Python Version Compatibility about pypdf

[mstamy2]

'unicode' as a keyword does not exist in python 3, as strings are Unicode. That change needs to be made.

[mstamy2]

extractText Issue #17 appears to be related. The method works relatively smoothly, but fails often in Python 3

[mstamy2]

The goal for branch Python3-3 is for all features of PyPDF2 to be compatible with all python versions from 2.5 - 3.3 (and soon 3.4). This may make the code a little 'less clean'. I pushed a few experimental changes in commit df5d5f2. The main error that persists is that PyPDF2 is unable to find the endstream marker when parsing streams. (In Python 3+). Also, we need to merge this branch with PyPDF2 1.18, which will cause many merge conflicts in pdf.py.

[mstamy2]

This project is nearly finished (although support for Python 2.5 is dropped). The only major compatibility issue that remains is that of encryption, which is listed in a separate issue.