364ec7e82fe842585e9980b93d6d
-  data.tar.gz: 147e95264502c2073444b6aa18d904905675311518c0aa8b44c8eb90a62cbed573b0927d5f64dd85be0e324fe6806d78af24f143592533524352d161358d3fb8
+  metadata.gz: 5869a4095f6d4bffaff7c06440425fca4009641d3b29ca7c8530a1513168546427db0f0a3481bfb3dd25d03ec9d423200b555f08b159dbf24c8d269f317a844a
+  data.tar.gz: a9fe702be95c54a4e2e290b4cb811b3ce42f83a68e88e679788cf70a0bdb03e28470246ba71797aedb1ebf7faa778484aedde3e74ac6764ebf5f8407eb7a7c75

data/jekyll-sanity-checker.gemspec

@@ -7,6 +7,7 @@ require 'jekyll_sanity_checker/version'
 Gem::Specification.new do |spec|
   spec.name          = 'jekyll-sanity-checker'
   spec.version       = JekyllSanityChecker::VERSION
+  spec.homepage      = 'https://github.com/fastly/jekyll-sanity-checker'
   spec.authors       = ['Edmund Huber']
   spec.email         = 'edmund@fastly.com'
   spec.license       = 'MIT'
@@ -23,6 +24,7 @@ Gem::Specification.new do |spec|
 
   spec.add_development_dependency 'bundler', '~> 1.6'
   spec.add_development_dependency 'minitest', '~> 5.0'
+  spec.add_development_dependency 'pry'
   spec.add_development_dependency 'rake'
   spec.add_development_dependency 'rubocop', '~> 0.25', '>= 0.25.0'
 end

data/lib/jekyll_sanity_checker.rb

@@ -2,10 +2,9 @@ require 'addressable/uri'
 require 'css_parser'
 require 'nokogiri'
 
-module JekyllSanityChecker
-  class CheckFailed < StandardError
-  end
+require 'jekyll_sanity_checker/errors'
 
+module JekyllSanityChecker
   def process
     super
 
@@ -23,43 +22,60 @@ module JekyllSanityChecker
     run_success = true
 
     Dir.glob("#{dest}/**/*.html") do |html|
-      run_success = run_success && JekyllSanityChecker.check_html(@config, html)
+      run_success = JekyllSanityChecker.check_html(@config, html) && run_success
     end
 
     Dir.glob("#{dest}/**/*.css") do |css|
-      run_success = run_success && JekyllSanityChecker.check_css(@config, css)
+      run_success = JekyllSanityChecker.check_css(@config, css) && run_success
     end
 
-    raise Jekyll::Errors::FatalException, 'checks failed' unless run_success
+    # If Jekyll is run with --watch , don't raise FatalException, just warn. Fixes #4 .
+    raise Jekyll::Errors::FatalException, 'checks failed' if !config.fetch('watch', false) && !run_success
+  end
+
+  def self.aggregate_errors(page, el = nil)
+    yield
+    true
+  rescue CheckFailed => e
+    if el.nil?
+      Jekyll.logger.error("problem on #{page}")
+    else
+      Jekyll.logger.error("problem on #{page} , line #{el.line}")
+      Jekyll.logger.error(el.to_s.strip)
+    end
+    Jekyll.logger.error(e.message)
+    false
   end
 
   def check_html(config, page)
     success = true
     doc = Nokogiri::HTML::Document.parse File.read(page)
+
     doc.css('a').each do |link|
-      begin
+      success = aggregate_errors(page, link) do
         # Does the link have an href atttribute?
-        raise CheckFailed, 'no href on <a>' unless link.keys.include? 'href'
+        raise LinkNoHref unless link.keys.include? 'href'
 
         # If it's a hash-only URL then that's fine.
         next if link['href'].start_with? '#'
 
-        check_resource(config, link['href'], %w(mailto), true)
-      rescue CheckFailed => e
-        Jekyll.logger.error("on #{page}, line #{link.line}, problem with '#{link}': #{e.message}")
-        success = false
-      end
+        # What's the URL of this page when it's deployed?
+        absolute_destination = File.expand_path(config['destination'])
+        from_url = File.expand_path(page)
+        raise 'terrible problem' unless from_url.slice!(absolute_destination) == absolute_destination
+
+        check_resource(config, link['href'], %w(mailto), true, from_url)
+      end && success
     end
+
     doc.css('img').each do |img|
-      begin
+      success = aggregate_errors(page, img) do
         # Does the img have a src attribute?
-        raise CheckFailed, 'no src on <img>' unless img.keys.include? 'src'
+        raise ImgNoSrc unless img.keys.include? 'src'
         check_resource(config, img['src'], %w(data))
-      rescue CheckFailed => e
-        Jekyll.logger.error("on #{page}, line #{img.line}, problem with '#{img}': #{e.message}")
-        success = false
-      end
+      end && success
     end
+
     success
   end
   module_function :check_html
@@ -72,12 +88,9 @@ module JekyllSanityChecker
       rule_set.each_declaration do |_, value, _|
         match = /url\((.*?)\)/.match value
         next if match.nil?
-        begin
+        success = aggregate_errors(stylesheet) do
           check_resource(config, match[1].tr('\'"', ''), %w(data))
-        rescue CheckFailed => e
-          Jekyll.logger.error("in #{stylesheet}, problem with selector '#{rule_set.selectors.join ' '}': #{e.message}")
-          success = false
-        end
+        end && success
       end
     end
     success
@@ -101,23 +114,23 @@ module JekyllSanityChecker
   end
   module_function :load_redirects
 
