[Server-devel] Notes on replacing bridging with bonding
Read http://www.linuxfoundation.org/en/Net:Bonding in depth - excellent docs! - and performed some tests on one of my XS sample machines, replacing br0 with bond0. It works and I will attempt the switchover. We'll need quite a bit of testing to feel confident with this though... This is roughly what I am doing: # mark the device as a bonding device # - for some reason TYPE does not work # cat /etc/modprobe.d/xs-bonding alias bond0 bonding # cat ifcfg-msh0 DEVICE=msh0 TYPE=Ethernet MASTER=bond0 SLAVE=yes # Will be brought up by wlanX ONBOOT=no IPV6INIT=no HOTPLUG=yes NM_CONTROLLED=no LINKDELAY=0 # this replaces the bridge - all the bonding opts # I can see are for actually bonded devices - any # useful options we should use...? # cat ifcfg-bond0 DEVICE=bond0 BOOTPROTO=static # BONDING_OPTS="" IPADDR=$XS_BR0_IPADDR NETMASK=255.255.254.0 NETWORK=$XS_BR0_NETWORK BROADCAST=$XS_BR0_BROADCAST IPV6INIT=no ONBOOT=yes cheers, m -- [EMAIL PROTECTED] [EMAIL PROTECTED] -- School Server Architect - ask interesting questions - don't get distracted with shiny stuff - working code first - http://wiki.laptop.org/go/User:Martinlanghoff ___ Server-devel mailing list [EMAIL PROTECTED] http://lists.laptop.org/listinfo/server-devel
Re: Font problems (affecting Java and others?)
Alan, Thanks very much for the detailed writeup of your findings (and for your efforts make OLPC's software distribution more friendly to people who like Java). I can't personally resolve any of the questions which you raise with any authority but I can direct you toward the people who might be able to -- those people tend to live in #fedora-devel and on [EMAIL PROTECTED] might. (In particular, you'll find the Fedora OpenJDK packagers among them.) Regards, Michael ___ Devel mailing list Devel@lists.laptop.org http://lists.laptop.org/listinfo/devel
Terminal Text Fragment Drop Support Patch
Hi, Terminal really should support dropping dev key commands, I think. Here's a patch to make it do so. Could we update the devkey page to explain frame-based cut and paste? Also, I am looking for suggestions on how to upload patches from Sugar. To upload this one I created a Write document, quit Write, and used vi to edit the newly-created datastore object. Then the Web file picker could choose it. -- The world is quiet here. tmpoPQWCs.odt Description: application/vnd.oasis.opendocument.text ___ Devel mailing list Devel@lists.laptop.org http://lists.laptop.org/listinfo/devel
Re: Font problems (affecting Java and others?)
Alan Eliasen wrote: > I'm working on making it possible to get Java to run on the OLPC, > and have a solution to what many of you may have seen: the problem that > any Java program uses a strange italicized serif font, no matter what > font you request. This odd font is used everywhere, and no other fonts > seem to work. Update: I have filed bug http://dev.laptop.org/ticket/8348 for this issue, if you'd rather respond there. I have also done further research, and OpenJDK's font configuration file points at many fonts that aren't distributed on the OLPC, including fonts for Bengali, Malayalam, Tamil, etc. These will need to be fixed or added also. -- Alan Eliasen | "Furious activity is no substitute [EMAIL PROTECTED]|for understanding." http://futureboy.us/ | --H.H. Williams ___ Devel mailing list Devel@lists.laptop.org http://lists.laptop.org/listinfo/devel
Font problems (affecting Java and others?)
