Re: Problem building -current userland
On Tue, 4 Dec 2012, Dave Anderson wrote: Problem solved; PEBCAK. I didn't fully understand what 'cvs update' was doing, and managed to create a source tree containing a mixture of old and current files. Apologies for the noise. Dave I recently upgraded to the 2 December 2012 amd64 snapshot, then at about 11am EST today (4 December) updated my source tree to -current. After compiling and installing a new kernel and rebooting, my attempt to rebuild userland aborted with a slew of errors in /usr/src/usr.sbin/smtpd/smtpd/../dns.c. In case this was a short-term glitch I re-updated my source tree at about 4pm EST and tried again, with the same result. Is this a known problem, or have I managed to screw something up? I'm using the same build procedure as I've successfully used before, and didn't see anything in the current version of the 'building from source' section of the FAQ that would require changing it or see anything relevant in 'following current'. The sequence of commands I used is: cd /usr/src cvs -z 9 -d$CVSROOT -q up -Pd cd /usr/src/sys/arch/`machine`/conf config GENERIC.MP cd ../compile/GENERIC.MP make clean make make install reboot cd /usr/obj touch junk mkdir -p .old mv * .old rm -rf .old cd /usr/src make obj cd /usr/src/etc env DESTDIR=/ make distrib-dirs cd /usr/src make build Thanks for any help, Dave -- Dave Anderson d...@daveanderson.com
Problem building -current userland
I recently upgraded to the 2 December 2012 amd64 snapshot, then at about 11am EST today (4 December) updated my source tree to -current. After compiling and installing a new kernel and rebooting, my attempt to rebuild userland aborted with a slew of errors in /usr/src/usr.sbin/smtpd/smtpd/../dns.c. In case this was a short-term glitch I re-updated my source tree at about 4pm EST and tried again, with the same result. Is this a known problem, or have I managed to screw something up? I'm using the same build procedure as I've successfully used before, and didn't see anything in the current version of the 'building from source' section of the FAQ that would require changing it or see anything relevant in 'following current'. The sequence of commands I used is: cd /usr/src cvs -z 9 -d$CVSROOT -q up -Pd cd /usr/src/sys/arch/`machine`/conf config GENERIC.MP cd ../compile/GENERIC.MP make clean make make install reboot cd /usr/obj touch junk mkdir -p .old mv * .old rm -rf .old cd /usr/src make obj cd /usr/src/etc env DESTDIR=/ make distrib-dirs cd /usr/src make build Thanks for any help, Dave -- Dave Anderson d...@daveanderson.com
Re: Problem building -current userland
rebuild userland aborted with a slew of errors in /usr/src/usr.sbin/smtpd/smtpd/../dns.c. In case this was a short-term glitch I re-updated my source tree at about 4pm EST and tried again, with the same result. Is this a known problem, or have I managed to screw something up? I built everything on amd64 around 2pm CST today. So definitely a cvs update or other issue. cd /usr/src cvs -z 9 -d$CVSROOT -q up -Pd Use cvs up -APd (-A updates to latest version without any release tags, since you mention following -current)
Re: Problem building -current userland
On Tue, 4 Dec 2012, Amit Kulkarni wrote: rebuild userland aborted with a slew of errors in /usr/src/usr.sbin/smtpd/smtpd/../dns.c. In case this was a short-term glitch I re-updated my source tree at about 4pm EST and tried again, with the same result. Is this a known problem, or have I managed to screw something up? I built everything on amd64 around 2pm CST today. So definitely a cvs update or other issue. cd /usr/src cvs -z 9 -d$CVSROOT -q up -Pd Use cvs up -APd (-A updates to latest version without any release tags, since you mention following -current) Thanks for the suggestion; I'll give it a try -- and double-check that the cvs update worked properly. Dave -- Dave Anderson d...@daveanderson.com
Re: Problem building -current userland
On Tue, Dec 4, 2012 at 5:53 PM, Dave Anderson d...@daveanderson.com wrote: Thanks for the suggestion; I'll give it a try -- and double-check that the cvs update worked properly. If it does fail again, please include the error messages and the run up to them. Without that, there's nothing anyone can say other than try again; good luck! Philip Guenther