Folks,

Just a few hours ago I discovered an apparent bug in SUnit that resulted my resource setUp being called twice - this may have some relationship w/what Rob is seeing.

In TestResource(class)>>new, the code is "super new initialize". Because in 3.7 (as I understand it) new now calls initialize ("^self basicNew initialize") TestResource>>initialize is called twice and my setUp is called twice.

I fixed by removal of the "initialize" message from TestResource(class)>>new.

I haven't done my homework on - validating this is the right fix, - seeing if there's unit tests for SUnit (!) - filing this as my 1st bug w/Squeak.

So I though I'd send this out as a start.

thanks, Steve Sanderson

-----Original Message----- From: squeak-dev-bounces@lists.squeakfoundation.org [mailto:squeak-dev-bounces@lists.squeakfoundation.org]On Behalf Of stephane ducasse Sent: Saturday, October 30, 2004 10:56 AM To: The general-purpose Squeak developers list Subject: Re: Unit tests failing in clean image

Hi

I should say that I like this kind of attitude :)

I just downloaded, unpacked and ran Squeak 3.7 for windows. I popped open the SUnit test runner and ran the tests. A bunch of them failed and some required me to manually intervene. Is this normal, or have I done something wrong?

If it is normal .... why? Is there anything I can do to fix it; other than just deleting all of the tests - which seems wrong. I'm a proponent of TDD and I can't live with a system where tests fail every time I run them.

What is need is a test server reporting bug and linking that to a mantis like bug report management system? but we do not have that :(

Now we should really identify the ones that can be fixed and fix them so if you can comment/participate this would be great because we are a bit flooded in this moment.

Stef

On Saturday, October 30, 2004, at 11:56 AM, stéphane ducasse wrote:

I should say that I like this kind of attitude :)

...

I just downloaded, unpacked and ran Squeak 3.7 for windows. I popped open the SUnit test runner and ran the tests. A bunch of them failed and some required me to manually intervene. Is this normal, or have I done something wrong?

If it is normal .... why? Is there anything I can do to fix it; other than just deleting all of the tests - which seems wrong. I'm a proponent of TDD and I can't live with a system where tests fail every time I run them.

What is need is a test server reporting bug and linking that to a mantis like bug report management system? but we do not have that :(

Now we should really identify the ones that can be fixed and fix them so if you can comment/participate this would be great because we are a bit flooded in this moment.

One thing we might consider is adding something like this Expected Failures enhancement to SUnit:

http://lists.squeakfoundation.org/pipermail/squeak-dev/2003-July/ 063931.html

Then we might categorize the long-standing bugs as expected failures, and then we'd be more alert to new bugs that appear. And the expected failures list would represent a to-do list of bugs for anyone to work on. (Ideally the expected failures should show up somewhere in the SUnit UI.)

Well, it's not a perfect solution, but IMO it *is* better than what we have now. Should we add this to the SUnit package?

- Doug