Sprint Projects/Resources for PyCon 2008 Sprints
Hi all, We'll being doing an Activities sprint after PyCon2008, March 17th (evening) to the afternoon of the 20th. We'll likely have 8-10 Python programmers of various skill levels available. Given the compressed time-frame, we should be able to either complete very small projects or work on existing projects. We've got a pretty good set of small project tasks outlined in the OLPC Austria wiki and the main OLPC wiki, but new programmers will likely be looking for projects they can join as well. I'm intending to spend any coding time I have (when I'm not helping the teams) working on Productive and/or the OLPCGames wrapper it uses. I can readily absorb about 3-4 programmers into those projects, but I expect there will be people would would rather work on something else, but who don't feel comfortable starting their own project or joining a new project. So if you: * are a project lead for an Activity o something where you already have code and can make use of (junior) collaborators * will be available on IRC during the sprint time-frame to act as a mentor for new coders o particularly in the evening on Sunday the 17th and then Monday daytime to get people started o right through until Wednesday would be useful * have some set of tasks that you feel a new coder could work on as a starting point for joining your project o if you have reasonably self-contained tickets in the tracker that's probably the best approach let me know so I can suggest your project to participants as a mentored introductory projects. I'll try to familiarize myself with your code-base before the sprints so that I can help out if people get stuck. Thanks all, Mike ___ Devel mailing list Devel@lists.laptop.org http://lists.laptop.org/listinfo/devel
Re: GLX available?
Alex Asimakopoulos wrote: Hello, i am watching this discussion, and i would like to ask you some things.. i would like to install OpenGL support just for testing reasons.. in olpc do i have to install mesa rpms found here: http://bfccomputing.com/downloads/olpc/mesa/ and stated here: http://dev.laptop.org/ticket/227 ... The Doom Activity includes OpenGL AFAICS. You should be able to crib the setup from there. HTH, Mike -- Mike C. Fletcher Designer, VR Plumber, Coder http://www.vrplumber.com http://blog.vrplumber.com ___ Devel mailing list Devel@lists.laptop.org http://lists.laptop.org/listinfo/devel
Re: GLX available?
[EMAIL PROTECTED] wrote: That won't work - prboom (working) doesn't use OpenGL. Hmm, interesting... might be useful to drop the lib/libGL* files from the .xo file, then. Apologies for the confusion, Mike On Feb 5 2008, Mike C. Fletcher wrote: ... The Doom Activity includes OpenGL AFAICS. You should be able to crib the setup from there. HTH, Mike ___ Devel mailing list Devel@lists.laptop.org http://lists.laptop.org/listinfo/devel
Re: I have an OLPCGames SVG problem
Kent Loobey wrote: I am new to both Python and SVG. I have taken some sample code from OLPCGames and made a test program. I have created an svg file using Inkscape. if I use the sample activity.svg file it displays okay but if I use the file I made it does not. What Linux software can create SVG files that olpcgames svgsprite can process? I use Inkscape for producing svg files. Can you tell *how* it is failing? Is there an exception showing up in the log viewer? Or is it just silently showing nothing? Is your size such that the image would show up on-screen? SVGSprite will try to guess a size based on the embedded declared size of the image, so it's possible your image is actually showing up, but you're just seeing the (blank) corner of it. Under the covers we're just using rsvg to do the rendering, so it should be able to handle most SVG files. Good luck, Mike -- Mike C. Fletcher Designer, VR Plumber, Coder http://www.vrplumber.com http://blog.vrplumber.com ___ Devel mailing list Devel@lists.laptop.org http://lists.laptop.org/listinfo/devel
Re: I have an OLPCGames SVG problem
Kent Loobey wrote: On Saturday 02 February 2008 06:51:39 Mike C. Fletcher wrote: ... At this point my problem seems to be one of color. I have since learned that the colors of my original images were not distinguishable. I seem to remember somewhere that you all are using a 16 bit palette. It could be that I am making images that are 32 bit and when it maps them to 16 bit they change. I say this because I made some new images with more colors in them and they now display with distorted color in svgspritetest.activity. 32 bit to 16 bit shouldn't cause any reasonably distinct colours to disappear into one another, 16 bit gives you 32 shades for each component, unless your graphics were really-pale-ivory on white they should still show up. The colors are just wrong, i.e., Red shows up as blue for example. This leads me to believe that all my images have been displayed but that I can't see some of them because they have been mapped to the same color as the background. That sounds like a problem in the array-handling code. The SVGSprite has probably only been tested on AMD64 machines (i.e. my workstation here (though I thought I'd run on the XO to test it)), could be we're seeing a problem with the translation code on 32-bit machines. There's also an explicit colour rotation going on for handling text on certain versions of pycairo/pygtk, but that shouldn't affect the rsvg renderer. Can you also tell me what environment you're running under? Desktop 32-bit Linux in either jhbuild or ubuntu packages is what I'm assuming. My initial images were 48x48 pixels. The activity.svg image included with svgspritetest.activity is 45x45. Okay, size isn't the problem, then. If you can send me a problem file I'll try to figure out what's going wrong. Take care, Mike -- Mike C. Fletcher Designer, VR Plumber, Coder http://www.vrplumber.com http://blog.vrplumber.com ___ Devel mailing list Devel@lists.laptop.org http://lists.laptop.org/listinfo/devel
Staffing of an OLPC Booth at PyCon, volunteers needed
Edward Cherlin wrote: 2008/1/29 Noah Kantrowitz [EMAIL PROTECTED]: ... We are holding open an OLPC booth, as someone had mentioned that they wanted one. Can anyone confirm this? A number of OLPC volunteers will be at PyCon, myself included. Does that work? We're told by the PyCon organizers that they require an official request and a commitment to staff the booth throughout the conference hours in order to get the space reserved. They have the entire Expo hall bookable, so we're competing with people paying for the privilege of displaying their wares. We need to get the request in ASAP to avoid losing the booth (i.e. yesterday-ish). We'd probably need 5-6 people to man the booth throughout the conference days without asking someone to spend the whole conference behind the desk. I'm certainly willing to take a half-day or more. We should have a good selection of XO's available. Would be nice if we had a table, maybe a couple of posters and the like, but realistically I don't have enough time booked for OLPC before the conference to get that done. If we can get 5-6 volunteers to commit to it I'm willing to put together an official request and move forward. Take care, Mike ___ Devel mailing list Devel@lists.laptop.org http://lists.laptop.org/listinfo/devel
Re: [Fwd: [PyCON-Organizers] OLPC Booth]
Noah Kantrowitz wrote: Anyone know the status on this? It wasn't me that requested it. Take care, Mike Subject: [PyCON-Organizers] OLPC Booth From: VanL [EMAIL PROTECTED] Date: Tue, 29 Jan 2008 06:00:47 -0600 To: [EMAIL PROTECTED] [EMAIL PROTECTED] To: [EMAIL PROTECTED] [EMAIL PROTECTED] CC: [EMAIL PROTECTED] Hello, We are holding open an OLPC booth, as someone had mentioned that they wanted one. Can anyone confirm this? -- Mike C. Fletcher Designer, VR Plumber, Coder http://www.vrplumber.com http://blog.vrplumber.com ___ Devel mailing list Devel@lists.laptop.org http://lists.laptop.org/listinfo/devel
Machines for the tutorial at PyCon
Seems we'll have plenty of machines available for the Sprints after PyCon. That's great. I'd like to have a number available on the tutorial day as well (the day before the conference starts). I have a B4 and can likely get a B2 back from one of my local developers for a few days, but I'd like to have at least 5-6 machines so that people can try out the software we'll be building as we go. Any people with laptops willing to loan them for a few hours on the day before the conference starts (first thing in the morning, March 13, 2008, in Chicago)? We'll need a reasonably stable environment on them (e.g. Update.x) so that tutorial-goers are not hitting regressions as they try to work through the tutorials. Shouldn't be all that much 'wear and tear' on the machines, mostly they're just so that people who learn best by muscle memory of the steps can learn how to work on the systems. Thanks all, Mike ___ Devel mailing list Devel@lists.laptop.org http://lists.laptop.org/listinfo/devel
Re: Animation/Python/PyGames vs battery charge
