lib/kramdown/converter/man.rb in kramdown-man-0.1.4 vs lib/kramdown/converter/man.rb in kramdown-man-0.1.5

@@ -11,32 +11,398 @@
     class Man < Base
 
       # Comment header
       HEADER = [
         ".\\\" Generated by kramdown-man #{::Kramdown::Man::VERSION}",
-        ".\\\" https://github.com/postmodern/kramdown-roff#readme"
+        ".\\\" https://github.com/postmodern/kramdown-man#readme"
       ].join("\n")
 
       # Typographic Symbols and their UTF8 chars
       TYPOGRAPHIC_SYMS = {
-        :ndash       => "--",
-        :mdash       => "—",
-        :hellip      => "…",
-        :laquo       => "«",
-        :raquo       => "»",
-        :laquo_space => "«",
-        :raquo_space => "»"
+        :ndash       => '\-\-',
+        :mdash       => '\[em]',
+        :hellip      => '...',
+        :laquo       => '\[Fo]',
+        :raquo       => '\[Fc]',
+        :laquo_space => '\[Fo]',
+        :raquo_space => '\[Fc]'
       }
 
       # Smart Quotes and their UTF8 chars
       SMART_QUOTES = {
-        :lsquo => "‘",
-        :rsquo => "’",
-        :ldquo => "“",
-        :rdquo => "”"
+        :lsquo => '\[oq]',
+        :rsquo => '\[cq]',
+        :ldquo => '\[lq]',
+        :rdquo => '\[rq]'
       }
 
