The control flow

The cfperl control flow is summarized briefly in the POD section of cfperl itself.

        
First, a top-level parser is applied, preprocessing everything into a
cfrun queue (a queue of actions, tagged with a section and some classes).

Second, the 'import', 'control' and 'groups' sections are parsed (first
the imports, then the control and group statements).  The actionsequence
(called a cfqueue) is defined. 

Third, the remaining statements (everything not processed in the second
step) are processed. 

The functions that handle these steps are:

• Top-level parser: parse_line()
• Import: load_file()
• Control and groups: dispatch()
• All other sections: dispatch()

All these steps except the top-level parsing are done inside cfrun(). In effect, cfrun() is the actual interpreter, whereas parse_line() is the preparation phase that sorts out sections and imports. Imports have to be done before interpretation, because their contents may affect interpretation.

Note that the documentation of the control flow is inside the program itself! It is stated in fairly simple terms, which should be understandable to someone experienced with cfengine.