00:04.19 | TimRiker | tell prathiba I'll have to send the email after I get home. gotta run. |
00:05.16 | TimRiker | kergoth: basics: set umask in apps start up so files are all world writeable. chmod go+rw /device/backpack |
00:05.35 | TimRiker | set umask on apache so dav files are world writable. |
00:06.15 | TimRiker | check uboot mmc update to see that a package file without a user partition leaves a valid user partition in tact. |
00:06.52 | TimRiker | give CUPPDs 2 mmc cards. one with sample user partition, one with no user partition (for recovery etc). |
00:07.00 | TimRiker | back later. |
00:30.05 | AlanSmith | kergoth: any ETA on the newest build? |
00:37.39 | kergoth | AlanSmith: not sure.. its still going, but looks like all the long builds are done, shouldnt be more than a few minutes |
00:38.35 | AlanSmith | No need to rush. I am toast and will put the package together at 6:30 tomorrow morning when I get in. |
00:38.39 | *** part/#pet AlanSmith (~a0384855@proxyle01.ext.ti.com) |
00:59.57 | kergoth | whee, build's done, time to go home |
01:01.49 | xrharris | anyone know the size limit of environment variables? |
01:01.58 | xrharris | the value that is |
01:56.13 | *** join/#pet a-atwood (~xf009976@proxyle01.ext.ti.com) |
02:15.57 | *** join/#pet a-atwood (~xf009976@proxyle02.ext.ti.com) |
05:51.53 | kergoth | bleh. |
09:42.31 | *** join/#pet a-atwood (~xf009976@proxyle02.ext.ti.com) |
10:47.43 | a-atwood | whoohoo! the build worked |
12:36.52 | *** join/#pet bmw_3 (~xf013426@proxyle01.ext.ti.com) |
13:33.41 | *** join/#pet a-atwood (~xf009976@proxyle01.ext.ti.com) |
15:16.49 | *** join/#pet TimRiker (~timr@proxyle01.ext.ti.com) |
15:16.49 | *** mode/#pet [+o TimRiker] by ChanServ |
15:49.29 | TimRiker | who created /etc/network/interfaces ? |
15:49.59 | TimRiker | 192.168.1.x is the most likely network to have internal conflicts. It should be changed. I requested 192.168.0.x |
15:50.07 | TimRiker | why did it get changed? |
15:51.49 | *** join/#pet xrharris (~xrharris@proxyle01.ext.ti.com) |
15:52.34 | TimRiker | xrharris: I got eth0 to start. I had to restart udbcpc. don't know why. |
15:52.48 | TimRiker | dns lookups are failing now... again, don't know why. |
15:54.08 | xrharris | bryanw: I have a few more tweaks I want to make to the /etc/apache/httpd.conf and then I want to make sure it gets incorporated into the next OS build |
15:54.58 | xrharris | bryanw: also, we need to incorporate a "chmod u+s /bin/busybox" somewhere in an init or start script |
15:56.14 | glytch | k |
16:00.09 | TimRiker | hmm. ifdown eth0;ifup eth0;ifdown eth0;ifup eth0 then dns works. |
16:01.35 | xrharris | very interesting, any ideas why? |
16:03.05 | a-atwood | kergoth: what is the status of the ftp and telnet services in the image? |
16:03.15 | TimRiker | usr is a really really really bad name for /mnt/usr |
16:03.49 | TimRiker | user, internal, home, would all be better. I'm partial to internal |
16:04.53 | TimRiker | a-atwood: I have the 22 image. no telnetd, no ftpd, etc. |
16:05.07 | TimRiker | apache with webdav should work. |
16:05.16 | TimRiker | I'm trying nfs now... |
16:08.30 | TimRiker | works great. some irq issues with the usb d-link card, but minor. |
16:23.57 | Neo|Work | kergoth: is there a plan to fix (or is it fixed) /dev/rtc support in the os build? |
16:25.17 | kergoth | TimRiker: i set them to the addresses handed me by gerry/bill/etc |
16:25.26 | kergoth | TimRiker: *shrugs* |
