------- Comment #11 from rob1weld at aol dot com 2009-01-27 22:45 ------- (In reply to comment #9) > Subject: Re: [trunk regression]?gcc trunk 143562 - Testsuite - gfortran > failing tests that worked previously > > > I think adding a printf() clone to libiberty is WAY overkill just to > silence one failing test.
Two alternatives are: 1. A POSIX compliant Testsuite to check gcc libraries. Any program, on any platform, using (new) gcc _must_ printf() the _same_ output under any circumstances. 2. Incorrect operation. Sometimes there is "one failing test" because the Testsuite Coverage is poor (and we hope more Tests will be forthcoming). RFE - Need Makefile to test coverage of Testsuite http://gcc.gnu.org/bugzilla/show_bug.cgi?id=38833 By the term "Testsuite Coverage" I mean both in that the tests that we do have do not fully exercise the features they test _and_ also that we do not test all possible features but instead we slack off and simply disable the Tests (as is being suggested in this Thread). I believe Bug 36443 should be investigated first but failing that is it not the place of Libiberty to fix what GNU believes is broken? Wave @ DJ (net-acquaintance of 25+ years), Rob -- http://gcc.gnu.org/bugzilla/show_bug.cgi?id=38946
