[ANN] Refactoring Browser (0.92)

danielv at netvision.net.il danielv at netvision.net.il
Mon Nov 4 22:14:53 UTC 2002


Thanks Steve, I don't know if people appreciate how much timely, clear
and methodical testing like that helps me improve things. 

I'm really glad to hear people can now run the tests and have them mean
something. I'll deal with the Object>>testing thing next release,
probably this weekend, so the tests can be completely green. 

I don't know why, I was under the impression the Object>>testing thing
wasn't part of the RB changes. Cleaning up such crufty additions to the
image remaining from the port work is something I haven't done for a
while and will soon go back to.

Please let me know of any further problems you find..

Daniel 

Steve Gilbert <steven.l.gilbert at lmco.com> wrote:
> danielv at netvision.net.il wrote:
> > 
> > Now on SM.
> > I'd appreciate it if someone ran the test suite, so I'm sure the new
> > packaging is working. I've just found message complaining about these
> > darn three files from almost a year ago, so I'm hoping this long
> > standing problem is now fixed permanently.
> 
> Daniel,
> 
> Testing Report
> 
> - Fresh 3.2 image #4956 for Win32, on an NT 4.0 box.
> - Installed SqueakMap using the simple script from the SqueakMap page at
> http://minnow.cc.gatech.edu/squeak/2726
> - Installed SARInstaller (since the RB did not appear as installable
> until I did so, only appeared as downloadable)
> - Installed the RB
> - Installed Ned Konz's TestRunnerEnh ( can't stand running long tests
> without it )
> - "TestRunner open" in a workspace
> - Refresh then Run All
> - Take a walk
> 
> Results:
> 585 run, 583 passed, 2 failed, 0 errors
> Failures:
> 	AllTestSelectorsFixTest>>testAllTestSelectors
> 	TestUUIDPrimitives>>testCreationNodeBased
> 
> Analysis:
> 	The AllTestSelectorsFixTest>>testAllTestSelectors is a failure due to
> something I've seen and commented on privately before.  This is the
> "infamous" Object>>testing  method that gets added to Object in the
> Refactory-RBAddonsReasonable category.  With this selector included a
> test looking for 2 method names beginning with "test*" becomes 3 and so
> fails.
> 
>    My second failure is not regarding an RB test, so you may not be
> interested.
> 
>    The TestUUIDPrimitives>>testCreationNodeBased failure is due to a
> problem that I've seen before and commented on to John McIntosh since he
> appears to have added it to the image.  I somehow get a different result
> than John had expected for his test.  It looks valid but doesn't match
> his expected pattern.
> 
> Overall, great work!  I'm going to try and be more helpful if I can find
> the time.  I still believe that the RB work is extremely important to
> the Squeak community.
> 
> Steve Gilbert



More information about the Squeak-dev mailing list