tags inside a tag, and individually do not form valid lists.
+
+ This function iterates through the Markdown tree and generates
+ self-contained lines of HTML that can be rendered individually.
+ """
+ nodes = get_markdown_element_tree(markdown_html)
+
+ for node in nodes:
+ if node.nodeType == node.ELEMENT_NODE:
+ if (node.tagName == 'div' and
+ node.attributes.get('class', 'codehilite')):
+ # This is a code block, which will consist of a bunch of lines
+ # for the source code. We want to split that up into
+ # individual lines with their own
tags.
+ for line in node.toxml().splitlines():
+ yield '%s
' % line
+ elif node.tagName in ('ul', 'ol'):
+ # This is a list. We'll need to split all of its items
+ # into individual lists, in order to retain bullet points
+ # or the numbers.
+ #
+ # For the case of numbers, we can set each list to start
+ # at the appropriate number so that they don't all say "1."
+ i = node.attributes.get('start', 1)
+
+ for child_node in node.childNodes:
+ if (child_node.nodeType == child_node.ELEMENT_NODE and
+ child_node.tagName == 'li'):
+ # This is a list item element. It may be multiple
+ # lines, but we'll have to treat it as one line.
+ yield '<%s start="%s">%s' % (
+ node.tagName, i, child_node.toxml(),
+ node.tagName)
+
+ i += 1
+ elif node.tagName == 'p':
+ # This is a paragraph, possibly containing multiple lines.
+ for line in node.toxml().splitlines():
+ yield line
+ else:
+ # Whatever this is, treat it as one block.
+ yield node.toxml()
+ elif node.nodeType == node.TEXT_NODE:
+ # This may be several blank extraneous blank lines, due to
+ # Markdown's generation from invisible markup like fences.
+ # We want to condense this down to one blank line.
+ yield '\n'
+
+
+def get_markdown_element_tree(markdown_html):
+ """Returns an XML element tree for Markdown-generated HTML.
+
+ This will build the tree and return all nodes representing the rendered
+ Markdown content.
+ """
+ doc = parseString('%s' % markdown_html)
+ return doc.childNodes[0].childNodes
+
+
+def render_markdown(text):
+ """Renders Markdown text to HTML.
+
+ The Markdown text will be sanitized to prevent injecting custom HTML.
+ It will also enable a few plugins for code highlighting and sane lists.
+ """
+ return markdown(text, **MARKDOWN_KWARGS)
+
+
+def render_markdown_from_file(f):
+ """Renders Markdown text to HTML.
+
+ The Markdown text will be sanitized to prevent injecting custom HTML.
+ It will also enable a few plugins for code highlighting and sane lists.
+ """
+ s = StringIO()
+ markdownFromFile(input=f, output=s, **MARKDOWN_KWARGS)
+ html = s.getvalue()
+ s.close()
+
+ return html
diff --git a/reviewboard/reviews/ui/markdownui.py b/reviewboard/reviews/ui/markdownui.py
index 45dda0b35fdcabcac8655498534240d503a69905..463472f8f14c0525cfb63bb36245e690f5410cc6 100644
--- a/reviewboard/reviews/ui/markdownui.py
+++ b/reviewboard/reviews/ui/markdownui.py
@@ -3,10 +3,11 @@ from __future__ import unicode_literals
import logging
from xml.dom.minidom import parseString
-from djblets.util.compat.six.moves import cStringIO as StringIO
-import markdown
+from pygments.lexers import TextLexer
from reviewboard.reviews.ui.text import TextBasedReviewUI
+from reviewboard.reviews.markdown_utils import (iter_markdown_lines,
+ render_markdown_from_file)
class MarkdownReviewUI(TextBasedReviewUI):
@@ -24,31 +25,16 @@ class MarkdownReviewUI(TextBasedReviewUI):
js_view_class = 'RB.MarkdownReviewableView'
def generate_render(self):
- buffer = StringIO()
- self.obj.file.open()
- markdown.markdownFromFile(input=self.obj.file, output=buffer,
- output_format='xhtml1', safe_mode='escape',
- extensions=['fenced_code', 'codehilite'])
- rendered = buffer.getvalue()
- buffer.close()
- self.obj.file.close()
+ with self.obj.file as f:
+ f.open()
+ rendered = render_markdown_from_file(f)
try:
- doc = parseString('%s' % rendered)
- main_node = doc.childNodes[0]
-
- for node in main_node.childNodes:
- for html in self._process_markdown_html(node):
- yield html
+ for line in iter_markdown_lines(rendered):
+ yield line
except Exception as e:
logging.error('Failed to parse resulting Markdown XHTML for '
'file attachment %d: %s' % (self.obj.pk, e))
- def _process_markdown_html(self, node):
- if (node.nodeType == node.ELEMENT_NODE and
- node.tagName == 'div' and
- node.attributes.get('class', 'codehilite')):
- for line in node.toxml().splitlines():
- yield '%s
' % line
- else:
- yield node.toxml()
+ def get_source_lexer(self, filename, data):
+ return TextLexer()
diff --git a/reviewboard/reviews/ui/text.py b/reviewboard/reviews/ui/text.py
index b0fe1425eb30332906a7b5dd4c9986bbd78c5657..dc48ab1bf4ebf9993e632e7c8c7e4bf1226f5805 100644
--- a/reviewboard/reviews/ui/text.py
+++ b/reviewboard/reviews/ui/text.py
@@ -91,11 +91,7 @@ class TextBasedReviewUI(FileAttachmentReviewUI):
data = self.obj.file.read()
self.obj.file.close()
- try:
- lexer = guess_lexer_for_filename(self.obj.filename, data)
- except ClassNotFound:
- lexer = TextLexer()
-
+ lexer = self.get_source_lexer(self.obj.filename, data)
lines = highlight(data, lexer, NoWrapperHtmlFormatter()).splitlines()
return [
@@ -103,6 +99,19 @@ class TextBasedReviewUI(FileAttachmentReviewUI):
for line in lines
]
+ def get_source_lexer(self, filename, data):
+ """Returns the lexer that should be used for the text.
+
+ By default, this will attempt to guess the lexer based on the
+ filename, falling back to a plain-text lexer.
+
+ Subclasses can override this to choose a more specific lexer.
+ """
+ try:
+ return guess_lexer_for_filename(filename, data)
+ except ClassNotFound:
+ return TextLexer()
+
def generate_render(self):
"""Generates a render of the text.