[antlr-interest] JUnit testing of ASTs

[stefan]

On Tuesday 21 December 2004 20:20, Paul J. Lucas wrote:
> 	Does anybody do JUnit regression testing on their parsers?  If
> 	so, how, exactly?
>
> 	For lack of anything better, *a* way of doing it is to have a
> 	set of files to parse and the pretty-printed output of the
> 	expected ASTs.  Then, when you run the tests, simply diff their
> 	output with the expected output.
>
> 	Anything better?
     Hello Paul,
                     you can do more than diff-ing pretty-printed AST. You can 
walk the AST and check the node type and node value versus a golden model. I 
personally used this kind of checking during parser verification.
     The golden model can be something like :
...............................................
// for source code :
   2 * (b + a)
// a possible expected/resulted AST would be :
[MUL, “*”]
|--[NUMERIC_LITERAL, “2”]
|--[PLUS, “+”]
|  |--[ID, “b”]
|  |--[ID, “a”]
................................................

      For this kind of testing you can look over Bogdan Mitu's tester (see 
files at antlr-interest at yahoo group).

      Stefan.

> 	- Paul
>
> _______________________________________________
> antlr-interest mailing list
> antlr-interest at antlr.org
> http://www.antlr.org/mailman/listinfo/antlr-interest