> On Fri, Dec 03, 2010 at 02:52:12PM -0200, Fernando Lemos wrote: > > The requirement of linking to boost::system is an implementation > > detail of boost::filesystem that package maintainers should *not* > > have to worry about.
[Roger Leigh] > We shouldn't have to worry about it, it should be an implementation > detail. Unfortunately, it isn't, and we do need to worry about it. This is exactly why we have 'pkg-config --libs foo', and why it sometimes makes sense for this to output multiple -l flags. I haven't actually checked, but it sounds like boost simply does not support 'pkg-config --libs' correctly. A straightforward bug that, as Fernando says, should not involve patching other packages. The DFSG means we don't have to work around bugs in another package, we can just fix them in the right place. Other packages just need to make sure they use pkg-config --libs, rather than hardcoding the particular -l they think they need. (And by the way, even without binutils-gold, you _still_ always want to explicitly link. Otherwise you lose the benefit of symbol versioning - ld does not import symbol versions from indirect dependencies. I found this out the hard way with an upstream that intentionally relied on indirect linking. I suppose boost doesn't use versioned symbols, so this is a more general point.) -- Peter Samuelson | org-tld!p12n!peter | http://p12n.org/ -- To UNSUBSCRIBE, email to debian-devel-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/20101203221758.gb13...@p12n.org
