| Beacon11 | Hey guys | 04:17 |
|---|---|---|
| persia | Hey Beacon11 | 04:18 |
| Beacon11 | I have a question | 04:18 |
| persia | which one? | 04:19 |
| Beacon11 | I have a via pico itx (really teeny motherboard) and I have a project I'm working on with it. Sounds big, but I have the university of idaho behind me. Basically, I'm making a really small tablet PC | 04:19 |
| Beacon11 | And initially (this project started a little over a year ago) I was running kubuntu on it | 04:20 |
| Beacon11 | I'm using a lilliput 629GL 7-inch touchscreen | 04:20 |
| Beacon11 | And a few days ago, I saw a blog post about ubuntu mobile | 04:20 |
| Beacon11 | I immediately threw away kubuntu after trying ubuntu mobile live | 04:20 |
| Beacon11 | But... now I'm having issues getting my touchscreen working | 04:21 |
| Beacon11 | And I'm not sure if its ubuntu mobile itself, or simply the fact that it's 8.10 now and something changed | 04:21 |
| Beacon11 | My initial question is this: should the calibration utility included work with my touchscreen? | 04:22 |
| persia | The calibration utility only works with touchscreens supported by the evtouch driver (I think). No idea if yours falls into that category. | 04:23 |
| Beacon11 | It does, evtouch is what I used to get it working previously | 04:23 |
| Beacon11 | But I couldn't get the utility to work before either :P | 04:23 |
| Beacon11 | So I ended up hacking xorg.conf to work correctly | 04:24 |
| Beacon11 | Yeah... took a while | 04:24 |
| persia | The utility has changed completely since 8.04. The version in 8.04 didn't work for any screens. | 04:24 |
| persia | You might need an .fdi file for your screen though, if it's sufficiently special. | 04:24 |
| Beacon11 | Hmmm... well I tested the calibration, and it then restarted X, and nothing changed. No changes to my xorg.conf either | 04:25 |
| Beacon11 | What exactly is the utility supposed to do? | 04:25 |
| Beacon11 | Oh | 04:25 |
| Beacon11 | Hmm | 04:25 |
| Beacon11 | You're over my head there– what's a .fdi file? | 04:25 |
| persia | I think the utility writes some hints for the size metric for the screen. I forget what fdi is supposed to mean, but it's a file to help HAL set the properties for a given piece of hardware. | 04:26 |
| persia | (and I'm talking over my head as well, so apologies if something is lost in translation) | 04:26 |
| Beacon11 | Don't worry about it, sounds like you know more than I do! | 04:27 |
| Beacon11 | Okay, I can work on that. However... even if the calibration isn't working like I need it to, in kubuntu I just basically followed the directions here http://www.conan.de/touchscreen/evtouch.html and it worked. I created a udev rule for it, but other than that, I followed those directions | 04:28 |
| Beacon11 | But now, when I make those changes to ubuntu mobile's xorg.conf (which is very sparse compared to what I'm used to seeing) x fails to restart | 04:29 |
| Beacon11 | What's different now? | 04:29 |
| persia | 8.04 vs. 8.10. If you grab the Kubuntu 8.10 RC, you'll get the same sparse xorg.conf. | 04:30 |
| Beacon11 | For instance, in the xorg.xonf generated by ubuntu-mobile, there IS no section "ServerLayout" | 04:30 |
| Beacon11 | Ah! | 04:30 |
| persia | X 1.5 is using XInput2 hotplug to determine the input devices from HAL. | 04:30 |
| Beacon11 | Okay, so this has completely changed | 04:30 |
| persia | Yep. Some stuff works *lots* better, like adding USB or bluetooth mice to laptops. Some stuff needs different sorts of hacking, like touchscreens. | 04:31 |
| Beacon11 | Hey, I'm all for improvement. So do you think that looking for the correct .fdi file for my monitor is my best bet? | 04:31 |
| persia | As the body of knowledge about hardware grows, the appropriate hints can be given to HAL so it works for everyone : that way once a known hack for xorg.conf exists, it can be given automatically to everyone with that hardware, rather than people passing around snippets on web pages. | 04:32 |
| persia | Unfortunately, the data collection part is still underway, so it is a little harder to get stuff that doesn't work by default working for now. | 04:32 |
| Beacon11 | Ooo, I think it's worth it | 04:32 |
| Beacon11 | What a great idea | 04:32 |
| persia | Well, .fdi file for the touchscreen : I suspect the display function works fine. | 04:32 |