-  def self.check_resource(config, src, shortcircuit_valid_schemes = [], loose_html = false)
+  def self.check_resource(config, src, shortcircuit_valid_schemes = [], loose_html = false, from_url = nil)
     # Parse the URL,
     begin
       uri = Addressable::URI.parse(src)
     rescue Addressable::URI::InvalidURIError
-      raise CheckFailed, 'not a URI'
+      raise InvalidURL, src
     end
 
     # Maybe the scheme tells us that the URI is okay.
     return if shortcircuit_valid_schemes.include? uri.scheme
 
     # Otherwise the scheme should be http or https.
-    raise CheckFailed, 'not a valid scheme' unless [nil, 'http', 'https'].include? uri.scheme
+    raise InvalidScheme, uri.scheme unless [nil, 'http', 'https'].include? uri.scheme
 
     # Only absolute URLs please.
     this_host = config['sanity_check']['host']
-    raise CheckFailed, "fully-qualified URL back to #{this_host}" if uri.host == this_host
+    raise FullyQualifiedURL, uri if uri.host == this_host
 
     # Resolve redirects..
     if @redirects.include? uri.path
@@ -134,25 +147,32 @@ module JekyllSanityChecker
     # And if it's a local resource,
     if uri.host.nil?
       # Check that it's not a relative url.
-      raise CheckFailed, 'relative URL' unless uri.path.start_with? '/'
+      raise RelativeURL.new(uri, from_url) unless uri.path.start_with? '/'
 
       parts = URI.unescape(uri.path).split('/').reject { |s| s.empty? }
       path = ([File.expand_path(config['destination'])] + parts).join(File::SEPARATOR)
 
       if loose_html
-        if File.directory?(path)
+        if self.file?(path) && File.directory?(path)
           # We treat links to directories special: if there is a directory/index.html, it is ok.
-          raise CheckFailed, 'directory reference without an index.html' unless File.file? "#{path}/index.html"
+          raise NoIndexHTML, path unless self.file? "#{path}/index.html"
         else
           # A bla.html can be referenced as bla.
-          raise CheckFailed, "file '#{path}' does not exist" unless File.file?(path) || File.file?("#{path}.html")
+          raise FileDoesNotExist, path unless self.file?(path) || self.file?("#{path}.html")
         end
       else
         # Check that the file exists.
-        raise CheckFailed, "file '#{path}' does not exist" unless File.file? path
+        raise FileDoesNotExist, path unless self.file? path
       end
     end
   end
+
+  def self.file?(fn)
+    # This relies on a not-obvious property of File.{dirname,basename}: for
+    # files, those functions behave as you would expect. For directories,
+    # File.dirname('lib') returns '.' and File.basename('lib') returns 'lib'.
+    Dir.entries(File.dirname(fn)).include? File.basename(fn)
+  end
 end
 