+      GLYPHS = {
+        'Ð' => '\[-D]',
+        'ð' => '\[Sd]',
+        'Þ' => '\[TP]',
+        'þ' => '\[Tp]',
+        'ß' => '\[ss]',
+        # Ligatures and Other Latin Glyphs
+        'ff' => '\[ff]',
+        'fi' => '\[fi]',
+        'fl' => '\[fl]',
+        'ffi' => '\[Fi]',
+        'ffl' => '\[Fl]',
+        'Ł' => '\[/L]',
+        'ł' => '\[/l]',
+        'Ø' => '\[/O]',
+        'ø' => '\[/o]',
+        'Æ' => '\[AE]',
+        'æ' => '\[ae]',
+        'Œ' => '\[OE]',
+        'œ' => '\[oe]',
+        'IJ' => '\[IJ]',
+        'ij' => '\[ij]',
+        'ı' => '\[.i]',
+        'ȷ' => '\[.j]',
+        # Accented Characters
+        'Á' => '\[\'A]',
+        'Ć' => '\[\'C]',
+        'É' => '\[\'E]',
+        'Í' => '\[\'I]',
+        'Ó' => '\[\'O]',
+        'Ú' => '\[\'U]',
+        'Ý' => '\[\'Y]',
+        'á' => '\[\'a]',
+        'ć' => '\[\'c]',
+        'é' => '\[\'e]',
+        'í' => '\[\'i]',
+        'ó' => '\[\'o]',
+        'ú' => '\[\'u]',
+        'ý' => '\[\'y]',
+        'Ä' => '\[:A]',
+        'Ë' => '\[:E]',
+        'Ï' => '\[:I]',
+        'Ö' => '\[:O]',
+        'Ü' => '\[:U]',
+        'Ÿ' => '\[:Y]',
+        'ä' => '\[:a]',
+        'ë' => '\[:e]',
+        'ï' => '\[:i]',
+        'ö' => '\[:o]',
+        'ü' => '\[:u]',
+        'ÿ' => '\[:y]',
+        'Â' => '\[^A]',
+        'Ê' => '\[^E]',
+        'Î' => '\[^I]',
+        'Ô' => '\[^O]',
+        'Û' => '\[^U]',
+        'â' => '\[^a]',
+        'ê' => '\[^e]',
+        'î' => '\[^i]',
+        'ô' => '\[^o]',
+        'û' => '\[^u]',
+        'À' => '\[`A]',
+        'È' => '\[`E]',
+        'Ì' => '\[`I]',
+        'Ò' => '\[`O]',
+        'Ù' => '\[`U]',
+        'à' => '\[`a]',
+        'è' => '\[`e]',
+        'ì' => '\[`i]',
+        'ò' => '\[`o]',
+        'ù' => '\[`u]',
+        'Ã' => '\[~A]',
+        'Ñ' => '\[~N]',
+        'Õ' => '\[~O]',
+        'ã' => '\[~a]',
+        'ñ' => '\[~n]',
+        'õ' => '\[~o]',
+        'Š' => '\[vS]',
+        'š' => '\[vs]',
+        'Ž' => '\[vZ]',
+        'ž' => '\[vz]',
+        'Ç' => '\[,C]',
+        'ç' => '\[,c]',
+        'Å' => '\[oA]',
+        'å' => '\[oa]',
+        # Accents
+        '˝' => '\[a"]',
+        '¯' => '\[a-]',
+        '˙' => '\[a.]',
+        '^' => '\[a^]',
+        '´' => "\\´",
+        '`' => '\`',
+        '˘' => '\[ab]',
+        '¸' => '\[ac]',
+        '¨' => '\[ad]',
+        'ˇ' => '\[ah]',
+        '˚' => '\[ao]',
+        '~' => '\(ti',
+        '˛' => '\[ho]',
+        '^' => '\(ha',
+        '~' => '\[ti]',
+        # Quotes
+        '„' => '\[Bq]',
+        '‚' => '\[bq]',
+        '“' => '\[lq]',
+        '”' => '\[rq]',
+        '‘' => '\[oq]',
+        '’' => '\[cq]',
+        "'" => '\(aq',
+        '"' => '\[dq]',
+        '«' => '\[Fo]',
+        '»' => '\[Fc]',
+        '‹' => '\[fo]',
+        '›' => '\[fc]',
+        # Punctuation
+        '.' => '\.',
+        '¡' => '\[r!]',
+        '¿' => '\[r?]',
+        '—' => '\[em]',
+        '–' => '\[en]',
+        '‐' => '\[hy]',
+        # Brackets
+        '[' => '\[lB]',
+        ']' => '\[rB]',
+        '{' => '\[lC]',
+        '}' => '\[rC]',
+        '⟨' => '\[la]',
+        '⟩' => '\[ra]',
+        '⎪' => '\[bv]',
+        '⎪' => '\[braceex]',
+        '⎡' => '\[bracketlefttp]',
+        '⎣' => '\[bracketleftbt]',
+        '⎢' => '\[bracketleftex]',
+        '⎤' => '\[bracketrighttp]',
+        '⎦' => '\[bracketrightbt]',
+        '⎥' => '\[bracketrightex]',
+        '╭' => '\[lt]',
+        '⎧' => '\[bracelefttp]',
+        '┥' => '\[lk]',
+        '⎨' => '\[braceleftmid]',
+        '╰' => '\[lb]',
+        '⎩' => '\[braceleftbt]',