| Beacon11 | Oh... yeah, you're right | 04:33 |
| Beacon11 | Also, I say this blog post http://www.umpcportal.com/2008/09/ubuntu-mobile-edition-news-and-first-boot-video | 04:33 |
| persia | Yeah. I'm entirely in favour of the idea, even though it breaks a few things. It's completely worth it. | 04:33 |
| Beacon11 | In the comments, ogra said to email him lshal output, and he'll add support in the next image | 04:33 |
| Beacon11 | Would you recommend that? | 04:33 |
| persia | I think he underestimated the number of people who would install Ubuntu Mobile :) I'd recommend filing a bug against xserver-xorg-input-evdev with your lshal output. | 04:34 |
| persia | ogra will still see it, but it also gets tracked, and you'll automatically get an email when it is closed. | 04:35 |
| Beacon11 | Ha! Well I consider that GOOD news | 04:35 |
| persia | I also :) | 04:35 |
| Beacon11 | Okay, well I'll definitely file that bug. Not having much luck with an fdi file for my touchscreen though... I may just have to wait | 04:35 |
| persia | Well, I think the HAL package has some docs on generating an .fdi file. If you look at the ones that come with the xserver-xorg-input-evdev package, you might be able to generate one for your device. | 04:36 |
| Beacon11 | Hey, good idea | 04:37 |
| persia | (attaching a working .fdi file to a bug is even more likely to get it supported than attaching lshal output only (but please also include lshal if attaching an .fdi file) | 04:37 |
| persia | This is the best part of open source :) One can fix one's own bugs. | 04:38 |
| Beacon11 | Hahaha, I can't even decide what the best part is | 04:38 |
| Beacon11 | Maybe that if you can't fix your own bug, you're nearly guaranteed someone else can ;) | 04:39 |
| persia | Personally, I think all the other good bits come from someone else fixing it before I noticed the bug :) | 04:39 |
| persia | (because they could fix their own bugs) | 04:39 |
| Beacon11 | True, true | 04:39 |
| Beacon11 | Hey, does umobile include all .fdi files? For instance... could I hunt one down somewhere on my system for another touchscreen? | 04:42 |
| persia | It doesn't include *all* .fdi files : the idea is to only include the ones each driver needs. The xserver-xorg-input-evtouch package contains a few touchscreen .fdi files. | 04:45 |
| Beacon11 | Okay, thanks :) | 04:47 |
| Beacon11 | Alright... in a bug report, I found an fdi file for an eGalax touchscreen from Samsung. My touchscreen, although from lilliput, also identifies as eGalax. Do you think this fdi file could potentially work? | 04:49 |
| persia | Quite possibly. Drop it in /etc/hal/fdi/policy and see if it helps. | 04:51 |
| Beacon11 | Alright | 04:52 |
| Beacon11 | An fdi file has a line that says this: <match key="info.product" contains="Touchkit Touch"> | 05:04 |
| Beacon11 | Would that info.product be the line printed in dmesg by any change? | 05:04 |
| persia | It would be the info.product entry from lshal. | 05:05 |
| Beacon11 | Ah! | 05:05 |
| Beacon11 | I think this is going to work | 05:05 |
| Beacon11 | So my touchscreen tells lshal what it's capable of, basically? | 05:07 |
| persia | Right. And then the drivers ask hal which devices they should manage. | 05:08 |
| persia | The .fdi files are the glue that makes this work. | 05:08 |
| Beacon11 | Nice! Actually... this is a lot cleaner than hacking the xorg.conf. I just put all the dirty details in the .fdi file. It almost looks exactly the same! | 05:11 |
| Beacon11 | Okay... so using the .fdi files, I don't need udev rules anymore do I? | 05:11 |
| Beacon11 | (for my touchscreen, that is) | 05:11 |
| Beacon11 | Because hal, using the fdi files, grabs it wherever it is anyway? | 05:12 |
| persia | You probably still need the udev rules to create the devices for HAL to monitor, unless your udev rule was just to get a static name for xorg.conf. | 05:37 |
| Beacon11 | I was thinking it was just to get a static name for xorg. But here's what I have, you can tell me if I'm right: | 05:44 |
| Beacon11 | KERNEL=="event*", SUBSYSTEM=="input", ATTRS{name}=="eGalax Inc.", SYMLINK+="input/evtouch_event" | 05:44 |
| persia | I think you can drop that. Try it without it first. | 05:44 |
| Beacon11 | Sure thing | 05:44 |
| persia | It ought take two of input/event* | 05:44 |
