xmlprettify can cause mangled output

[fridgecow]

During chapter loads, xmlprettify is called to format the output nicely, and in doing so it strips the text and tail attributes from elements. Unfortunately this can have the unintended consequence of producing mangled epubs from reasonable HTML.

For example, this HTML:

<div>1234 <i>5</i> 6789</div>

should produce output exactly like the input,

<div>1234 <i>5</i> 6789</div>

but actually looks like:

  <div>1234
  <i>5</i>6789
</div>

Which will be rendered differently since there's no space after the 5.

Removing the xmlprettify call from Chapter._render makes the output correct again.

[clach04]

From bs4 docs, https://www.crummy.com/software/BeautifulSoup/bs4/doc/#pretty-printing, prettify() is for debugging purposes only.

Since it adds whitespace (in the form of newlines), prettify() changes the meaning of an HTML document and should not be used to reformat one. The goal of prettify() is to help you visually understand the structure of the documents you work with.

https://github.com/search?q=repo%3Awcember%2Fpypub%20%20%20prettify&type=code