On Fri, Mar 03, 2017 at 01:29:26PM +0100, Dag-Erling Smørgrav wrote: > "Bjoern A. Zeeb" <bzeeb-li...@lists.zabbadoz.net> writes: > > Dag-Erling Smørgrav <d...@des.no> writes: > >> + /* split address if necessary */ > >> + err = EAI_SYSTEM; > >> + if ((sep = strchr(addr, ':')) != NULL) { > >> + len = snprintf(hbuf, sizeof(hbuf), > >> + "%.*s", (int)(sep - addr), addr); > > I believe this code is what broke > > fetch http://[::1]:6666/ > > just to give an example; and the printf traces will not reveal this > > but “addr” at this point has no more addr:port in it given the > > function arguments, right? > > I think you're right, I'll look into it. It means replacing strchr() > with a loop. Is there a ticket for this bug, and if not, could you > please open one? >
This causes breakage on the cluster for everything using libfetch: On the jenkins cluster we cannot run anymore fetch inside i386 chroots # chroot ufs fetch -v http://pkg.freebsd.org resolving server address: pkg.freebsd.org:80 fetch: http://pkg.freebsd.org: Non-recoverable resolver failure What is specific is the host only have ipv6 is amd64 and the chroot is i386 I can't figure out the issue can you have a look? Best regards, Bapt
signature.asc
Description: PGP signature