| Beacon11 | So I see all these input.x11_options.* in my lshal. Are those default values? If I don't have an option in my fdi, it'll resort to those values? | 05:48 |
| Beacon11 | For instance: in my lshal, I see input.x11_options.taptimer = '30' (string) | 05:50 |
| Beacon11 | Does that mean that a line like the following is unneeded in my fdi file? <merge key="input.x11_options.taptimer" type="string">30</merge> | 05:50 |
| persia | I suspect that comes from your .fdi file. | 05:51 |
| persia | If you get that without a .fdi file, then yes, you don't need it. | 05:51 |
| Beacon11 | I do get that without an fdi file | 05:52 |
| Beacon11 | Okay, cool | 05:52 |
| persia | Then you probably don't need it. | 05:52 |
| Beacon11 | So now, say I have this fdi file. The calibration utility writes to it, I'm assuming? | 05:54 |
| Beacon11 | So the fact that I had no fdi file means the calibration utility had nothing to write to...? | 05:54 |
| persia | I think the calibration utility writes to somewhere else, although I'll admit to not really understanding it. | 05:54 |
| Beacon11 | Okay, no problem | 05:55 |
| === merriam_ is now known as merriam | ||
| Beacon11 | Ha! Persia, you're a lifesaver. My touchscreen is now working :) | 06:27 |
| Beacon11 | Thank you so much for your help | 06:28 |
| Beacon11 | I will now file a bug report | 06:28 |
| persia | Beacon11, Thanks for porting Ubuntu to work with your hardware :) | 06:32 |
| Beacon11 | I can't wait for this ubuntu-mobile thing to go further... it's exactly what I needed to bring my project to the next step | 06:32 |
| Beacon11 | So anything I can do to help ;) | 06:32 |
| persia | Well, this week, not so much. Archive freeze is tomorrow or so, and then there's release processing, etc. | 06:34 |
| persia | Next release prep doesn't start until mid November or so. | 06:34 |
| Beacon11 | All I care about is that work is continuing | 06:34 |
| persia | At that point, if you want to test the upcoming version, and help submit patches to fix even more bugs, I'm sure they would be accepted. | 06:35 |
| Beacon11 | Nice | 06:35 |
| persia | I don't see any reason why it should stop. It seems popular, and the release team is willing to distribute it. As long as we all test and submit patches, it should become what we all want. | 06:35 |
| Beacon11 | Good, I was worried about the user base | 06:35 |
| persia | Well, since it's never been released, that's probably pretty small right now. I expect it will get larger. | 06:36 |
| Beacon11 | Good | 06:36 |
| Beacon11 | It's awfully well done for a beta | 06:36 |
| persia | I've been showing it off on some of my devices, and people are interested. You might do the same, if you want to get more people involved. | 06:36 |
| Beacon11 | There are only a few complaints I have (and the touchscreen isn't one of them) | 06:36 |
| persia | Well, it's based on a strong foundation :) | 06:36 |
| Beacon11 | Yeah! | 06:37 |
| Beacon11 | Actually | 06:37 |
| persia | What sort of complaints? I could probably point you in the right direction for them. | 06:37 |
| Beacon11 | I'm a computer engineering major at the university of idaho, and I've been asking advise from several professors, who have requested to see the project ;) | 06:37 |
| Beacon11 | Oh, just beta ones, like how polished the web browser is (with the button to switch from drag to select, etc) but then with the previewer, that little toolbar isn't there and I need to use a fingernail to scroll | 06:38 |
| Beacon11 | Teeny things | 06:38 |
| persia | Dunno how much of that will be fixed for release : it's probably worth filing bugs. I think there's only a few packages getting updated between now and release. | 06:39 |
| Beacon11 | Alright, I'll mention it. They aren't really bugs even... just preferences | 06:40 |
| Beacon11 | I kinda feel guilty :P | 06:40 |
| Beacon11 | But for me, that little toolbar at the bottom of the screen for the web browser would be very handy to have for other applications | 06:40 |
| persia | If you think they're just preferences ask someone in #ubuntu-bugs to set them to wishlist after filing them. | 06:41 |
| Beacon11 | Ah, good idea | 06:41 |
| persia | That's actually part of the browser : it would have to be done per-program. I agree that it'd be neat to pull the iconbar from the top to the bottom though, although I don't know how that would be done globally. | 06:41 |