I'm working on making it possible to get Java to run on the OLPC, and have a solution to what many of you may have seen: the problem that any Java program uses a strange italicized serif font, no matter what font you request. This odd font is used everywhere, and no other fonts seem to work. This is focusing on the OpenJDK 1.6.0.0 distribution as found in yum. (That's "yum install java-1.6.0-openjdk") I've also considered gcj, but gcj's implementations of things like AWT or Swing is far less complete than OpenJDK currently. My OLPC is running Joyride 2346. I was glad to see that despite what our Wiki says about Java, it's actually relatively easy to make graphical Java programs at least run on the OLPC, to switch to its window, etc. (I'm not talking about making it a full-blown activity, which is dependent on a lot of other things happening, such as not having to set X properties on windows, bug http://dev.laptop.org/ticket/5271 and having a clean implementation of the services currently built in Python, and making a C/Java binding to them.) (By the way, I'm also experimenting with "java-gnome" which creates GTK bindings for gnome, so Java interfaces can look more like other Sugar applications, but due to several compounding bugs in our yum distribution, I can't build that on the OLPC, and can't emulate the OLPC due to bug http://dev.laptop.org/ticket/7813 . I'll file those later.) I know that there may be involved workarounds for some of these bugs, but I think it's important for everyone to know which bugs are seriously hampering developers who want to contribute, but don't have the time to research and duplicate lots of workarounds. BTW, is there any progress on 7813? The problem is that our font.properties files are not set up correctly for the fonts distributed on the OLPC. I've verified that adding one file fixes the problem, but there are several ways of fixing the problem and I wanted input from others on the proper way to do it. Here's a page that gives background on how fonts are loaded in Java: http://java.sun.com/j2se/1.5.0/docs/guide/intl/fontconfig.html (It's Java 1.5 but it still seems to work the same in OpenJDK 1.6) When looking at that page, note that Joyride 2346 sets the following: JavaHome=/usr/lib/jvm/jre-1.6.0-openjdk OS=Fedora Version=9 Thus, it searches for a working font configuration in the following order: /usr/lib/jvm/jre-1.6.0-openjdk/lib/fontconfig.Fedora.9.properties /usr/lib/jvm/jre-1.6.0-openjdk/lib/fontconfig.Fedora.9.bfc /usr/lib/jvm/jre-1.6.0-openjdk/lib/fontconfig.Fedora.properties * /usr/lib/jvm/jre-1.6.0-openjdk/lib/fontconfig.Fedora.bfc /usr/lib/jvm/jre-1.6.0-openjdk/lib/fontconfig.9.properties /usr/lib/jvm/jre-1.6.0-openjdk/lib/fontconfig.9.bfc /usr/lib/jvm/jre-1.6.0-openjdk/lib/fontconfig.properties /usr/lib/jvm/jre-1.6.0-openjdk/lib/fontconfig.bfc The one marked with * is the first one that exists. The .bfc files are a binary format. In the above directory, there exists a file /usr/lib/jvm/jre-1.6.0-openjdk/lib/fontconfig.Fedora.properties.src By copying this file to the filename "fontconfig.Fedora.properties" in the same directory, we can create a file that is found before the .bfc file and test our changes. Inside the new "fontconfig.Fedora.properties" file, at the end, there are lines like: filename.DejaVu_Sans=/usr/share/fonts/dejavu/DejaVuSans.ttf Unfortunately, those font files don't exist, but instead of, say DejaVuSans.ttf, we have files named DejaVuLGCSans.ttf. Fixing the "fontconfig.fedora.properties" to change filenames of all DejaVu... to DejaVuLGC... fixes these problems, and Java implementations now display correctly. Now, how is the proper way to fix it? Here are a few alternatives: 1.) Just add the fixed properties file as demonstrated above to the OLPC's OpenJDK distribution. Doing this is nice because it doesn't affect any existing files, and it's user-tweakable as opposed to the compiled .bfc file. 2.) Create symbolic links in the fonts directory, for example, DejaVuSans.ttf would be a symbolic link to DejaVuLGCSans.ttf. I don't know if this might help other applications. Note that Fedora 9 usually contains /usr/share/fonts/dejavu/ *AND* /usr/share/fonts/dejavu-lgc/ directories while the OLPC just contains the former. 3.) Add the missing fonts? 4.) Do some unspecified font fallback trick? 5.) Change "osname" (probably fetched from the Java property os.name) in the OpenJava JVM from "Fedora" to something more specific to the OLPC, like "OLPC". This might allow us to push an OLPC configuration file upstream to the OpenJDK project. I'm not sure where os.name is set, or how it is detected, though. Which of these (or some other alternative) is the best way to fix our Java font problem? I'll file a bug on this. Is there anyone in
Re: Short circuting USB port?
