tartrazine/lexers/haskell.xml

<lexer>
  <config>
    <name>Haskell</name>
    <alias>haskell</alias>
    <alias>hs</alias>
    <filename>*.hs</filename>
    <mime_type>text/x-haskell</mime_type>
  </config>
  <rules>
    <state name="escape">
      <rule pattern="[abfnrtv&#34;\&#39;&amp;\\]">
        <token type="LiteralStringEscape"/>
        <pop depth="1"/>
      </rule>
      <rule pattern="\^[][\p{Lu}@^_]">
        <token type="LiteralStringEscape"/>
        <pop depth="1"/>
      </rule>
      <rule pattern="NUL|SOH|[SE]TX|EOT|ENQ|ACK|BEL|BS|HT|LF|VT|FF|CR|S[OI]|DLE|DC[1-4]|NAK|SYN|ETB|CAN|EM|SUB|ESC|[FGRU]S|SP|DEL">
        <token type="LiteralStringEscape"/>
        <pop depth="1"/>
      </rule>
      <rule pattern="o[0-7]+">
        <token type="LiteralStringEscape"/>
        <pop depth="1"/>
      </rule>
      <rule pattern="x[\da-fA-F]+">
        <token type="LiteralStringEscape"/>
        <pop depth="1"/>
      </rule>
      <rule pattern="\d+">
        <token type="LiteralStringEscape"/>
        <pop depth="1"/>
      </rule>
      <rule pattern="\s+\\">
        <token type="LiteralStringEscape"/>
        <pop depth="1"/>
      </rule>
    </state>
    <state name="root">
      <rule pattern="\s+">
        <token type="Text"/>
      </rule>
      <rule pattern="--(?![!#$%&amp;*+./&lt;=&gt;?@^|_~:\\]).*?$">
        <token type="CommentSingle"/>
      </rule>
      <rule pattern="\{-">
        <token type="CommentMultiline"/>
        <push state="comment"/>
      </rule>
      <rule pattern="\bimport\b">
        <token type="KeywordReserved"/>
        <push state="import"/>
      </rule>
      <rule pattern="\bmodule\b">
        <token type="KeywordReserved"/>
        <push state="module"/>
      </rule>
      <rule pattern="\berror\b">
        <token type="NameException"/>
      </rule>
      <rule pattern="\b(case|class|data|default|deriving|do|else|family|if|in|infix[lr]?|instance|let|newtype|of|then|type|where|_)(?!\&#39;)\b">
        <token type="KeywordReserved"/>
      </rule>
      <rule pattern="&#39;[^\\]&#39;">
        <token type="LiteralStringChar"/>
      </rule>
      <rule pattern="^[_\p{Ll}][\w\&#39;]*">
        <token type="NameFunction"/>
      </rule>
      <rule pattern="&#39;?[_\p{Ll}][\w&#39;]*">
        <token type="Name"/>
      </rule>
      <rule pattern="(&#39;&#39;)?[\p{Lu}][\w\&#39;]*">
        <token type="KeywordType"/>
      </rule>
      <rule pattern="(&#39;)[\p{Lu}][\w\&#39;]*">
        <token type="KeywordType"/>
      </rule>
      <rule pattern="(&#39;)\[[^\]]*\]">
        <token type="KeywordType"/>
      </rule>
      <rule pattern="(&#39;)\([^)]*\)">
        <token type="KeywordType"/>
      </rule>
      <rule pattern="\\(?![:!#$%&amp;*+.\\/&lt;=&gt;?@^|~-]+)">
        <token type="NameFunction"/>
      </rule>
      <rule pattern="(&lt;-|::|-&gt;|=&gt;|=|'([:!#$%&amp;*+.\\/&lt;=&gt;?@^|~-]+))(?![:!#$%&amp;*+.\\/&lt;=&gt;?@^|~-]+)">
        <token type="OperatorWord"/>
      </rule>
      <rule pattern=":[:!#$%&amp;*+.\\/&lt;=&gt;?@^|~-]*">
        <token type="KeywordType"/>
      </rule>
      <rule pattern="[:!#$%&amp;*+.\\/&lt;=&gt;?@^|~-]+">
        <token type="Operator"/>
      </rule>
      <rule pattern="\d+[eE][+-]?\d+">
        <token type="LiteralNumberFloat"/>
      </rule>
      <rule pattern="\d+\.\d+([eE][+-]?\d+)?">
        <token type="LiteralNumberFloat"/>
      </rule>
      <rule pattern="0[oO][0-7]+">
        <token type="LiteralNumberOct"/>
      </rule>
      <rule pattern="0[xX][\da-fA-F]+">
        <token type="LiteralNumberHex"/>
      </rule>
      <rule pattern="\d+">
        <token type="LiteralNumberInteger"/>
      </rule>
      <rule pattern="&#39;">
        <token type="LiteralStringChar"/>
        <push state="character"/>
      </rule>
      <rule pattern="&#34;">
        <token type="LiteralString"/>
        <push state="string"/>
      </rule>
      <rule pattern="\[\]">
        <token type="KeywordType"/>
      </rule>
      <rule pattern="\(\)">
        <token type="NameBuiltin"/>
      </rule>
