libstdc++/8197: std::sin(float) causes undefined reference to sinf

Phil Edwards
Sun Oct 27 14:34:00 GMT 2002

Answering because I happen to have just checked my mail:

On Sun, Oct 27, 2002 at 11:41:47PM +0100, Gabriel Dos Reis wrote:
>   Do we have infrastructure to export symbols based on target
> configurations?

We need to, but we don't yet.

> If not, does telling linker-map.gnu to export
> non-existing symbols an error?

No, it's no problem.  See for example the multiple entries for size_t:
only one of them will actually be used, depending on the architecture.

> If not, is it documented to work as
> "expected"?

Probably not documented, but anything marked as exported that doesn't
exist in the final library is just ignored.  That is, the symbols may
be exported, but they aren't expected.  :-)


I would therefore like to posit that computing's central challenge, viz. "How
not to make a mess of it," has /not/ been met.
                                                 - Edsger Dijkstra, 1930-2002

More information about the Libstdc++ mailing list