| 00:26.59 | ljp | my 6000 rom gui doesnt respond to the ts ;( |
| 00:27.50 | CIA-7 | 03kergoth 07 * r1.1771 10packages/matrixssl/ (3 files in 3 dirs): Add matrixssl 1.1.1. |
| 00:27.55 | kergoth | could be using either the COLLIE or CORGI event types |
| 00:28.00 | kergoth | try changing the raw module in ts.conf |
| 00:31.03 | kergoth | also see what happens if you call ts_print_raw |
| 00:31.12 | kergoth | and make sure TSLIB_TSDEVICE points to a device that exists in /dev |
| 00:33.38 | bug1 | ljp, you may need to load one of the -raw modules in your config file |
| 00:33.59 | kergoth | tslib is looking pretty solid right now |
| 00:34.12 | kergoth | should release |
| 00:34.54 | kergoth | bug1: any ideas on the EV_SYN mess? its really annoying to have that as a compile time option |
| 00:35.10 | kergoth | pb suggested either watching for the first EV_SYN event to turn on that processing, or a seperate -raw module |
| 00:35.23 | bug1 | kergoth, yea, i was about mention maybe something should be done about that |
| 00:35.45 | kergoth | i'm leaning towards the former |
| 00:36.10 | bug1 | a seperate ev-syn module probably makes sense, there is minimal shared code between them IIRC |
| 00:36.45 | bug1 | i dont have an EV_SYN device to test, so i didnt touch it |
| 00:37.03 | kergoth | well, with both we're just collecting each event in. the difference is how we know when we've got a matched set of x/y/pressure |
| 00:37.14 | bug1 | i suspect it may have similar issues of the cursor disapearing as the non-EV_SYN code had |
| 00:40.12 | bug1 | i guess there should be a way to build it with both ev-syn and non-ev-syn support, rather than one of the other |
| 00:40.31 | kergoth | yeah, just needs to be reworked a bit |
| 00:41.15 | bug1 | actually another thing to do before release, if no -raw module is loaded apps with segfault when they try and use, NULL pointer there somewhere, should probably chack we have a raw device after the config file is parsed |
| 00:41.22 | kergoth | i wish we could ask the event device if the module supports EV_SYN or something |
| 00:42.14 | kergoth | bug1: not a bad idea. quick fix would be to make ts_read and ts_read_raw check for == NULL |
| 00:42.52 | kergoth | i wonder if i should pull in my linear pressure calibration stuff |
| 00:43.16 | *** join/#oe Cwiiis (~chris@82-43-43-168.cable.ubr02.croy.blueyonder.co.uk) |
| 00:43.17 | kergoth | its kinda nice to bring the pressure values into a known range, so apps could actually have some clue about what the value means |
| 00:43.59 | bug1 | hmm, the check_fd doesnt look to be used anywhere |
| 00:44.06 | kergoth | yeah i saw that |
| 00:44.18 | kergoth | i'd hate to call that on every read.. perhaps the first one? |
| 00:44.27 | bug1 | the touchscreen i use doesnt have pressure messurments |
| 00:44.50 | kergoth | yeah, thats fairly common. but for those that do, and its useful, it could be worthwhile |
| 00:45.06 | kergoth | the problem is, it has to hook into the actual calibration step (ts_calibrate), because it needs access to those original numberes |
| 00:45.10 | kergoth | s/numberes/numbers/ |
| 00:45.56 | bug1 | maybe that check_fd should be called when the module is loaded, |
| 00:46.29 | bug1 | kergoth, seems reasonable to put it in then |
| 00:48.31 | kergoth | well, check_fd is bound to ts_open(), not parse time |
| 00:48.41 | kergoth | and right now there's no hook into a module at open |
| 01:08.40 | Cwiiis | kergoth - Do you know if it's necessary to have an external arm-linux-gcc 2.95 to compile opie-image for oz? |
| 01:17.51 | kergoth | Cwiiis: the external 2.95 toolchain is needed to build the zaurus kernels |
| 01:17.55 | kergoth | so yep |
| 01:18.36 | Cwiiis | kergoth - Ok, thanks |
| 01:26.21 | Cwiiis | kergoth - What do I need to define/add in OEMAKE to define that I have an external 2.95 toolchain? (sorry to ask, I forgot to write it down earlier..) |
| 01:27.10 | kergoth | read the comments in local.conf.sample |
| 01:28.53 | Cwiiis | ok, thanks |
| 01:29.07 | W8TVI | how do you echo hex chars to the serial port? |
| 01:30.08 | kergoth | W8TVI: man echo |
| 01:30.16 | W8TVI | k |
| 01:30.28 | kergoth | echo -e '\011' > /dev/ttyS2 would send ascii code 011 (octal) |
| 01:30.33 | kergoth | other than that, write code. |