Noah Kantrowitz wrote: Kent Loobey wrote: I have set up some test animation code. Normally games try to take all the cycles they can get. I am trying to preserve as much battery energy as I can. So I am setting a specific frame rate and sleeping beyond what it takes to maintain that frame rate. Do you think this will actually reduce the drain on the XO battery? In other words What does the XO do when apps sleep? If SDL is actually using a real timed sleep, it would help, but I don't think it does. This is often too inaccurate for games, so we may need to look at that ourselves. Because pygame is polling driven (instead of using async callbacks like GTK) it is unlikely that it will ever get as nice a battery life as GTK-based activities. There are two types of sleep in Pygame, one is: pygame.time.wait( milliseconds ) which does a process sleep. The other is: pygame.time.delay( milliseconds ) which does a busy-loop to provide accurate time-keeping. OLPCGames also now has a function which combines wait with a time-based sleep state where the program is completely suspended if there are no events for a given period. The same is true of Pippy's (independent) wrapper. That said, lots of games, particularly ones which are simulation based, may need to run all the time. The XO is also *not* currently (AFAIK) doing aggressive suspend. That is, while we would like to get to the point where the machine can suspend and resume in 10ms range, it currently takes too long to shut down between frames of a game. That is, while eventually micro-sleep may show up, AFAIK we don't yet have support for it on the laptop, so the processor is likely running at 100% power (even if in a busy loop in the kernel) (AFAIK there is no frequency scaling available in the processor). We might get some minimal benefit from having less load, but if we're waking up 10 or 15 times a second we're not going to see the kind of power benefits we'd see with multi-second or multi-minute suspending. We still want to have the games waiting when they are not needed. Eventually we hope to get aggressive suspend such that between frames we can go to sleep, but until then the power benefits are in allowing other processes to run, and in shutting down the game entirely after a period of inactivity. HTH, Mike -- Mike C. Fletcher Designer, VR Plumber, Coder http://www.vrplumber.com http://blog.vrplumber.com ___ Devel mailing list Devel@lists.laptop.org http://lists.laptop.org/listinfo/devel
Re: B2s and OFW
Mike C. Fletcher wrote: ... I just upgraded a B2-2 machine that we're wanting to reassign to another developer. Seem to have bricked the machine. Used an auto-reinstallation image (machine appeared to have the original B2-2 firmware and OS image 216 to start), upgrade to Q2D07 seemed to go fine. Unplugged, removed battery, waited, replaced battery, replaced power... no battery-charging light any more, no response on the power button at all. The machine has been sitting unused for many, many months. Had power applied for about 1/2 hour before the upgrade attempt, charging light was showing properly then. And this morning, with the battery removed, it booted and upgraded itself fine. Adding the battery back while it's running seems to show it charging. So, guess this was just a fluke. Sorry for the noise, Mike -- Mike C. Fletcher Designer, VR Plumber, Coder http://www.vrplumber.com http://blog.vrplumber.com ___ Devel mailing list Devel@lists.laptop.org http://lists.laptop.org/listinfo/devel
Re: B2s and OFW
Mitch Bradley wrote: ... I don't have a B2-2, so I can't test that. But I don't know of any reason why it wouldn't work. I just loaded Q2D07 on a B1. It seems to be working okay from the firmware side of things - I can do copy-nand network accesses and dir nand:\ and things like that. But the kernel is hanging somewhere in the early startup - no kernel messages on either the screen or the serial port. I'm not entirely surprised; IIRC that B1 was flaky the last time I tested it, which was some time ago. Oh, I see what is happening. This B1 has a bad 14 MHz clock - the OFW diags caught the problem. So my best guess is that Q2D07 is good on everything B and later. I just upgraded a B2-2 machine that we're wanting to reassign to another developer. Seem to have bricked the machine. Used an auto-reinstallation image (machine appeared to have the original B2-2 firmware and OS image 216 to start), upgrade to Q2D07 seemed to go fine. Unplugged, removed battery, waited, replaced battery, replaced power... no battery-charging light any more, no response on the power button at all. The machine has been sitting unused for many, many months. Had power applied for about 1/2 hour before the upgrade attempt, charging light was showing properly then. It's possibly just a coincidence that it's bricked on this update (might have done it on any update, the machine has been sitting unused for a long time, there could be a battery-discharge issue or the like). Anyway, if someone has advice on how to un-brick (or debug) it would be appreciated, Mike -- Mike C. Fletcher Designer, VR Plumber, Coder http://www.vrplumber.com http://blog.vrplumber.com ___ Devel mailing list Devel@lists.laptop.org http://lists.laptop.org/listinfo/devel
Controlling the Glide/Pressure sensor
I have a student who's interested in doing a term project on the UI for the track sensor. I've put together this quick summary. Deadline looms for starting the project, so if people have don't do that or we've already done that feedback, please speak up ASAP. Background: * XO has two different devices, resistive glide-sensor and pressure-sensitive tablet o Both of these are currently showing up as core pointer events in X AFAIK o Changes between pressure and glide-sensor activity have the potential to cause jumps of the cursor (absolute versus relative mode) * There is currently no UI to map the pressure-sensitive tablet's area into a particular area on the screen (nor, AFAIK an API to accomplish the mapping) o Use case: use the entire drawing area to draw into a small area of a drawing in Paint * Activities currently do not have control over the mapping of the area o Use case: in a penmanship course, collect samples of the child's letters in special widget areas within a test, focusing a new area should remap the pen to that area Trackpad UI Design Requirements: * API for configuring the resistive/pressure sensor allowing control of all the major features o Note that there will likely be some X11 hacking here, to get the pointer mapping working * Standard UI controls for redirecting input areas o Standard GTK UI for positioning, and scaling o Standard GTK widget for e.g. handwritten text entry, provide access as a bitmap (or a series of strokes optional) + Allow for capturing all data (full resolution) or just scaled data as configuration option o Intuitive (HIG-compliant) standard mechanisms for controlling the various configuration parameters o A 6 year old should be able to figure out how to direct their drawings, written text and the like into the desired areas o Standard feedback on where the tablet area is bounded on screen when drawing with the tablet * System level (possibly on Sugar's Frame) trigger to bring up the control mechanisms (optional) o Most pen-aware applications will likely use internal logic and the API to determine position and the like, but a general trigger to the functionality should be useful for non-pen-aware activities * Paint Controls o Work with Paint's authors to provide intuitive controls to make using the pen/tablet intuitive within the context of paint Obviously we would need to find a machine to work on to make the project feasible. I'll see if we can repurpose one that's local to the task. Discussions welcome, Mike -- Mike C. Fletcher Designer, VR Plumber, Coder http://www.vrplumber.com http://blog.vrplumber.com ___ Devel mailing list Devel@lists.laptop.org http://lists.laptop.org/listinfo/devel
Re: Mounting a USB drive (windows format)
Jeffrey Kesselman wrote: On Dec 21, 2007 5:00 PM, Mike C. Fletcher [EMAIL PROTECTED] wrote: ... Side note: kqemu does work under 64-bit Linux (it's what I use to emulate on my AMD64 box). Do you run qemu or qemu-x86_64 ? qemu-x86_64, particularly I run: qemu-system-x86_64 -kernel-kqemu -m 1024 -hda $1 -net user -net nic,model=rtl8139 -soundhw es1370 as a script. Qemu will indeed work on a 64 bit (Win32) system with kqemu, but its not operating as a 64 bit app. qemu-x86_64, at least in the build pointed at by the OLPC wiki, is not built with kqemu support and throws an error when youy try to use the -kernel-kqemu flag Ah, I'm on Gentoo, so I *always* use a custom build of *everything*, my version does support -kernel-kqemu. http://wiki.laptop.org/go/Developers/FAQ#How_do_I_mount_a_USB_drive.3F Thanks! I'll give it a shot! Good to hear it worked for you. If you can send me the upgraded v6 VMX I'll try to produce both a 5.x and a 6.x vmx file from the image conversions so that others don't need to do that work. Have fun, Mike -- Mike C. Fletcher Designer, VR Plumber, Coder http://www.vrplumber.com http://blog.vrplumber.com ___ Devel mailing list Devel@lists.laptop.org http://lists.laptop.org/listinfo/devel
Re: Project Hosting request: LiveBackup_XO-LiveCD
[EMAIL PROTECTED] wrote: 1. Project name : LiveBackup_XO-LiveCD 2. Existing website, if any : 3. One-line description : Live-CD build based on the LiveBackup Framework 4. Longer description : Live-CD's are created from official OLPC builds. : Unmodified images are converted to squashfs : filesystems and embedded into a : framework which provides everything : needed to run a live system. ... please create a shell account, because we want to use dev.laptop.org to distribute the Live-CD This LiveCD should be very helpful going forward and seems likely to be easily maintained. Having it available on the OLPC servers would be a good thing and I'm strongly for having the project created (I suggested applying for the project in order to have that happen). Take care, Mike -- Mike C. Fletcher Designer, VR Plumber, Coder http://www.vrplumber.com http://blog.vrplumber.com ___ Devel mailing list Devel@lists.laptop.org http://lists.laptop.org/listinfo/devel
Re: clean installation of 643.