| Beacon11 | I'm thinking make that toolbar more of an applet. That way, one could make it launch with certain programs | 06:42 |
| Beacon11 | Welll... applet... I guess I'm thinking more of a plugin | 06:43 |
| Beacon11 | Used to ruby web apps :P | 06:43 |
| persia | I guess. A lot of programs I use have a button bar on the top on the desktop. I'd think it would be easier just to find a way to pull that to the bottom. | 06:43 |
| Beacon11 | True | 06:43 |
| Beacon11 | Uh ohh... I had a working fdi file, but it wasn't quite on. So I ran the calibrator... and now it's all messed up. So I thought "Okay, hopefully it just edited my fdi file and I can copy back over it." But no... the fdi file is the same. How do I make it forget the calibration? :P | 06:46 |
| Beacon11 | I can't find much info about that | 06:47 |
| persia | I don't know. Maybe xorg.conf? | 06:48 |
| Beacon11 | Nope. Aw dang! | 06:48 |
| Beacon11 | It's gonna be somewhere that's gonna make me feel stupid, I know it | 06:48 |
| persia | Fastest way is probably to look at the source for the calibration tool. | 06:50 |
| Beacon11 | To be honest I've never done that before. How would I go about this? | 06:51 |
| Beacon11 | The evtouch home page? | 06:52 |
| persia | mkdir evtouch-source; cd evtouch-source; apt-get source xserver-xorg-input-evtouch | 06:52 |
| Beacon11 | Thank you :) | 06:52 |
| Beacon11 | Never done the apt-get source, I'll remember that | 06:53 |
| persia | The goal is to make it easy, to encourage patches :) | 06:53 |
| Beacon11 | I assume you've used this utility, correct? | 07:25 |
| persia | Only a couple times. I never investigated it in depth. One of my touchscreens isn't evtouch, and the other just works. | 07:27 |
| Beacon11 | Oh, gotcha | 07:27 |
| persia | I do a lot of apt-get source though, if you have questions about that :) | 07:27 |
| Beacon11 | Hmm.... this code prints to the screen, the utility doesn't | 07:28 |
| persia | That's very confusing. That code is what generates the utility. | 07:28 |
| Beacon11 | It prints to a relative output file, and the screen. I have a hunch the final utility grabs the screen output and does something with it | 07:28 |
| Beacon11 | Where is that code... *searching* | 07:28 |
| Beacon11 | Because that was only the c code, not the sh | 07:29 |
| persia | That sounds like a reasonable theory. | 07:29 |
| Beacon11 | Dang... what does "mknod /tmp/ev_calibrate p" do? | 07:30 |
| Beacon11 | Nevermind, man pages | 07:32 |
| persia | I think it creates a named pipe (man mknod may be helpful). | 07:32 |
| Beacon11 | Well... I know that with this code, it requires the option Calibrate to be 1, and the reporting mode must be raw. However, I read elsewhere (now for some reason I can't remember where... maybe the bug reports) that the utility takes care of that | 07:41 |
| Beacon11 | If that's the case... then the code for that isn't exactly the same. Right? | 07:41 |
| persia | I'm not sure. | 07:42 |
| Beacon11 | Okay... so the code communicates using fifos, which is what the mknod did | 07:54 |
| Beacon11 | Ah | 08:08 |
| Beacon11 | The calibration utility USES this code | 08:08 |
| Beacon11 | But it is indeed seperate code | 08:08 |
| Beacon11 | separate... | 08:08 |
| Beacon11 | The utility is calibrate_touchscreen, which calls ev_calibrate. ev_calibrate is the compiled version of this code | 08:09 |
| Beacon11 | So... I need the source for calibrate_touchscreen. How would I do that? | 08:09 |
| Beacon11 | Oh... it's just a shell | 08:10 |
| Beacon11 | Duh | 08:10 |
| Beacon11 | Oh clever | 08:12 |
| Beacon11 | Too bad I haven't done more of this :S | 08:17 |
| persia | It's a great time to get started : you can get familiar with the processes, and be all prepared for archive open :) | 08:18 |
| Beacon11 | No kidding! I still can't figure out where it puts that info though! It seemed like everything in the program were tmps, but there's a fifo that remains constant in all of these– tmp/ev_calibrate. The program reads from it, then closes it. These scripts create it, call the program, them remove it. I can't figure out anything beyond that, really :P | 08:21 |
| persia | Odd. It has to store it somewhere. | 08:22 |
| Beacon11 | Yeah... it tosses the screen output of the program to /dev/null, it seems | 08:23 |