16:26.10 | *** join/#pet xrharris (~xrharris@proxyle02.ext.ti.com) |
16:27.42 | Neo|Work | kergoth: also on an unrelated not, prelink worked in terms of not crashing but I still got the problem when prelinking qtopia (out of size something or other) |
16:28.27 | kergoth | k |
16:28.39 | kergoth | could be something as simple as us forgetting to -fPIC a lib |
16:28.52 | kergoth | and also, are you guys still inadvertantly linking in libstdc++, or did you fix that? |
16:30.34 | Neo|Work | actually it is being linked by mistake |
16:30.35 | kergoth | xrharris: done. (busybox suid) |
16:30.37 | Neo|Work | might be fixed now |
16:31.07 | Neo|Work | yes, it should be fixed |
16:31.19 | bryanw | xrharris: i'm in now |
16:31.46 | Neo|Work | as of 2004/01/22 01:49:57 (UTC I suppose) |
16:35.02 | Neo|Work | Must say the device is looking really nice (played some with it, using hte buttons and such) |
16:42.59 | *** join/#pet dez (~dez@proxyle01.ext.ti.com) |
16:45.25 | TimRiker | kergoth: I changed it on the wiki and in cvs. |
16:45.33 | kergoth | TimRiker: k |
16:46.27 | TimRiker | looks like the first dns server handed out by dhcp is down. |
16:46.27 | kergoth | TimRiker: what about /mnt/usr, theh apps team doesnt have any dependency on that location directly, correct? |
16:47.19 | TimRiker | kergoth: need to no have /var and /tmp in flash. |
16:47.45 | TimRiker | not yet, but the device to device is using some hardcoded paths that might include that. |
16:47.53 | kergoth | hmm, k |
16:48.06 | Neo|Work | kergoth: if I make a /etc/init.d/hwclock, can you add it? |
16:48.13 | Neo|Work | or do you have one already? |
16:48.46 | kergoth | Neo|Work: i have both a syslogd and hwclock init scsripts. both belong with busybox, and neither are included atm |
16:48.49 | kergoth | oversight on my part |
16:48.59 | kergoth | s/both a/both/ |
16:49.04 | Neo|Work | ok |
16:50.31 | TimRiker | color on virtual console is meesed up. probably a symtom of bad frame buffer code someplace. |
16:50.36 | Neo|Work | but does it modprobe omap-rtc (actually add omap-rtc to /etc/modules would work) and ln -s /dev/misc/omap-rtc /dev/rtc? :) |
16:51.47 | TimRiker | kergoth: where does /etc/passwd come from? |
16:52.07 | kergoth | TimRiker: the initial one came from debian's base-passwd. then modified |
16:52.14 | kergoth | TimRiker: need something changed there? |
16:52.48 | TimRiker | needs a "default" user added and we can likely remove gnats,list,irc |
16:52.59 | TimRiker | default should be 1000:1000 imho |
16:53.29 | TimRiker | so add a default group in /etc/group as 1000 too. |
16:53.44 | kergoth | k |
16:53.54 | TimRiker | want me to do it? |
16:54.07 | kergoth | yeah, go ahead |
16:54.34 | Neo|Work | yep, hwclock does work |
16:57.02 | Neo|Work | although on at least the devboard the hwclock loses time if you pull the power |
16:58.06 | TimRiker | kergoth: k. changed base-passwd-3.5.4/passwd.master base-passwd-3.5.4/group.master |
16:58.52 | TimRiker | oh, I added a eth1 and eth2 dhcp entry in interfaces. |
16:59.03 | TimRiker | not auto. for usb ether devices. |
16:59.32 | kergoth | didnt gerry change the naming to usb#? should confirm with him that eth# is correct |
17:03.07 | TimRiker | kergoth: need a /home/default where do I create that? |
17:03.39 | kergoth | TimRiker: probably base-files, just edit the .oe and add an install -d to its install task |
17:03.46 | TimRiker | k |
