Set of syntax rules and a PL to allow for easy writing of test procedures.


Header: value of any kind
Other-header: name1=value1; name2=value2
Another-one: something else

After an empty line, body starts.  (Just like MIME.)  The body is
basically just Markdown with some modifications discussed later.

#case name#
1. instruction
   > expected result (or check point, or oracle)
   > another expected result {fam>10} {os~win*}
2. cond. instruction
   * yes: opt 1
   * no: opt 2
0. Assertion (specifically zero, can be anywhere)
3. another instruction


Basic form is similar to MIME. First headers, then, after one empty line, the body, up to the EOF. Just like in MIME, headers are only for meta-data--they are not visible in the final output.

Inside header, various parameter ranges like OSes, architectures, product editions or hardware types can be declared. Later, presence of entities (e.g. instructions) can be conditioned using these.

This is also meant to allow for generation of set of ready procedures with variations of parameters.


  • if assertion -or- instruction fails

    • IN is generated
    • rest of test case is skipped
  • if expected result fails

    • IN is generated
    • TC goes on


  • IN - incident note

    • note itself
    • bug report against tested subject?
    • bug report against test documentation?
  • TC - test case

  • ST - sugar trail (trail of test cases)


See examples/