Missing block in output replaced with md5-stuffstuffstuff

[mingp]

We have been using the django-markdown-deux markdown Django template tag (and therefore indirectly python-markdown2) occasionally throughout our Django app.

Today, we added a new usage of the markdown Django template tag and saw some unusual behavior in the output rendering. A large block of text/formatting had been removed and replaced with md5-013e64a4a229d1926c2bd523e0c5d77b. Please see attached input and output.

We have found a temporary workaround by including an extra paragraph, containing only  , where the cutoff would otherwise occur.

Possibly related to #127. We're seeing the same symptoms but from different input cases.

Appreciate if you have any insight into this issue. Thanks in advance!

Relevant excerpt of Markdown input: https://www.refheap.com/025bf00f2235b9c7e2601a43a
Relevant excerpt of rendered HTML output: https://www.refheap.com/3eac75d0a0d170170997ed318

[nicholasserra]

Hello, thanks for the bug report. I'll check it out. Pull requests are also welcome if you think you have a solution for this. Even issuing a PR with some example failing tests would be helpful. Thanks!

[william-index]

Note on this and note to self:
Line 147 in markdown2.py is currently

return 'md5-' + md5(SECRET_SALT + s.encode("utf-8")).hexdigest()

This is not being unhashed where it needs to be it seems?
I fixed the problem locally for my purposes by replacing with

return s.encode("utf-8")

But I suspect this has other larger repercussions and will introduce more issues along the way.

[ZuluPro]

I had the same problem with <img>'s src attribute,

All _ was replaced by md5-stuffstuffstuff
My quick fix is the following, I subclass markdown2.Markdown and in postprocess:

for key, value in markdown2.g_escape_table.iteritems():
    text = text.replace(value, key)

markdown2.g_escape_table is g_escape_table = dict([(ch, _hash_text(ch) )for ch in '\\*_{}>#+-.!'])`

So you can unhash the escaped characters.