G'day Timothy, The symptoms you describe are also consistent with battery voltage falling below the critical threshold as a result of the powering up of the host-powered hub, keyboard and mouse. But only if there was no AC adaptor connected. Was there one connected at the time? -- James Cameronmailto:[EMAIL PROTECTED] http://quozl.netrek.org/ ___ Devel mailing list Devel@lists.laptop.org http://lists.laptop.org/listinfo/devel
Re: [Server-devel] Work in progress - F9 install ISO
On Wed, Sep 3, 2008 at 9:04 PM, Douglas Bagnall <[EMAIL PROTECTED]> wrote: > In an unrelated goose chase, I removed sqlalchemy from idmgr today. > Not tested yet though. A sqlachemy-free idmgr would be a welcome fix for the F9 port... once it's tested! :-) cheers, m -- [EMAIL PROTECTED] [EMAIL PROTECTED] -- School Server Architect - ask interesting questions - don't get distracted with shiny stuff - working code first - http://wiki.laptop.org/go/User:Martinlanghoff ___ Server-devel mailing list [EMAIL PROTECTED] http://lists.laptop.org/listinfo/server-devel
Short circuting USB port?
Hello List, My name is Timothy Hobbs and I have an OLPC XO-1 G1G1 laptop. Last night, I came down stairs with a headset plugged in to the audio ports(both a speaker and headphones. I plugged in my keyboad+mouse, which are attached to an unpowerd "Targus USB mobil port replicator with ethernet"(to the usb port directly bellow the audio ports), and I heard a little popping noise in my headphones, and the OLPC was instantly off! I started up the computer again, and plugged in the port replicator, and again the OLPC turned off. Now I know that with just a speaker plugged in, plugging in the replicator is fine, but I've been to ffraid to experiment much. School starts up in 2 weeks and I can't loose my laptop to a short circut right now. Possible steps to reproduce: Plug in a speaker and microphone into the audio ports. Plug in a usb device to the port just below the audio ports. I don't really know what could be causing this, it is very surprising bug. One oddnes about my system is that a while back the camera stopped working(on test it gives me error code -1) but that is on the other side. The other oddness is that I keep my olpc up on a windowsill, when I use the keyboard, so there may be a sligt downward pressure on the usb port when the replicator is plugged in, as you can imagine that might be pushing the end of the usb port up into the audio ports(lever action) But I don't think that is likely. Anyways, I hope that this is just a really weird problem with my individual laptop, which can be ignored. But I wanted to get this bug out anyways. Timothy Hobbs ___ Devel mailing list Devel@lists.laptop.org http://lists.laptop.org/listinfo/devel
cerebro documentation
Hello world, The following document provides documentation on Cerebro's design, scalability properties, implementation and potential applications: http://cerebro.mit.edu/cerebro-final.pdf For the impatient, Chapter 3 provides a detailed technical and theoretical analysis of Cerebro's system architecture and design goals. Chapter 4 provides a description of Cerebro's implementation and its various modules. Chapter 5 provides an account of potential applications. This includes documentation on how the Xavier activity can be used for file-sharing, presence information, chat, etc. Chapter 6 evaluates Cerebro's performance by providing experimental results on a testbed with tens of nodes. With scalable regards, Pol -- Polychronis Ypodimatopoulos Graduate student Viral Communications MIT Media Lab Tel: +1 (617) 459-6058 http://www.mit.edu/~ypod/ ___ Devel mailing list Devel@lists.laptop.org http://lists.laptop.org/listinfo/devel
Re: Wellington testers + Activities vs 8.2-759
Hi Martin, > Well, I get a "3 block" puzzle. Perhaps I don't understand the rules > completely, but clicking on the blocks didn't do anything interesting. > Other testers suggested I tried the larger games / blocksets and with > those I managed to get some implosions going, though I didn't manage > to solve one. The way the game works, you can only remove blocks if they are in a contiguous group of size *three or more*. There are other score-based games (like SameGnome) that allow you to remove blocks in groups as small as one or two, but Implode is intended to emphasize logic over accumulation of score, and the three-piece rule leads to puzzles that are more obviously solvable. I've tried to make the game "discoverable" by only highlighting the removable groups as the mouse cursor passes over them. There should always be a series of removals that clears the board. It sounds like you might have tried unsuccessfully to remove groups of one or two on the easy level and didn't try removing a larger group. Is this the case? If so, does the game perhaps need some sort of tutorial or hint mode? If you actually encountered a puzzle where it didn't let you click on *any* blocks on the board right after the board was generated, that would be a bug. The easy puzzles should be easily solvable -- even random clicking on blocks works for many of them. Thanks, -- Joe ___ Devel mailing list Devel@lists.laptop.org http://lists.laptop.org/listinfo/devel