X-Git-Url: http://nitlanguage.org

diff --git a/lib/markdown/markdown.nit b/lib/markdown/markdown.nit
index add43e1..0d264cb 100644
--- a/lib/markdown/markdown.nit
+++ b/lib/markdown/markdown.nit
@@ -31,12 +31,94 @@ import template
 class MarkdownProcessor
 
 	# `MarkdownEmitter` used for ouput.
-	var emitter: MarkdownEmitter is noinit
+	var emitter: MarkdownEmitter is noinit, protected writable
+
+	# Work in extended mode (default).
+	#
+	# Behavior changes when using extended mode:
+	#
+	# * Lists and code blocks end a paragraph
+	#
+	#   In normal markdown the following:
+	#
+	#		This is a paragraph
+	#		* and this is not a list
+	#
+	#   Will produce:
+	#
+	#		<p>This is a paragraph
+	#		* and this is not a list</p>
+	#
+	#	When using extended mode this changes to:
+	#
+	#		<p>This is a paragraph</p>
+	#		<ul>
+	#		<li>and this is not a list</li>
+	#		</ul>
+	#
+	# * Fences code blocks
+	#
+	#   If you don't want to indent your all your code with 4 spaces,
+	#   you can wrap your code in ``` ``` ``` or `~~~`.
+	#
+	#	Here's an example:
+	#
+	#		```
+	#		fun test do
+	#			print "Hello World!"
+	#		end
+	#		```
+	#
+	# * Code blocks meta
+	#
+	#   If you want to use syntax highlighting tools, most of them need to know what kind
+	#   of language they are highlighting.
+	#   You can add an optional language identifier after the fence declaration to output
+	#   it in the HTML render.
+	#
+	#		```nit
+	#		import markdown
+	#
+	#		print "# Hello World!".md_to_html
+	#		```
+	#
+	#   Becomes
+	#
+	#		<pre class="nit"><code>import markdown
+	#
+	#		print "Hello World!".md_to_html
+	#		</code></pre>
+	#
+	# * Underscores (Emphasis)
+	#
+	#   Underscores in the middle of a word like:
+	#
+	#		Con_cat_this
+	#
+	#	normally produces this:
+	#
+	#		<p>Con<em>cat</em>this</p>
+	#
+	#   With extended mode they don't result in emphasis.
+	#
+	#		<p>Con_cat_this</p>
+	#
+	# * Strikethrough
+	#
+	#   Like in [GFM](https://help.github.com/articles/github-flavored-markdown),
+	#   strikethrought span is marked with `~~`.
+	#
+	#		~~Mistaken text.~~
+	#
+	#   becomes
+	#
+	#		<del>Mistaken text.</del>
+	var ext_mode = true
 
 	init do self.emitter = new MarkdownEmitter(self)
 
 	# Process the mardown `input` string and return the processed output.
-	fun process(input: String): Streamable do
+	fun process(input: String): Writable do
 		# init processor
 		link_refs.clear
 		last_link_ref = null
@@ -52,17 +134,21 @@ class MarkdownProcessor
 
 	# Split `input` string into `MDLines` and create a parent `MDBlock` with it.
 	private fun read_lines(input: String): MDBlock do
-		var block = new MDBlock
+		var block = new MDBlock(new MDLocation(1, 1, 1, 1))
 		var value = new FlatBuffer
 		var i = 0
+
+		var line_pos = 0
+		var col_pos = 0
+
 		while i < input.length do
 			value.clear
 			var pos = 0
 			var eol = false
 			while not eol and i < input.length do
+				col_pos += 1
 				var c = input[i]
 				if c == '\n' then
-					i += 1
 					eol = true
 				else if c == '\t' then
 					var np = pos + (4 - (pos.bin_and(3)))
@@ -70,18 +156,20 @@ class MarkdownProcessor
 						value.add ' '
 						pos += 1
 					end
-					i += 1
 				else
 					pos += 1
 					value.add c
-					i += 1
 				end
+				i += 1
 			end
+			line_pos += 1
 
