Comment 5 for bug 750585

Out of interest, why is it necessary to have co-installable -dev packages? Is that only for convenience, to avoid having e. g. a i386 pbuilder/build chroot on amd64? So far I thought that multiarch was pretty much a runtime-only thing.

Anyway, for enabling build-time multiarch I agree that linux-libc-dev is a pretty big blocker, so getting this solved for natty would be nice. The actual linux patch looks fine to me.

For limiting the breakage, would it be reasonable to ship a /usr/include/asm symlink which points to the "main" architecture? Or would that hide potential bugs too much?

Thanks for preparing this so well with the full archive grep, which convinces me that the impact is fairly limited (and actually only affects the packages which rightfully look at asm/*). I'm not too worried about resulting build failures, though, as these are easy enough to fix when we have to update; the runtime impact is harder, but then again only things like strace and other debuggers will look at this. One potential corner case that comes to my mind is a "os.path.join(prefix, 'include', 'asm') construction, but I admit that looking at asm includes from python is rather contrived.

So +1 from me. However, I'd appreciate if Colin could have a look at this as well, to get more eyes on it.