fix spelling

[schettino72]

on doit docs:

• the spell checker uses a dictionary.txt file where you can put words that you want to spell checker to ignore.
• I also put most python code in separate files, so the spell checker is not run on code. (also advantage of being able to run pyflakes on docs snippets.

[vlcinsky]

I am aware of spelling not being perfect. My priority was to collect ideas and draft the outline first and spelling was getting too often into my way.

But the time for proper spelling comes. I will modify it.

[vlcinsky]

Fixed all spelling errors.

However, i am not sure, if I did it properly as I simply added into dictionary.txt all what was acceptable but in many cases there were:

• thing from code
• split keywords, e,g. "dep" originating from "file_dep"

If there is any way to mark these things better or in place, I would try, otherwise we can close it.

[vlcinsky]

I think I got an idea for spelling solution.

https://pypi.python.org/pypi/sphinxcontrib-spelling/2.3.0

It shall be aware of complete documentation structure, code, other text components, and even allow listing special acceptable terms within rst files.

I have used it in past (with little problem to spell Czech text, but fixed that). I will try.

[vlcinsky]

Spelling modified to use sphinxcontrib.spelling.

However, it requires package libenchant1c2a to be installed. On ubuntu I usually do: sudo apt-get install libenchant1c2a but I have no idea, where to put it into ci-circle configuration file.

[vlcinsky]

Differences from original spelling solution:

• sphinx one better recognize free text and special constructs (as code listing, keywords, variables) so it does not report so many false positives. dictionary.txt shrinked from 96 lines down to 17.
• it has a bit different dependencies (I think).
• it is run as new target. Some instructions say to use $ make spell but I it fails on my box.
• spelling is now managed via sphinx-build -b spell as shown in dodo.py. It works well for me.

conf.py currently declares spelling_language="en_GB" as this is my preferred one (over US).