-			var line = new MDLine(value.write_to_string)
+			var loc = new MDLocation(line_pos, 1, line_pos, col_pos)
+			var line = new MDLine(loc, value.write_to_string)
 			var is_link_ref = check_link_ref(line)
 			# Skip link refs
 			if not is_link_ref then block.add_line line
+			col_pos = 0
 		end
 		return block
 	end
@@ -162,7 +250,7 @@ class MarkdownProcessor
 	#
 	# Markdown allows link refs to be defined over two lines:
 	#
-	# 	[id]: http://example.com/longish/path/to/resource/here
+	#	[id]: http://example.com/longish/path/to/resource/here
 	#		"Optional Title Here"
 	#
 	private var last_link_ref: nullable LinkRef = null
@@ -219,12 +307,14 @@ class MarkdownProcessor
 		if value[leading] == '#' then return new LineHeadline
 		if value[leading] == '>' then return new LineBlockquote
 
-		if value.length - leading - trailing > 2 then
-			if value[leading] == '`' and md.count_chars_start('`') >= 3 then
-				return new LineFence
-			end
-			if value[leading] == '~' and md.count_chars_start('~') >= 3 then
-				return new LineFence
+		if ext_mode then
+			if value.length - leading - trailing > 2 then
+				if value[leading] == '`' and md.count_chars_start('`') >= 3 then
+					return new LineFence
+				end
+				if value[leading] == '~' and md.count_chars_start('~') >= 3 then
+					return new LineFence
+				end
 			end
 		end
 
@@ -286,63 +376,72 @@ class MarkdownProcessor
 			c2 = ' '
 		end
 
+		var loc = text.pos_to_loc(pos)
+
 		if c == '*' then
 			if c1 == '*' then
 				if c0 != ' ' or c2 != ' ' then
-					return new TokenStrongStar(pos, c)
+					return new TokenStrongStar(loc, pos, c)
 				else
-					return new TokenEmStar(pos, c)
+					return new TokenEmStar(loc, pos, c)
 				end
 			end
 			if c0 != ' ' or c1 != ' ' then
-				return new TokenEmStar(pos, c)
+				return new TokenEmStar(loc, pos, c)
 			else
-				return new TokenNone(pos, c)
+				return new TokenNone(loc, pos, c)
 			end
 		else if c == '_' then
 			if c1 == '_' then
 				if c0 != ' ' or c2 != ' 'then
-					return new TokenStrongUnderscore(pos, c)
+					return new TokenStrongUnderscore(loc, pos, c)
 				else
-					return new TokenEmUnderscore(pos, c)
+					return new TokenEmUnderscore(loc, pos, c)
+				end
+			end
+			if ext_mode then
+				if (c0.is_letter or c0.is_digit) and c0 != '_' and
+				   (c1.is_letter or c1.is_digit) then
+					return new TokenNone(loc, pos, c)
+				else
+					return new TokenEmUnderscore(loc, pos, c)
 				end
 			end
 			if c0 != ' ' or c1 != ' ' then
-				return new TokenEmUnderscore(pos, c)
+				return new TokenEmUnderscore(loc, pos, c)
 			else
-				return new TokenNone(pos, c)
+				return new TokenNone(loc, pos, c)
 			end
 		else if c == '!' then
-			if c1 == '[' then return new TokenImage(pos, c)
-			return new TokenNone(pos, c)
+			if c1 == '[' then return new TokenImage(loc, pos, c)
+			return new TokenNone(loc, pos, c)
 		else if c == '[' then
-			return new TokenLink(pos, c)
+			return new TokenLink(loc, pos, c)
 		else if c == ']' then
-			return new TokenNone(pos, c)
+			return new TokenNone(loc, pos, c)
 		else if c == '`' then
 			if c1 == '`' then
-				return new TokenCodeDouble(pos, c)
+				return new TokenCodeDouble(loc, pos, c)
 			else
-				return new TokenCodeSingle(pos, c)
+				return new TokenCodeSingle(loc, pos, c)
 			end
 		else if c == '\\' then
 			if c1 == '\\' or c1 == '[' or c1 == ']' or c1 == '(' or c1 == ')' or c1 == '{' or c1 == '}' or c1 == '#' or c1 == '"' or c1 == '\'' or c1 == '.' or c1 == '<' or c1 == '>' or c1 == '*' or c1 == '+' or c1 == '-' or c1 == '_' or c1 == '!' or c1 == '`' or c1 == '~' or c1 == '^' then
-				return new TokenEscape(pos, c)
+				return new TokenEscape(loc, pos, c)
 			else
-				return new TokenNone(pos, c)
+				return new TokenNone(loc, pos, c)
 			end
 		else if c == '<' then
-			return new TokenHTML(pos, c)
+			return new TokenHTML(loc, pos, c)
 		else if c == '&' then
-			return new TokenEntity(pos, c)
-		else if c == '^' then
-			if c0 == '^' or c1 == '^' then
-				return new TokenNone(pos, c)
-			else
-				return new TokenSuper(pos, c)
-			end
+			return new TokenEntity(loc, pos, c)
 		else
-			return new TokenNone(pos, c)
+			if ext_mode then
+				if c == '~' and c1 == '~' then
+					return new TokenStrike(loc, pos, c)
+				end
+			end
+			return new TokenNone(loc, pos, c)
 		end
 	end
 
