Re: upcoming DejaGnu release

[Franco Carbognani]

Ben Elliston wrote:

Rob and I have discussed the idea of preparing a new DejaGnu 1.5.0
release and I would like to start preparing for that.  First, I would
like to put out a "last call for patches" that will end on June 30.

After that, I intend to spend some time exercising DejaGnu over a
range of DejaGnu-tested packages, including Windows hosts and embedded
targets.  Would anyone be willing to assist with this?

My intention is to have something ready around September.

Thanks, Ben
  

Dear Ben and Rob
We are currently proceeding with the integration project we have mentioned to you and that is supposed to finish at the end of September.
We are trying to understand if the wrapping capabilities of dejagnu would be enough for our purposes.
We feel anyhow that adding the TAT API into dejagnu would be the best and proper solution so we keep it as
our tentative final target.
There should be no licencing probem since TAT is already being distributed under a LGPL licence but we are
trying to clear this completely out with the other projects currently using and contributing to TAT

Supposing we are then going to contribute to dejagnu by adding the TAT based API, how this should be sincronized with your
release planning?

Obviously we are willing to assist in release testing, let us know how.

Best Regards

Franco Carbognani
Geoffrey Guilbon

Franco Carbognani wrote:
  

> We are currently thinking, within our EGO software group, to start a
> project wich aim is the merging of TAT within dejagnu in a way that
> preserve backward compatibility with the existing TAT test suites.
    

...
  

> We would like wery much this to happen within the official dejagnu
> project so that our contribution become an enhancement of the tool
> Would this be possible? In which way that shoud be arraged?
    

  This is entirely possible. DejaGnu supports "wrapping" existing
testing tools in a way that they use the DejaGnu testing framework at
the top level, but other testing frameworks underneath. This may be
easier that reimplementing TAT within DejaGnu. Another thought would be
to add an additional TAT based API to DejaGnu where it would run tests
directly. The only issue with this approach is the resulting code would
need to be GPL'd to be redistributed with DejaGnu.

  A big advantage of using DejaGnu is then your tests can plug into any
higher level test run analyzing software.

    - rob -

  

-- 
Franco Carbognani
European Gravitational Observatory (EGO)
via Amaldi
56021 Cascina (PI) Italy
Tel: +39 050 752 308 Fax: +39 050 752 566 
Email: address@hidden
www.ego-gw.it