test/regression.txt case seems to have been accidentally neutered since cmark fork

Looks like the regression test has itself had a regression, most likely due to someone's git/editor/.editorconfig settings stripping extra control chars from ends of line. In the following test case, those control chars are critical:

cmark-gfm/test/regression.txt

Lines 3 to 13 in c32ef78

    
           Issue #113: EOL character weirdness on Windows 
        
           (Important: first line ends with CR + CR + LF) 
        
           ```````````````````````````````` example 
        
           line1 
        
           line2 
        
           . 
        
           <p>line1</p> 
        
           <p>line2</p> 
        
           ````````````````````````````````

The change happens here in the history: b6e462f#r103721889

You can also diff against the source:

https://github.com/commonmark/cmark/blob/7195c6735f29be947ddc41f86c9ddfc8621d33b9/test/regression.txt#L3-L12

I will submit a PR shortly.

	Issue #113: EOL character weirdness on Windows
	(Important: first line ends with CR + CR + LF)

	```````````````````````````````` example
	line1

	line2
	.
	<p>line1</p>
	<p>line2</p>
	````````````````````````````````