@@ -365,15 +464,23 @@ end
 # The emitter use a `Decorator` to select the output format.
 class MarkdownEmitter
 
+	# Kind of processor used for parsing.
+	type PROCESSOR: MarkdownProcessor
+
 	# Processor containing link refs.
-	var processor: MarkdownProcessor
+	var processor: PROCESSOR
+
+	# Kind of decorator used for decoration.
+	type DECORATOR: Decorator
 
 	# Decorator used for output.
 	# Default is `HTMLDecorator`
-	var decorator: Decorator = new HTMLDecorator is writable
+	var decorator: DECORATOR is writable, lazy do
+		return new HTMLDecorator
+	end
 
 	# Create a new `MarkdownEmitter` using a custom `decorator`.
-	init with_decorator(processor: MarkdownProcessor, decorator: Decorator) do
+	init with_decorator(processor: PROCESSOR, decorator: DECORATOR) do
 		init processor
 		self.decorator = decorator
 	end
@@ -390,9 +497,7 @@ class MarkdownEmitter
 	fun emit_in(block: Block) do block.emit_in(self)
 
 	# Transform and emit mardown text
-	fun emit_text(text: Text) do
-		emit_text_until(text, 0, null)
-	end
+	fun emit_text(text: Text) do emit_text_until(text, 0, null)
 
 	# Transform and emit mardown text starting at `from` and
 	# until a token with the same type as `token` is found.
@@ -446,7 +551,7 @@ class MarkdownEmitter
 	end
 
 	# Append `e` to current buffer.
-	fun add(e: Streamable) do
+	fun add(e: Writable) do
 		if e isa Text then
 			current_buffer.append e
 		else
@@ -455,18 +560,18 @@ class MarkdownEmitter
 	end
 
 	# Append `c` to current buffer.
-	fun addc(c: Char) do current_buffer.add c
+	fun addc(c: Char) do add c.to_s
 
 	# Append a "\n" line break.
-	fun addn do current_buffer.add '\n'
+	fun addn do add "\n"
 end
 
 # A Link Reference.
 # Links that are specified somewhere in the mardown document to be reused as shortcuts.
 #
-# Example:
-#
-#    [1]: http://example.com/ "Optional title"
+# ~~~raw
+# [1]: http://example.com/ "Optional title"
+# ~~~
 class LinkRef
 
 	# Link href
@@ -489,62 +594,67 @@ end
 # Default decorator used is `HTMLDecorator`.
 interface Decorator
 
+	# Kind of emitter used for decoration.
+	type EMITTER: MarkdownEmitter
+
 	# Render a ruler block.
-	fun add_ruler(v: MarkdownEmitter, block: BlockRuler) is abstract
+	fun add_ruler(v: EMITTER, block: BlockRuler) is abstract
 
 	# Render a headline block with corresponding level.
-	fun add_headline(v: MarkdownEmitter, block: BlockHeadline) is abstract
+	fun add_headline(v: EMITTER, block: BlockHeadline) is abstract
 
