parslet-css 0.0.1

data/lib/parslet-css.rb

@@ -0,0 +1,19 @@
+# encoding: UTF-8
+require 'parslet'
+
+class ParsletCSS
+  VERSION = "0.0.1"
+
+  autoload :Parser, 'parslet-css/parser'
+
+  def self.compile(str)
+    parser = ParsletCSS::Parser.new
+    parser.parse(str)
+  rescue Parslet::ParseFailed => error
+    puts parser.root.error_tree
+  end
+end
+
+if __FILE__ == $0
+  ParsletCSS.compile(ARGF.read)
+end

data/lib/parslet-css/parser.rb

@@ -0,0 +1,153 @@
+# encoding: UTF-8
+class ParsletCSS::Parser < Parslet::Parser
+  rule(:selectors) {
+    ignore >> selector_chars >>
+    ((space? >> comma >> space? | space) >> selector_chars).repeat
+  }
+  rule(:selector_chars) {
+    match('[a-zA-Z0-9\-\._~\/*"$^:=#>+|\[\]]').repeat(1) >>
+    (str('(') >> match['a-zA-Z0-9\-+:*\[\]'].repeat(1) >> str(')')).maybe
+  }
+
+  rule(:lcurly) { ignore >> str('{') >> ignore }
+  rule(:declarations) { (ignore >> declaration >> semicolon? >> ignore).repeat }
+  rule(:declaration) {
+    name >> ignore >> str(':') >> ignore >> property_values
+  }
+
+  rule(:property_values) {
+    property_value >> (semicolon.absent? >> space >> property_value).repeat(0, 5)
+  }
+  # TODO: be more precise for margin, padding... property values
+  rule(:property_value) {
+    font_height | property_value_keywords | uri | percent | color | size |
+    float | font_family_list
+  }
+  rule(:property_value_keywords) {
+    str('no-repeat') | str('scroll') | str('inherit') |
+    str('baseline') | str('block') | str('both') | str('bold') |
+    str('inline-block') | str('!important') | str('inline')
+  }
+
+  # URL
+  # http://labs.apache.org/webarch/uri/rev-2002/rfc2396bis.html#collected-abnf
+  rule(:uri) { str('url(') >> quote? >> url.as(:url) >> quote? >> str(')') }
+  rule(:url) { protocol? >> domain? >> path }
+  rule(:protocol) {
+    (str('https') | str('http') | str('ftp') | str('gopher') | str('mailto') |
+     str('news') | str('nntp') | str('telnet') | str('wais') | str('file') |
+     str('prospero')) >> str('://')
+  }
+  rule(:protocol?) { protocol.maybe }
+  rule(:domain) { match('[a-zA-Z0-9\-\._\/]').repeat >> (str(':') >> integer).maybe }
+  rule(:domain?) { domain.maybe }
+  rule(:path) { path_chars >> query? }
+  rule(:path_chars) { (str('/').maybe >> match('[a-zA-Z0-9\-\._~]')).repeat }
+  rule(:query) {
+    str('?') >> (query_key_value >> (query_sep >> query_key_value).repeat).maybe
+  }
+  rule(:query_sep) { str('&') | semicolon }
+  rule(:query?) { query.maybe }
+  rule(:query_key_value) { name >> str('=') >> name }
+
+  rule(:percent) { sign? >> integer >> str('%') }
+  rule(:size) { sign? >> (float | integer) >> length_unit }
+  rule(:length_unit) {
+    str('em') | str('ex') | str('px') | str('in') |
+    str('cm') | str('mm') | str('pt') | str('pc')
+  }
+
+  # http://www.w3.org/TR/2002/WD-css3-fonts-20020802/#font-family-prop
+  rule(:font_family_list) {
+    font_family >> ((comma >> space? >> font_family).repeat).maybe
+  }
+  # Font family names containing whitespace [link to syntax module] should be quoted.
+  rule(:font_family) {
+    quoted | name
+  }
+  rule(:font_height) {
+    size >> str('/') >> size
+  }
+
+  # COLORS
+  rule(:color) { color_keywords | hex_value | rgb }
+  rule(:color_keywords) {
+    str('aqua') | str('black') | str('blue') | str('fuchsia') | str('gray') |
+    str('green') | str('lime') | str('maroon') | str('navy') | str('olive') |
+    str('purple') | str('red') | str('silver') | str('teal') | str('white') |
+    str('yellow') | str('transparent')
+  }
+  rule(:hex_value) {
+    str('#') >> (digit_hex.repeat(6,6) | digit_hex.repeat(3,3))
+  }
+  rule(:digit_hex) { match('[0-9a-fA-F]') }
+  rule(:rgb) {
+    (str('rgb(') | str('rgba(') | str('hsl(')) >>
+    space? >> rgb_value >> space? >> comma >>
+    space? >> rgb_value >> space? >> comma >>
+    (space? >> rgb_value >> space? >> comma).maybe >>
+    space? >> rgb_value >> space? >> str(')')
+  }
+  rule(:rgb_value) { percent | float | integer }
+
+  # TODO: http://www.w3.org/TR/css3-2d-transforms/
+
+  rule(:rcurly) { ignore >> str('}') >> ignore }
+
+  rule(:space) { match('\s').repeat(1) }
+  rule(:space?) { space.maybe }
+  rule(:quote) { str('"') | str("'") }
+  rule(:quote?) { quote.maybe }
+  rule(:name) { match['_a-zA-Z0-9-'].repeat(1) }
+  rule(:quoted) { quote >> (quote.absent? >> any).repeat >> quote }
+  rule(:integer) { match('[0-9]').repeat }
+  rule(:float) { integer >> str('.') >> integer }
+  rule(:semicolon) { str(';') }
+  rule(:semicolon?) { semicolon.maybe }
+  rule(:comma) { str(',') }
+  rule(:sign) { str('+') | str('-') }
+  rule(:sign?) { sign.maybe }
+  rule(:comment) { space? >> (str('/*') >> (str('*/').absent? >> any).repeat >> str('*/')) >> space? }
+  rule(:ignore) { comment | space? }
+
+  # @charset
+  # http://www.w3.org/TR/CSS21/syndata.html#charset
+  # http://www.iana.org/assignments/character-sets
+  @@charsets_file = File.join(File.dirname(__FILE__), '..', '..', 'data', 'iana_character_sets.txt')
+  @@charsets = open(@@charsets_file).read.split
+  def charsets
+    Parslet::Atoms::Alternative.new(*@@charsets.map {|c| str(c)})
+  end
+  rule(:charset) {
+    str('@charset') >> space >> quote >>
+    charsets >>
+    quote >> semicolon >> ignore
+  }
+
+  # @import
+  # http://www.w3.org/TR/CSS2/cascade.html#at-import
+  rule(:import) {
+    str('@import') >> space >> (quote >> name >> str('.css') >> quote | uri) >>
+    media_type_list.maybe >> semicolon >> ignore
+  }
+
+  # @media
+  # http://www.w3.org/TR/CSS2/media.html#media-intro
+  rule(:media) {
+    ignore >> str('@media') >> media_type_list >> space? >>
+    lcurly >> ruleset.repeat >> rcurly
+  }
+  rule(:media_type_list) {
+    space >> media_types >> ((comma >> space? >> media_types).repeat).maybe
+  }
+  rule(:media_types) {
+    str('all') | str('braille') | str('embossed') | str('handheld') |
+    str('print') | str('projection') | str('screen') | str('speech') |
+    str('tty') | str('tv')
+  }
+
+  rule(:ruleset) { selectors >> lcurly >> declarations >> rcurly }
+  # TODO @namespace at-rule
+  rule(:stylesheet) { charset.maybe >> import.repeat.maybe >> (media | ruleset).repeat }
+  root :stylesheet
+end

