notebook stores HTML escaped text in the file

[minrk]

The notebook frontend escapes various plaintext things for HTML output, but it shouldn't save the transformed versions to files, which it does currently.

This code:

In [1]: u'hi'
Out[1]: u'hi'

In []: |

Is stored in the ipynb as:

    {
     "cell_type": "code", 
     "collapsed": false, 
     "input": [
      "u'hi'"
     ], 
     "language": "python", 
     "outputs": [
      {
       "output_type": "pyout", 
       "prompt_number": 1, 
       "text": [
        "u'hi'"
       ]
      }
     ], 
     "prompt_number": 1
    }, 
    {
     "cell_type": "code", 
     "collapsed": true, 
     "input": [], 
     "language": "python", 
     "outputs": [], 
     "prompt_number": " "
    }

There should be zero html escapes in .ipynb files outside explicit HTML fields.

[minrk]

The issue appears to be the use of utils.fixConsole when handling output. We clobber the actual data with the html-escaped version, which we shouldn't do. We should preserve the original, and never persist the escaped version.

[ellisonbg]

This is now fixed.