 	# Render a paragraph block.
-	fun add_paragraph(v: MarkdownEmitter, block: BlockParagraph) is abstract
+	fun add_paragraph(v: EMITTER, block: BlockParagraph) is abstract
 
 	# Render a code or fence block.
-	fun add_code(v: MarkdownEmitter, block: BlockCode) is abstract
+	fun add_code(v: EMITTER, block: BlockCode) is abstract
 
 	# Render a blockquote.
-	fun add_blockquote(v: MarkdownEmitter, block: BlockQuote) is abstract
+	fun add_blockquote(v: EMITTER, block: BlockQuote) is abstract
 
 	# Render an unordered list.
-	fun add_unorderedlist(v: MarkdownEmitter, block: BlockUnorderedList) is abstract
+	fun add_unorderedlist(v: EMITTER, block: BlockUnorderedList) is abstract
 
 	# Render an ordered list.
-	fun add_orderedlist(v: MarkdownEmitter, block: BlockOrderedList) is abstract
+	fun add_orderedlist(v: EMITTER, block: BlockOrderedList) is abstract
 
 	# Render a list item.
-	fun add_listitem(v: MarkdownEmitter, block: BlockListItem) is abstract
+	fun add_listitem(v: EMITTER, block: BlockListItem) is abstract
 
 	# Render an emphasis text.
-	fun add_em(v: MarkdownEmitter, text: Text) is abstract
+	fun add_em(v: EMITTER, text: Text) is abstract
 
 	# Render a strong text.
-	fun add_strong(v: MarkdownEmitter, text: Text) is abstract
+	fun add_strong(v: EMITTER, text: Text) is abstract
 
-	# Render a super text.
-	fun add_super(v: MarkdownEmitter, text: Text) is abstract
+	# Render a strike text.
+	#
+	# Extended mode only (see `MarkdownProcessor::ext_mode`)
+	fun add_strike(v: EMITTER, text: Text) is abstract
 
 	# Render a link.
-	fun add_link(v: MarkdownEmitter, link: Text, name: Text, comment: nullable Text) is abstract
+	fun add_link(v: EMITTER, link: Text, name: Text, comment: nullable Text) is abstract
 
 	# Render an image.
-	fun add_image(v: MarkdownEmitter, link: Text, name: Text, comment: nullable Text) is abstract
+	fun add_image(v: EMITTER, link: Text, name: Text, comment: nullable Text) is abstract
 
 	# Render an abbreviation.
-	fun add_abbr(v: MarkdownEmitter, name: Text, comment: Text) is abstract
+	fun add_abbr(v: EMITTER, name: Text, comment: Text) is abstract
 
 	# Render a code span reading from a buffer.
-	fun add_span_code(v: MarkdownEmitter, buffer: Text, from, to: Int) is abstract
+	fun add_span_code(v: EMITTER, buffer: Text, from, to: Int) is abstract
 
 	# Render a text and escape it.
-	fun append_value(v: MarkdownEmitter, value: Text) is abstract
+	fun append_value(v: EMITTER, value: Text) is abstract
 
 	# Render code text from buffer and escape it.
-	fun append_code(v: MarkdownEmitter, buffer: Text, from, to: Int) is abstract
+	fun append_code(v: EMITTER, buffer: Text, from, to: Int) is abstract
 
 	# Render a character escape.
-	fun escape_char(v: MarkdownEmitter, char: Char) is abstract
+	fun escape_char(v: EMITTER, char: Char) is abstract
 
 	# Render a line break
-	fun add_line_break(v: MarkdownEmitter) is abstract
+	fun add_line_break(v: EMITTER) is abstract
 
 	# Generate a new html valid id from a `String`.
 	fun strip_id(txt: String): String is abstract
@@ -594,7 +704,11 @@ class HTMLDecorator
 	end
 
 	redef fun add_code(v, block) do
-		v.add "<pre><code>"
+		if block isa BlockFence and block.meta != null then
+			v.add "<pre class=\"{block.meta.to_s}\"><code>"
+		else
+			v.add "<pre><code>"
+		end
 		v.emit_in block
 		v.add "</code></pre>\n"
 	end
@@ -635,10 +749,10 @@ class HTMLDecorator
 		v.add "</strong>"
 	end
 