Bert Freudenberg wrote: On Dec 4, 2007, at 9:25 , Yoshiki Ohshima wrote: Thank you, John, The autoinstaller is not working properly. This is a known problem, but given that the preferred method of upgrade is now over the network, it has not been a priority. ... For me at least, the XO's wireless is still totally unreliable. So I'd appreciate if the download-to-usb upgrade process was still supported. Having just updated my XO last night via USB I'd suggest that it's useful too (olpc-upgrade wasn't even on my machine as far as I could see, so it wasn't really an option). BTW, work around for me was to do something like the following (keep in mind, I am *not* sure what the implications of doing this are, this is so the people who know can say egads, don't do that if I'm doing something wrong): Plug in the USB Key (MP3 player), hold X to get the firmware and then the backup done. When the machine starts showing the failed to boot unhappy-face, interrupt and go to the OFW prompt (ok). Manually boot from the USB key... boot u:\boot\OLPC.FTH which lets the upgrade continue (at least for me). That suggests it's probably just something to do with timing or similar issues that prevented the initial attempt to read the file from USB. HTH, Mike -- Mike C. Fletcher Designer, VR Plumber, Coder http://www.vrplumber.com http://blog.vrplumber.com ___ Devel mailing list Devel@lists.laptop.org http://lists.laptop.org/listinfo/devel
Re: boot failed after updating MP to joyride-357 (activation?)
Tomeu Vizoso wrote: Hi, After rebooting, a sad face appeared, 'BOOT FAILED' was written at the top left and under the laptop icon was a lock. Booting the alternate image (pressing the 'O' button while powering on) worked fine. What did I wrong? Whatever it was, I had the same symptoms (on my never-as-far-as-I-know-locked B4). I worked around it by starting the OFW prompt and manually telling the machine to boot u:\boot\OLPC.FTH which then completed the upgrade process. Your mileage may vary. HTH, Mike -- Mike C. Fletcher Designer, VR Plumber, Coder http://www.vrplumber.com http://blog.vrplumber.com ___ Devel mailing list Devel@lists.laptop.org http://lists.laptop.org/listinfo/devel
Re: WSJ
Edward Cherlin wrote: On Nov 25, 2007 10:57 PM, Albert Cahalan [EMAIL PROTECTED] wrote: Mike C. Fletcher writes: If we are serious in our goal of educating children, we need to do a few things, and some of this requires a readjustment and a detachment from the question of which hardware goes where. Does the hardware matter? Yes. Yes, the low power consumption, wireless mesh networking, camera and microphone, and non-toxic construction matter. The known set of ports matters. Knowing exactly how much memory is available matters for some purposes. But those are all nice-to-haves, compared with being able to run the software at all. We can already run a nice subset of the laptop software from a Live CD on almost any x86 computer, including Macs. It would not take much effort to create a new ISO file with an up-to-date image. A closed platform has some advantages. The fact is, though, that we no longer have a closed platform. There *will* be Classmates and EEEs out in the field. Children should *not* be blocked from our educational content just because they got the wrong kind of computer[1]. Despite the special hardware, most of the XO is pretty much bog standard at the coding level. It's an i586 class processor with a web cam accessed via v4l and a microphone. It has a very high resolution screen (which requires a bit of special coding, but the emulators can run at 72 dpi already with a small hack). It has some game buttons, we need to support those, but they're mapped to extended keyboard keys by default anyway. Resource compression ratios and fixed screen-sizing are fine, but to reach a few million extra kids it's probably worth re-running the converter and doing a bit of testing on the screen layout. The most difficult changes to adapt to are probably in the aggressive suspend stuff. Most machines are going to go nuts if the OS tries to suspend and resume them every fraction of a second... so we might have to use a less aggressive suspend... maybe set an integer somewhere... and then as the other machines get aggressive-suspend kernels and test them, we can ratchet down the number for them too. Standardized hardware means I can count on having things, and I can optimize. There is a microphone. There is an input that can measure DC voltage levels. There is a camera that does 640x480. The side of the camera with the user's face is quite predictable. There are 3DNow! instructions. JPEG compression levels can be made appropriate to the display. Images can be in ideal sizes. The kernel and X server don't need to ship a zillion drivers. The control panel can be managable. Again, for a few million children, it's probably worth porting to another machine. It becomes 3 set hardware platforms instead of 1. It requires resources that we're short on, no question, but it isn't a project on the order of a new OS, it's a project on the order of a few weeks of a guru Fedora sysadmin to figure out what's needed and compile and package that. we should port to the other inexpensive laptops, if a country decides to go with EEEs or Classmates, we should be in there offering an EEE or Classmate-optimised Sugar + Activities + Content that they can load onto those machines This is a mixed bag. It dilutes the message. The message is already diluted, or perhaps polluted. Intel and Microsoft have seen to that. It is vital that XO software run unaltered but with fewer features on alternate computers that lack some of its hardware. Then the fact that free software that runs on their machines is actually better on our cheaper XO laptop is a major market advantage. The XO is iconic, and it is powerful as a delivery mechanism for the message of making education available to the underprivileged children around the world. But in the same way as the crank was iconic, and was a powerful delivery mechanism for the message, the XO is just part of the message. The XO is already winning the battle to get inexpensive computers to the forefront of manufacturers' minds. In doing so, it has convinced those manufacturers to build their own machines. We should not be so tied up in a single message (particularly one about selling laptops) that we subvert our own goals. There is something like 50% of the world without reliable power, and the XO was designed for that 50% of the world. The fact is that none of the other manufacturers are going for that 50% yet. There's plenty of room in the pool. For the other 50%, a Classmate or an EEE might be the choice of their government. If we abandon those children's education on the alter of the purity of our message, what are we really about? It's an education project, not a laptop project. That is our proper message. It's right there on the about page. It's the message we should be unwilling to subvert. Choose a laptop model, we have one that is designed for this kind of deployment
Re: WSJ
Mitch Bradley wrote: Mike C. Fletcher wrote: It's not About the Hardware: In principle, that is true. In practice, it is the hardware that has been responsible for all the attention. If the project had been just a software framework to support constructivist education, the worldwide response would have been ho hum, yet another program/operating system/GUI/whatyoumaycallit. Agreed. The hardware was key in forming a vision. The idea of a laptop so inexpensive that it could be given away to the children in the developing world en-mass was an idea that caught the minds of the world. So too was the crank (impractical as it would have been in that position). The hardware captured the imagination, and I'm still of the belief that it's a marvel of engineering. It has generated excitement and enthusiasm and has made the project fly. The physical joy of the design is inspiring. The sturdy presence of the device is comforting. The XO is marvelous. I still think we need to get the XO hardware out to the 50% that the manufacturers aren't targeting. We need the XO hardware to reach the children in the back woods. But I would suggest we need to bring Asus and Intel/Classmate into the project so that we can reach those whom the hardware will not reach. The message we will do what is necessary to give the children access is powerful. We've already shown that, given an absence of a tool, we'll put together a multi-year engineering effort to provide that tool. That's a powerful message, and the XO is the iconic result of that effort. What I'm suggesting is that now what is necessary is to partner/connect/work with these other organizations to make sure we aren't fragmenting our efforts. I love the cute hardware... Mike -- Mike C. Fletcher Designer, VR Plumber, Coder http://www.vrplumber.com http://blog.vrplumber.com ___ Devel mailing list Devel@lists.laptop.org http://lists.laptop.org/listinfo/devel
Re: WSJ
