Fauxton: do not slice off the error text from erlang errors

[robertkowalski]

As the erlang errors contain two linebreaks, and the parsing
function is using linebreaks to separate the messages, the content
was sliced off, which looked like there would not be enough space
for the message. In fact it got lost during the parsing.

This tries to normalize the data before and after the parsing,
it removes the linebreaks of the erlang errors so it is parsed
right. As the other spaces are no &nspb;-entities it does not hurt
that the are removed.

Before:

Now:

[deathbearbrown]

+1 :3

[garrensmith]

+1 @robertkowalski, since you are a committer you can merge this in yourself :)

[robertkowalski]

@deathbearbrown already merged for me, thank you! thanks for taking a look, highly appreciated! :)