Annotation of GNUtools/cc/TESTS.FLUNK, revision 1.1.1.1

1.1       root        1: This is a collection of things that test suites have
                      2: said were "wrong" with GCC--but that I don't agree with.
                      3: 
                      4: First, test suites sometimes test for compatibility with
                      5: traditional C.  GCC with -traditional is not completely
                      6: compatible with traditional C, and in some ways I think it
                      7: should not be.
                      8: 
                      9: * K&R C allowed \x to appear in a string literal (or character
                     10: literal?)  even in cases where it is *not* followed by a sequence of
                     11: hex digits.  I'm not convinced this is desirable.
                     12: 
                     13: * K&R compilers allow comments to cross over an inclusion boundary (i.e.
                     14: started in an include file and ended in the including file).
                     15: I think this would be quite ugly and can't imagine it could
                     16: be needed.
                     17: 
                     18: Sometimes tests disagree with GCC's interpretation of the ANSI standard.
                     19: 
                     20: * One test claims that this function should return 1.
                     21: 
                     22:     enum {A, B} foo;
                     23: 
                     24:     func (enum {B, A} arg)
                     25:     {
                     26:       return B;
                     27:     }
                     28: 
                     29: I think it should return 0, because the definition of B that
                     30: applies is the one in func.
                     31: 
                     32: * Some tests report failure when the compiler does not produce
                     33: an error message for a certain program.
                     34: 
                     35: ANSI C requires a "diagnostic" message for certain kinds of invalid
                     36: programs, but a warning counts as a diagnostic.  If GCC produces
                     37: a warning but not an error, that is correct ANSI support.
                     38: When test suites call this "failure", the tests are broken.
                     39: 

unix.superglobalmegacorp.com

This archive runs on limited infrastructure. Preserving old code on modern bandwidth. Automated agents are requested to crawl responsibly.