Bernardo Innocenti wrote: On 11/25/07 18:58, Mike C. Fletcher wrote: * we should port to the other inexpensive laptops, if a country decides to go with EEEs or Classmates, we should be in there offering an EEE or Classmate-optimised Sugar + Activities + Content that they can load onto those machines o we should also port to the thin-client-style setups seen in e.g. Canonical's deployments of computing labs in the developing world That would be a good idea, but we clearly lack internal resources. Yup. All the code is out there and I bet everybody in the Sugar team would be glad to help whoever wants to port it to the Classmate or any other laptop. Which is where I come in, I suppose. I need to find someone who wants to do the port. Ellis has suggested they're willing to donate an EEE and maybe even do some load it and see if it runs tests for us once we produce an image for them. I just need to find some Fedora Guru who wants to help change the world to do the work. I unfortunately don't know all that many Fedora Gurus (just one, actually, and he's already working on the project). If people know a sysadmin Fedora Guru who they think would be interested, let me know. ... AFAIK, both the mainstream desktops were far too bloated to be usable on the XO. I can easily believe it, as the latest versions are almost unusable on my 1GHz iBook G4, too. Porting Gnome or KDE to the laptop would require help from the respective teams to further optimize them and stripping down some advanced features and some configurability. I seem to have confused a lot of people on that point, I was referring to running Sugar on other machines that normally run Gnome/KDE, not running Gnome/KDE on the XO. * we need to get our installation requirements on non-Fedora Linux down to a package-level installation o (and have this be supported and maintained (preferably internally)) o (also very useful for encouraging developers) I wonder what's the status of Debian and Ubuntu for running on the OLPC. Once the platform part works well out of the box, installing sugar should be just a matter of using alien. Debian and Ubuntu are the farthest along here AFAIK. They have packages. The packages don't seem to like living alongside a jhbuild installation (hosed my laptop when I tried that), but they seemed to run the packaged apps fine. I need a few days I don't have to sit down and test that further. At this point, not even Fedora officially supports the OLPC out of the box! I'd like to see our kernel and platform bits go upstream and appear in all mainstream Linux distros. Even if we cannot afford to put resources on this, I'm sure the core developers would be glad to answer questions on IRC and on this list. Yes, again, we need sysadmin-guru people to do this kind of work. We need to see supporting such people as important, but the core team is likely too busy to do much work on it. If we are serious about educating children, and we truly believe that the software and content we are creating is key to allowing children to learn well, then we need to make that software and content available for all of the projects that are sending computers out in the service of educating children. I couldn't agree more on the general principle, but operating systems and desktops are just a small subset of educational content, and not even a very interesting one for teaching to little children. We shouldn't let ourselves (as OLPC) get distracted in porting 20 different Linux distros to the laptop when we're missing a good astronomy and chemisrtry activity. I'm not suggesting that we need to core developers to do this work. For any given Linux distribution, porting should be just a matter of getting someone from the distro who is interested in porting. Most of the work will be quite mechanical I would think, just a matter of figuring out how your distribution deals with SRPMs as foreign packages and using those to build your native packages. Then all the fun of conflict negotiation and the like, of course. What we might need from the Core devs is a way to kick off a Sugar session as a desktop shell from GDM/KDM/XDM (i.e. the multi-user stuff), and some thought on whether running in a multi-user environment is going to cause problems somewhere. I don't know the mechanics of that interaction all that well, but I'm guessing it's a pretty trivial amount of code in the core. Standardisation and Application Availability: [...] As much as I appreciate software diversity and the network effect it enables, I'm not convinced there's much educational value in providing the full range of Fedora (or Debian) packages. Our target audience is, for the most part, really not interested in traditional GUI applications and even less in UNIX
Re: WSJ
centres. Teacher Training: Again, a quoted problem is teacher training concerns. Peru did an intensive program in the pilot programme where teachers were given one-on-three training by the deployment team. Uruguay AFAIK just handed the laptops out. Nepal has a teacher training programme under development AFAIK. Experience seems to suggest that the students are learning the machines reasonably quickly with other students leading the way, but that's no reason not to make materials available. Addressing the Issue: * We need to have the Teacher Training documented as it is done o What was done o What was covered o Where did teachers have problems o How long did the training need o Where were teachers starting from * In effect, we need to collect together a model training curriculum for teachers that countries can use as a starting point o yes I'm aware of the implications of cultural imperialism here, but again, we offer a model that can be adapted, used or rejected We should probably also be asking Teachers and Countries to share their class plans and curricula so that other teachers and countries can see what's involved in the integration effort. Anyway, just don't want us to throw away a chance to improve ourselves... we are an education project and should take every opportunity that presents itself to learn. Mike [1] http://online.wsj.com/article/SB119586754115002717.html?mod=googlenews_wsj [2] As Nicholas has explicitly stated on a number of occasions. -- Mike C. Fletcher Designer, VR Plumber, Coder http://www.vrplumber.com http://blog.vrplumber.com ___ Devel mailing list Devel@lists.laptop.org http://lists.laptop.org/listinfo/devel
Re: Consistent sound
Eben Eliason wrote: I, too, find this a very nice idea. Not only does it encourage culturally appropriate sounds, but it simply encourages context appropriate sounds in general. Right now there is a lack of audible feedback in a number of the games on the system, and said feedback would really enhance the experience, especially for kids. This would give developers a simple way to accomplish this without having to locate, record, package, and play their own sounds unless they want to, which also saves some space. Of course, I too lack any technical knowledge in this area, but am interested to see where it leads. The basic concept seems to be that of a system notification, as seen on all modern desktops. We'd need to integrate with Sugar and Bitfrost, and implement the generic emotional notifications for games and the like, but otherwise the implementation should be quite familiar. That should make it fairly straightforward to implement. I have a request from a consulting course at U of T to act as a client on a project for one of their students. I'll propose this as the project if people are amenable. Assumptions: * DBUS Interface o This lets non-python activities use the same interface o Python wrapper can be provided o A pipe-level interface might also be useful for games written by new coders (open pipe, write happy 2.5s\n) * Sugar Control Panel extensions to customise the sound-scape for each user (just think how quickly many people just *have* to shut off the Windows start sound) o Alternately, a GUI on the daemon that allows for customisation * Visual Notification options when muted o For the deaf/hard-of-hearing/classroom use * Severely restricted environment for the daemon/service o Access to sound files, preferably just those in the default set(s) plus those explicitly loaded by the user into the application's work-space via the configuration tool o Access to sound hardware o Access to current volume setting (read-only, likely) * Support Localisation for default sound-sets * Classic Notification Set (window actions, system startup/shutdown, that kind of thing) o Likely taken from an existing free-software system to start quickly * Emoticon Notification Set (emotional content, e.g. for games) with intensity setting * Sound implementation(s) o GStreamer sources + OGG files + Wav files o CSound scores * Visual notification implementation o Will require some coordination with the Sugar peoples to provide a non-obtrusive overlay notification mechanism Have fun, Mike On 11/19/07, Samuel Klein [EMAIL PROTECTED] wrote: On Nov 19, 2007 12:31 PM, Gerard J. Cerchio [EMAIL PROTECTED] wrote: Hi, I am just beginning to get involved with the OLPC so please forgive me if this topic is already covered. I am also new to python so my learning curve is somewhat steep. I am attempting to build a go game activity starting with the connect_activity. I would like to produce sounds for various game play events such as victory, loss, atari, etc. I would also like the sounds the child hears from the OLPC be consistent and culturally appropriate. I would posit that if all the activities made consistent utterances to the child, the value of the OLPC learning experience would be enhanced. I would like to suggest a way to give a this consistent localized sound personality to the OLPC through the csound object. I propose that there be a simple csound method: emote( emotion, intensity ) where emotion is a string index into a table of localized sounds intensity is an integer that regulates the degree of the emotion Sample emotion strings would be: win - produces a reward sound appropriate to the locale lose - opposite of win yes - indicate acceptance no - opposite of no warn - indicate more thought may be required approval - encourage disapproval - discourage Maybe this covers more than just emotions; I have no comment on how to implement this in csound sugar, but it is a charming idea. SJ ___ Devel mailing list Devel@lists.laptop.org http://lists.laptop.org/listinfo/devel ___ Devel mailing list Devel@lists.laptop.org http://lists.laptop.org/listinfo/devel -- Mike C. Fletcher Designer, VR Plumber, Coder http://www.vrplumber.com http://blog.vrplumber.com ___ Devel mailing list Devel@lists.laptop.org http://lists.laptop.org/listinfo/devel
Re: [design] Lack of built-in serial and // port ?