17:04.18 | TimRiker | usb0/usbd0 are set to 192.168.0.x and eth[0-2] are non-auto but dhcp |
17:04.46 | kergoth | cool |
17:06.27 | TimRiker | what knows about /mnt/"usr" ? fstab alone? I think backpack searches /mnt/*/backpack |
17:07.14 | kergoth | the apps start script (*cough*hack*cough*) also knows about it. hema requested that /mtn/usr/backpack replace /device/backpack, via symlink. i suggested bind mounts as a better implementation |
17:07.22 | kergoth | is that the way it should be done? |
17:07.33 | TimRiker | no. |
17:07.45 | TimRiker | /mnt/*/backpack is how it should be done. |
17:08.37 | kergoth | are the apps searching there currently? |
17:09.11 | TimRiker | where is the apps start script so I can change that too? |
17:09.25 | kergoth | PET/build/rootfs_conf is where mark put it iirc |
17:10.14 | TimRiker | how do I get /home/default to be 1000:1000 ? flags to install in the oe? |
17:12.03 | kergoth | TimRiker: probably easiest to just chown it in the do_install of base-files, so it ends up that way in the tarball. i run the whole schebang in fakeroot anyway |
17:12.14 | TimRiker | k. changed the script too. |
17:12.23 | kergoth | k |
17:12.41 | *** join/#pet richardw (~richardw@proxyle01.ext.ti.com) |
17:12.48 | TimRiker | install takes user and group options. the install -d get's run in the fakeroot too, yes? |
17:13.17 | TimRiker | the script change what the /mnt/internal issue... not the /home/default one yet. |
17:13.52 | TimRiker | richardw: your vc fix works, but not completely. |
17:14.03 | TimRiker | I can switch consoles but colors are broken now. |
17:14.07 | a-atwood | ~lart bmw_3 |
17:14.17 | TimRiker | also what's up with the spastic cursor? |
17:15.28 | TimRiker | kergoth: install -u 1000 -g 1000 -d ${D}/home/default ok? |
17:15.50 | kergoth | TimRiker: looks fine |
17:16.15 | kergoth | spastic cursor.. david hedbor reported touchscreen being jittery, i'm looking into it today |
17:16.30 | kergoth | probably just a matter of tweaking the config for the filters |
17:16.32 | TimRiker | not that one.... the console blinking cursor. |
17:16.36 | kergoth | ohh, that |
17:16.38 | Neo|Work | heh |
17:16.52 | kergoth | TimRiker: the odd horizontal lines flickering through it? i was wondering about that |
17:17.04 | TimRiker | yeah, that. |
17:17.15 | TimRiker | and the ls -l /dev not being in color |
17:17.19 | kergoth | TimRiker: hmm, just for the heck of it, try doing /etc/init.d/dpm stop |
17:19.33 | *** join/#pet dez (~dez@proxyle01.ext.ti.com) |
17:21.20 | kergoth | TimRiker: opinion on the rtc situation? quickest fix is ln -sf /dev/misc/omap-rtc /dev/rtc. better fix is probably name it /dev/misc/rtc, and patch hwclock to use that |
17:22.25 | TimRiker | kergoth: no change to the cursor with dpm off. |
17:22.38 | Neo|Work | but that's just me |
17:22.50 | Neo|Work | btw, there's a second "rtc" something module |
17:22.55 | Neo|Work | does the device have two RTC's? |
17:23.15 | Neo|Work | /lib/modules/2.4.20_mvlcee30-toto/kernel/drivers/char/pcf8563-rtc.o |
17:23.55 | Neo|Work | that one doesn't seem to add any device driver however |
17:24.15 | xrharris | TimRiker: i had to restart my device and I cannot change the MAC address with "ifconfig eth0 hw ether 08:00:28:32:16:99" |
17:24.28 | xrharris | TimRiker: it tells me the device is busy |
17:25.56 | TimRiker | kergoth: I don't care what you do for /device/backpack for pilot. I meant that that dir will not exist post pilot. bind mount is fine. hema just asked. |
