Boost logo

Boost :

From: williamkempf_at_[hidden]
Date: 2001-03-02 16:08:41


--- In boost_at_y..., Phlip <pplumlee_at_o...> wrote:
> Proclaimed williamkempf_at_h... from the mountaintops:
> > --- In boost_at_y..., Phlip <pplumlee_at_o...> wrote:
> > > Proclaimed williamkempf_at_h... from the mountaintops:
> > > > The code is not Boostified in any way. It was written adhoc
in a
> > > > hurry
> > >
> > > Does it have unit tests?
> >
> > Not really. The main.cpp file illustrates the full functionality
and
> > is what I used for testing, but it doesn't hold up to the full
> > criteria of a "unit test". I'd suggest first just looking at the
> > general functionality illusrtated in main.cpp to judge the
concept.
> > If that looks promising then we can get into more serious
evaluation
> > and I'll work on supplying unit tests (this will give the
Boost:Test
> > library time to be included, which will simplify writing the unit
> > tests). However, I expect there to be some ideas about ways to
> > modify the interface before that time.
>
> Just an ideological nit:
>
> http://c2.com/cgi/wiki?CodeUnitTestFirst

The point's well taken, and I know about this concept. However,
there's a bit of disconnect here. I have unit tested the code by
using the executable generated by main.cpp and running various
command lines with known expected output against it. This achieved
the same results, but it means I don't have any formal unit tests to
supply here.

For development purposes the unit tests are very much needed, and
exist in an informal manner for me. For evaluation purposes I'm not
sure they are important at all at this stage.

Bill Kempf


Boost list run by bdawes at acm.org, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk