Re: [Sugar-devel] XO on Fedora 20 (was Re: [GSoC] Porting To Python3)
These are fantastic news! Thanks Daniel for working on this Gonzalo On Wed, May 7, 2014 at 9:04 PM, Daniel Narvaez dwnarv...@gmail.com wrote: As an update, I have patched xorg-x11-drv-dove and built rpms for it http://shell.sugarlabs.org/~dnarvaez/oob/f20-xo1.75/ Now building an image with those. On 7 May 2014 15:18, Martin Abente martin.abente.lah...@gmail.com wrote: Great! I will try your oob branch for 1.5, I do have XOs 1.5 for testing :) On Wed, May 7, 2014 at 9:01 AM, Daniel Narvaez dwnarv...@gmail.comwrote: On 7 May 2014 01:44, James Cameron qu...@laptop.org wrote: On Tue, May 06, 2014 at 04:24:39PM +0200, Daniel Narvaez wrote: [...] And with the XO stuck on Fedora 18 we might not have good enough introspection to make the port compatible with it. If anybody would like to work on moving olpc-os-builder to something more recent, feel free. It isn't something OLPC is looking at right now, but it would be helpful to the users. Not the die hard 0.98 users, of course. ;-) I'm giving that a try. I was able to build a Fedora 20 image for XO 1.5 https://github.com/dnarvaez/olpc-os-builder I don't have hardware to test that though... I'm now trying to build for 1.75 which is harder but I can actually test. I need to rebuild the X driver but I think that will require some patching, let's see if I can get it to work... ___ Devel mailing list de...@lists.laptop.org http://lists.laptop.org/listinfo/devel ___ Sugar-devel mailing list Sugar-devel@lists.sugarlabs.org http://lists.sugarlabs.org/listinfo/sugar-devel -- Daniel Narvaez -- Gonzalo Odiard SugarLabs - Software for children learning ___ Sugar-devel mailing list Sugar-devel@lists.sugarlabs.org http://lists.sugarlabs.org/listinfo/sugar-devel
Re: [Sugar-devel] Api freeze break request
If we agree then, I will give it a last review and push it. On Wed, May 7, 2014 at 11:06 PM, Walter Bender walter.ben...@gmail.comwrote: Of all the reasons for an except, adding stability to the Journal seems to be at the top of the list. +1 from me. -walter On Tue, May 6, 2014 at 10:42 AM, Martin Abente martin.abente.lah...@gmail.com wrote: I have been testing and reviewing this PR since the last week and I think it would be beneficial for 0.102. On Mon, May 5, 2014 at 6:33 PM, Gonzalo Odiard godi...@sugarlabs.orgwrote: Is a good idea as plan B :) On Mon, May 5, 2014 at 7:25 PM, James Cameron qu...@laptop.org wrote: I don't think I qualify as a developer; no recent coding. I'm in favour of your change, as it is not a public library ABI break, but an enhancement; compatibility is retained. However, if you can't get traction, rather than enhance the ABI, duplicate the new mime.py implementation into src/jarabe/journal/model.py in pr 367 for the release, and for next version move it from model.py to mime.py. This will fix the bug, avoid ABI change, and generate more work later. -- James Cameron http://quozl.linux.org.au/ -- Gonzalo Odiard SugarLabs - Software for children learning ___ Sugar-devel mailing list Sugar-devel@lists.sugarlabs.org http://lists.sugarlabs.org/listinfo/sugar-devel ___ Sugar-devel mailing list Sugar-devel@lists.sugarlabs.org http://lists.sugarlabs.org/listinfo/sugar-devel -- Walter Bender Sugar Labs http://www.sugarlabs.org ___ Sugar-devel mailing list Sugar-devel@lists.sugarlabs.org http://lists.sugarlabs.org/listinfo/sugar-devel
[Sugar-devel] API-string freeze break request
Hello Everyone, Ad-hoc networks have been used intensively in Australia lately, thus some issues have been reported specifically about how ad-hoc auto-connect logic works. Therefore, I have submitted two late enhancements to ad-hoc network management in Sugar [1]. One of these changes requires to add a new gsettings key in order to leave the auto-connect functionality as optional (instead of fixed as it is today) [2]. This breaks the API and string freeze breaks, so I would like to hear other developers feedback about if we should make an exception for this one. Thanks for your time, tch. Refs: 1. https://github.com/sugarlabs/sugar/pull/375 2. https://github.com/tchx84/sugar/commit/46c78e4b4e1a92a157f33d635567577f5270dc35 ___ Sugar-devel mailing list Sugar-devel@lists.sugarlabs.org http://lists.sugarlabs.org/listinfo/sugar-devel
Re: [Sugar-devel] XO on Fedora 20 (was Re: [GSoC] Porting To Python3)
Hello Daniel, I built the image for XO 1.5 but it freezes during boot. It reaches to the stage: Starting Wait for Wait for Plymouth Boot Screen to Quit... Right after that I see these two messages: dcon_freeze_store: 1 dcon_source_switch to DCON Then the screen slowly and gradually turns gray. Any idea what this could be? Any suggestion for debugging it? Regards, Martin. On Thu, May 8, 2014 at 8:16 AM, Gonzalo Odiard godi...@sugarlabs.orgwrote: These are fantastic news! Thanks Daniel for working on this Gonzalo On Wed, May 7, 2014 at 9:04 PM, Daniel Narvaez dwnarv...@gmail.comwrote: As an update, I have patched xorg-x11-drv-dove and built rpms for it http://shell.sugarlabs.org/~dnarvaez/oob/f20-xo1.75/ Now building an image with those. On 7 May 2014 15:18, Martin Abente martin.abente.lah...@gmail.comwrote: Great! I will try your oob branch for 1.5, I do have XOs 1.5 for testing :) On Wed, May 7, 2014 at 9:01 AM, Daniel Narvaez dwnarv...@gmail.comwrote: On 7 May 2014 01:44, James Cameron qu...@laptop.org wrote: On Tue, May 06, 2014 at 04:24:39PM +0200, Daniel Narvaez wrote: [...] And with the XO stuck on Fedora 18 we might not have good enough introspection to make the port compatible with it. If anybody would like to work on moving olpc-os-builder to something more recent, feel free. It isn't something OLPC is looking at right now, but it would be helpful to the users. Not the die hard 0.98 users, of course. ;-) I'm giving that a try. I was able to build a Fedora 20 image for XO 1.5 https://github.com/dnarvaez/olpc-os-builder I don't have hardware to test that though... I'm now trying to build for 1.75 which is harder but I can actually test. I need to rebuild the X driver but I think that will require some patching, let's see if I can get it to work... ___ Devel mailing list de...@lists.laptop.org http://lists.laptop.org/listinfo/devel ___ Sugar-devel mailing list Sugar-devel@lists.sugarlabs.org http://lists.sugarlabs.org/listinfo/sugar-devel -- Daniel Narvaez -- Gonzalo Odiard SugarLabs - Software for children learning ___ Sugar-devel mailing list Sugar-devel@lists.sugarlabs.org http://lists.sugarlabs.org/listinfo/sugar-devel
[Sugar-devel] [ASLO] Release TurtleBots-24
Activity Homepage: http://activities.sugarlabs.org/addon/4434 Sugar Platform: 0.86 - 0.100 Download Now: http://activities.sugarlabs.org/downloads/file/28946/turtlebots-24.xo Release notes: -uses TurtleBlocks v202+ -new translations -force close bb -add block for vel with cc motors -update PyUSB to Beta 2 -update Rodi commit -remove binaries of bluetooth -update extras commit with wedo translations -update NXT commit -com_usb.py not use interface parameter in write/read -update TB and WeDo commits -only load patterns object data one time -fix return type of pattern plugin -add followme example -update .phony of makefile -move _split_module to common functions -allow change client host and port from command line -only uses argv in main call -use same icon -add generic modules to pybot chotox -bluez for nxt -fix permissions -restart udev -reorder one block in followme -fix type in followme -add rodi (juky) plugin -add replace in control to prevent conflicts -add help command in pybot client and server Sugar Labs Activities http://activities.sugarlabs.org ___ Sugar-devel mailing list Sugar-devel@lists.sugarlabs.org http://lists.sugarlabs.org/listinfo/sugar-devel
Re: [Sugar-devel] XO on Fedora 20 (was Re: [GSoC] Porting To Python3)
It would be good to start the kernel with systemd.unit=multi-user.target so that (hopefully) we just get a console instead of dcon freezing/X startup... I'm not sure how to easily pass kernel arguments on the XO though. Does anyone know? On 8 May 2014 18:55, Martin Abente martin.abente.lah...@gmail.com wrote: Hello Daniel, I built the image for XO 1.5 but it freezes during boot. It reaches to the stage: Starting Wait for Wait for Plymouth Boot Screen to Quit... Right after that I see these two messages: dcon_freeze_store: 1 dcon_source_switch to DCON Then the screen slowly and gradually turns gray. Any idea what this could be? Any suggestion for debugging it? Regards, Martin. On Thu, May 8, 2014 at 8:16 AM, Gonzalo Odiard godi...@sugarlabs.orgwrote: These are fantastic news! Thanks Daniel for working on this Gonzalo On Wed, May 7, 2014 at 9:04 PM, Daniel Narvaez dwnarv...@gmail.comwrote: As an update, I have patched xorg-x11-drv-dove and built rpms for it http://shell.sugarlabs.org/~dnarvaez/oob/f20-xo1.75/ Now building an image with those. On 7 May 2014 15:18, Martin Abente martin.abente.lah...