| 01:30.43 | W8TVI | the trouble with google is knowing what to google for |
| 01:30.48 | kergoth | heh |
| 01:33.35 | W8TVI | hmmm.... I guess I could convert the hex to octal |
| 01:41.07 | W8TVI | what am I doing wrong? # echo "\001\125\252\000\061\061\061\064\061" |
| 01:41.07 | W8TVI | \001\125\252\000\061\061\061\064\061 |
| 01:41.07 | W8TVI | # |
| 01:41.29 | kergoth | 20:29 < kergoth> echo -e '\011' > /dev/ttyS2 would send ascii code 011 (octal) |
| 01:41.37 | kergoth | and read the man page again. |
| 01:42.06 | W8TVI | d'oh! read it as if -e DISabled that |
| 01:42.14 | *** join/#oe bipolar (~bflong@dsl-66-216-161-056.dejazzd.com) |
| 01:45.05 | Cwiiis | kergoth - When I define I have the external 2.95 toolchain, everything else seems to use it - Is that what should happen? |
| 01:45.18 | kergoth | Cwiiis: you didnt rename the binaries. |
| 01:45.36 | kergoth | Cwiiis: you need to move arm-linux-gcc to arm-linux-gcc-2.95, and arm-linux-ld to arm-linux-ld-2.95 |
| 01:45.42 | kergoth | in /usr/local/arm/2.95.3/bin/ |
| 01:45.50 | kergoth | then itll use it only for the zaurus kernels |
| 01:46.04 | kergoth | its a gross hack, but it gets the job done |
| 01:46.41 | Cwiiis | kergoth - ok, thanks.. Is there a wiki page on building oz using oe that has all this that I'm missing? |
| 01:46.59 | kergoth | dunno, not that i'm aware of |
| 01:47.03 | kergoth | feel free to create one if there isnt |
| 01:47.06 | kergoth | we need better docs |
| 01:48.08 | Cwiiis | ok, if I get this building, I'll see if I can get a page together detailing the troubles I've had.. (and how to overcome them) |
| 01:48.25 | Cwiiis | Does it need to be in /usr/local/arm/... or is that just your particular setup? |
| 01:48.39 | kergoth | whose toolchain are you using? |
| 01:48.40 | Cwiiis | (I have mine in /opt y'see, probably a stupid question, but just making sure) |
| 01:48.49 | Cwiiis | Used the rpm's from the ZUG |
| 01:48.50 | kergoth | when you get a toolchain, its bound to the location it was built for. |
| 01:49.02 | kergoth | no idea whose those are. they're an unknown quantity |
| 01:49.12 | kergoth | i'd personally recommend the one that goes in /usr/local/arm/2.95.3/ |
| 01:49.16 | Cwiiis | They work, just checking that I don't need to change the path or anything |
| 01:49.19 | kergoth | its on the ftp.arm.linux.org.uk ftp site |
| 01:49.25 | Cwiiis | ah, ok, I'll check that out |
| 01:49.35 | kergoth | no, "They work" is meaningless unless you've tested it seriously |
| 01:49.44 | kergoth | what you mean is, they've worked for you so far |
| 01:49.53 | Cwiiis | I've compiled opie with them if that's serious enough |
| 01:49.59 | kergoth | the one on ftp.arm.linux.org.uk is known sane in all situations, including the zaurus kernels |
| 01:50.07 | Cwiiis | ok, I'll use those |
| 01:50.23 | kergoth | its possible the ones you have right now will work, but i'm advising going with what we know works |
| 01:50.33 | kergoth | then once you've got it going, feel free to experiment to your hearts contnet |
| 01:50.34 | kergoth | heh |
| 01:50.59 | Cwiiis | hehe, ok, I'll use those - probably a good idea anyway |
| 02:00.30 | Cwiiis | fingers crossed this will be the successful attempt |
| 02:14.54 | epos | i broke my 5500 stylus tonigth one more time. normally cianocrilate fix it, but this time i lost the two black parts in the garden's grass.. using old (and robust) NEC MobilPro 450 stylus now, heh. re :) |
| 02:16.13 | kergoth | those damn 5500 styluses are so easy to break |
| 02:16.23 | kergoth | i had one of those pen/styluses from cross for a while |
| 02:16.33 | kergoth | worked nice till i lost it |
| 02:16.34 | kergoth | :) |
| 02:17.00 | epos | s/cianocrilate/cyanoacrylate |
| 02:18.23 | epos | kergoth: heh, are this styluses standard ? |
| 02:18.52 | kergoth | there is no standard that i can tell. that one did happen to fit into the Z stylus slot, but was a bit too long, so stuck out a bit |
| 02:20.47 | epos | ok |
| 02:21.24 | bug1 | if we did XLIBS_SRC_FILE="${XLIBS_MIRROR}/${PN}-${PV}.tar.bz2", and a freedesktop.class that sets SRC_URI=${XLIBS_SRC_FILE} then we wouldnt even need to specify SRC_URI in all the xlibs |
| 02:22.36 | bug1 | the freedesktop.class could also "inherit autotools pkgconfig" and then oe's would just need to inhert freedesktop.class |
