When using private, static constants in a class declaration, and using these constants in an inline function that returns a ternary expression , gcc produces linker errors. This seems to be a problem going back to at least 4.0, and I can reproduce it against 4.4.1. I have not tried anything earlier than 4.0 or later than 4.4.1.
Created attachment 19039 [details] reduced test case This program doesn't link, though it should. keima:src dvander$ g++ test.cpp -o test test.cpp: In function ‘int main()’: test.cpp:28: warning: format ‘%d’ expects type ‘int’, but argument 2 has type ‘size_t’ Undefined symbols: "X::LENGTH_MASK", referenced from: X::dependentLength() const in ccdhJ0JB.o "X::DEPENDENT_LENGTH_MASK", referenced from: X::dependentLength() const in ccdhJ0JB.o ld: symbol(s) not found collect2: ld returned 1 exit status
You don't have a definition of the static const variable which you need for this to be valid C++.
(In reply to comment #2) > You don't have a definition of the static const variable which you need for > this to be valid C++.
(In reply to comment #3) > (In reply to comment #2) > > You don't have a definition of the static const variable which you need for > > this to be valid C++. Heh, oops, hit 'Commit' accidentally. This example does have a definition for the static member variable.
I'll also note that the compiler seems to accept it, as I get a .o file. The linking step botches.
It doesn't have any definition, it does have a *declaration*. This is pretty basic C++, by the way. Just add out of class: const size_t X::DEPENDENT_LENGTH_MASK; ... ... and everything will be fine. The compiler doesn't diagnose that, right, because the actual diagnosis, as for any undefined entity, eg, functions, happens only at link time, the linker tells you: as happens for functions, you could well have the definitions in another file.
(In reply to comment #6) > This is pretty basic C++, by the way. Mmm hmm. SO basic that it warranted special clarification in 9.4.2-4: "The member shall still be defined in a namespace scope if it is used in the program and the namespace scope definition shall not contain an initializer."
It is basic, yes, a point worth making with people insisting that we do have a serious bug, thus reopening the PR at will, without trusting the competence of the maintainers and wasting some of our time.
*** Bug 42241 has been marked as a duplicate of this bug. ***
*** Bug 42330 has been marked as a duplicate of this bug. ***
*** Bug 43594 has been marked as a duplicate of this bug. ***
*** Bug 44215 has been marked as a duplicate of this bug. ***
*** Bug 44528 has been marked as a duplicate of this bug. ***