, this must often be provided by burning petrol or diesel) * dead-on-arrival technology * integration costs * shipping costs (it costs a *lot* to collect and ship old computers, chap at a conference a few weeks ago was mentioning that he'd spent a quarter of a million shipping old machines to Africa from North America... hmm, should have asked how many machines, old machines tend to be large and heavy, particularly desktops and their monitors) which is not to say it's not worth doing (indeed, it has helped in many areas), but you need to balance that against the positives when you decide which path to take. The OLPC project went down a different path, but working with projects that took the reuse path is a desirable outcome. So when you consider the target audience (poor countries, rural zone, poor people), I think this is a design mistake. Or at least it's reducing a lot the possibility of hacking and recycling obsolete hardware lying around. Always open to the possibility that we've made a critical mistake. The trade-offs we've made are many, and there are going to be thousands of examples of things where someone has to work around a limitation in the project. That said, moving forward we're going to have to provide bridges to various other technologies. Some of those bridges might include providing a pointer to a cheap USB-serial adapter that schools or ministries of education can order. Similarly for parallel port adapters, though I'm more skeptical about those myself. Maybe and probably do you have already talk about this matter ? Indeed, but doesn't hurt to re-iterate the conversation for those just joining in, Mike -- Mike C. Fletcher Designer, VR Plumber, Coder http://www.vrplumber.com http://blog.vrplumber.com ___ Devel mailing list Devel@lists.laptop.org http://lists.laptop.org/listinfo/devel
Could we get a feasibility assessment on this from hardware/production?
We've been having a long-ish thread over on security regarding a possibility mitigation strategy for black/grey-market resale of educational laptops while selling laptops in G1G1 or similar approaches. The proposed mechanism is as follows: * For the G1G1 (or other non-educational/government purchased) runs, alter the configuration file which controls the colours to be applied to the laptop lids, so that the X-O symbol on the lid of all personally-purchased laptops will be a single distinctive colour (suggest white, to increase difficulty in painting over the symbol, but any distinctive colour would be fine) the thread on the security list is long and goes into a lot of details, but we don't yet have feedback from those on the physical manufacturing side as to whether we can make this change at all, and whether it would be disruptive to the runs. With thanks, Mike -- Mike C. Fletcher Designer, VR Plumber, Coder http://www.vrplumber.com http://blog.vrplumber.com ___ Devel mailing list Devel@lists.laptop.org http://lists.laptop.org/listinfo/devel
Re: stderr/stdout msgs in sugar
Victor Lazzarini wrote: Hello everyone, when debugging an activity how does one get stderr/stdout messages from Python in sugar. It is quite hard to debug blindly, not knowing what errors Python is spewing out. Start your activity from a developer's console (ALT-=) using: sugar-activity YourActivityName and the console should be treated as stdin/out/err for your activity. http://wiki.laptop.org/go/Developers#How_do_I_test_a_Sugar_activity.3F HTH, Mike -- Mike C. Fletcher Designer, VR Plumber, Coder http://www.vrplumber.com http://blog.vrplumber.com ___ Devel mailing list Devel@lists.laptop.org http://lists.laptop.org/listinfo/devel
Re: Vmware image
ankita prasad wrote: Hi, I downloaded the OLPC-595.zip vmware image for the XO from http://dev.laptop.org/pub/virtualbox/. Everything works but for my usb devices, which are not recognized at all. When I plug a usb device and move it to my VM, i get an message saying usb 2.0 might not be supported(not verbatim) and the device may not behave as expected. ls -l on /proc/bus/usb shows no files. Why is this happening? Do i need to install usb 2.0 support? If so, how? You don't mention the host platform or the VMWare version you are using. I'm going to assume you are running on VMWare 5.x on Linux. I've never managed to get USB devices working with any virtual machine on my boxes, btw, and that didn't change just now when I went to map a usb key drive into my vm. Apparently people do it all the time, but it doesn't work *here* anyway. Anyway, reading around at: http://forums.pcper.com/showthread.php?t=430890 http://www.vmware.com/community/thread.jspa?threadID=94556tstart=0 seems to suggest that various versions of vmware just don't support USB 2.0 at all and suggest using e.g. a USB 1.1 hub in order to interface, or to use Player 2.0, which apparently does support USB 2.0. Good luck, and HTH, Mike -- Mike C. Fletcher Designer, VR Plumber, Coder http://www.vrplumber.com http://blog.vrplumber.com ___ Devel mailing list Devel@lists.laptop.org http://lists.laptop.org/listinfo/devel
Re: Power envelope (was Re: radio off guarantee?)
John Watlington wrote: Suppose I'm someplace where I don't expect or want to do any mesh networking. How much would turning off the radio help battery life? Last I checked, the effect was very small. There will be occasional scans as the unit hunts around for nearby radios. One could save more by making those scans back off rather than provide a UI element. We see 700 to 800 mW consumed by the mesh interface, and as with most WiFi interfaces, receiving consumes as much power as transmitting. Isn't that about 1/2 of our total power budget? .7 to .8 W on a 2W machine? I'd thought the Marvell chip was supposed to be down in the .3W range. Obviously haven't been following the hardware closely enough. (http://wiki.laptop.org/go/Power_Management is where I got the .3W idea). Is that .7W something we'll be able to bring down with software to reduce collisions or the like? University of Toronto's system's group has algorithms that optimise for power savings by reducing collisions, if that would help. From the same page, we'd only last 20 hours at ~.7W draw with a *full* charge (and with hand-charging a full charge likely won't be there, especially at the end of the day), which means that the machines are going to be dead each morning (having drained their batteries keeping up an unused network all night). With a 40 hour period (.3W) we were possibly going to have some juice left in the morning (need to be at 1/4 power when you shut off for the night), but that becomes less likely with a 20 hour period (need to be at 1/2 power when you shut off for the night). While it makes sense to turn off the wireless networking interface on developer machines, we are hesitant to add this to the UI. We are really relying on laptops to extend the mesh away from the school. If we really are spending half our power budget on the mesh network I would imagine that kids will want to be able to turn it off. Yes, meshing is good, but if you could double your battery life while you're at home reading, that would be very worthwhile too. What about *only* keeping the mesh up (continuously) if you are actually routing packets? That is, wake up the mesh interface every few minutes, check to see if your new routing structure makes you a link that is desirable, and only stay on if that's the case. (I realise I'm talking about long-term projects here). You'd need the machines to be able to queue up messages to go out so that they could leap at the request and say yes, please stay up when a linking machine pops on to check for need (implying from activity requests on the local machine would be likely be sufficient, but a simple UI might be workable too). If the network is inactive for X period, go into periodic sleep mode again to save energy. Even if you had to wake up the processor for a few seconds to run the code that decides whether to sleep, it's only about double the power of running the mesh for those seconds, and it could save you the entire power budget for a few minutes. Of course, in the field it may be that the mesh is so densely used that it's never going to go down with that algorithm, but it seems like something we need to investigate if we really are losing that much power to the interface. Just a thought, Mike -- Mike C. Fletcher Designer, VR Plumber, Coder http://www.vrplumber.com http://blog.vrplumber.com ___ Devel mailing list Devel@lists.laptop.org http://lists.laptop.org/listinfo/devel
Re: Auto-building emulator downloads
C. Scott Ananian wrote: We've been running and building qemu images on xs-dev; you might consider doing root-stuff on a machine which is not dev. xs-dev already has qemu, etc installed. Seems like the building of the VMWare/VirtualBox images has been temporarily shunted off to too busy status [1], so I guess it's up to me to do get it arranged in the near term. I've modified the image-building script so that it knows the build numbers (instead of using timestamps). I'd like to have the script cron'd up so that it runs every X period and picks up the latest devel releases. I can only to run it every week-or-so on my workstation due to a limited-bandwidth account between me and the OLPC servers (both very slow and having a low total transfer cap). Current script only requires a qemu install, no root required. bzr branch http://www.vrplumber.com/bzr/buildemulator Scott, picking up on your (possible, implied) offer, any chance we could run that conversion/upload script on xs-dev? It would require that we have a no-password-required ssh key on xs-dev that would give that account access to my dev.laptop.org account to do the uploads. Other than that there shouldn't be too many issues (maybe storage space, but we can change the script to delete the results after upload to fix that, I leave the results around so that I can test the vmdk's and see if they work). If it's not practical to run on xs-dev, no problem, I'll run the script here each week. Current images are not running under VMWare or VirtualBox[2], but the old 557 image is running with the conversion process as currently written (it no longer creates VMI files), so it seems like we have a real regression there, rather than a problem with the conversion. I'm producing .zip files for download so that both Win32 and Linux/Unix users can access them easily. Don't worry too much about upgrading -- as long as your developer tools are in /home/olpc they will be preserved across updates by the upgrade magic which is going to land on Monday. --scott Is that magic already in-place? If so, how is it getting triggered, and does it go to the latest version? Or is it only going to stable versions? That is, can a developer just run their VMWare image and have it automatically update itself to the latest build over the network? Anyway, have fun, Mike [1] https://dev.laptop.org/ticket/3112 [2] https://dev.laptop.org/ticket/3503 -- Mike C. Fletcher Designer, VR Plumber, Coder http://www.vrplumber.com http://blog.vrplumber.com ___ Devel mailing list Devel@lists.laptop.org http://lists.laptop.org/listinfo/devel
