On Jun 22, 2006, at 12:25 PM, Laurynas Biveinis wrote:
So I guess that there is no baseline as such, and the only meaningful
way to use the results is to compare testsuite outputs with and
without your changes?
Depends upon what you're doing...  For the normal regression test  
while developing  work for gcc, it is customary to have 2 runs and  
compare them.
If you're doing up a new port, and want it to work well, then the  
absolute, there should be no unexpected failures is a good place to  
aim for.
You can consider each email of results to the gcc tests results list  
to be a baseline, in that case, we have oodles of baselines.


Reply via email to