back from OLS. it wasn't quite what I expected but all in all it was fine. Ottawa and Toronto are very nice. "Try it on winter!" some said. I agree, three months without a single day of sun should be really boring.
nothing special these days, working in different small projects.
ok, I've been cursing D-Link for almost a year and I was wrong. It's not their fault my DWL-G650 doesn't works but yenta_socket driver (O2Micro oz6812 CardBus controller). I just got another Atheros based cardbus from 3com and it didn't worked as DWL-G650 didn't. tx works flawlessly but there're lots of lost packets in rx because of frame errors and it happens on windows too (windows xp service pack 2, latest drivers from D-Link and 3com). In linux is easy to notice it:
root@matthew:~/data# ping 192.168.67.1
PING 192.168.67.1 (192.168.67.1): 56 data bytes
64 bytes from 192.168.67.1: icmp_seq=0 ttl=64 time=1.5 ms
wrong data byte #20 should be 0x14 but was 0x0
0 10 18 1 18 19 1a 1b 1c 1d 1e 1f 20 21 22 23 24 25 26 27 28 29 2a 2b 2c 2d 2e 2f 30 31 32 33
34 35 36 37 0 0 0 0 84 dc 5 8 ff ff ff ff
64 bytes from 192.168.67.1: icmp_seq=4 ttl=64 time=1.3 ms
wrong data byte #20 should be 0x14 but was 0x0
0 10 18 1 18 19 1a 1b 1c 1d 1e 1f 20 21 22 23 24 25 26 27 28 29 2a 2b 2c 2d 2e 2f 30 31 32 33
34 35 36 37 0 0 0 0 84 dc 5 8 ff ff ff ff
I've been busy working on thumbs. I'm almost finishing the html template parser, one of the hardest features to be implemented in thumbs 1.0. also, I got obexftpd to take a look on it and try to make it functional (seems its latest update was in 2003 or so). it's a bit sad don't find some tools to play with bluetooth without using kde or gnome. let's see if I can help on this.
ols: got my visa and everything's set. I'll fly to Sao Paulo then to Toronto without stopping in uncle sam's lands. from Toronto I'll get a train to Ottawa in the same day. It'll be 6 hours by train and I hope I'll be able to take lots of pictures in the way.
Today I noticed how different is to program when you're very tired and when you're awake for some hours. yesterday at night I was trying to solve a problem in a game engine I've been working on. The problem was so huge and so complicated that I created a text file to try to organize my ideas. And then I went to bed. In the next morning while I was going to work (my company pays a cab from our office to the client, wheeeeee) I took a look in the problem without much hope. Damn, the "huge" was suddenly so simple that I almost solved it entirely while I was in the cab. It's strange that usually I don't feel tired and I have to measure how fast things are going to know if it's time to rest or not.
today at work I finally installed git and made a simple script so I can keep my tree in sync with Marcelo's 8xx tree. When I got some spare time I'll try to learn how it works. For now, quilt + svn seems to be a lot sufficient (make patches in quilt, test all of them and when I'm sure, make changesets based on those patches. While it's not done, the "patches/" directory are stored also in a svn repository, so I can work in different machines and the patches are always in sync). I have to say, quilt simply rox. Then, to finish the day, I gave svk a try. Amazing how intuitive it is and how useful it will be in my everyday's cab sessions 8)
spent some days to get 2.6 running in a mpc860 based board. while 2.4 runs fine, I couldn't manage to make it work until today. with help of some fine people in #mklinux we figured the problem: this board has registers mapped in physical address 0xFF000000. to make use of them, the 8xx port maps it to the same virtual address, i.e. 0xFF000000 virtual. this is not a problem in 2.4 but in 2.6 it is because kernel by default reserves an area beginning in 0xFF100000 for allocating consistent DMA areas. guess what: I hit in a WARN_ON() on dma_alloc_init() and interrupts just don't work (hangs trying to register inet socket types as it depends on rcu, which works using tasklets). easy, wasn't it? :)
I just finished to take a look in my uncle's computer. amazing, he got a computer about two weeks ago and it was full of trojans and all sorts of crap. and he's using firefox, firebird and AVG anti virus. seems another don't-open-strange-attachments problem.
