Contributed by tbert on from the one-giant-step-for-softwarekind dept.
Our fifth b2k13 hackathon report comes from Sebastian Reitenbach (sebastia@), who writes:
Most of the time during the hackathon I spent working on the GNUstep related ports. My goal was to upgrade at least i386 and amd64 to compile with clang and enable the non fragile ABI, and upgrade libobjc2 to the latest release. Most of the bits were already prepared the months before. Two ports were still known to not compile, www/sope and www/sogo. After I got those problems sorted out, and the ports prepared for the libobjc2 upgrade, I was able to start upgrading the GNUstep ports step by step, upgrade gnustep-base, and libobjc2.
Then I prepared everything to build with clang and the non fragile ABI, but since I knew that there were problems seen in the past on macppc, I first had to try it there. The problems were still there, so changed the gnustep module and related ports not using the module to only build with clang and the enabled non fragile ABI on amd64 and i386, leaving macppc compiling with gcc. With this in, I then also upgraded some of the other GNUstep ports, and got some new ones imported.
Other time I spent a bit with trying to figuring out what might be wrong with my i386 notebook since the ACPI change that went into the tree just before the hackathon. With ACPI enabled it now takes about 2 hours to boot up to be able to log on. With some hints from Stefan Sperling (stsp@) I at least figured out it's a interrupt storm coming from acpi0. A kernel with ACPI_DEBUG enabled then was able at least to give Mike Larkin (mlarkin@) a clue, that there might be a stuck GPE.
Overall, this was a fairly productive hackathon for me. Many thanks and flowers go to Stefan Sperling (stsp@) for organizing this event!
Thank your Sebastian for sharing your report. And by the way, where did you ever find the patience to wait two hours for something to boot?
(Comments are closed)