Multiple configurations with various mains

[Gary Feldman]

I didn't get any good answers to my previous question about unit testing,
so let me rephrase the question:

As near as I can see, it's possible to add arbitrary configurations to a C#
project in Visual Studio 2003. And it's possible to have multiple main
functions within a project (but not within a module), and use different
configurations to build different versions with different mains.

What I'm less clear about is how easy or difficult it may be for one
configuration to build an executable, and another to build a library. Is
this a reasonable thing to attemp?

And finally, is this a reasonable, simple approach to doing unit testing,
assuming that there is no need for complicated results analysis? The sole
purpose is to be able to write a module and test it in isolation before
moving on to the next module, not to do large scale regression testing on
the entire project.

Thanks,

Gary

[Peter Rilling]

First, it is rather easy to switch between a library and an application.
All you do is change the property setting that controls this.

Second, don't unit test this way. Rather you should download something like
nUnit that will provide an actual framework for you to test with.

http://nunit.org/

"Gary Feldman" <ga************ *@ziplink.stopa llspam.net> wrote in message
news:1g******** *************** *********@4ax.c om...

I didn't get any good answers to my previous question about unit testing,
so let me rephrase the question:

As near as I can see, it's possible to add arbitrary configurations to a C# project in Visual Studio 2003. And it's possible to have multiple main
functions within a project (but not within a module), and use different
configurations to build different versions with different mains.

What I'm less clear about is how easy or difficult it may be for one
configuration to build an executable, and another to build a library. Is
this a reasonable thing to attemp?

And finally, is this a reasonable, simple approach to doing unit testing,
assuming that there is no need for complicated results analysis? The sole
purpose is to be able to write a module and test it in isolation before
moving on to the next module, not to do large scale regression testing on
the entire project.

Thanks,

Gary

[Gary Feldman]

On Fri, 22 Aug 2003 12:04:04 -0700, "Peter Rilling"
<pe***@nospam.r illing.net> wrote:

First, it is rather easy to switch between a library and an application.
All you do is change the property setting that controls this.

Second, don't unit test this way. Rather you should download something like
nUnit that will provide an actual framework for you to test with.

http://nunit.org/

The limited quality of the nUnit documentation doesn't do much to convince
me that it's a better approach. Telling me it's a "unit testing framework
for all .Net languages" doesn't tell me anything about what it does and
how. And as near as I can tell, it has no Visual Studio integration.

But to clariry what I said in my original note, I'm not interested in
suggestions to use this or that. I'm interested in comparisons between
them.

Gary