Jean-Louis Gassée visits us!
Day three at LinuxWorld Expo 2008 started off with Scott McCreary dropping his car off at my sisters' apartment, and catching a ride to the Moscone Center with me. Despite nearly running over a few pedestrians, we made it there with plenty of time to get ready. Jorge Mare had to leave for home the evening before, so it was just going to be Scott and me this day. I had updated my laptop with a slightly newer revision the night before, and spent some time getting it setup to run live queries before the show started (which seemed to be broken for some reason before the rebuild.)
Special Visitors
It started off like the other days, didn't seem to slow down as much as I expected on the last day. We did have a couple of interesting visitors on this day indeed. Amy Bonner from IDG stopped by our booth to say hello. Amy helped us secure the booth space after we were turned down for a space in the .Org pavilion. She said she was really happy we could make it, and shared some ideas with us for next year's .Org submission. We gave her a complimentary T-shirt for helping us out this year. It was great to finally meet her in person, and we snapped a shot of her standing in front of the booth.
Being a go-getter kind of person has, on occasion, actually gotten me somewhere besides into a mess. With having significantly more free time than usual because of being on summer vacation, I decided to work on a document which combined two RFCs I have already written, which can be found here and here. Knowing how it seems like discussions on R2 usability seem to be both endless and unproductive, I decided to put some the ideas into code before publishing it in an effort to demonstrate that most, if not all, of the ideas I propose are practical, reasonable, and worth implementing for the second version of the community’s beloved OS. The first of these to see public eye is the Filer.
Back to the Moscone Center Haiku and ReactOS at LW2008Today, Scott McCreary was nice enough to swing by and pick me up on his way to the conference. We cruised by my company's corporate headquarters to pick up a package I was expecting, and then went directly to the Moscone Center from there.
Before the conference was under way, we re-arranged the layout in the booth a little bit, putting Art and the ReactOS machine up in front next to ours a bit more.
Trekking back to the Moscone Center
Haiku Booth - Day 1At 8:15AM, my sister gave me a ride over to the Moscone Center. We picked up Art Yerkes (from ReactOS) on our way, and proceeded down to the exhibit floor once we got there.
We arrived an hour early, and started with a few finishing touches on the booth. I think it turned out pretty well :)
LinuxWorld Day 0I set out on my 3+ hour trip to San Francisco at 9:30AM the morning of August 4th. My trip included stopping and picking up Jorge Mare (a.k.a. Koki) during my drive along with the rest of the supplies and equipment that he was providing for the event. My car currently has no air conditioning, and the majority of my trip was in 95F (35C) heat - so I was quite sweaty when I arrived at Jorge's house.
As the m68k port is getting shape, maybe some of you want to give a hand, so here is how to set up the environment. After explaining the choice of the target platform we’ll start with the build system, then the emulator to debug on the chosen platform.
Time for a quick update.
As with r26676, a first buggy xsi semaphore implementation is now present in Haiku (buggy because there is another patch waiting to be reviewed and commited that fixes some issues, but there might be some more coming).
It is now possible to download, compile and run the file system benchmark suite bonnie++. The version I've used is 1.03d, which has been suggested by the author of the suite. With this suite it is possible to test the file system implementation, plus the way Haiku works under low memory and heavy I/O operations.
Unlike BeOS, our kernel includes some pieces of C++ code, which sometimes give a headache when it comes making sense of a stack crawl from the kernel debugger, since symbols are mangled when linked into binaries, which means we must Decode__12CrypticCNamesPCc. I recalled seeing some gcc4 private API to demangle symbols into human-friendly names, but the code doing that, from libsupc++, has been written without concern for the inhabitants of the Kernel Debugging Land, using calls to malloc, realloc and free... But I still wanted to get nicer names, so I didn't give up. I also wanted to be able to get assembler dumps since not everyone has a serial cable to make use of the gdb stub.
Code_Swarm for Haiku videoToday I received an email from Fredrik Holmqvist (TQH, of Bezilla fame) about a video that he recently created and posted on Vimeo. Titled “Code_Swarm for Haiku,” this is a video generated using Code_Swarm, a technology that allow visualizing the activity on a software repository. The video that TQH created was generated from the Haiku subversion commit messages, and shows the period starting from the time Haiku moved to Subversion up until revision 26538.
Another few weeks have passed since my last update, so here is another short report on the status of my project.
I hope everyone is enjoying the results of my work that have already been integrated into the main source trunk. Now I am pleased to announce that also the second part of my project is almost finished.
