jip@alw.nih.gov (john powell) (06/25/91)
From conftest (make log) %%%% Executing the configuration tester. ./conftest conftest: Checking configuration for sys type ``sun4_41'', full name ``SUN4_4.3b sd'' Sys type sun4_41 is most like 4.3 BSD. conftest: Checking for problems in the allsys.h/system.h/site.h files. AFS_ENV is defined, but WHITEPAGES_ENV is not. Either define WHITEPAGES_ENV or undefine AFS_ENV . Where does it say that if I define AFS_ENV that WHITEPAGES_ENV must be defined? -- John Powell phone: (301) 496-2963 Building 12A, Room 2031 FAX: (301) 402-0007 National Institutes of Health uucp: uunet!jip%alw.nih.gov Bethesda, MD 20892 Internet: jip@alw.nih.gov
Craig_Everhart@TRANSARC.COM (06/25/91)
I believe that the test in conftest is correct as it stands. Certainly it says that AFS_ENV implies WHITEPAGES_ENV in the file Gary quotes. It is an oversight that this implication isn't in the README file. I suppose, though, that if you wanted to turn off AMS_ENV entirely, you wouldn't need WHITEPAGES_ENV, even if AFS_ENV were defined. We never designed for AFS_ENV being defined without WHITEPAGES_ENV also being defined. I expect that if you try to build the world this way, you'll encounter link errors, and if you solve them, you may encounter run-time errors. Why not build with WHITEPAGES_ENV on? It's simple enough. Craig