@gmail.comwrote: Great! I will try your oob branch for 1.5, I do have XOs 1.5 for testing :) On Wed, May 7, 2014 at 9:01 AM, Daniel Narvaez dwnarv...@gmail.comwrote: On 7 May 2014 01:44, James Cameron qu...@laptop.org wrote: On Tue, May 06, 2014 at 04:24:39PM +0200, Daniel Narvaez wrote: [...] And with the XO stuck on Fedora 18 we might not have good enough introspection to make the port compatible with it. If anybody would like to work on moving olpc-os-builder to something more recent, feel free. It isn't something OLPC is looking at right now, but it would be helpful to the users. Not the die hard 0.98 users, of course. ;-) I'm giving that a try. I was able to build a Fedora 20 image for XO 1.5 https://github.com/dnarvaez/olpc-os-builder I don't have hardware to test that though... I'm now trying to build for 1.75 which is harder but I can actually test. I need to rebuild the X driver but I think that will require some patching, let's see if I can get it to work... ___ Devel mailing list de...@lists.laptop.org http://lists.laptop.org/listinfo/devel ___ Sugar-devel mailing list Sugar-devel@lists.sugarlabs.org http://lists.sugarlabs.org/listinfo/sugar-devel -- Daniel Narvaez -- Gonzalo Odiard SugarLabs - Software for children learning -- Daniel Narvaez ___ Sugar-devel mailing list Sugar-devel@lists.sugarlabs.org http://lists.sugarlabs.org/listinfo/sugar-devel
Re: [Sugar-devel] XO on Fedora 20 (was Re: [GSoC] Porting To Python3)
On Thu, May 08, 2014 at 12:55:32PM -0400, Martin Abente wrote: I built the image for XO 1.5 but it freezes during boot. It reaches to the stage: Starting Wait for Wait for Plymouth Boot Screen to Quit... Right after that I see these two messages: dcon_freeze_store: 1 dcon_source_switch to DCON Are these messages on display or on serial port? If on display, attach a serial terminal and look for more messages. Because once the DCON is frozen, you won't see any display updates. Then the screen slowly and gradually turns gray. The screen does this in the event that the video output from the processor stops. With the DCON frozen, the screen should not do this. I think the DCON has been unfrozen, but the video output has stopped. Any idea what this could be? Any suggestion for debugging it? Attach serial terminal and enquire as to the state of processes responsible for video output. Usually the X server. Check the X server logs. Alternatively, if serial terminal cannot be used (e.g. no cable), build the image with a preloaded SSH key, and a preselected network. Then once the wireless LED shows network connection, use SSH to log in and debug through that channel. -- James Cameron http://quozl.linux.org.au/ ___ Sugar-devel mailing list Sugar-devel@lists.sugarlabs.org http://lists.sugarlabs.org/listinfo/sugar-devel
Re: [Sugar-devel] XO on Fedora 20 (was Re: [GSoC] Porting To Python3)
On Thu, May 08, 2014 at 11:49:39PM +0200, Daniel Narvaez wrote: It would be good to start the kernel with systemd.unit=multi-user.target so that (hopefully) we just get a console instead of dcon freezing/X startup... I'm not sure how to easily pass kernel arguments on the XO though. Does anyone know? Yes. Stop Open Firmware at the ok [1] prompt, then edit the olpc.fth file using microEmacs [2]: ok emacs int:\boot\olpc.fth Scroll down to the extra kernel parameters section: \ extra kernel parameters here to boot-file and change it to systemd.unit=multi-user.target to boot-file Carefully see that the slash has been removed, and that a space is kept after the first double quote character. The slash is a comment character [3]. Write the file with Control/X Control/S, then exit with Control/X Control/C. The file is reset by fs-update. You can also customise olpc.fth in olpc-os-builder. You can also manually pass arguments, but this is more typing and isn't remembered: ok int:\boot\vmlinuz to boot-device ok int:\boot\initrd.img to ramdisk ok systemd.unit=multi-user.target to boot-file ok boot References: 1. http://wiki.laptop.org/go/Ok 2. http://wiki.laptop.org/go/Forth_Lesson_13#Text_editor 3. http://wiki.laptop.org/go/Forth_Lesson_2#Comment_to_End_of_Line -- James Cameron http://quozl.linux.org.au/ ___ Sugar-devel mailing list Sugar-devel@lists.sugarlabs.org http://lists.sugarlabs.org/listinfo/sugar-devel
Re: [Sugar-devel] XO on Fedora 20 (was Re: [GSoC] Porting To Python3)