+        '⎪' => '\[braceleftex]',
+        '╮' => '\[rt]',
+        '⎫' => '\[bracerighttp]',
+        '┝' => '\[rk]',
+        '⎬' => '\[bracerightmid]',
+        '╯' => '\[rb]',
+        '⎭' => '\[bracerightbt]',
+        '⎪' => '\[bracerightex]',
+        '⎛' => '\[parenlefttp]',
+        '⎝' => '\[parenleftbt]',
+        '⎜' => '\[parenleftex]',
+        '⎞' => '\[parenrighttp]',
+        '⎠' => '\[parenrightbt]',
+        '⎟' => '\[parenrightex]',
+        # Arrows
+        '←' => '\[<-]',
+        '→' => '\[->]',
+        '↔' => '\[<>]',
+        '↓' => '\[da]',
+        '↑' => '\[ua]',
+        '↕' => '\[va]',
+        '⇐' => '\[lA]',
+        '⇒' => '\[rA]',
+        '⇔' => '\[hA]',
+        '⇓' => '\[dA]',
+        '⇑' => '\[uA]',
+        '⇕' => '\[vA]',
+        '⎯' => '\[an]',
+        # Lines
+        '|' => '\[ba]',
+        '│' => '\[br]',
+        '_' => '\[ul]',
+        '‾' => '\[rn]',
+        '_' => '\[ru]',
+        '¦' => '\[bb]',
+        '/' => '\[sl]',
+        '\\' => '\e',
+        # Text markers
+        '○' => '\[ci]',
+        '·' => '\[bu]',
+        '‡' => '\[dd]',
+        '†' => '\[dg]',
+        '◊' => '\[lz]',
+        '□' => '\[sq]',
+        '¶' => '\[ps]',
+        '§' => '\[sc]',
+        '☜' => '\[lh]',
+        '☞' => '\[rh]',
+        '@' => '\[at]',
+        '#' => '\[sh]',
+        '↵' => '\[CR]',
+        '✓' => '\[OK]',
+        # Legal Symbols
+        '©' => '\[co]',
+        '®' => '\[rg]',
+        '™' => '\[tm]',
+        # Currency symbols
+        '$' => '\[Do]',
+        '¢' => '\[ct]',
+        '€' => '\[eu]',
+        '€' => '\[Eu]',
+        '¥' => '\[Ye]',
+        '£' => '\[Po]',
+        '¤' => '\[Cs]',
+        'ƒ' => '\[Fn]',
+        # Units
+        '°' => '\[de]',
+        '‰' => '\[%0]',
+        '′' => '\[fm]',
+        '″' => '\[sd]',
+        'µ' => '\[mc]',
+        'ª' => '\[Of]',
+        'º' => '\[Om]',
+        # Logical Symbols
+        '∧' => '\[AN]',
+        '∨' => '\[OR]',
+        '¬' => '\[no]',
+        '¬' => '\[tno]',
+        '∃' => '\[te]',
+        '∀' => '\[fa]',
+        '∋' => '\[st]',
+        '∴' => '\[3d]',
+        '∴' => '\[tf]',
+        '|' => '\[or]',
+        # Mathematical Symbols
+        '½' => '\[12]',
+        '¼' => '\[14]',
+        '¾' => '\[34]',
+        '⅛' => '\[18]',
+        '⅜' => '\[38]',
+        '⅝' => '\[58]',
+        '⅞' => '\[78]',
+        '¹' => '\[S1]',
+        '²' => '\[S2]',
+        '³' => '\[S3]',
+        '+' => '\[pl]',
+        '-' => '\-',
+        '−' => '\[mi]',
+        '∓' => '\[-+]',
+        '±' => '\[+-]',
+        '±' => '\[t+-]',
+        '·' => '\[pc]',
+        '⋅' => '\[md]',
+        '×' => '\[mu]',
+        '×' => '\[tmu]',
+        '⊗' => '\[c*]',
+        '⊕' => '\[c+]',
+        '÷' => '\[di]',
+        '÷' => '\[tdi]',
+        '⁄' => '\[f/]',
+        '∗' => '\[**]',
+        '≤' => '\[<=]',
+        '≥' => '\[>=]',
+        '≪' => '\[<<]',
+        '≫' => '\[>>]',
+        '=' => '\[eq]',
+        '≠' => '\[!=]',
+        '≡' => '\[==]',
+        '≢' => '\[ne]',
+        '≅' => '\[=~]',
+        '≃' => '\[|=]',
+        '∼' => '\[ap]',
+        '≈' => '\[~~]',
+        '≈' => '\[~=]',
+        '∝' => '\[pt]',
+        '∅' => '\[es]',
+        '∈' => '\[mo]',
+        '∉' => '\[nm]',
+        '⊂' => '\[sb]',
+        '⊄' => '\[nb]',
+        '⊃' => '\[sp]',
+        '⊅' => '\[nc]',
+        '⊆' => '\[ib]',
+        '⊇' => '\[ip]',
+        '∩' => '\[ca]',
+        '∪' => '\[cu]',
+        '∠' => '\[/_]',
+        '⊥' => '\[pp]',
+        '∫' => '\[is]',
+        '∫' => '\[integral]',
+        '∑' => '\[sum]',
+        '∏' => '\[product]',
