I always forgot to forward to the list, instead of the author... > Dirk, what's the best way to deal with that? Blocking the bugs? > Reassigning them to dpkg? Raising the dpkg bugs to "serious"? I'm not > sure what is the correct procedure here. But as this affects every > package that wants a) move to openmpi or b) builds packages for more > than one MPI flavor, the u-a bugs should be fixed soon. (Which I work > on.)
Hi Manuel, thanks a lot for this detailed reply. You don't have to block or raise severity of this bug, since as Sylvestre wrote, I think in my case the real bug is missing g++. I have no idea why it's missing and why it builds in pbuilder, but nevertheless, the libmesh is in incoming, so tomorrow it hits unstable, then I'll upload a new revision with g++ builddepends and we'll see if it compiles. If it does, then I think this bug is not the problem you are describing. If I could see the config.log form buildhosts, I'd know where the problem is, but currently all I can do is guess. So just leave this bug as is and I'll respond later, in two or three days, how it goes. Thanks a lot, Ondrej -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]