[LLVMbugs] [Bug 19496] New: [modules] should probably not diagnose conflict between unimported default argument and new default argument

[bugzilla-daemon at llvm.org]

http://llvm.org/bugs/show_bug.cgi?id=19496

            Bug ID: 19496
           Summary: [modules] should probably not diagnose conflict
                    between unimported default argument and new default
                    argument
           Product: clang
           Version: unspecified
          Hardware: PC
                OS: All
            Status: NEW
          Severity: normal
          Priority: P
         Component: Modules
          Assignee: unassignedclangbugs at nondot.org
          Reporter: richard-llvm at metafoo.co.uk
                CC: dgregor at apple.com, llvmbugs at cs.uiuc.edu
    Classification: Unclassified

Suppose we have:

m1.h: (empty)
m2.h: void f(int = 0);
module.map: module m { module m1 { header "m1.h" } module m2 { header "m2.h" }
}
tu.cc: import m1; void f(int = 0);

Right now, we'll reject this due to the redefinition of the default argument
for 'f'. We should (probably) allow this, because the default argument is not
visible in tu.cc (so it's not redundant).

Further, we should check that the default argument matches that from the
invisible declaration, as is required by [dcl.fct.default]p4 ("the accumulated
sets of default arguments at the end of the translation units shall be the
same").

Finally, we might want to warn whenever this happens ("default argument for 1st
parameter of function 'foo' is also provided by module 'blah'; consider
importing that module instead of redefining the default argument") since it
indicates layering confusion / non-modular code.

-- 
You are receiving this mail because:
You are on the CC list for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/llvm-bugs/attachments/20140420/c47f2f23/attachment.html>