CRLF regression with fenced code

[btrask]

Thank you for making quick work of #68, but I'm afraid that's not the end of it.

Example input:

```
test
```

Output:

<pre><code>test
```</code></pre>

This happens with or without the --hardbreaks option.

[jgm]

Can't reproduce.

% perl -pe 's/\n/\r\n/g' | ./cmark
```
test
```
^D
<pre><code>test
</code></pre>

[jgm]

Did your input have no line ending character (cr, lf, crlf) after the final line?
With that I can reproduce.
That's because the close-fence scanner is looking for a final line ending character.
There used to be some code that added a line ending character to a line if none was present; I assume that's what got messed up in the last batch of changes.

[btrask]

Oh, you're right. Sorry for jumping to conclusions!

I briefly tested a few other things and I didn't find any other problems. Thank you!

[jgm]

I found two more issues by modifying line endings in the
test suite. See #73.

+++ Ben Trask [Aug 08 15 13:31 ]:

Oh, you're right. Sorry for jumping to conclusions!

I briefly tested a few other things and I didn't find any other
problems. Thank you!

—
Reply to this email directly or [1]view it on GitHub.

References

1. CRLF regression with fenced code #71 (comment)

[jgm]

See also #72.

[btrask]

Great work! Thanks for making line endings a priority. :)