 module Jekyll

data/lib/jekyll_sanity_checker/errors.rb

@@ -0,0 +1,84 @@
+module JekyllSanityChecker
+  class CheckFailed < StandardError
+    def message
+    end
+  end
+
+  class LinkNoHref < CheckFailed
+    def message
+      "A link (<a>) should always contain an 'href' attribute, even if it's only 'href=\"#\"'."
+    end
+  end
+
+  class ImgNoSrc < CheckFailed
+    def message
+      "Every <img> should have a 'src' attribute."
+    end
+  end
+
+  class InvalidURL < CheckFailed
+    def initialize(url)
+      @url = url
+    end
+
+    def message
+      "'#{@url}' is not a valid URL."
+    end
+  end
+
+  class InvalidScheme < CheckFailed
+    def initialize(scheme)
+      @scheme = scheme
+    end
+
+    def message
+      "'#{@scheme}' is not an appropriate scheme in this context."
+    end
+  end
+
+  class FullyQualifiedURL < CheckFailed
+    def initialize(uri)
+      @uri = uri
+    end
+
+    def message
+      "'#{@uri}' is unnecessarily fully-qualified. It may as well be written '#{@uri.path}'."
+    end
+  end
+
+  class NoIndexHTML < CheckFailed
+    def initialize(path)
+      @path = path
+    end
+
+    def message
+      "There is no '#{@path}/index.html', so this link won't work."
+    end
+  end
+
+  class RelativeURL < CheckFailed
+    def initialize(uri, from_url)
+      @uri = uri
+      @from_url = from_url
+    end
+
+    def message
+      if @from_url
+        did_you_mean_url = Addressable::URI.join(@from_url, @uri)
+        "'#{@uri}' is a relative URL, which are not recommended. Did you mean '#{did_you_mean_url}' ?"
+      else
+        "'#{@uri}' is a relative URL, which are not recommended."
+      end
+    end
+  end
+
+  class FileDoesNotExist < CheckFailed
+    def initialize(path)
+      @path = path
+    end
+
+    def message
+      "The file '#{@path}' does not exist."
+    end
+  end
+end

data/lib/jekyll_sanity_checker/version.rb

@@ -1,3 +1,3 @@
 module JekyllSanityChecker
-  VERSION = '1.0.0'
+  VERSION = '1.0.2'
 end

data/spec/fixtures/a-missing-href/test.html

@@ -0,0 +1 @@
+<a></a>

data/spec/fixtures/fixed-bug-spaces-in-urls/test.html

@@ -0,0 +1 @@
+<a href="/test%20with%20spaces.html"></a>

data/spec/fixtures/missing-file-with-case-sensitivity/test.html

@@ -0,0 +1 @@
+<a href="/OTHER.html"></a>

data/spec/fixtures/relative-url/a/test.html

@@ -0,0 +1 @@
+<a href="b/other.html"></a>

data/spec/lib/sanity_checker.rb

@@ -23,7 +23,7 @@ $capture_error = CaptureError.new # rubocop:disable Style/GlobalVars
 Jekyll.logger = $capture_error # rubocop:disable Style/GlobalVars
 Jekyll.logger.log_level = :error
 
-def refute_sane(page, error)
+def configure_jekyll(page)
   config_hash = {
     'source' => "spec/fixtures/#{page}",
     'destination' => 'spec/output',
@@ -31,28 +31,38 @@ def refute_sane(page, error)
       'host' => 'www.computers.com'
     }
   }
-  site = Jekyll::Site.new(Jekyll.configuration(config_hash))
+  Jekyll::Site.new(Jekyll.configuration(config_hash))
+end
+
+def refute_sane(page, error)
+  site = configure_jekyll(page)
   assert_raises Jekyll::Errors::FatalException do
     site.process
   end
   assert_match error, $capture_error.captured # rubocop:disable Style/GlobalVars
 end
 
+def assert_sane(page)
+  site = configure_jekyll(page)
+  site.process
+end
+
 describe 'Jekyll sanity checker' do
   it 'warns about fully-qualified URLs' do
-    refute_sane 'fully-qualified-url', /fully-qualified URL back to/
+    refute_sane 'fully-qualified-url', /is unnecessarily fully-qualified/
   end
 
   it 'warns about unrecognizable schemes' do
-    refute_sane 'unrecognizable-scheme', /not a valid scheme/
+    refute_sane 'unrecognizable-scheme', /not an appropriate scheme/
   end
 
   it 'warns about relative URLs' do
-    refute_sane 'relative-url', /relative URL/
+    # Jekyll logger appends a space. Don't know why.
+    refute_sane 'relative-url', "'b/other.html' is a relative URL, which are not recommended. Did you mean '/a/b/other.html' ? "
   end
 
   it 'warns about missing index.html' do
-    refute_sane 'no-index-html', /without an index\.html/
+    refute_sane 'no-index-html', /There is no.*index\.html/
   end
 
   it 'warns about missing files' do
@@ -60,11 +70,11 @@ describe 'Jekyll sanity checker' do
   end
 
   it 'warns about <img> without src' do
-    refute_sane 'img-missing-src', /no src on <img>/
+    refute_sane 'img-missing-src', /should have a 'src'/
   end
 
   it 'warns about <img> with fully-qualified URL src' do
-    refute_sane 'img-fully-qualified-url', /fully-qualified URL back to/
+    refute_sane 'img-fully-qualified-url', /is unnecessarily fully-qualified/
   end
 
   it 'warns about <img> with relative URL src' do
@@ -80,6 +90,22 @@ describe 'Jekyll sanity checker' do
   end
 
   it 'warns about malformed URLs in <a>' do
-    refute_sane 'a-invalid-uri', /not a URI/
+    refute_sane 'a-invalid-uri', /not a valid URL/
+  end
+
+  it 'warns about <a> missing href attribute' do
+    refute_sane 'a-missing-href', /should always contain/
+  end
+
+  # Because this test could only fail on case-insensitive filesystems, e.g. OSX
+  # with HFS+, if you're concerned that you might have broken related code, or
+  # you want to do a full test run, you should also run this test on one such
+  # vulnerable system.
+  it 'warns about missing files in spite of a case-insensitive filesystem' do
+    refute_sane 'missing-file-with-case-sensitivity', /does not exist/
+  end
+
+  it 'can handle spaces in links' do
+    assert_sane 'fixed-bug-spaces-in-urls'
   end
 end

