Moved line-endings cleanup from transform to convert method so it runs prior to textPreprocessors. Raw HTML with CRLF line endings now works properly. Also added a test. Fixes [1908691].

3 files changed, 16 insertions, 7 deletions

diff --git a/markdown.py b/markdown.py
index 97038ba..3dc8f00 100644
--- a/markdown.py
+++ b/markdown.py
@@ -1282,16 +1282,10 @@ class Markdown:
         self.top_element.setAttribute('class', 'markdown')
         self.doc.appendChild(self.top_element)
 
-        # Fixup the source text
-        text = self.source
-        text = text.replace("\r\n", "\n").replace("\r", "\n")
-        text += "\n\n"
-        text = text.expandtabs(TAB_LENGTH)
-
         # Split into lines and run the preprocessors that will work with
         # self.lines
 
-        self.lines = text.split("\n")
+        self.lines = self.source.split("\n")
 
         # Run the pre-processors on the lines
         for prep in self.preprocessors :
@@ -1727,6 +1721,11 @@ class Markdown:
             message(CRITICAL, 'UnicodeDecodeError: Markdown only accepts unicode or ascii  input.')
             return u""
 
+        # Fixup the source text
+        self.source = self.source.replace("\r\n", "\n").replace("\r", "\n")
+        self.source += "\n\n"
+        self.source = self.source.expandtabs(TAB_LENGTH)
+
         for pp in self.textPreprocessors:
             self.source = pp.run(self.source)
 
diff --git a/tests/misc/CRLF_line_ends.html b/tests/misc/CRLF_line_ends.html
new file mode 100644
index 0000000..8e6849f
--- /dev/null
+++ b/tests/misc/CRLF_line_ends.html
@@ -0,0 +1,5 @@
+<p>foo
+</p>
+<div>
+bar
+</div>
\ No newline at end of file
diff --git a/tests/misc/CRLF_line_ends.txt b/tests/misc/CRLF_line_ends.txt
new file mode 100644
index 0000000..7d5e324
--- /dev/null
+++ b/tests/misc/CRLF_line_ends.txt
@@ -0,0 +1,5 @@
+foo

+

+<div>

+bar

+</div>