Revise convert_project.py

[mattmccutchen-cci]

In today's status meeting, someone (Mike?) reported a bunch of behavior of convert_project.py that seemed inappropriate, at least to newcomers to 3C. We decided that someone who hasn't previously used convert_project.py should run it and try to evaluate it from a newcomer's perspective to make a list of what we want to change, and I offered to do so. Mike is to send me some guidance about where to start.

[mwhicks1]

Here are things that I saw:

• The default usage of script adds arguments that we don't want: -dump-stats and -dump-intermediate. We would like -alltypes to be included as an argument, in most cases.
• The parameters that specify the paths of 3c and the Checked C includes files are a little surprising/irritating. I wonder if we could infer what these things are, e.g., just take them from the current environment PATH and CPATH.
• I'm not convinced that we want to immediately run the script after generating it. I think that should be the exception, not the rule as it is now.
• Right now the script assumes that a compilation database is present in the project directory. This perhaps is as it should be. But I also wonder if that DB should be passed as an argument; I expect most programs won't have it there.