I use nAnt all the time for all our companies projects. But I must admit I am a bit of a lazy user, simply pointing nAnt to our solution files to build debug or release versions depending on choice, a clean up task, another task to create install packages, nUnit testing, and nCover tests as well. Has really streamlined our release procedures (down to 25 minutes from almost 1&1/2 hours, and could get less than that with automating copying of files to release server and install packages to various offices) and have also coupled it with CruiseControl for continuous integration. A really nice combination of projects.
 
Is shifting nUnit2 task into nAntContrib project then and getting it out of the mainstream project an idea?
 
Angus.
-----Original Message-----
From: Bart Read [mailto:[EMAIL PROTECTED]
Sent: Monday, 22 August 2005 6:45 PM
To: Angus Grant; nant-users@lists.sourceforge.net
Cc: Alan Wallace
Subject: RE: [Nant-users] nUnit2 task: Remove from project

I can see the point behind this, and to an extent I agree: if a feature is broken and there’s no intention of fixing it then it should simply be removed, because otherwise people (like me – see my earlier post) are going to waste a lot of time trying to get something to work that just won’t. But I have to be honest, and say that using <exec> is really a last resort. I’d pretty much rather write my own custom task than do this, especially if it’s something that’s going to be used by other people in the company. However, since many people are going to want to do at least some automated testing, and since this is something that should be encouraged (how tired are we all of using buggy software?), I really do think it’s worth the effort of fixing NUnit2 for the 1.0 release.

 

Just my two-pence worth. To bring some balance to this though, and to be entirely fair, I should also point out that NAnt is a great piece of software that has made building our software a *lot* easier. We were using a commercial product before (naming no names), and to be honest there’s just no comparison.

 

Thanks,

 

 

Bart

 

 

======================

Bart Read

Software Engineer

Red Gate Software Ltd

+44 (0) 870 160 0037 ext 31

======================


From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED]] On Behalf Of [EMAIL PROTECTED]
Sent: 19 August 2005 00:25
To: nant-users@lists.sourceforge.net
Subject: [Nant-users] nUnit2 task: Remove from project

 

To in-charge-type-dudes,

There seems to be a lot of emails about the nUnit2 task not working properly. And generally the reply is a canned "nUnit2 task is not really supported. You are better off using <exec> and the nunit executable"

For the 1.0 release, would it be an idea to remove the nUnit2 task? People are going to keep using it and striking problems unless either it is removed and they can only use <exec> or it is improved to be of releasable nature. I don't think it will matter how may faq entries, warnings, compiler warnings, etc, etc you put up. If someone sees an "nUnit2 task", and they are trying to implement nUnit testing they will jump on it.

Angus.

Reply via email to