gcc6, libgcc, and the ABI issue reprised

classic Classic list List threaded Threaded
1 message Options
Reply | Threaded
Open this post in threaded view

gcc6, libgcc, and the ABI issue reprised

Ken Cunningham
There is expressed concern about the improvements to portconfigure.tcl to put forth gcc6 for PPC instead of clang-3.4 (which always fails) when the default compilers were blacklisted.

I don't see there being such a concern. I'd hope to put some of this to rest with a few examples.

1. the port built with the default compilers previously.

All is well. Nothing will change. The ABI is correct. Nothing to see here, folks.

2. the port did not build with the default compilers, clang-3.4 was called in, and the port failed to build.

No worries. Once we update the compiler blacklisting correctly (*gcc-3.* *gcc-4.*) or something similar that we agree on, the port will most likely now build, using gcc6. The CXX11 ABI=0 fiag will be used on c++ builds. I can verify this now works on gtk3 as you would hope it would.

3. The port previously used the cxx11 1.1 PortGroup, and built.

No worries. All is well. The ABI is correct. Nothing to see here.

4. The port previously whitelisted gcc6 by itself internally, was a c++ port, but did not use the cxx11 1.1 PortGroup, and built with gcc6 against the wrong ABI.

I don't know how many of these there might be, but the same port is still there, built the same way it always was.
Nothing has changed in any way from how things already were with the port the past many years. If nobody has complained about it after all these years, presumably nobody will. If they do, we revbump it and it will rebuild against the correct ABI (as it always should have been doing, but wasn't).  If it's a c++ port, it will build with the new (correct) ABI now. If it's not a c++ port, nothing will change. The next time the port is revbumped, it will build correctly with the correct ABI.

The way I see it, it's all good.