Account for code spans at start of line.

Previously we didn't test for this because we did not allow raw blocks to have any indent. Now, that we allow up to 3 spaces of indent, we need to confirm those 3 chars are actually whitespace. Tests added. Note: the case where the code span is not on the first line is still failing. I havn't worked out why but ran out of time today.

Author: waylan

markdown/htmlparser.py

@@ -65,11 +65,30 @@ def close(self):
  self.cleandoc.append(self.md.htmlStash.store(''.join(self._cache)))
  self._cache = []
 
+ @property
+ def line_offset(self):
+ """Returns char index in self.rawdata for the start of the current line. """
+ if self.lineno > 1:
+ return re.match(r'([^\n]*\n){{{}}}'.format(self.lineno-1), self.rawdata).end()
+ return 0
+
+ def at_line_start(self):
+ """
+ Returns True if current position is at start of line.
+
+ Allows for up to three blank spaces at start of line.
+ """
+ if self.offset == 0:
+ return True
+ if self.offset > 3:
+ return False
+ # Confirm up to first 3 chars are whitespace 
+ return self.rawdata[self.line_offset:self.offset].strip() == ''
+
  def handle_starttag(self, tag, attrs):
  self.stack.append(tag)
 
- line, col = self.getpos()
- if col < 4 and self.md.is_block_level(tag) and not self.inraw:
+ if self.at_line_start() and self.md.is_block_level(tag) and not self.inraw:
  # Started a new raw block
  self.inraw = True
  if len(self.cleandoc):
@@ -84,12 +103,7 @@ def handle_starttag(self, tag, attrs):
 
  def handle_endtag(self, tag):
  # Attempt to extract actual tag from raw source text
- if self.lineno > 1:
- # Find start position: char index for end of line at self.lineno + self.offset
- start = re.match(r'([^\n]*\n){{{}}}'.format(self.lineno-1), self.rawdata).end() + self.offset
- else:
- # On first line. Just use self.offset for start position.
- start = self.offset
+ start = self.line_offset + self.offset
  m = parser.endendtag.search(self.rawdata, start)
  if m:
  text = self.rawdata[start:m.end()]
@@ -122,11 +136,10 @@ def handle_data(self, data):
 
  def handle_empty_tag(self, data, is_block):
  """ Handle empty tags (`<data>`). """
- line, col = self.getpos()
  if self.inraw:
  # Append this to the existing raw block
  self._cache.append(data)
- elif col < 4 and is_block:
+ elif self.at_line_start() and is_block:
  # Handle this as a standalone raw block
  self.cleandoc.append(self.md.htmlStash.store(data))
  # Insert blank line between this and next line.

tests/test_syntax/blocks/test_html_blocks.py

@@ -74,8 +74,8 @@ def test_raw_span(self):
 
  def test_code_span(self):
  self.assertMarkdownRenders(
- '`<em>code span</em>`',
- '<p><code>&lt;em&gt;code span&lt;/em&gt;</code></p>'
+ '`<p>code span</p>`',
+ '<p><code>&lt;p&gt;code span&lt;/p&gt;</code></p>'
  )
 
  def test_raw_empty(self):
@@ -227,6 +227,25 @@ def test_raw_surrounded_by_text_without_blank_lines(self):
  )
  )
 
+ # TODO: Fix this. Not sure why its failing...
+ def test_multiline_markdown_with_code_span(self):
+ self.assertMarkdownRenders(
+ self.dedent(
+ """
+ A paragraph with a block-level
+ `<p>code span</p>`, which is
+ at the start of a line.
+ """
+ ),
+ self.dedent(
+ """
+ <p>A paragraph with a block-level
+ <code>&lt;p&gt;code span&lt;/p&gt;</code>.
+ More <em>Markdown</em> text.</p>
+ """
+ )
+ )
+
  # Note: The blank line between the tags is a change in behavior.
  def test_raw_one_line_followed_by_text(self):
  self.assertMarkdownRenders(
@@ -645,6 +664,12 @@ def test_raw_comment_one_line_with_tag(self):
  '<!-- <tag> -->'
  )
 
+ def test_comment_in_code_span(self):
+ self.assertMarkdownRenders(
+ '`<!-- *foo* -->`',
+ '<p><code>&lt;!-- *foo* --&gt;</code></p>'
+ )
+
  # Note: this is a change in behavior for Python-Markdown only in that a blank line is added.
  # While it does not match the reference implementation, there is no difference in rendering.
  def test_raw_comment_one_line_followed_by_text(self):