On Tue, 2005-03-29 at 14:52 -0800, Joe Buck wrote: > Daniel Berlin wrote: > > >IE if we added a very large warning to the submission page that said > > >"PLEASE NOTE: BY SUBMITTING A TESTCASE, YOU AGREE THAT WE HAVE THE RIGHT > > >TO CREATE, USE, AND PUBLISH EITHER YOUR VERBATIM TESTCASE OR A > > >DERIVATIVE UNDER GCC'S CURRENT LICENSE" > > >or something of the sort, they would be very hard pressed to win in a > > >real lawsuit in court. > > On Tue, Mar 29, 2005 at 05:21:02PM -0500, Robert Dewar wrote: > > I think that would be a goof ides, I would say GPL explicitlty, since > > several different licenses are used in GCC. > > Freudian slip? It seems that this may well be a "goof idea". > > Then we can't accept any test cases from free software that has a > GPL-incompatible license. Our only requirement for a testcase is that it > can be distributed with gcc, and that modified forms can also be > distributed.
Right. I was going for that, but was in the middle of doing something else, hence the reason "something of the sort". Anyway, I don't seriously believe we should add that because i think it might cause a bit of an uproar. I think we should just keep on doing what we do, which is produce small minimized testcases that are well within the "de minimis" limits for copyright, or even if they aren't, the tests don't actually do anything useful from a program perspective. The number of questionable testcases we get is small enough that taking this approach, and handling outliers on a case by case basis seems correct to me.