data/test/parser_test.rb

@@ -0,0 +1,176 @@
+# encoding: UTF-8
+require 'minitest/autorun'
+require 'parslet-css'
+
+describe ParsletCSS::Parser do
+  before do
+    @parser = ParsletCSS::Parser.new
+  end
+
+  describe "import parse" do
+    it "can import other css" do
+      @parser.parse('@import "mystyle.css";')
+      @parser.parse('@import url("mystyle.css");')
+      @parser.parse('@import url("bluish.css") projection, tv;')
+      @parser.parse('@import url("fineprint.css") print; @import url("bluish.css") projection, tv; body {}')
+    end
+  end
+
+  describe "charset" do
+    it '@charset rule must place the rule at the very beginning of the style sheet' do
+      @parser.parse('@charset "ISO-8859-1";')
+      @parser.parse('@charset "UTF-8"; @import "mystyle.css"; body {}')
+    end
+  end
+
+  describe '@media rule' do
+    it 'valid statement' do
+      @parser.parse('@media print { body { font-size: 10pt } }')
+      @parser.parse('@media screen, print { body { font-size: 13px } }')
+      @parser.parse('/* print */ @media print { body { font-size: 10pt } }
+                     /* screen */ @media screen { body { font-size: 13px } }')
+    end
+  end
+
+  describe "parse success" do
+    it "parse with percent" do
+      @parser.parse("body { height: 100%; width: 100%; }")
+    end
+    it "with url" do
+      @parser.parse("body { background: url(/images/plop.png); }")
+      @parser.parse("body { background: url(https://localhost:3000/images/plop.png); }")
+      @parser.parse("body { background: url(/~spk/images/plop.png?size=30;toto=tata); }")
+      @parser.parse("body { background: url(/~spk/images/plop.png?size=30&toto=tata); }")
+      u = @parser.parse("body { background: url(/up.png); } nav { background: url(/nav.png) }")
+      assert_equal(2, u.size)
+    end
+    it "with comments" do
+      @parser.parse("body { /* comment */ padding: 0; /* comment */}")
+    end
+    it "with no space" do
+      @parser.parse(":focus{outline:0;/*comment*/}")
+      @parser.parse(":focus{outline:0}")
+    end
+    it "with no last semicolon" do
+      @parser.parse("body { height: 100%; width: 100% }")
+    end
+    it "with one property and no semicolon" do
+      @parser.parse("body { width: 100% }")
+    end
+    it "with blank property" do
+      @parser.parse("body { }")
+    end
+    it 'with color' do
+      @parser.parse("body { color: #1a171b; }")
+      @parser.parse("body { color: #fff; }")
+    end
+    it 'with font family' do
+      @parser.parse("body { font: 12px monospace, serif; }")
+      @parser.parse("body { font: bold 14px/28px Arial; }")
+    end
+    it 'with multiple property values' do
+      @parser.parse("body { margin: 10px 10px 10px 10em;}")
+      @parser.parse("body { background: url(/images/ui/down.png) no-repeat scroll 140px 0px transparent; }")
+      @parser.parse("body { background: url(/images/ui/up.png) no-repeat scroll 140px 0px #1a171b; }")
+    end
+    it "float size" do
+      @parser.parse(".date_selector .nav { width: 17.5em; }")
+    end
+    it 'property priority' do
+      @parser.parse(".date_selector .nav { width: 17.5em !important; }")
+    end
+
+    describe('CSS selectors') do
+      # http://www.w3.org/TR/CSS2/selector.html
+      it "CSS2 w3c examples" do
+        # 5.2.1 Grouping
+        @parser.parse("h1, h2, h3 { font-family: sans-serif }")
+        # 5.6 Child selectors
+        @parser.parse("body > P { line-height: 1.3 }")
+        @parser.parse("div ol>li p { font: 2px }")
+        # 5.7 Adjacent sibling selectors
+        @parser.parse("math + p { text-indent: 0 } ")
+        @parser.parse("h1 + h2 { margin-top: -5mm }")
+        @parser.parse("h1.opener + h2 { margin-top: -5mm }")
+        # 5.8.1 Matching attributes and attribute values
+        @parser.parse("h1[title] { color: blue; }")
+        @parser.parse("span[class=example] { color: blue; }")
+        @parser.parse('span[hello="Cleveland"][goodbye="Columbus"] { color: blue; }')