17:26.16 | kergoth | TimRiker: k, wanted to make sure |
17:32.59 | bmw_3 | TimRiker: OT question -- If one is using Outlook under Crossover Office, does the baseline linux distribution have anything running that will check for virii? |
17:33.40 | richardw | TimRiker: i doubt that switching VC and having color screen ever worked. |
17:35.09 | richardw | There is a very specific sequence for shutting down the LCD in order to turn it back on in the same boot. If its not followed bad things happen. I sweated over this back in inferno. Others did the MV porting, that code is likely not quite right. |
17:35.44 | *** join/#pet hema (~cgiirc@proxyle02.ext.ti.com) |
17:39.20 | kergoth | TimRiker: hmm, my irc channel scrollback isnt long enough. what changes were necessary to umask, etc again? |
17:44.24 | Neo|Work | <PROTECTED> |
17:44.30 | Neo|Work | <TimRiker> set umask on apache so dav files are world writable. |
17:44.31 | Neo|Work | <TimRiker> check uboot mmc update to see that a package file without a user partition leaves a valid user partition in tact. |
17:44.31 | Neo|Work | <TimRiker> give CUPPDs 2 mmc cards. one with sample user partition, one with no user partition (for recovery etc). |
17:44.52 | Neo|Work | you mean that? |
17:44.55 | kergoth | yep |
17:44.56 | kergoth | thanks |
17:45.06 | Neo|Work | I have about forever backlog. |
17:45.11 | Neo|Work | actually I think I have it set to no limit |
17:51.43 | hema | TimRiker: Tim, you said in start apps script, do, chmod go+rw /device/backpack. Shouldn't this be chmod go+rw /mnt/internal/backpack, since that is what the apps are going to use? |
17:52.42 | hema | TimRiker: Also, I talked to Jian and tiinstallos correctly handles "not erasing" user partition, if we want to update just OS & Apps. I'll test this as soon as we have a new build with correct mounting of backpack. |
17:53.50 | TimRiker | kergoth: link in /dev/rtc is fine for now. I'll think on it more |
17:53.56 | kergoth | TimRiker: k |
17:54.27 | TimRiker | xrharris: you can't change the hw ether address if the interface is up. |
17:55.25 | TimRiker | bmw_3: no virus protection under cxoffice, but not a lot that would be possible. all the dll's etc are owned by root and the user does not have write access to them |
17:55.45 | TimRiker | richardw: color console used to work with jians first images. |
17:56.17 | bmw_3 | TimRiker: True, but I didn't want to archive "badd stuff" and then access it later under Windoze and get bit |
17:56.57 | TimRiker | kergoth: umask for apache to 000 and umask for apps to 000 and backpack/ to file 666, dirs 777 |
17:57.21 | richardw | ...I added a bounded wait for DMA done, maybe its not long enough...I was seeing lockups in the LCD off code. I'll up the wait time, if that doesn't do it then its working before was probably luck. |
17:57.46 | Neo|Work | also shouldn't the chmod be recursive? |
17:57.48 | TimRiker | chmod go+rX on backpack should work to get x only where needed. |
17:58.07 | kergoth | k. umasks are in, will add the chmod now |
17:58.09 | TimRiker | yes. chmod -R go+rX /mnt/internal/backpack |
17:58.29 | kergoth | 666/777 would be go=rwX not go+rX. minor nit. ;) |
17:58.29 | kergoth | k |
17:58.30 | Neo|Work | gi+rwX |
17:58.33 | Neo|Work | ? |
17:58.36 | Neo|Work | go+rwX even |
17:58.56 | TimRiker | ah.. yeah go+rwX |
17:59.34 | kergoth | k, done |
17:59.37 | TimRiker | not sure = works with X. I know the + will work. = should too but you might test first. ;-) |
