Contributed by tbert on from the tug-on-it-until-it-unravels dept.
Ken Westerback (krw@) writes:Don't tell anybody, but my main reason to come to r2k12 was that it was the perfect excuse to finally try the new Frankfurt-Paris rail connection, zipping through the French countryside at a comfortable 320 km/h. Yippee!
Officially, I had been invited to help deal with the fallout in the ports tree from the rthreads switch, but most of that had already been cleared up before the event. So I did what I usually do at hackathons: update one of those basic libraries a few thousand ports depend on, especially after there have been some API changes. This time around it was libtiff 4.0, which went smoothly and required only a few ports to be fixed as revealed by a bulk package build.
While the machines were crunching, I worked through landry@'s lists of bulk build failures on alpha and sparc64 and fixed the low-hanging fruit there. The remaining failures on these architectures are now dominated by toolchain problems, i.e., issues with cc/as/ld.
Finally, there was the usual miscellany of minor things that came up during other work and aren't worth listing separately, partial work that didn't quite gel into something commit-worthy, and taking time to test other people's changes like making matthew@ cry when I immediately broke his attempt to implement IUTF8 for canonical tty input.
Still more hackathon reports coming in, and we'll have them up as soon as we can bang them into shape!The main thing I worked on, with guenther, was fixing the closef() function in kern/kern_descrip.c to not use a tsleep() to ensure all threads are done witha file before actually closing it. This should be committed soon.
I sacrificed my laptop ssd while doing this and had to limp to the end of the hackathon on guenther's kind loan of a i386 laptop.
In addition I worked with Kurt and Espie (i.e. pointing out the problem to them, :-)) to fix the gcc-42,-java build so that the OpenBSD m4 was used by bison, fixing parallel dpb builds.
The room was tight enough to be an esprite de corps building environment without being an excessive test of french air conditioning.
The Indian, Chinese, and Moroccan food in Paris was excellent. As we were in the academic universe, we did not get to sample expensive French cuisine outside of the delicious morning croissants. Although the French themselves often slipped away for lunch to mysterious venues unknown to the rest of us!
While Belgian beer was heavily represented there were several excellent Parisianne microbrewery offerings.
Note to museum visitors: the Louve is closed Tuesdays. And you can climb to the top of the Arc de Triumph.
All in all an excellent venue and productive hackathon.
(Comments are closed)
By Raymond Lillard (rlillard) rlillard@sonic.net on
Comments
By Marc Espie (espie) on
Sorry, that's a hackathon exclusive.
If you meet OpenBSD developers, you can often recognize them by the T-shirts they wear...
Comments
By Miod Vallat (miod) on
Except for the shirtless ones.
Comments
By Marc Espie (espie) on
>
> Except for the shirtless ones.
So says Mr. "Miod et travaux"...
By phessler (phessler) on http://theapt.org
>
> Sorry, that's a hackathon exclusive.
>
> If you meet OpenBSD developers, you can often recognize them by the T-shirts they wear...
Even hackers that don't go to the hackathon don't get one.
By sneaker (sneaker) on
>
> Sorry, that's a hackathon exclusive.
>
> If you meet OpenBSD developers, you can often recognize them by the T-shirts they wear...
Ahh, so it's not so difficult after all. Just become a developer!
By sneaker (sneaker) on
>
> Sorry, that's a hackathon exclusive.
>
> If you meet OpenBSD developers, you can often recognize them by the T-shirts they wear...
Ahh, so it's not so difficult after all. Just become a developer!