Initial documentation

[arporter]

I've incorporated Pearu's old documentation in the new structure. I've also added extra auto-generated documentation using sphinx.

[coveralls]

Coverage remained the same at 80.286% when pulling 8f8368e on initial_doc into 79dfb7e on master.

[coveralls]

Coverage remained the same at 80.286% when pulling f957bef on initial_doc into 79dfb7e on master.

[arporter]

Just a note to say that I used sphinx-quickstart to re-generate the Makefile and conf.py files for the documentation which is why there are so many differences.

[coveralls]

Coverage remained the same at 80.286% when pulling 211a43d on initial_doc into 79dfb7e on master.

[rupertford]

I'm afraid I don't like the documentation as it is. It is simply a list of all the classes and methods and their interfaces which is not going to be much use to most people.

What I think would be useful in the first instance is a simple introduction on how to parse a fortran code using fparser ...

1. a quick overview section (which you pretty much have)

2. a section which shows how to parse a program, output the code again, see the internal structure and walk the tree, all using the "default" parser. Most of this can be generated automatically from the doc strings of the appropriate classes, it just needs to be pulled together with some additional text.

3. A section repeating the above for the f2003 parser.

4. Optionally an appendix with all the current autogenerated API stuff.

[coveralls]

Coverage increased (+0.6%) to 80.924% when pulling 65c8922 on initial_doc into 79dfb7e on master.

[coveralls]

Coverage increased (+0.6%) to 80.924% when pulling 9dbd469 on initial_doc into 79dfb7e on master.

[coveralls]

Coverage increased (+0.6%) to 80.924% when pulling 782943d on initial_doc into 79dfb7e on master.

[arporter]

I've removed all auto-generated content and re-structured the document slightly to give the Fortran66-90 and Fortran2003 parsers their own sections. The latter is rather sparse as I have to confess I got habakkuk working on a trial-and-error basis and don't have a deep understanding of the workings of the parser.
Ready for second review...

[coveralls]

Coverage increased (+0.6%) to 80.924% when pulling 0c70be1 on initial_doc into 79dfb7e on master.

[rupertford]

It's better now and I think it's good enough as an initial document so am approving.

However, for the record, I think it still reads mostly like a reference guide as it is based around the structure of the python packages. I think would be improved by starting from the perspective of what users would want to do i.e. how to parse fortran code (as a section), then how to iterate over the tree (as a section).

Some minor points you can ignore if you want ...

I don't see why we include information about CHAR_BIT in the api.py file as it is not referenced or explained anywhere.

I think the names of the sections for the two parsers should be made consistent. I think renaming the first "Parser for Fortran66..." should be renamed to "Fortran66... parser".