+        @parser.parse('a[rel~="copyright"] {}')
+        @parser.parse('a[href="http://www.w3.org/"] {}')
+        @parser.parse("*[lang=fr] { display : none }")
+        @parser.parse('*[lang|="en"] { color : red }')
+        @parser.parse('DIALOGUE[character=juliet] { voice-family: "Vivien Leigh", victoria, female }')
+        # 5.8.3 Class selectors
+        @parser.parse("*.pastoral { color: green }")
+        @parser.parse("H1.pastoral { color: green }")
+        @parser.parse('p.marine.pastoral { color: green }')
+        # 5.9 ID selectors
+        @parser.parse("h1#chapter1 { text-align: center }")
+        # 5.11 Pseudo-classes
+        @parser.parse('div > p:first-child { text-indent: 0 }')
+        @parser.parse('p:first-child em { font-weight : bold }')
+        @parser.parse('* > a:first-child {} /* A is first child of any element */')
+        @parser.parse('a:first-child {} /* Same */')
+        @parser.parse(':link { color: red }')
+        @parser.parse('a.external:visited { color: blue }')
+        @parser.parse('a:focus:hover { background: white }')
+        # 5.11.4 The language pseudo-class: :lang
+        @parser.parse("html:lang(fr-ca) { quotes: '« ' ' »' }")
+        @parser.parse(":lang(fr) > Q { quotes: '« ' ' »' }")
+        # 5.12.1 The :first-line pseudo-element
+        @parser.parse("p:first-line { text-transform: uppercase }")
+        @parser.parse('p:first-letter { font-size: 3em; font-weight: normal }')
+        # 5.12.3 The :before and :after pseudo-elements
+        @parser.parse('p.special:before {content: "Special! "}')
+        @parser.parse('p.special:first-letter {color: #ffd800}')
+      end
+
+      # http://www.w3.org/TR/css3-selectors/
+      it "CSS3 w3c examples" do
+        @parser.parse('object[type^="image/"] {}')
+        @parser.parse('a[href$=".html"] {}')
+        @parser.parse('tr:nth-child(2n+1) {} /* represents every odd row of an HTML table */')
+        @parser.parse('p:nth-child(4n+1) { color: navy; }')
+        @parser.parse('button:not([DISABLED]) {}')
+        @parser.parse('p::first-line { text-transform: uppercase }')
+      end
+    end
+  end
+
+  describe "parse fail" do
+    # http://www.w3.org/TR/CSS21/syndata.html#parsing-errors
+    @raises = [
+      {:msg => "with extra semicolon", :css => "body { height: 100%; ; width: 100%; }"},
+      {:msg => "with extra curly", :css => "body { height: 100%; width: 100%; }}"},
+      {:msg => "& is not valid token", :css => "h3, h4 & h5 {color: red }"},
+      {:msg => "1 malformed declaration missing ':', value", :css => "p { color:green; color }"},
+      {:msg => "2 malformed declaration missing value", :css => "p { color:green; color: }"},
+      {:msg => "unexpected tokens { }", :css => "p { color:green; color{;color:maroon} }"},
+      {:msg => "ruleset with unexpected at-keyword @here", :css => "p @here {color: red}"},
+      {:msg => "at-rule with unexpected at-keyword @bar", :css => "@foo @bar;"},
+      {:msg => "ruleset with unexpected right brace", :css => "}} {{ - }}"},
+      {:msg => "ruleset with unexpected right parenthesis", :css => ") ( {} ) p {color: red }"},
+
+      # http://www.w3.org/TR/CSS2/syndata.html#at-rules
+      {:msg => "must ignore any '@import' rule that occurs inside a block",
+        :css => '@import "subs.css"; h1 { color: blue } @import "list.css";'},
+      {:msg => 'non valid charset', :css => '@charset "none";'},
+      {:msg => '@charset rule not at the beginning of the style sheet',
+        :css => '@import "awesome.css"; @charset "UTF-8";'},
+
+      {:msg => 'identifier must not be empty. (Otherwise, the selector is invalid.)',
+        :css => 'html:lang() {}'},
+    ]
+    @raises.each do |r|
+      class_eval do
+        it r[:msg] do
+          assert_raises Parslet::ParseFailed do
+            @parser.parse(r[:css])
+          end
+        end
+      end
+    end
+
+  end
+end

