Testing Camlp4-generated code

Testing Camlp4-generated code

[Matt Gushee]

I have developed a tool that generates a module using Camlp4, and I
would like to be able to test the output. The structure is fairly small
and simple, so it should be sufficient to verify that a generated module
is equivalent to a certain hand-coded module ... but is there a reliable
way to do that?

If it makes a difference, the output consists mainly of two values: a
data structure which is of an abstract type, and an object. The former
is implemented as follows:

   type data = S of string | L of string list | D of t
   and t = {
     data : (string, data) Hashtbl.t;
     mutable locked : bool;
   }

and the latter is an immediate object, but it inherits from a class
defined in another module. The object is associated with two instances
of the above type, has public methods that get and set values, and may
contain one or more sub-objects similar to itself.

Thanks in advance for any suggestions.

-- 
Matt Gushee
Englewood, CO, USA

Re: [Caml-list] Testing Camlp4-generated code

[Martin Jambon]

On Sat, 5 Nov 2005, Matt Gushee wrote:

> I have developed a tool that generates a module using Camlp4, and I
> would like to be able to test the output. The structure is fairly small
> and simple, so it should be sufficient to verify that a generated module
> is equivalent to a certain hand-coded module ... but is there a reliable
> way to do that?

My 2 cents:
   camlp4o pa_yourextension.cmo pr_o.cmo input.ml > auto.ml
   camlp4o manual.ml > manual2.ml
   diff auto.ml manual2.ml

Of course if you have a lot of automatic identifiers, that's not very 
convenient. But at least you will get the same indentation if your 
identifiers have the same length.


> If it makes a difference, the output consists mainly of two values: a
> data structure which is of an abstract type, and an object. The former
> is implemented as follows:
>
>   type data = S of string | L of string list | D of t
>   and t = {
>     data : (string, data) Hashtbl.t;
>     mutable locked : bool;
>   }
>
> and the latter is an immediate object, but it inherits from a class
> defined in another module. The object is associated with two instances
> of the above type, has public methods that get and set values, and may
> contain one or more sub-objects similar to itself.
>
> Thanks in advance for any suggestions.
>
> -- 
> Matt Gushee
> Englewood, CO, USA
>
> _______________________________________________
> Caml-list mailing list. Subscription management:
> http://yquem.inria.fr/cgi-bin/mailman/listinfo/caml-list
> Archives: http://caml.inria.fr
> Beginner's list: http://groups.yahoo.com/group/ocaml_beginners
> Bug reports: http://caml.inria.fr/bin/caml-bugs
>

--
Martin Jambon, PhD

Re: [Caml-list] Testing Camlp4-generated code

[Martin Jambon]

On Sun, 6 Nov 2005, Martin Jambon wrote:

> On Sat, 5 Nov 2005, Matt Gushee wrote:
>
>> I have developed a tool that generates a module using Camlp4, and I
>> would like to be able to test the output. The structure is fairly small
>> and simple, so it should be sufficient to verify that a generated module
>> is equivalent to a certain hand-coded module ... but is there a reliable
>> way to do that?
>
> My 2 cents:
>  camlp4o pa_yourextension.cmo pr_o.cmo input.ml > auto.ml
>  camlp4o manual.ml > manual2.ml

I meant reindenting:
    camlp4o pr_o.cmo manual.ml > manual2.ml

>  diff auto.ml manual2.ml
>
> Of course if you have a lot of automatic identifiers, that's not very 
> convenient. But at least you will get the same indentation if your 
> identifiers have the same length.
>
>
>> If it makes a difference, the output consists mainly of two values: a
>> data structure which is of an abstract type, and an object. The former
>> is implemented as follows:
>>
>>   type data = S of string | L of string list | D of t
>>   and t = {
>>     data : (string, data) Hashtbl.t;
>>     mutable locked : bool;
>>   }
>> 
>> and the latter is an immediate object, but it inherits from a class
>> defined in another module. The object is associated with two instances
>> of the above type, has public methods that get and set values, and may
>> contain one or more sub-objects similar to itself.
>> 
>> Thanks in advance for any suggestions.
>> 
>> -- 
>> Matt Gushee
>> Englewood, CO, USA
>> 
>> _______________________________________________
>> Caml-list mailing list. Subscription management:
>> http://yquem.inria.fr/cgi-bin/mailman/listinfo/caml-list
>> Archives: http://caml.inria.fr
>> Beginner's list: http://groups.yahoo.com/group/ocaml_beginners
>> Bug reports: http://caml.inria.fr/bin/caml-bugs
>> 
>
> --
> Martin Jambon, PhD
>
> _______________________________________________
> Caml-list mailing list. Subscription management:
> http://yquem.inria.fr/cgi-bin/mailman/listinfo/caml-list
> Archives: http://caml.inria.fr
> Beginner's list: http://groups.yahoo.com/group/ocaml_beginners
> Bug reports: http://caml.inria.fr/bin/caml-bugs
>

--
Martin Jambon, PhD
Store and share your bioinformatics tips at http://wikiomics.org

Re: [Caml-list] Testing Camlp4-generated code

[Hendrik Tews]

Matt Gushee <matt@gushee.net> writes:

   I have developed a tool that generates a module using Camlp4, and I
   would like to be able to test the output. The structure is fairly small
   and simple, so it should be sufficient to verify that a generated module
   is equivalent to a certain hand-coded module ... 

In addition to Martin Jambon's suggestion you can also compare
the internal camlp4 abstract syntax trees of the generated and
the handcoded module:

- use camlp4 with a printer that marshal's the ast into a file
  for both, the generated and the handwritten module
- in a separate program marshal both files back and compare them

You probably want a handwritten equality test in the second step
because of the location info in the ast.

Bye,

Hendrik