On Thursday, June 19, 2014 10:28 CEST, Stuart Henderson <st...@openbsd.org> wrote: > On 2014/06/19 10:02, Sebastian Reitenbach wrote: > > > > On Wednesday, June 18, 2014 22:02 CEST, Stuart Henderson > > <st...@openbsd.org> wrote: > > > > > On 2014/06/18 21:54, Sebastian Reitenbach wrote: > > > > > > > > On Wednesday, June 18, 2014 20:23 CEST, Stuart Henderson > > > > <st...@openbsd.org> wrote: > > > > > > > > > On 2014/06/18 18:54, Sebastian Reitenbach wrote: > > > > > > Hi, > > > > > > > > > > > > not sure if this is known already, but this happens to me on i386: > > > > > > > > > > Check you don't have old libs around, especially make sure you've > > > > > followed > > > > > the kerberos removal on current.html. > > > > > > > > I did not upgrade, it happens on a fresh clean install. > > > > as it turned out, i have DEBUG= defined in my /etc/mk.conf, which is > > triggering > > the build failure. > > > > attached patch allows building alpine with DEBUG in mk.conf. > > > > Don't know if this is the most clever way, but looking at how other ports > > handle RAND_egd they just removed it. > > Also I'm unsure which of the many SUBPACKAGE to bump REVISION ;) > > It seems rather strange that setting DEBUG should trigger building code > that uses RAND_egd .. it indeed is, but I haven't seen any obvious #ifdef around it. There are some #ifdefs in that file, but nothing that immediately seems to relate to DEBUG
Sebastian