      <rule pattern="[][(),;`{}]">
        <token type="Punctuation"/>
      </rule>
    </state>
    <state name="import">
      <rule pattern="\s+">
        <token type="Text"/>
      </rule>
      <rule pattern="&#34;">
        <token type="LiteralString"/>
        <push state="string"/>
      </rule>
      <rule pattern="\)">
        <token type="Punctuation"/>
        <pop depth="1"/>
      </rule>
      <rule pattern="qualified\b">
        <token type="Keyword"/>
      </rule>
      <rule pattern="([\p{Lu}][\w.]*)(\s+)(as)(\s+)([\p{Lu}][\w.]*)">
        <bygroups>
          <token type="NameNamespace"/>
          <token type="Text"/>
          <token type="Keyword"/>
          <token type="Text"/>
          <token type="Name"/>
        </bygroups>
        <pop depth="1"/>
      </rule>
      <rule pattern="([\p{Lu}][\w.]*)(\s+)(hiding)(\s+)(\()">
        <bygroups>
          <token type="NameNamespace"/>
          <token type="Text"/>
          <token type="Keyword"/>
          <token type="Text"/>
          <token type="Punctuation"/>
        </bygroups>
        <push state="funclist"/>
      </rule>
      <rule pattern="([\p{Lu}][\w.]*)(\s+)(\()">
        <bygroups>
          <token type="NameNamespace"/>
          <token type="Text"/>
          <token type="Punctuation"/>
        </bygroups>
        <push state="funclist"/>
      </rule>
      <rule pattern="[\w.]+">
        <token type="NameNamespace"/>
        <pop depth="1"/>
      </rule>
    </state>
    <state name="module">
      <rule pattern="\s+">
        <token type="Text"/>
      </rule>
      <rule pattern="([\p{Lu}][\w.]*)(\s+)(\()">
        <bygroups>
          <token type="NameNamespace"/>
          <token type="Text"/>
          <token type="Punctuation"/>
        </bygroups>
        <push state="funclist"/>
      </rule>
      <rule pattern="[\p{Lu}][\w.]*">
        <token type="NameNamespace"/>
        <pop depth="1"/>
      </rule>
    </state>
    <state name="funclist">
      <rule pattern="\s+">
        <token type="Text"/>
      </rule>
      <rule pattern="[\p{Lu}]\w*">
        <token type="KeywordType"/>
      </rule>
      <rule pattern="(_[\w\&#39;]+|[\p{Ll}][\w\&#39;]*)">
        <token type="NameFunction"/>
      </rule>
      <rule pattern="--(?![!#$%&amp;*+./&lt;=&gt;?@^|_~:\\]).*?$">
        <token type="CommentSingle"/>
      </rule>
      <rule pattern="\{-">
        <token type="CommentMultiline"/>
        <push state="comment"/>
      </rule>
      <rule pattern=",">
        <token type="Punctuation"/>
      </rule>
      <rule pattern="[:!#$%&amp;*+.\\/&lt;=&gt;?@^|~-]+">
        <token type="Operator"/>
      </rule>
      <rule pattern="\(">
        <token type="Punctuation"/>
        <push state="funclist" state="funclist"/>
      </rule>
      <rule pattern="\)">
        <token type="Punctuation"/>
        <pop depth="2"/>
      </rule>
    </state>
    <state name="comment">
      <rule pattern="[^-{}]+">
        <token type="CommentMultiline"/>
      </rule>
      <rule pattern="\{-">
        <token type="CommentMultiline"/>
        <push/>
      </rule>
      <rule pattern="-\}">
        <token type="CommentMultiline"/>
        <pop depth="1"/>
      </rule>
      <rule pattern="[-{}]">
        <token type="CommentMultiline"/>
      </rule>
    </state>
    <state name="character">
      <rule pattern="[^\\&#39;]&#39;">
        <token type="LiteralStringChar"/>
        <pop depth="1"/>
      </rule>
      <rule pattern="\\">
        <token type="LiteralStringEscape"/>
        <push state="escape"/>
      </rule>
      <rule pattern="&#39;">
        <token type="LiteralStringChar"/>
        <pop depth="1"/>
      </rule>
    </state>
    <state name="string">
      <rule pattern="[^\\&#34;]+">
        <token type="LiteralString"/>
      </rule>
      <rule pattern="\\">
        <token type="LiteralStringEscape"/>
        <push state="escape"/>
      </rule>
      <rule pattern="&#34;">
        <token type="LiteralString"/>
        <pop depth="1"/>
      </rule>
    </state>
  </rules>
</lexer>