API Reference and Tutorials please check/update your libraries
I've been collecting pointers to (and generating) documentation for the various components in the system into what I'm intending to become central locations for developer documentation on the project. If you see some resource that developers need or would find useful that I've missed, please add it. There are two pages, the first is for tutorial-style documentation pointers: http://wiki.laptop.org/go/Tutorials if you have or know about a developer's tutorial that isn't there, please add it. Eventually I'd like to get a lot more sample code and tutorials in there. The second is a collection of pointers to reference documentation for developers: http://wiki.laptop.org/go/API_Reference At present a lot of this documentation is actually auto-generated pydoc stuff, particularly for the in-house stuff. For much of the GTK-related stuff this is of little value due to the lack of method signatures or useful docstrings. If you are the owner of any of these packages (e.g. abiword, hippo, hulahop) and have documentation available online, please update your package's entry to point to the documentation. Note: the various Sugar services and the Sugar GUI Shell are not documented because they are not importable and are using top-level executable code. We should probably fix that structuring at some point. Scripts to auto-generate the documentation from within an image are available via: bzr branch http://www.vrplumber.com/bzr/autodoc cd autodoc ./builddocs.py for those who want to make the pydoc-generated documentation their primary documentation format. It's not pretty, but at least it's automatic :) . Have fun all, Mike -- Mike C. Fletcher Designer, VR Plumber, Coder http://www.vrplumber.com http://blog.vrplumber.com ___ Devel mailing list Devel@lists.laptop.org http://lists.laptop.org/listinfo/devel
Re: Auto-building emulator downloads
John (J5) Palmieri wrote: ... There's a few: * Qemu installation (for the qemu-img tool that converts to VMWare disks) * VirtualBox installation (for the vboxmanage tool that converts to VirtualBox disks) No need for this as VB can run VMWare disk fine. Duh! Reading the feature-list of the package would be a good thing I suppose. Thanks for the heads up on that. I already have scripts in pilgrim to do a sdk build which outputs a vmware image with vmware and VB meta files. Let's build on that and I can do sdk builds along with regular builds. Yay! J5 rocks. I just need to know what people think should be in a development image. BTW the build servers run as root and have qemu on them. Let's start with just the raw official images for now. I expect to have a few groups needing input into the SDK, including software and content developers (Mel can likely help with collecting input there). For right now we probably need to start with make something available and let people tell us what they are missing and/or tell us what they are doing to make the environment more useful for themselves. Over time we can flesh out what the developer/content-developer SDK overlay is composed of. Continuous-builds-are-good-for-iterative-development-ly yr's, Mike -- Mike C. Fletcher Designer, VR Plumber, Coder http://www.vrplumber.com http://blog.vrplumber.com ___ Devel mailing list Devel@lists.laptop.org http://lists.laptop.org/listinfo/devel
Auto-building emulator downloads
We are getting very close to having fully-working VMWare and VirtualBox emulation. Qemu still seems to have some problems with network and sound, but those seem solvable. We are probably to the point where we can consider auto-building emulator packages for download from the official images. It's not that the building is all that complex, but it's a lot of fiddly steps for people to have to go through for every release they want to try. To start addressing that, I've hacked together a quick script that downloads the official images and uses qemu and VirtualBox to generate VirtualBox and VMWare images. It also has commented-out code that attempts to merge in an overlay directory before doing the conversion, but so far I haven't got that to actually work (mostly because merging the config file I want to merge (the olpc.collabora.co.uk server setting in .sugar/default/config) seems to kill some logic so that resulting images won't display the initial login prompt). Eventually we'd want the overlay to have all the developer-specific stuff people need, developer documentation links, scripts for setting up links with the host machine for development, all that fun stuff. That is, we'd want a developer to be able to download the image and just start working. Super-cool extra points if we can arrange to have a separate disk as a *run-time* overlay so that the developers can update the image to current build while keeping their work and settings (that's beyond my Fedora-foo, I'm afraid). What we'd need to work this is a box somewhere with fast access (for sharing the eventual products), qemu and VirtualBox installed, and the ability to have the user running the script mount/unmount loopback files (for the overlay-integration stuff). At the moment it's a bit hacky about determining versions (just checks the md5 sum against everything it has downloaded), I suppose it could download the ftp-over-http directory listing to determine available builds. VMWare is already handled fairly straightforwardly, the script produces files from a template file, though the settings in that template are just the values I happen to use (a more stable set might be useful). VirtualBox doesn't *appear* to have an easy way to ship a machine+disk image set, it wants the images registered separately from the machines. I don't have all that much experience with the package, though, so it may have a way to set up a simple download. Anyway, the script, the current VMWare template and the rest is available via bzr with: bzr branch http://www.vrplumber.com/bzr/buildemulator Have fun all, Mike -- Mike C. Fletcher Designer, VR Plumber, Coder http://www.vrplumber.com http://blog.vrplumber.com ___ Devel mailing list Devel@lists.laptop.org http://lists.laptop.org/listinfo/devel
Re: Auto-building emulator downloads
Bert Freudenberg wrote: On Aug 26, 2007, at 12:01 , Mike C. Fletcher wrote: We are getting very close to having fully-working VMWare and VirtualBox emulation. Very cool. Could you make a working VirtualBox image of the latest stable release available? That in itself would be incredibly useful. Which brings me back to the need for a server with a reasonable connection on which to run the conversion. My home cable connection gets ~ 4Kbps upstream (which we discovered with uploading the Developer's Images, stupid throttling), which makes even a 200MB upload painful (14 some-odd hours). I'm thinking we can just put a box at 1cc (or wherever) to do the conversions. Installing Qemu takes a few minutes on any reasonable Linux machine and the script will produce the VMWare image within a matter of seconds from the downloaded file (all automated), *much* faster than my uploading it from way out here. VirtualBox is the one that has both network and sound working, btw. VMWare only has networking at the moment (just a missing driver). BTW, stable release 542 does *not* have the drivers, so it will *not* work. Have fun, Mike -- Mike C. Fletcher Designer, VR Plumber, Coder http://www.vrplumber.com http://blog.vrplumber.com ___ Devel mailing list Devel@lists.laptop.org http://lists.laptop.org/listinfo/devel
Re: Auto-building emulator downloads
Ivan Krstić wrote: On Aug 26, 2007, at 6:28 PM, Mike C. Fletcher wrote: Which brings me back to the need for a server with a reasonable connection on which to run the conversion. I just gave you a shell on dev. Things you put into ~/public_html/virtualbox will appear at http://dev.laptop.org/pub/virtualbox . Cheers, Okay, I guess we'll try uploading from here for now. Not as bad as before, looks like about an hour per 200MB file. Still, building them somewhere not behind a cable modem would be a good idea. Have fun, Mike -- Mike C. Fletcher Designer, VR Plumber, Coder http://www.vrplumber.com http://blog.vrplumber.com ___ Devel mailing list Devel@lists.laptop.org http://lists.laptop.org/listinfo/devel
Re: Auto-building emulator downloads
Ivan Krstić wrote: On Aug 26, 2007, at 7:07 PM, Mike C. Fletcher wrote: Still, building them somewhere not behind a cable modem would be a good idea. Right, I assumed giving you a shell on dev would let you convert the images right on that machine. Are there other requirements? There's a few: * Qemu installation (for the qemu-img tool that converts to VMWare disks) * VirtualBox installation (for the vboxmanage tool that converts to VirtualBox disks) * The overlay stuff also requires: o ability to mount a loopback partition (to modify the ext3 image with the overlay) o ability to do an rsync over top of the mounted image (requires super-user access AFAICS) The first two might be something you could run on dev (though I'd be hesitant to increase your attack surface for the machine that way), but the overlay operations are not the kind of thing you want running on a publicly accessible server AFAICS, just too many security issues, really. You'd want to run it on an old box somewhere with nothing sensitive on it and then push the images up to dev I would think. Enjoy yourself, Mike -- Mike C. Fletcher Designer, VR Plumber, Coder http://www.vrplumber.com http://blog.vrplumber.com ___ Devel mailing list Devel@lists.laptop.org http://lists.laptop.org/listinfo/devel
Anti-theft traceback on booting 553
Trying to boot build 553 in a VirtualBox environment (testing fix in [1]), I see a failure in the activation code (I'm copying this manually): /antitheft.py, line 17, in run from initutil import activate_fail ImportError: cannot import name activate_fail which suggests that even if I fix the error it's just going to fail anyway, it's just not telling me nicely. Any suggestions on how to get past this point while running on a virtual machine? I've tried mapping a USB key into the VirtualBox client with a build 553 upgrade/install image, but it didn't do the activation, producing the same traceback. Almost there, he thinks, Mike [1] http://dev.laptop.org/ticket/917 -- Mike C. Fletcher Designer, VR Plumber, Coder http://www.vrplumber.com http://blog.vrplumber.com ___ Devel mailing list Devel@lists.laptop.org http://lists.laptop.org/listinfo/devel
Re: Power manager specification... (request for comments).