| Beacon11 | But although I can program, my in-depth linux experience is still relatively weak. I think I'm losing something there | 08:23 |
| * persia finds https://lists.ubuntu.com/archives/intrepid-changes/2008-September/007575.html which provides a clue, and suggests looking in the changelog for more | 08:24 | |
| Beacon11 | Hmm | 08:24 |
| Beacon11 | No way :P | 08:25 |
| Beacon11 | Oh man, it's right there! | 08:26 |
| Beacon11 | \/etc/evtouch/config | 08:26 |
| persia | changelogs are cool :) | 08:27 |
| Beacon11 | (don't irc often enough to know how to bypass the forward slash) | 08:27 |
| persia | space | 08:27 |
| Beacon11 | Oh :P | 08:27 |
| Beacon11 | Hmm, the log also says that it includes .fdi files to eGalax screens. Why wasn't one installed for my setup? | 08:28 |
| persia | It was : remember you said you found a .fdi file for the Samsung eGalax screen? | 08:28 |
| persia | Apparently that file didn't also work for yours. | 08:29 |
| Beacon11 | Online, though | 08:29 |
| Beacon11 | It wasn't included in the installation | 08:29 |
| persia | You might want to look under /usr/share/hal/fdi/policy ... | 08:29 |
| Beacon11 | Ooo | 08:30 |
| Beacon11 | Quite a collection, actually | 08:30 |
| Beacon11 | Hey persia... I tweaked the fdi file until it works well. Should I just commend on this bug or file my own? https://bugs.launchpad.net/ubuntu/+source/xf86-input-evtouch/+bug/261873 | 09:14 |
| ubottu | Ubuntu bug 261873 in xf86-input-evtouch "make evtouch devices work with hal-input in intrepid" [High,Fix released] | 09:14 |
| Beacon11 | comment, rather | 09:14 |
| elegos | hello there | 09:35 |
| elegos | any news about mounting Ubuntu-Mobile under eeePC? | 09:35 |
| elegos | does the ubuntu-mobile appear in the list of the netinstall disk? | 09:57 |
| persia | Beacon11, I'd recommend a new bug. | 10:05 |
| elegos | here I am | 10:06 |
| persia | elegos, ubuntu-mobile doesn't appear in the list of the netinstall disk. | 10:06 |
| elegos | u.u ok | 10:06 |
| persia | Other people have reported success with installing on an Eee, so I'm not sure what's different for you. | 10:07 |
| elegos | is there a way to alternatively install ubuntu mobile? | 10:07 |
| elegos | reading from the net I read something about JeOS but they were still speaking about Hardy | 10:07 |
| elegos | uhm | 10:08 |
| elegos | if I install a command line system | 10:08 |
| elegos | may I then install ubuntu-mobile as a metapackage? | 10:08 |
| persia | Sure. That should work. Or install all of ubuntu-desktop first. | 10:12 |
| elegos | I think that I can't install intrepid on 4GB :P | 10:13 |
| elegos | not a problem: dhclient eth0 && sudo apt-get install ubuntu-mobile :) | 10:13 |
| persia | intrepid should be fine for 4G. | 10:16 |
| popey | 14:43:15 < ogra> persia, popey had to install lbm on a 700 iirc | 10:24 |
| popey | a 900 actually :) | 10:24 |
| elegos | the netinstall asked me if I want to use the "recompiled" packages... are those the backports I need? | 10:36 |
| elegos | hey! there is ubuntu mobile as choice in the netinstall !!! :D | 10:45 |
| * persia is pleasantly surprised, and wonders how it works. | 10:46 | |
| * elegos is hoping it'll work :P | 10:47 | |
| === asac_ is now known as asac | ||
| elegos | the installation is stuck on the configuration of "scrollkeeper" (or something like that)... the HD is doing something, but it's about 5 minutes there... what should it be? | 11:44 |
| elegos | no problem... it's still goiing on :P | 11:47 |
| persia | That step takes a frustratingly long time. Just wait. | 11:47 |
| elegos | on eeePC is also longer then a normal PC | 11:51 |
| persia | Depends on the normal PC :) | 11:53 |
| ian_brasil | mo mention of mobile on http://news.softpedia.com/news/Ubuntu-8-10-Release-Candidate-Screenshot-Tour-96353.shtml | 13:17 |
| ian_brasil | :( | 13:18 |
| persia | That's OK. There wasn't a lot of mention of Mobile in the RC announcement. We need to get the wiki pages organised for the real release, and add a proper note. That should get some attention. | 13:20 |
| persia | Personally, I'm glad not to have too much press before we have a decent landing site. | 13:20 |
| elegos | what is the package I have to install for the Atheros wifi card? | 13:24 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!