> OK. Attached is the possible hack/fix for the build-error. This > patch will provide the limited support for the Fixtures and one > should be able to build the code successfully. NOTE: For future > development, I would suggest that test writers will follow the > same procedure as mentioned in this patch, to keep it compatible > with nocheck flag incase one plans to use the fixture. Hope this > will help.
Great, now at least it will compile with no-check. But now, what happens when running the tests? I guess that check fixtures had some purpose. what happens if the fixture is not enabled? Aleks, can you test the latest CVS version of check? If it is able to run our tests compiled mingw32 with FORK disabled, I would say to drop nocheck and to stop improving it. I cannot do the testing myself this week, because I am in the middle of the mountains with a non-stable creepy and shitty gprs connection. I will be back in the civilized world by the next week. -- Jose E. Marchesi [email protected] GNU Project http://www.gnu.org