-	redef fun add_super(v, text) do
-		v.add "<sup>"
+	redef fun add_strike(v, text) do
+		v.add "<del>"
 		v.add text
-		v.add "</sup>"
+		v.add "</del>"
 	end
 
 	redef fun add_image(v, link, name, comment) do
@@ -750,9 +864,31 @@ class HTMLDecorator
 	private var allowed_id_chars: Array[Char] = ['-', '_', ':', '.']
 end
 
+# Location in a Markdown input.
+class MDLocation
+
+	# Starting line number (starting from 1).
+	var line_start: Int
+
+	# Starting column number (starting from 1).
+	var column_start: Int
+
+	# Stopping line number (starting from 1).
+	var line_end: Int
+
+	# Stopping column number (starting from 1).
+	var column_end: Int
+
+	redef fun to_s do return "{line_start},{column_start}--{line_end},{column_end}"
+end
+
 # A block of markdown lines.
 # A `MDBlock` can contains lines and/or sub-blocks.
 class MDBlock
+
+	# Position of `self` in the input.
+	var location: MDLocation
+
 	# Kind of block.
 	# See `Block`.
 	var kind: Block = new BlockNone(self) is writable
@@ -805,7 +941,14 @@ class MDBlock
 
 	# Split `self` creating a new sub-block having `line` has `last_line`.
 	fun split(line: MDLine): MDBlock do
-		var block = new MDBlock
+		# location for new block
+		var new_loc = new MDLocation(
+			first_line.location.line_start,
+			first_line.location.column_start,
+			line.location.line_end,
+			line.location.column_end)
+		# create block
+		var block = new MDBlock(new_loc)
 		block.first_line = first_line
 		block.last_line = line
 		first_line = line.next
@@ -814,6 +957,9 @@ class MDBlock
 			last_line = null
 		else
 			first_line.prev = null
+			# update current block loc
+			location.line_start = first_line.location.line_start
+			location.column_start = first_line.location.column_start
 		end
 		if first_block == null then
 			first_block = block
@@ -1032,6 +1178,9 @@ 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
@@ -1183,6 +1332,9 @@ end
 # A markdown line.
 class MDLine
 
+	# Location of `self` in the original input.
+	var location: MDLocation
+
 	# Text contained in this line.
 	var value: String is writable
 
@@ -1409,10 +1561,10 @@ class LineOther
 		var was_empty = line.prev_empty
 		while line != null and not line.is_empty do
 			var t = v.line_kind(line)
-			if v.in_list and t isa LineList then
+			if (v.in_list or v.ext_mode) and t isa LineList then
 				break
 			end
-			if t isa LineCode or t isa LineFence then
+			if v.ext_mode and (t isa LineCode or t isa LineFence) then
 				break
 			end
 			if t isa LineHeadline or t isa LineHeadline1 or t isa LineHeadline2 or
@@ -1551,7 +1703,8 @@ class LineFence
 		else
 			block = v.current_block.split(v.current_block.last_line.as(not null))
 		end
-		block.kind = new BlockFence(block)
+		var meta = block.first_line.value.meta_from_fence
+		block.kind = new BlockFence(block, meta)
 		block.first_line.clear
 		var last = block.last_line
 		if last != null and v.line_kind(last) isa LineFence then
@@ -1689,7 +1842,10 @@ end
 # Some tokens have a specific markup behaviour that is handled here.
 abstract class Token
 
-	# Position of `self` in markdown input.
+	# Location of `self` in the original input.
+	var location: MDLocation
+
+	# Position of `self` in input independant from lines.
 	var pos: Int
 