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: jekyll-sanity-checker
 version: !ruby/object:Gem::Version
-  version: 1.0.0
+  version: 1.0.2
 platform: ruby
 authors:
 - Edmund Huber
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2014-12-09 00:00:00.000000000 Z
+date: 2014-12-11 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: css_parser
@@ -92,6 +92,20 @@ dependencies:
     - - ~>
       - !ruby/object:Gem::Version
         version: '5.0'
+- !ruby/object:Gem::Dependency
+  name: pry
+  requirement: !ruby/object:Gem::Requirement
+    requirements:
+    - - '>='
+      - !ruby/object:Gem::Version
+        version: '0'
+  type: :development
+  prerelease: false
+  version_requirements: !ruby/object:Gem::Requirement
+    requirements:
+    - - '>='
+      - !ruby/object:Gem::Version
+        version: '0'
 - !ruby/object:Gem::Dependency
   name: rake
   requirement: !ruby/object:Gem::Requirement
@@ -139,22 +153,28 @@ files:
 - Rakefile
 - jekyll-sanity-checker.gemspec
 - lib/jekyll_sanity_checker.rb
+- lib/jekyll_sanity_checker/errors.rb
 - lib/jekyll_sanity_checker/version.rb
 - spec/fixtures/a-invalid-uri/test.html
+- spec/fixtures/a-missing-href/test.html
 - spec/fixtures/css-url-missing-file/test.css
+- spec/fixtures/fixed-bug-spaces-in-urls/test with spaces.html
+- spec/fixtures/fixed-bug-spaces-in-urls/test.html
 - spec/fixtures/fully-qualified-url/test.html
 - spec/fixtures/img-fully-qualified-url/test.html
 - spec/fixtures/img-missing-file/test.html
 - spec/fixtures/img-missing-src/test.html
 - spec/fixtures/img-relative-url/test.html
 - spec/fixtures/img-relative-url/thing
+- spec/fixtures/missing-file-with-case-sensitivity/other.html
+- spec/fixtures/missing-file-with-case-sensitivity/test.html
 - spec/fixtures/missing-file/test.html
 - spec/fixtures/no-index-html/no-index/crumb
 - spec/fixtures/no-index-html/test.html
-- spec/fixtures/relative-url/test.html
+- spec/fixtures/relative-url/a/test.html
 - spec/fixtures/unrecognizable-scheme/test.html
 - spec/lib/sanity_checker.rb
-homepage: 
+homepage: https://github.com/fastly/jekyll-sanity-checker
 licenses:
 - MIT
 metadata: {}
@@ -181,17 +201,22 @@ summary: A Jekyll plugin that checks the generated HTML and CSS for common mista
 test_files:
 - jekyll-sanity-checker.gemspec
 - spec/fixtures/a-invalid-uri/test.html
+- spec/fixtures/a-missing-href/test.html
 - spec/fixtures/css-url-missing-file/test.css
+- spec/fixtures/fixed-bug-spaces-in-urls/test with spaces.html
+- spec/fixtures/fixed-bug-spaces-in-urls/test.html
 - spec/fixtures/fully-qualified-url/test.html
 - spec/fixtures/img-fully-qualified-url/test.html
 - spec/fixtures/img-missing-file/test.html
 - spec/fixtures/img-missing-src/test.html
 - spec/fixtures/img-relative-url/test.html
 - spec/fixtures/img-relative-url/thing
+- spec/fixtures/missing-file-with-case-sensitivity/other.html
+- spec/fixtures/missing-file-with-case-sensitivity/test.html
 - spec/fixtures/missing-file/test.html
 - spec/fixtures/no-index-html/no-index/crumb
 - spec/fixtures/no-index-html/test.html
-- spec/fixtures/relative-url/test.html
+- spec/fixtures/relative-url/a/test.html
 - spec/fixtures/unrecognizable-scheme/test.html
 - spec/lib/sanity_checker.rb
 has_rdoc: 

data/spec/fixtures/relative-url/test.html

@@ -1 +0,0 @@
-<a href="relative"></a>