| 02:22.40 | bug1 | that sound reasonable ? |
| 02:23.31 | bug1 | maybe im lazy |
| 02:24.41 | bug1 | i was thinking of doing oe's for all the stable releases of xlibs, to cut back on the number cvs packages needed |
| 02:45.40 | *** join/#oe treke|laptop (~ggilbert@lsanca2-ar29-4-41-064-125.lsanca2.elnk.dsl.genuity.net) |
| 02:47.20 | treke|laptop | evening |
| 02:47.52 | kergoth | bug1: you dont want to use ${PN} in SRC_URI |
| 02:48.02 | kergoth | bug1: because if you then inherit native in a -native .oe, the sources will be wrong |
| 02:48.15 | kergoth | bug1: thats why we usually repeat the packagename all over, when it seems unnecessary |
| 02:48.30 | kergoth | adding the stable xlibs packages sounds like a good idea |
| 02:50.36 | bug1 | kergoth, ok |
| 02:54.10 | bug1 | ive been trying for ages to get oe to build native apps, ccant workout how to doit under gentoo |
| 02:54.26 | kergoth | what do you mean? |
| 02:54.50 | kergoth | targeting native (TARGET_SYS == BUILD_SYS), or -native .oe files? |
| 02:54.56 | kergoth | they behave differently |
| 02:55.03 | bug1 | i want to try and build a GPE that runs on my thlon |
| 02:55.26 | kergoth | okay, targeting native, you need to force autoconf configures to believe you're crosscompiling |
| 02:55.32 | kergoth | set TARGET_VENDOR |
| 02:55.38 | kergoth | that way TARGET_SYS wont match BUILD_SYS |
| 02:55.42 | kergoth | then itll work |
| 02:55.51 | bug1 | TARGET_ARCH has to be "i386" to compile glibc as far as i can work out |
| 02:56.22 | kergoth | i686 works as well, last i checked. i486 and i586 have issues, as reported by cosmicpenguin |
| 02:57.29 | bug1 | TARGET_ARCH has to match one of the directoreis under sysdeps from what i can tell... but im a bit lost |
| 02:57.40 | kergoth | pb would be the one to ask about that |
| 02:58.53 | bug1 | but TARGET_ARCH is also making its way into the cross-toolchain, so i would have to rename my i686-pc-linux-gnu-* stuff i think |
| 02:59.18 | bug1 | i will have a look at TARGET_VENDOR |
| 02:59.39 | kergoth | well, you could tell OE that your build environment has a toolchain so it wont build it, then override vendor, or override prefix directly |
| 02:59.47 | kergoth | or just let oe build itself a "crosscompiler" |
| 03:32.58 | W8TVI | how do you view data from a serial port on the zaurus? |
| 03:33.05 | kergoth | ? |
| 03:33.10 | kergoth | its linux. |
| 03:33.13 | kergoth | install minicom or something |
| 03:33.18 | kergoth | or just cat /dev/blah > file |
| 03:33.23 | kergoth | or just cat /dev/blah |
| 03:33.27 | kergoth | or install uucp |
| 03:33.36 | kergoth | or some other serial communications app |
| 03:33.49 | W8TVI | I tried cat /dev/ttyS2 and nothing comes up |
| 03:34.02 | kergoth | well then. |
| 03:44.12 | bug1 | hmm, i think i686 works because glibc oe changes TARGET_ARCH to i386, so glibc is really being compiled optimised for i386 if you set TARGET_ARCH="i686" |
| 03:44.30 | bug1 | ithink to detangle it we need a kernel-headers oe |
| 03:45.33 | kergoth | ? |
| 03:45.56 | kergoth | glibc's .oe changes TARGET_ARCH the *shell* variable |
| 03:46.05 | bug1 | ah |
| 03:46.06 | kergoth | whose scope is only within *that* shell script. unpack. |
| 03:46.17 | kergoth | do_configure, if you look at hte run. script, still passes the real one along to configure |
| 03:56.01 | W8TVI | what the hell am I doing wrong? |
| 04:10.55 | bug1 | i think TARGET_ARCH needs to be resolved so in do_compile as well, so it can do "touch ${S}/sysdeps/${TARGET_ARCH}/framestate.c" |
| 04:18.33 | *** join/#oe mithro (~tim@dsl1-83.gw1.adl1.airnet.com.au) |
| 05:44.59 | mjc | mmm |
| 05:45.14 | mjc | is there a decent asm package oin oe? |
| 05:48.00 | *** join/#oe [wobble]_ (~martin@p50832537.dip0.t-ipconnect.de) |
| 05:52.15 | mithro | mjc: asm package? |
| 05:52.24 | mithro | package which uses a decent amount of asm? |
| 05:52.51 | *** join/#oe sit1 (~user@210-55-151-200.dialup.xtra.co.nz) |
| 06:48.41 | mjc | mithro: sorry, asm compiler package |
| 06:51.23 | mithro | ahh |
| 06:52.02 | bug1 | mv: cannot stat `/home/bug1/oe/tmp/base/gcc-cross-3.3.3-r0/image//home/bug1/oe/tmp/cross/*/lib/libgcc_s.so.*': No such file or directory |
| 06:52.18 | bug1 | funny looking path.... |
| 06:52.37 | mithro | bug1: why? |
| 06:52.48 | mithro | o wait :) |