lib/markdown: allow to ignore the MDlocation on Tokens
[nit.git] / lib / markdown / markdown.nit
index d928228..d63f3e4 100644 (file)
@@ -133,6 +133,14 @@ class MarkdownProcessor
        # ~~~
        var ext_mode = true
 
+       # Disable attaching MDLocation to Tokens
+       #
+       # Locations are useful for some tools but they may
+       # cause an important time and space overhead.
+       #
+       # Default = `false`
+       var no_location = false is writable
+
        init do self.emitter = new MarkdownEmitter(self)
 
        # Process the mardown `input` string and return the processed output.
@@ -168,8 +176,9 @@ class MarkdownProcessor
                                var c = input[i]
                                if c == '\n' then
                                        eol = true
+                               else if c == '\r' then
                                else if c == '\t' then
-                                       var np = pos + (4 - (pos.bin_and(3)))
+                                       var np = pos + (4 - (pos & 3))
                                        while pos < np do
                                                value.add ' '
                                                pos += 1
@@ -396,11 +405,16 @@ class MarkdownProcessor
                        c2 = ' '
                end
 
-               var loc = new MDLocation(
-                       current_loc.line_start,
-                       current_loc.column_start + pos,
-                       current_loc.line_start,
-                       current_loc.column_start + pos)
+               var loc
+               if no_location then
+                       loc = null
+               else
+                       loc = new MDLocation(
+                               current_loc.line_start,
+                               current_loc.column_start + pos,
+                               current_loc.line_start,
+                               current_loc.column_start + pos)
+               end
 
                if c == '*' then
                        if c1 == '*' then
@@ -417,7 +431,7 @@ class MarkdownProcessor
                        end
                else if c == '_' then
                        if c1 == '_' then
-                               if c0 != ' ' or c2 != ' 'then
+                               if c0 != ' ' or c2 != ' ' then
                                        return new TokenStrongUnderscore(loc, pos, c)
                                else
                                        return new TokenEmUnderscore(loc, pos, c)
@@ -609,10 +623,12 @@ class MarkdownEmitter
        end
 
        # Append `c` to current buffer.
-       fun addc(c: Char) do add c.to_s
+       fun addc(c: Char) do
+               current_buffer.add c
+       end
 
        # Append a "\n" line break.
-       fun addn do add "\n"
+       fun addn do addc '\n'
 end
 
 # A Link Reference.
@@ -634,7 +650,7 @@ class LinkRef
 
        # Create a link with a title.
        init with_title(link: String, title: nullable String) do
-               self.link = link
+               init(link)
                self.title = title
        end
 end
@@ -646,6 +662,11 @@ interface Decorator
        # Kind of emitter used for decoration.
        type EMITTER: MarkdownEmitter
 
+       # Render a single plain char.
+       #
+       # Redefine this method to add special escaping for plain text.
+       fun add_char(v: EMITTER, c: Char) do v.addc c
+
        # Render a ruler block.
        fun add_ruler(v: EMITTER, block: BlockRuler) is abstract
 
@@ -753,8 +774,11 @@ class HTMLDecorator
        end
 
        redef fun add_code(v, block) do
-               if block isa BlockFence and block.meta != null then
-                       v.add "<pre class=\"{block.meta.to_s}\"><code>"
+               var meta = block.meta
+               if meta != null then
+                       v.add "<pre class=\""
+                       append_value(v, meta)
+                       v.add "\"><code>"
                else
                        v.add "<pre><code>"
                end
@@ -1168,6 +1192,26 @@ abstract class Block
                        block = block.next
                end
        end
+
+       # The raw content of the block as a multi-line string.
+       fun raw_content: String do
+               var infence = self isa BlockFence
+               var text = new FlatBuffer
+               var line = self.block.first_line
+               while line != null do
+                       if not line.is_empty then
+                               var str = line.value
+                               if not infence and str.has_prefix("    ") then
+                                       text.append str.substring(4, str.length - line.trailing)
+                               else
+                                       text.append str
+                               end
+                       end
+                       text.append "\n"
+                       line = line.next
+               end
+               return text.write_to_string
+       end
 end
 
 # A block without any markdown specificities.
@@ -1208,6 +1252,9 @@ end
 class BlockCode
        super Block
 
+       # Any string found after fence token.
+       var meta: nullable Text
+
        # Number of char to skip at the beginning of the line.
        #
        # Block code lines start at 4 spaces.
@@ -1234,9 +1281,6 @@ end
 class BlockFence
        super BlockCode
 
-       # Any string found after fence token.
-       var meta: nullable Text
-
        # Fence code lines start at 0 spaces.
        redef var line_start = 0
 end
@@ -1909,7 +1953,7 @@ end
 abstract class Token
 
        # Location of `self` in the original input.
-       var location: MDLocation
+       var location: nullable MDLocation
 
        # Position of `self` in input independant from lines.
        var pos: Int
@@ -1918,7 +1962,7 @@ abstract class Token
        var char: Char
 
        # Output that token using `MarkdownEmitter::decorator`.
-       fun emit(v: MarkdownEmitter) do v.addc char
+       fun emit(v: MarkdownEmitter) do v.decorator.add_char(v, char)
 end
 
 # A token without a specific meaning.
@@ -2102,6 +2146,7 @@ abstract class TokenLinkOrImage
                                        if pos == -1 then return -1
                                end
                        end
+                       if pos < start then return -1
                        if md[pos] != ')' then return -1
                else if md[pos] == '[' then
                        pos += 1
@@ -2300,18 +2345,11 @@ redef class Text
                        if c == '\\' and pos + 1 < length then
                                pos = escape(out, self[pos + 1], pos)
                        else
-                               var end_reached = false
-                               for n in nend do
-                                       if c == n then
-                                               end_reached = true
-                                               break
-                                       end
-                               end
-                               if end_reached then break
+                               for n in nend do if c == n then break label
                                out.add c
                        end
                        pos += 1
-               end
+               end label
                if pos == length then return -1
                return pos
        end