I still don't understand what those flight companies do in web based sales. If I check for a flight to/from Toronto I got prices like US$2000,00 and up. Now, ask one of those travel agencies. US$1000,00. well, anyway seems I'm really going to OLS this year, despite all cabal and stuff.
ok, a big "doh" for me: for some time I've been wondering what is wrong with qemu. in my old tbird it took too much time to boot. after booting, the system were quite usable. then I put the same image to run in a sempron 2800+. same behaviour. then I tried with my new workstation thinking "ok, now it _has_ to boot fast". same thing. some hours later, while doing something else I figured out what was happening: I always boot using -nographic option, so only serial output is visible. as lilo isn't configured to use serial port, I didn't noticed it's configured to wait some seconds before booting :PPP
spent the entire holidays writing a platform game called "cabal". nice way to spend some time :)
got an amd64 to replace my workstation and it works as charm, no problems so far (via chipset, not nforce, is it a coincidence?). memtest86 reports nothing less than 1.7Gb/s of memory bandwith and _16G/s_ of L1 cache. and this beast has 512k of L2 cache. sweeeeeeet
while trying to give wifi access to my friend's notebook, noticed it locked up after inserting the card. one of those cardmgr irq/io fatal conflicts. I wonder if this is fixed already with hotplug stuff.
ok, after days collecting more reasons to be pissed, I got two days of complete fury. swimming saves and saved me once again but I couldn't escape the collateral effect of being too much nervous or depressive: I got sick. after two days of pills, lots of orange juice I'm back.
I've been playing in embedded world again. got an application developed to run without any operational system to modify to run on Linux. the problem (or the source of true fun) is that it has parts which assumes real time behaviour. to make the port easier, I'm trying to keep as much code in user level as possible and seems RTAI allows me to do that. I made some tests with other solutions (no, I didn't tried RTLinux. two words: patents suck) too but there were unacceptable latencies. and all I want is a 10ms tick with +-1.5ms.
I replaced my server by a faster machine. I got a sempron 2800+, 1GB ram and an Asus a7n8x-x. the first problem I had was power supply. no, a el cheapo ATX power supply can't handle this machine plus 3 ide disks and a cdrw unit. so I started looking for a decent power supply and got a thermal take as it's the best we have around here. the great surprise is that while using nforce2 IDE I got oopses every 30min _with_ data corruption. so that's it. my three disks are using promise PCI card now and I'm reinstalling a system that had been untouched since 2001 or so. yup, I'm pissed.
I was been working for days in DCCP until I notice I was not actually having fun doing it. then I decided to have a break and play a bit with uinputd and thumbs. had a fun and very productive afternoon. sometimes I wonder how far the desire to learn should lead me without losing the real sense in all this. ok, some may say the big deal behind open source development is increase your skills and get a better job. I don't disagree but I don't get this far thinking like that. I used to have fun and things was a lot easier that way.
got a cell phone (sony ericsson t610) with bluetooth and ir. finally a cell phone that isn't a island. seems there's some problems with via usb ehci chipset support. I'll investigate this and fill a bug report later. also, I bought an encore bluetooth usb adapter and I got it up and running as soon as I plug it in usb port and install bluez package. then I looked at windoze drivers cd and thought to myself how someone can use such complicated operational system... kudos to everyone involved in hotplug and bluetooth support.
New HTML Parser: The long-awaited libxml2 based HTML parser code is live. It needs further work but already handles most markup better than the original parser.
Keep up with the latest Advogato features by reading the Advogato status blog.
If you're a C programmer with some spare time, take a look at the mod_virgule project page and help us with one of the tasks on the ToDo list!