Module and Function Dependencies

Concerning terminology

We say that a module A is dependent on another module B (A -> B) if there is at least one function call from A to B.
A cyclic dependency appears, when B is also dependent directly (B -> A) or indirectly (e.g. B -> C -> A) from A.
Note that it is possible to have a cyclic dependency among the modules while having no cyclic dependencies among the functions.
For example, a function call from A:foo to B:foo, and from B:foo2 to A:foo2 implies a cyclic dependency on the module level.
If one wants to have a deeper analysis and pays more attention to the concerning functions, a function level query should be done.
In our previous example, no function level cycle appears, unless A:foo calls B:foo, and B:foo calls A:foo.

The following examinations can be done considering dependencies:

• Checking whether there are cycles, if so, listing them out
• Printing out the cycles, meaning the modules/functions will not be rep-

resented by their proper graph node, but with their names
(for instance, instead of {'$gn', module,3} , module test will be printed, if the graph node stands for that exact module).

• Checking for cycle from one or more nodes as starting points
• Drawing the dependency graph
• Drawing the dependency graph from a starting node
• Drawing the cyclic part of the dependency graph if one exists (you can

also give a cyclic node as a starting node)

Dependency analysis can be done in two ways: by the means of ri interface or using the web interface.
To call the desired query, the user should give a proplist, stating the different requirements.
The two interface functions are:

1. ri:draw_dep/1 - for drawing
2. ri:print_dep/1 - for listing, printing out to the standard output
   

The options and the keys for the functions are:

• {level, Level}| Level = mod | func Stating the level of the query, module or function level.
• {type, Type} Type = all | cycles}

The investigation should be done on the whole graph/table, or just on the

cycle part (if it exists).
When listing out the cycles, "all" gives back the result in their graph node form, while "cycles" returns with their proper names (see examples above).