+        '∐' => '\[coproduct]',
+        '∇' => '\[gr]',
+        '√' => '\[sr]',
+        '√' => '\[sqrt]',
+        '⌈' => '\[lc]',
+        '⌉' => '\[rc]',
+        '⌊' => '\[lf]',
+        '⌋' => '\[rf]',
+        '∞' => '\[if]',
+        'ℵ' => '\[Ah]',
+        'ℑ' => '\[Im]',
+        'ℜ' => '\[Re]',
+        '℘' => '\[wp]',
+        '∂' => '\[pd]',
+        'ℏ' => '\[-h]',
+        'ℏ' => '\[hbar]',
+        # Greek glyphs
+        'Α' => '\[*A]',
+        'Β' => '\[*B]',
+        'Γ' => '\[*G]',
+        'Δ' => '\[*D]',
+        'Ε' => '\[*E]',
+        'Ζ' => '\[*Z]',
+        'Η' => '\[*Y]',
+        'Θ' => '\[*H]',
+        'Ι' => '\[*I]',
+        'Κ' => '\[*K]',
+        'Λ' => '\[*L]',
+        'Μ' => '\[*M]',
+        'Ν' => '\[*N]',
+        'Ξ' => '\[*C]',
+        'Ο' => '\[*O]',
+        'Π' => '\[*P]',
+        'Ρ' => '\[*R]',
+        'Σ' => '\[*S]',
+        'Τ' => '\[*T]',
+        'Υ' => '\[*U]',
+        'Φ' => '\[*F]',
+        'Χ' => '\[*X]',
+        'Ψ' => '\[*Q]',
+        'Ω' => '\[*W]',
+        'α' => '\[*a]',
+        'β' => '\[*b]',
+        'γ' => '\[*g]',
+        'δ' => '\[*d]',
+        'ε' => '\[*e]',
+        'ζ' => '\[*z]',
+        'η' => '\[*y]',
+        'θ' => '\[*h]',
+        'ι' => '\[*i]',
+        'κ' => '\[*k]',
+        'λ' => '\[*l]',
+        'μ' => '\[*m]',
+        'ν' => '\[*n]',
+        'ξ' => '\[*c]',
+        'ο' => '\[*o]',
+        'π' => '\[*p]',
+        'ρ' => '\[*r]',
+        'ς' => '\[ts]',
+        'σ' => '\[*s]',
+        'τ' => '\[*t]',
+        'υ' => '\[*u]',
+        'ϕ' => '\[*f]',
+        'χ' => '\[*x]',
+        'ψ' => '\[*q]',
+        'ω' => '\[*w]',
+        'ϑ' => '\[+h]',
+        'φ' => '\[+f]',
+        'ϖ' => '\[+p]',
+        'ϵ' => '\[+e]',
+        # Card symbols
+        '♣' => '\[CL]',
+        '♠' => '\[SP]',
+        '♥' => '\[HE]',
+        '♡' => '\[u2661]',
+        '♦' => '\[DI]',
+        '♢' => '\[u2662]'
+      }
+
+      # Regular expression to convert unicode characters into glyphs
+      GLYPH_REGEXP = Regexp.union(GLYPHS.keys)
+
       #
       # Initializes the converter.
       #
       # @param [Kramdown::Element] root
       #   The root of the markdown document.
@@ -322,12 +688,13 @@
       #   The roff output.
       #
       def convert_p(p)
         children = p.children
 
-        if (children.length >= 2) &&
-           (children.first.type == :em || children.first.type == :codespan)
+        if ((children.length >= 2) && (children.first.type == :codespan ||
+                                       children.first.type == :em       ||
+                                       children.first.type == :strong))
           newline = children.find_index { |el|
             el.type == :text && el.value.start_with?("\n")
           }
 
           if newline
@@ -377,11 +744,11 @@
       #
       # @return [String]
       #   The roff output.
       #
       def convert_codespan(codespan)
-        "\\fB\\fC#{codespan.value}\\fR"
+        "\\fB#{codespan.value}\\fR"
       end
 
       #
       # Converts a `kd:a` element.
       #
@@ -390,11 +757,11 @@
       #
       # @return [String]
       #   The roff output.
       #
       def convert_a(a)
-        href = a.attr['href']
+        href = escape(a.attr['href'])
         text = convert_children(a.children)
 
         case href
         when /^mailto:/
           email = href[7..-1]
@@ -434,10 +801,10 @@
       #
       # @return [String]
       #   The escaped text.
       #
       def escape(text)
-        text.gsub('\\','\&\&').gsub('-','\\-')
+        text.gsub(GLYPH_REGEXP) { |char| GLYPHS[char] }
       end
 
     end
   end
 end