 	# Character found at `pos` in the markdown input.
@@ -1899,7 +2055,7 @@ abstract class TokenLinkOrImage
 				comment = lr.title
 			end
 		else
-		var tid = name.write_to_string.replace("\n", " ").to_lower
+			var tid = name.write_to_string.replace("\n", " ").to_lower
 			if v.processor.link_refs.has_key(tid) then
 				var lr = v.processor.link_refs[tid]
 				link = lr.link
@@ -2038,17 +2194,19 @@ class TokenEscape
 	end
 end
 
-# A markdown super token.
-class TokenSuper
+# A markdown strike token.
+#
+# Extended mode only (see `MarkdownProcessor::ext_mode`)
+class TokenStrike
 	super Token
 
 	redef fun emit(v) do
 		var tmp = v.push_buffer
-		var b = v.emit_text_until(v.current_text.as(not null), pos + 1, self)
+		var b = v.emit_text_until(v.current_text.as(not null), pos + 2, self)
 		v.pop_buffer
 		if b > 0 then
-			v.decorator.add_super(v, tmp)
-			v.current_pos = b
+			v.decorator.add_strike(v, tmp)
+			v.current_pos = b + 1
 		else
 			v.addc char
 		end
@@ -2163,6 +2321,7 @@ redef class Text
 	# Safe mode can be activated to limit reading to valid xml.
 	private fun read_xml(out: FlatBuffer, start: Int, safe_mode: Bool): Int do
 		var pos = 0
+		var is_valid = true
 		var is_close_tag = false
 		if start + 1 >= length then return -1
 		if self[start + 1] == '/' then
@@ -2180,7 +2339,11 @@ redef class Text
 			pos = read_xml_until(tmp, pos, ' ', '/', '>')
 			if pos == -1 then return -1
 			var tag = tmp.write_to_string.trim.to_lower
-			if tag.is_html_unsafe then
+			if not tag.is_valid_html_tag then
+				out.append "&lt;"
+				pos = -1
+			else if tag.is_html_unsafe then
+				is_valid = false
 				out.append "&lt;"
 				if is_close_tag then out.add '/'
 				out.append tmp
@@ -2203,7 +2366,11 @@ redef class Text
 			if pos == -1 then return -1
 		end
 		if self[pos] == '>' then
-			out.add '>'
+			if is_valid then
+				out.add '>'
+			else
+				out.append "&gt;"
+			end
 			return pos
 		end
 		return -1
@@ -2275,6 +2442,14 @@ redef class Text
 		return tpl.write_to_string.to_lower
 	end
 
+	private fun is_valid_html_tag: Bool do
+		if is_empty then return false
+		for c in self do
+			if not c.is_alpha then return false
+		end
+		return true
+	end
+
 	# Read and escape the markdown contained in `self`.
 	private fun escape(out: FlatBuffer, c: Char, pos: Int): Int do
 		if c == '\\' or c == '[' or c == ']' or c == '(' or c == ')' or c == '{' or
@@ -2288,6 +2463,35 @@ redef class Text
 		return pos
 	end
 
+	# Extract string found at end of fence opening.
+	private fun meta_from_fence: nullable Text do
+		for i in [0..chars.length[ do
+			var c = chars[i]
+			if c != ' ' and c != '`' and c != '~' then
+				return substring_from(i).trim
+			end
+		end
+		return null
+	end
+
+	# Init a `MDLocation` instance at `pos` in `self`.
+	private fun pos_to_loc(pos: Int): MDLocation do
+		assert pos <= length
+		var line = 1
+		var col = 0
+		var i = 0
+		while i <= pos do
+			col += 1
+			var c = self[i]
+			if c == '\n' then
+				line +=1
+				col = 0
+			end
+			i +=1
+		end
+		return new MDLocation(line, col, line, col)
+	end
+
 	# Is `self` an unsafe HTML element?
 	private fun is_html_unsafe: Bool do return html_unsafe_tags.has(self.write_to_string)
 
@@ -2311,7 +2515,7 @@ redef class String
 	#    var md = "**Hello World!**"
 	#    var html = md.md_to_html
 	#    assert html == "<p><strong>Hello World!</strong></p>\n"
-	fun md_to_html: Streamable do
+	fun md_to_html: Writable do
 		var processor = new MarkdownProcessor
 		return processor.process(self)
 	end