> Most likely you haven't missed anything.  The whole tools build is a race 
> against platform vendors being creative, and in this case Apple might have 
> been creative with strlcpy().
> 
> I just updated src-netbsd, so you might want to pull in a new version and see 
> if you got lucky.  I didn't really see anything that would fix the above in 
> the diffs, so probably not (but worth a shot).

Didn’t help, as you correctly suspected.

> That failing, try to do a "regular" NetBSD build on your host.  See 
> instructions here:
> https://wiki.netbsd.org/tutorials/how_to_build_netbsd-current/
> 
> Just building "tools" is enough.

That did work:

===> Summary of results:
         build.sh command:    ./build.sh -O 
/Users/rickp/src/netbsd-test/src/../obj -U -j 8 tools
         build.sh started:    Tue Aug  2 17:40:11 BST 2016
         NetBSD version:      7.99.35
         MACHINE:             ia64
         MACHINE_ARCH:        ia64
         Build platform:      Darwin 15.5.0 x86_64
         HOST_SH:             /bin/sh
         No $TOOLDIR/bin/nbmake, needs building.
         Bootstrapping nbmake
         MAKECONF file:       /etc/mk.conf (File not found)
         TOOLDIR path:        
/Users/rickp/src/netbsd-test/src/../obj/tooldir.Darwin-15.5.0-x86_64
         DESTDIR path:        
/Users/rickp/src/netbsd-test/src/../obj/destdir.ia64
         RELEASEDIR path:     /Users/rickp/src/netbsd-test/src/../obj/releasedir
         Created 
/Users/rickp/src/netbsd-test/src/../obj/tooldir.Darwin-15.5.0-x86_64/bin/nbmake
         Updated makewrapper: 
/Users/rickp/src/netbsd-test/src/../obj/tooldir.Darwin-15.5.0-x86_64/bin/nbmake-ia64
         Tools built to 
/Users/rickp/src/netbsd-test/src/../obj/tooldir.Darwin-15.5.0-x86_64
         build.sh ended:      Tue Aug  2 17:46:34 BST 2016
===> .

Is the suggestion to use this toolchain for rumprun now?

> If that build fails too, file an upstream bug report to NetBSD.  Even if you 
> do come up with a patch, it needs to trickle into our repo via upstream, so 
> an upstream bug report is the right thing to do.
> 
> If that build doesn't fail, we need to figure out where the difference is -- 
> there should be zero differences, but computers are oh so good at screwing 
> people.

Well, yes, there’s always the chance that something is a bit crazy on this 
machine…

Cheers,
Rick

Attachment: signature.asc
Description: Message signed with OpenPGP using GPGMail

Reply via email to