metadata

@@ -0,0 +1,76 @@
+--- !ruby/object:Gem::Specification
+name: parslet-css
+version: !ruby/object:Gem::Version
+  version: 0.0.1
+  prerelease: 
+platform: ruby
+authors:
+- Laurent Arnoud
+autorequire: 
+bindir: bin
+cert_chain: []
+date: 2011-06-12 00:00:00.000000000Z
+dependencies:
+- !ruby/object:Gem::Dependency
+  name: parslet
+  requirement: &14494580 !ruby/object:Gem::Requirement
+    none: false
+    requirements:
+    - - ~>
+      - !ruby/object:Gem::Version
+        version: '1.2'
+  type: :runtime
+  prerelease: false
+  version_requirements: *14494580
+- !ruby/object:Gem::Dependency
+  name: minitest
+  requirement: &14493660 !ruby/object:Gem::Requirement
+    none: false
+    requirements:
+    - - ~>
+      - !ruby/object:Gem::Version
+        version: '2.0'
+  type: :development
+  prerelease: false
+  version_requirements: *14493660
+description: CSS parser with Parslet
+email: laurent@spkdev.net
+executables: []
+extensions: []
+extra_rdoc_files:
+- README.markdown
+files:
+- LICENSE
+- README.markdown
+- Gemfile
+- data/iana_character_sets.txt
+- data/character-sets
+- lib/parslet-css/parser.rb
+- lib/parslet-css.rb
+- test/parser_test.rb
+homepage: http://github.com/spk/parslet-css
+licenses: []
+post_install_message: 
+rdoc_options: []
+require_paths:
+- lib
+required_ruby_version: !ruby/object:Gem::Requirement
+  none: false
+  requirements:
+  - - ! '>='
+    - !ruby/object:Gem::Version
+      version: '0'
+required_rubygems_version: !ruby/object:Gem::Requirement
+  none: false
+  requirements:
+  - - ! '>='
+    - !ruby/object:Gem::Version
+      version: '0'
+requirements: []
+rubyforge_project: 
+rubygems_version: 1.7.2
+signing_key: 
+specification_version: 3
+summary: CSS parser with Parslet grammar tool
+test_files:
+- test/parser_test.rb