Hi, On May 8, 2014 at 4:55 PM James Cameron qu...@laptop.org wrote: On Thu, May 08, 2014 at 12:55:32PM -0400, Martin Abente wrote: I built the image for XO 1.5 but it freezes during boot. It reaches to the stage: Starting Wait for Wait for Plymouth Boot Screen to Quit... Right after that I see these two messages: dcon_freeze_store: 1 dcon_source_switch to DCON Are these messages on display or on serial port? If on display, attach a serial terminal and look for more messages. Because once the DCON is frozen, you won't see any display updates. Then the screen slowly and gradually turns gray. The screen does this in the event that the video output from the processor stops. With the DCON frozen, the screen should not do this. I think the DCON has been unfrozen, but the video output has stopped. Any idea what this could be? Any suggestion for debugging it? Attach serial terminal and enquire as to the state of processes responsible for video output. Usually the X server. Check the X server logs. Can you not hold down the X key while powering on to suppress the splash screen? Alternatively, if serial terminal cannot be used (e.g. no cable), build the image with a preloaded SSH key, and a preselected network. Should you not be able to use ctl+alt+F2(F3,F4) to access a command prompt in the other ttys? Then once the wireless LED shows network connection, use SSH to log in and debug through that channel. Jerry ___ Sugar-devel mailing list Sugar-devel@lists.sugarlabs.org http://lists.sugarlabs.org/listinfo/sugar-devel
Re: [Sugar-devel] XO on Fedora 20 (was Re: [GSoC] Porting To Python3)
On Thu, May 08, 2014 at 06:07:44PM -0500, Jerry Vonau wrote: Can you not hold down the X key while powering on to suppress the splash screen? Yes, but it might be frame buffer driver related too. Should you not be able to use ctl+alt+F2(F3,F4) to access a command prompt in the other ttys? Yes, but it might be frame buffer driver related too. Both ideas worth trying, but I guess it depends on the kernel. -- James Cameron http://quozl.linux.org.au/ ___ Sugar-devel mailing list Sugar-devel@lists.sugarlabs.org http://lists.sugarlabs.org/listinfo/sugar-devel
Re: [Sugar-devel] API-string freeze break request
G'day, I've reviewed the patches. I don't think this is worth breaking API and string freeze, as both are new features. I'd like to see them go through the new feature design process. In reviewing the patches, two more new features came to mind: - use higher bandwidth 5 GHz channels for ad-hoc, not only the channels 1, 6, and 11, - secured ad-hoc. With ad-hoc enabled a remote attacker need only transmit packets that prevent use of an access point, and the laptop will revert to unsecured ad-hoc, and begin to reveal useful data to an attacker. -- James Cameron http://quozl.linux.org.au/ ___ Sugar-devel mailing list Sugar-devel@lists.sugarlabs.org http://lists.sugarlabs.org/listinfo/sugar-devel
Re: [Sugar-devel] API-string freeze break request
Personally I think the odds that someone would do a disassociation attack to monitor or SSH into the laptops of Sugar-age students are low enough that it isn't worth designing for. I do not know of any consumer product which currently tries to guard against being forced to search for another network to use. But it would be useful to have the ability to lock to a network (AP, ad-hoc, or otherwise), and not resort to searching NetworkManager's known/preferred list if communication is temporarily lost. 5 GHz ad-hoc is slightly tricky in that there are only a few channels (36/40/44/48) one can manually select; the rest may only be selected via automatic means. According to Wikipedia[*] they seem to be reasonably well accepted -- but in some countries they may only be used indoors (so no under the tree). [*] https://en.wikipedia.org/wiki/List_of_WLAN_channels On Thu, May 8, 2014 at 9:12 PM, James Cameron qu...@laptop.org wrote: G'day, I've reviewed the patches. I don't think this is worth breaking API and string freeze, as both are new features. I'd like to see them go through the new feature design process. In reviewing the patches, two more new features came to mind: - use higher bandwidth 5 GHz channels for ad-hoc, not only the channels 1, 6, and 11, - secured ad-hoc. With ad-hoc enabled a remote attacker need only transmit packets that prevent use of an access point, and the laptop will revert to unsecured ad-hoc, and begin to reveal useful data to an attacker. -- James Cameron http://quozl.linux.org.au/ ___ Sugar-devel mailing list Sugar-devel@lists.sugarlabs.org http://lists.sugarlabs.org/listinfo/sugar-devel ___ Sugar-devel mailing list Sugar-devel@lists.sugarlabs.org http://lists.sugarlabs.org/listinfo/sugar-devel