Walter Bender wrote: Lets please be careful not to over-engineer. While Mike makes good points, we have this wonderful human social network we can depend upon as well. E.g., If I am downloading something from your machine, I can ask you to hold on a second until I finish. Let's take advantage of the fact that the kids are in the same community/school most of the time and not worry so much about corner cases until we have some more breathing room. Perhaps I was unclear, I was referring to another program *on the same machine* needing the machine to remain active in order to complete a task. That is, things such as: * playing music in the background while you read * downloading a file from the school server which you need for homework while you are reading the assignment/textbook * recording an audio stream while you are taking notes in write (i.e. interviewing someone) * upgrading the OS/downloading updates while you are working in class that is, all single-user on a single-machine use cases. No social networks involved. These are all fairly common use cases. We see the same basic inhibit_suspend operation in media-players on Linux or Windows wrt screen savers. You don't want your machine going to sleep in the middle of recording your interview, so the recording activity needs a way to say hey, I'm working, don't go to sleep right now (assume we've got a signed activity that can do background record for a moment). I'm not saying that the implementation needs to be there today, I'm just saying that we likely need to eventually move toward an implementation that uses need-based signaling rather than direct manipulation of the whole computer state by any activity. Something as simple as a decorator icon on the home view, (e.g. a bed with a circle-plus-slash over it for western audiences) should be sufficient to let the user know which activity is inhibiting suspension. Hope that's clearer, Mike ... On the original topic of the thread (what the power manager should do): I'm guessing eventually we'll want some of the logic currently in the read activity to migrate into HardwareManager. That is, allow for signaling inhibit_suspend( ) and allow_suspend()[3], rather than directly setting suspend, such that a given activity can declare that it must be allowed to continue processing in the back-end. Then you'd want something like suggest_suspend() so that a foreground activity can tell the system hey, I don't expect to do anything for a second or two, if no-one objects, feel free to suspend. From there, a second level does a suggest-suspend from Sugar (or whoever) on no-cpu, no-network (other than the autonomous routing), no-input, for a given period. No opinion on where/how to put that. HardwareManager should likely send dbus events so that activities can watch for resume, suggested-suspend, or what have you and adjust behaviour accordingly. Example usage scenario: switch a per-second clock-updating timer to a per-minute timer. Hope this helps, Mike [1] http://dev.laptop.org/git.do?p=projects/read-activity;a=blob;f=readactivity.py;h=3eeb858cc5ea1dc67a60faee90628100479509be;hb=HEAD [2] http://dev.laptop.org/git.do?p=hardware-manager;a=blob;f=hardwaremanager.py;h=3154b17553621cc41fa947cbff2756372e6e37ec;hb=HEAD [3] with allow-suspend happening automatically after a short-ish timeout if the activity doesn't re-assert the inhibition -- Mike C. Fletcher Designer, VR Plumber, Coder http://www.vrplumber.com http://blog.vrplumber.com ___ Devel mailing list Devel@lists.laptop.org http://lists.laptop.org/listinfo/devel
Re: Power manager specification... (request for comments).
Jim Gettys wrote: ... There seems to be no way to manually adjust the backlight level in ebook mode. What would be the policy for it? Leave it on all the time or dim down after some idle time? Also CPU (and wireless?) should go to suspend for most of time in ebook mode? We have code in at least our PDF viewer to literally do exactly this; as soon as the page is rendered and it is idle, it puts the machine to sleep. For the activity developers in the audience: This is (I believe) in the GIT read-activity/readactivity.py[1] module. The implementation uses the HardwareManager[2] service (set_kernel_suspend) exposed over dbus to directly send the kernel into suspended state. AFAICS there's no allowance for tracking whether something *else* might need the machine to be alive (e.g. a download or the like being done in the background). The same Manager object has controls for various operations such as changing brightness and the like. On the original topic of the thread (what the power manager should do): I'm guessing eventually we'll want some of the logic currently in the read activity to migrate into HardwareManager. That is, allow for signaling inhibit_suspend( ) and allow_suspend()[3], rather than directly setting suspend, such that a given activity can declare that it must be allowed to continue processing in the back-end. Then you'd want something like suggest_suspend() so that a foreground activity can tell the system hey, I don't expect to do anything for a second or two, if no-one objects, feel free to suspend. From there, a second level does a suggest-suspend from Sugar (or whoever) on no-cpu, no-network (other than the autonomous routing), no-input, for a given period. No opinion on where/how to put that. HardwareManager should likely send dbus events so that activities can watch for resume, suggested-suspend, or what have you and adjust behaviour accordingly. Example usage scenario: switch a per-second clock-updating timer to a per-minute timer. Hope this helps, Mike [1] http://dev.laptop.org/git.do?p=projects/read-activity;a=blob;f=readactivity.py;h=3eeb858cc5ea1dc67a60faee90628100479509be;hb=HEAD [2] http://dev.laptop.org/git.do?p=hardware-manager;a=blob;f=hardwaremanager.py;h=3154b17553621cc41fa947cbff2756372e6e37ec;hb=HEAD [3] with allow-suspend happening automatically after a short-ish timeout if the activity doesn't re-assert the inhibition -- Mike C. Fletcher Designer, VR Plumber, Coder http://www.vrplumber.com http://blog.vrplumber.com ___ Devel mailing list Devel@lists.laptop.org http://lists.laptop.org/listinfo/devel
Re: Python pyc/pyo (Was: Inflation)
Bernardo Innocenti wrote: ... No other dynamic language needs that pre-tokenized junk. Is Python so slow at parsing? Has anybody measured the hit? There is some minor improvement in speed from the cached byte code. That said, there's no reason to *ship* them, you can run a compileall script on installation (doesn't help with filesystem space usage though). You only need one of pyc or pyo, choose whether we run with optimisation on or not and use that version of the compiled cache. IIRC there are flags to the Python executable that can skip generating the byte code if you wanted to test it. I'm guessing you'll see a noticeable but not huge slowdown in startup time for activities. Alternately you can just try deleting all of the pyc and pyos and seeing if the activities slow down in starting up. You'd also want to time the whole-system boot, since Sugar is written in Python as well. Someone told me that Python 2.5 can now also read library files from an archive, and this is supposed to reduce startup time a lot. (I don't see why it should, on any decent filesystem). My understanding is that slowdown is seen due to large number of stats during module import. Python does run-time lookup of each module, and with more involved PYTHONPATHs that can be 90 or 120 stats per import (you have to check for py, pyc, pyo, so and pyd versions (as well as un-decorated names for packages) for each name in the import and paths of 30 directories can happen readily). Multiply by a few thousand imports in a really big application and it can start to add up. In contrast, the zip importer has loaded the zip directory table already on the top-level resolution, so can do simple in-memory hash lookups for each name from that point on. That's about the same effort required as a second import of the same module name. HTH, Mike -- Mike C. Fletcher Designer, VR Plumber, Coder http://www.vrplumber.com http://blog.vrplumber.com ___ Devel mailing list Devel@lists.laptop.org http://lists.laptop.org/listinfo/devel
Re: System update spec proposal
Christopher Blizzard wrote: On Wed, 2007-06-27 at 12:39 -0400, Mike C. Fletcher wrote: Could we get a summary of what the problem is: The main objection to vserver from all the kernel hackers (and those of us that have to support them!) is that it's a huge patch Okay, so a set of smaller, more targeted patches is preferred. that touches core kernel bits Isn't that going to have to happen in all cases where you can provide total isolation of the various elements? I'd think that adding the capability-base control to the major sub-systems would require some modifications. Of course, how core and how extensively are probably the real issue. and it has no plans to make it upstream. That sounds like a serious problem for maintainability, yes. Alright, so from my understanding of vserver it sounds like what we want is something along these lines: * chroot fixes o probably the least intrusive and most widely useful part of the system is patches to patch a couple of security holes in the chroot system (fchdir hole and the mknod hole IIRC) o goal of this is to make the chroot perform reliably and reasonably securely to provide file-system isolation o might be possible to pull just this work out of vserver * overlay/COW filesystem o goal here is to allow for a tree of overlays with read/only read/write support on each layer o aufs is the closest project at the moment? * hardware access capabilities and rate limiting o memory o cpu o disk io o network interfaces where we would want each of those patches to be as small, maintainable and elegant as possible, with all targeted for upstream inclusion. I'm guessing that it's the third set that cause the patch-size to balloon, as it seems rather involved. The first two alone, though, should provide quite a lot of the functionality we want and be reasonably general in their interest. So I guess we'd want three kernel hackers (or so) to work on the three projects simultaneously. First is probably a small project. Second is likely just a matter of massaging the code. Third is an involved project that would likely have to be working with SELinux and the like to try to integrate everything. Anyway, that's just the way it sounds to me, Mike -- Mike C. Fletcher Designer, VR Plumber, Coder http://www.vrplumber.com http://blog.vrplumber.com ___ Devel mailing list Devel@lists.laptop.org http://lists.laptop.org/listinfo/devel
