[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[LTP] Re: TET [Mea Culpa]




I'm going to respond to myself on this, I have to issue a mea culpa.  It
appears what the open group does is release older versions of TETware
under the Artistic license, which is GPL compatible.  There is no free
or open source issue with those versions of TET.  I'm assuming we would
use those versions.

Now we can talk about its functionality.  I've perused the users guide
and the programmers guide.  The good news is that our tests would not
have to be modified to work in TET.  A test can be 'API-incompatible'
with TET and can communicate its results to TET through the exit
status.  Again, 0 == PASS and !0 == FAIL.  However, if a test is written
to 'API-compatible' with TET, it seems to rely heavily on the
framework.  I'm afraid that this dependency might have a consequence of
limiting the usability of any tests [outside the framework].  For those
that use TET, maybe you can describe TET/test dependency issues.  My
biggest fear is that we write tests to TET, then something better comes
along, and were stuck with TET.

--aaron

Aaron Laffin wrote:
> 
> Hi John,
> 
> I hope we can work togther to improve the quality of Linux through our
> testing work.  That aside, I need some help determining if TET can fit
> into an open source project such as LTP or even the Linux Project
> itself.  It appears to me that TET is not open source [1] and is not
> free software [2].  Correct me if I'm wrong [3].
> 
> Taking off my corporate hat and putting on my open source advocate hat,
> I seriously question building open source software around non-open
> source software.  If we build our testware around TET, no matter how
> "standard" or functionally excellent TET is, our testware will always be
> at the mercy of the functionality and availability of TET.  As a long
> time open source advocate, I believe that building our project around
> TET would set us up for failure.
> 
> One of the things that I'm really interested in is making the tests and
> testware easy enough to use and run that developers, Linux kernel
> developers, might run our tests on their modifications before submitting
> patches.  I think using non-redistributable software (TET) would put a
> serious kink in that possibility and many others.
> 
> Thoughts, please,
> --aaron
> 
> [1] http://www.opensource.org/osd.html
> [2] http://www.fsf.org/philosophy/free-sw.html
> [3] http://tetworks.opengroup.org/LICENSE
> 
> John George/Beaverton/IBM wrote:
> >
> > Has anyone considered using The Test Environment Toolkit (TET) as a harness
> > for
> > the LTP?  TET3 is the latest version and is freely available.
> >
> > TET seems to be close to an industry standard and is used by many other
> > test
> > suites (e.g. vsx, vsth, and other standards tests).  The suite is very
> > robust
> > and provides easy to read result summaries.
> >
> > IBM is developing a set of linux system call tests using TET3 and would
> > like to
> > contribute these tests to the LTP.  We would also be willing to participate
> > in
> > converting the existing functionality to TET.
> >
> > The following site provides detailed information about the product and I
> > hope
> > you will seriously consider using TET for the LTP.
> >
> > http://tetworks.opengroup.org
> >
> > John George
> > Software Engineer
> > IBM NUMA-Q
> 
> --
> Aaron Laffin
> Silicon Graphics, Inc. OS Test Development
> Email: alaffin@sgi.com Voice: 651-683-5756
> USA/MN/CRP/F5233/SSBU

-- 
Aaron Laffin
Silicon Graphics, Inc. OS Test Development
Email: alaffin@sgi.com Voice: 651-683-5756
USA/MN/CRP/F5233/SSBU