be able to update formula, etc.

in lieu of the hard-coded dropbox csv files, I'm thinking about hosting them on github itself instead so that we can reference to themselves on github for the testing.

The path within .gitmodules points to public/jasmine, which doesn't exist, this prevents fresh clones from initialising the submodule. The correct path is lib/jasmine

Github pages doesn't load, I tested this on a fork and it fixed things, http://pld.github.io/bamboojs/

@larryweya can I go ahead and do this?

The link in the web page no longer exists. The example also leaves something to be desired since there's very little output. I used a cached good_eats CSV and this flow worked for me,

var dataset = new bamboo.Dataset({url: 'http://formhub.org/mberg/exports/good_eats/csv/good_eats_2013_03_23_13_25_14.csv'});

var dataset = new bamboo.Dataset({id: dataset.load_from_url()['id']});
dataset.query_info();
dataset.info;

but I'm sure it could be more literate.

Right now, the test works by querying a hard-coded bamboo dataset id, and checking that it exists. It has been fixed in the past by simply updating the hardcoded dataset id.

I think this is a bad test; it depends on an external environment.

Proposal to fix it:

• first test that dataset creation using csv works
• test that getting an id of a dataset created using csv works
• then, for a dataset created using a csv, get the id, and create a new dataset object using that id. test that this second object is defined, its info is queryable, etc.

ref: https://github.com/modilabs/bamboo_js/blob/f928292749f6bdeef00fe51503e409bbb2841a6a/src/spec.coffee#L20

See SEL-Columbia/pybamboo#27 and the discussion. Obviously js will be slightly different based on async, but bamboo clients should be in sync as much as possible.