Andy Hayden

Which is to say, this PR also fixes #10.

@UnknownProgrammer I don't follow what your advocating here going forward? (and whether it fixes #10?) > but you have to know the encoding of the byte string to have success....

If you're using non-unicode in a utf-coded document isn't that a bug in the python file? This is the bit I don't quite understand: why is non-valid unicode ending up...

It's unclear if the stderr part is https://github.com/LightTable/Python/blob/134751c32a86699c06ccafe46ae73a17a0e2d012/py-src/ltipy.py#L222 or https://github.com/LightTable/Python/blob/134751c32a86699c06ccafe46ae73a17a0e2d012/py-src/ltipy.py#L199 (those are my guesses). The print_function bug is different... scoping of exec? Edit: I can reproduce this locally: ``` In...

This could probably just be in a try-except block (i.e. not mention `unicode`).

@kenny-evitt I think that encode is actually incorrect here. it should **decode**, otherwise if you pass bytes it becomes u"b'string'" which I don't think is not what you want?

@goulu since this is never going to be included in pdfminer, can we/you fork this project and push it to pypi. I think it would make sense for [pattern](github.com/pattern3/pattern) to...

In porting pattern, I used [pdfminer for python 2 and pdfminer3k for python 3](https://github.com/pattern3/pattern/blob/0de58651e5308cb7b1408738a4dd4d60cfc54cec/setup.py#L22), I needed the following few lines for compat: https://github.com/pattern3/pattern/blob/master/pattern/web/__init__.py#L4234-L4255. ~~It could be a fairly light wrapper...

An easier solution may be to just have a lightweight wrapper of both pdfminer and pdfminer3k, normalizing the api. Happy to put one together quickly now and see if it...

I just saw @goulu pushed this branch to pypi as [pdfminer.six](https://github.com/goulu/pdfminer)... and it's working now, great!. Since pdfminer3k hasn't been updated in a few years, perhaps pushing people towards using...