Re: System update spec proposal
Alexander Larsson wrote: On Tue, 2007-06-26 at 15:03 -0400, Mike C. Fletcher wrote: ... I'm not sure what you mean here exactly. Discovery is done using avahi, a well known protocol which we are already using in many places on the laptop. The actual downloading of file uses http, which is a well known protocol with many implementations. The only thing i'm missing atm is a way to tell which ip addresses to prefer downloading from since they are close. This information is already availible in the mesh routing tables in the network driver (and possibly the arp cache), and its just a question of getting this info and using it to drive what servers to pick for downloading. Ah, somehow in the discussions I'd come under the impression that the only way the system would be allowed to work was a single-hop network link on the mesh. If we already have the information and can always have a fallback, even if it means going a number of hops across the network. Looking back I see that was actually a discussion on bandwidth characteristics that wasn't intended to imply an absolute requirement. I'm reasonably happy with the approach of using Avahi and only using the network topology to inform the decision on which server to use. That said, I would be more comfortable if the fallback included a way for the laptop to check a well-known machine every X period (e.g. in Ivan's proposal) and if there's no locally discovered source, use a publicly available source as the HTTP source by default Hmm, interestingly I see using our own tool as a disadvantage, not a huge one, but a disadvantage nonetheless, in that we have more untested code on the system (and we already have a lot), and in this case, in a critical must-never-fail system. For instance, what happens if the user is never connected to another XO or school server, but only connects to a (non-mesh) WiFi network? Does the mesh-broadcast upgrade discovery protocol work in that case? Avahi works fine for these cases too. Of course, since it was originally created for normal networks. However, if you never come close to another OLPC machine, then we won't find a machine to upgrade against. Sorry, should have been clearer that the later case (not coming close to another OLPC) was the one I was concerned about. I realise such situations will represent only a small percentage of children, but a small percentage of 50,000,000 or so users is a huge number of people to have to teach how to manually upgrade their machines. Its quite trivial to make it pull from any http server on the net, but that has to be either polled (which I don't like) or initiated manually (which might be fine). I'd advocate that the piece of Ivan's proposal wherein a central mechanism allows even a completely isolated machine to find and update automatically is a good idea. It's a fairly trivial proposal that way, in the same check to see if we've been stolen download 4 bytes (or so) telling us the currently available version. If we can't get it locally after X period, try to get it from the server (using whatever protocol, be it your own, rsync, BitTorrent or Telepathy (cute, I was actually writing telepathy there and then realised it was the name of our networking library)). That is, I'd like to see a robust, automatic mechanism for triggering the laptop's search and a fallback position that allows for resolution even in the isolated-user case that doesn't require user intervention. Enjoy, Mike -- Mike C. Fletcher Designer, VR Plumber, Coder http://www.vrplumber.com http://blog.vrplumber.com ___ Devel mailing list Devel@lists.laptop.org http://lists.laptop.org/listinfo/devel
Re: Ivan's XO Field Upgrade Proposal
Dan Williams wrote: ... I'm not arguing simplicity; just that we have to be aware of the implications of having lots of XOs pulling from the server with some overlap with this method, but we don't with XO-XO. We just have to make the tradeoffs clear, and understand them. Agreed. Given the simplicity of setting it up, the approach should be given consideration. I would judge that having this robust and simple from the start (when we'll be having lots of updates and a far greater chance for failures to happen) to be more important than bandwidth pre-optimisation. At a later point, if we can be sure that the core system software image is always distinct from the sensitive data and user data overlays, we can readily provide a service at some point that allows the laptops themselves to advertise an rsync-based source running in a special chroot with just those (read-only) layers. The overlay manager could initiate an rsync service on the laptop in response to a tubes request (or whatever) in order to allow for XO-to-XO sharing at some point. Have fun, Mike -- Mike C. Fletcher Designer, VR Plumber, Coder http://www.vrplumber.com http://blog.vrplumber.com ___ Devel mailing list Devel@lists.laptop.org http://lists.laptop.org/listinfo/devel
Re: Ivan's XO Field Upgrade Proposal
John (J5) Palmieri wrote: On Tue, 2007-06-26 at 12:32 -0400, Dan Williams wrote: ... We Also have to remember the countries want control over when the boxes update. At least that was the impression I got at the country meetings. This is a concern, no? In cases of regime change and the like, can a particular user opt out of the upgrades somehow? Although it's a good idea to have the laptops automatically update themselves by default I'm somewhat concerned if a new regime coming in can disable all of the laptops in the country, or force them to install something that the users themselves don't want. This concern is one of the reasons I want to allow for overlay file systems for the Core Operating System images, the install may be downloaded and automatically installed, but if it does something nasty the child can disable that overlay and potentially decide to move to a different update source (e.g. change countries) using the security configuration UI mechanism (or whatever). On a less tin-foil-hat note, I would be most comfortable if all system updates were notifying the user and allowing them (again, likely part of the security UI) to temporarily delay an update. 30 kids in the middle of a video-capture-and-editing activity for the school play that's taxing their systems and their network to the limit are going to be a little upset if the board of education decides that's the moment to start a whole-system update. Just a thought, Mike -- Mike C. Fletcher Designer, VR Plumber, Coder http://www.vrplumber.com http://blog.vrplumber.com ___ Devel mailing list Devel@lists.laptop.org http://lists.laptop.org/listinfo/devel
Re: System update spec proposal
Ivan Krstić wrote: On Jun 26, 2007, at 2:21 PM, Christopher Blizzard wrote:. Also I would have appreciated it if you had given direct feedback to Alex instead of just dropping your own proposal from space. It's a crappy thing to do. Let's not make this about approach on a public mailing list, please. Actually, while I found the response a bit harsh, could I suggest that what the project needs is *more* public discussion all around, not necessarily about approach, but about half-formed plans, ideas and rationale. Having discussion move offline into some private channel is a good way to prevent anyone outside the offices from knowing *why* things are happening and having things blow up when the decisions appear to come from on high. For instance: * Whole projects are surfacing after weeks or months of development. Papers and implementations are starting and stopping without anyone knowing what's going on or, more importantly, knowing *why* they have been done. Witness the immediate counter-proposals to Alex's implementation of the point-to-point protocol. * VServer only appeared in public discussions yesterday or so AFAIK, yet it's apparently already the chosen path for doing the system compartmentalization. We need more draft-level discussions, more discussion of plans, rationales, ideas and approaches. The discussions don't need to be long, they don't need to be formal and well structured, they just need to be sufficient to let people have an idea what they need to write for in a few months. I realise we're on a very tight schedule, but it's extremely difficult to help with the project if we don't know what's going on inside it. Assume good will and proper intentions on the part of all people until *proven* wrong (repeatedly), and only then assume ignorance of the proper path until proved wrong, and only then assume misguidance yet a thirst for knowledge until proved wrong. Even if proved wrong many times, attempt to find a way to solve the issue politely and respectfully. We are all working to make a better world, and there should be no egos involved if we are doing things right. Anyway, just my thoughts, Mike -- Mike C. Fletcher Designer, VR Plumber, Coder http://www.vrplumber.com http://blog.vrplumber.com ___ Devel mailing list Devel@lists.laptop.org http://lists.laptop.org/listinfo/devel
