I committed the 64bit support for the linux base ports (disabled by default, check the commit message), but this broke the INDEX build. Portmgr was faster than me to revert it. All errors are mine. I think most of the work is done, I just need to find out what the correct way is to handle this make/fmake difference (malformed conditional).
Day: July 26, 2015
Essen Hackathon 2015 – How it looks like
For those curious how it looks like:
Clockwise starting from the bottom left: flo, rodrigo, dru, Tux (in the background, acting as a guard-dog^Wpenguin of the Linuxhotel), bcr, lme, kp, brueffer.
From another angle:
Clockwise starting from the bottom: flo, rodrigo, dru, bcr, lme, kp, brueffer.
Not on the pictures is Sean Chittenden.
Essen Hackathon Status report – 2nd day
I had a look at the open PR’s for a quick-win and found one where the dependencies where incomplete. Fixed.
Then I reviewed Alan Jude’s patch for 64bit linux_base-c6 ports (on amd64). Looks good so far. Just a few minor issues. I took the time to get familiar with reviews.FreeBSD.org and the arc command line tool, applied the patch to my source tree, worked a while on merge-conflicts, added some minor changes, and validated the download of the 32bit RPM’s of the linux_base-c6 port.
In between I also discussed/reviewed some fixes for docs with Dru, signed some PGP keys, and served as a source for a funny picture (at least what geeks/nerds consider a funny picture). I also checked how to allow multi-cast in jails. There is a PR with a patch inside, but it’s IPv6 only. I did something similar for IPv4 and compiled a kernel. No compile time issues, but as the system where I can easily test this is at home, I prefer to be in front of the box in case it panics (that tells something about my confidence level of my patch… no idea if what I do there is actually correct… ENOCLUE about the network code in the kernel).
TODO for the last day of the Hackathon:
- validate all RPM’s (download / distinfo) of the ports which changed
- validate the install/deinstall of the 32bit version of the ports for regression
- validate the 64bit install/deinstall for at least the linux base port (more if time permits tomorrow)