libjava holding regression tester hostage

Fri Mar 25 21:09:00 GMT 2005

On Friday, March 25, 2005, at 12:14 PM, Andrew Pinski wrote:
> Actually it is not libjava which holding the regression tester hostage
> but libstdc++.

Ah, I now see why it is libstdc++... the tester doesn't check the 
return status of make check--target-libjava, so the fact it is non-zero 
doesn't matter much, though, it does prevent make check from working as 
well as we'd like, requiring make -k check instead... :-( Not a big 
deal, and my fingers do it that way anyway.
Sorry for the confusion.