On Tue, Mar 5, 2013 at 6:34 PM, Matthew Knepley <knepley at gmail.com> wrote:
> I was asking about the second block. I don't understand why we would ever >> want the C header to be broken when include from C++. >> > > Broken? My interpretation was that this define was turning on extern C for > the interface. The meaning of configureExternC changed from clanguage=C++ with-c-support to clanguage=C here: https://bitbucket.org/petsc/petsc-dev/commits/4b6962c351cf8206afe8cedb48495b4fffc8fa18 Anyway, "broken" would be when clanguage=C (so the library automatically extern "C"), but the header is include in a C++ project and extern "C" is not used. I started using PETSc back when this did not work automatically. -------------- next part -------------- An HTML attachment was scrubbed... URL: <http://lists.mcs.anl.gov/pipermail/petsc-dev/attachments/20130305/c4327d8b/attachment.html>