17:59.56 | Neo|Work | technically speaking, chmod g+rwX and chmod g+s for all directories probably would work just as well, if the group was set to the group apache uses |
18:00.12 | kergoth | = does work with X, but if i use = itll remove the +x on non-dirs if any exist |
18:00.16 | Neo|Work | Never seen /'=' used |
18:00.22 | TimRiker | bmw_3: your windows checking should catch it no? |
18:00.32 | Neo|Work | ah |
18:00.34 | Neo|Work | doh |
18:00.43 | Neo|Work | makes sense now |
18:01.06 | TimRiker | kergoth: sounds safe to me. |
18:01.45 | kergoth | TimRiker: = does, or + does? :) |
18:01.49 | TimRiker | hmm... might not be safe in the future with backpack/.apps/ or whatever is coming. |
18:02.04 | kergoth | yeah, i'll leave it + for now |
18:02.09 | Neo|Work | man, for having used Linux/UNIX since '93 it's funny how I never used 'fdisk -l' or chmod x=y |
18:02.11 | TimRiker | kergoth: agreed. |
18:02.22 | TimRiker | or cat /proc/partitions ? |
18:02.30 | kergoth | hehe |
18:03.07 | Neo|Work | ah |
18:03.17 | Neo|Work | X actually doesn't just set 'x' for directories |
18:03.43 | Neo|Work | it also does it if "some user" already has execute permissions |
18:03.46 | kergoth | oh, or already has execute permission for some user |
18:03.47 | kergoth | cool |
18:03.57 | kergoth | so =rwX would be safe after all |
18:04.22 | Neo|Work | : 0 neotron@cn014xf015460l1 ls -l dfqew ~/ |
18:04.22 | Neo|Work | <PROTECTED> |
18:04.26 | Neo|Work | : 0 neotron@cn014xf015460l1 chmod g=rX dfqew ~/ |
18:04.26 | Neo|Work | : 0 neotron@cn014xf015460l1 ls -l dfqew ~/ |
18:04.26 | Neo|Work | <PROTECTED> |
18:04.47 | Neo|Work | so yes, seems like it |
18:04.49 | kergoth | heh, interesting |
18:05.19 | richardw | TimRiker: so I made the easy change, next kernel source pull may work. if not some more effort is needed to get the screen clean. |
18:05.20 | Neo|Work | I never knew that either |
18:05.20 | kergoth | TimRiker: you planning any more changes in cvs, or should i start a build? |
18:05.41 | TimRiker | kergoth: you might want to chown -R default:default /mnt/internal/backpack while you're at it. ;-) |
18:05.58 | kergoth | TimRiker: heh, preparing for the non-root are we? |
18:06.56 | TimRiker | richardw: nice. thanx. |
18:07.08 | TimRiker | kergoth: at some point, yes. ;-) |
18:07.13 | kergoth | :) |
18:07.17 | kergoth | 12:05 < treke> greg's first law of corruption. Everyone is corrupt. |
18:08.09 | TimRiker | kergoth: do we start anything on vc/2 ? |
18:08.22 | TimRiker | 2 is probably a better choice than 5 for vc. |
18:08.36 | TimRiker | can I change that in cvs? |
18:08.41 | kergoth | probably. and it looks like i have a getty spawning on vc/0, that should be vc/1. |
18:08.52 | kergoth | err no, its vc/1 |
18:08.54 | TimRiker | correct. |
18:08.55 | kergoth | remembered incorrectly |
18:08.57 | kergoth | :) |
18:09.01 | TimRiker | vc/1 is correct. |
18:09.17 | kergoth | k, go ahead and change to vc/2, i'm about to start a build |
18:09.53 | kergoth | out of curiosity, if the apps die atm, the unit is probably useless eh? we start the apps as a once in inittab, not respawn, and itll never return to vc/1 from vc/blah |
18:10.27 | TimRiker | kergoth: changed. |
18:10.43 | TimRiker | yes, change that to a respawn |
18:10.58 | kergoth | k, you want to change the apps script to remove the & on the call to start? ;) |
18:11.06 | kergoth | otherwise we'll get a nice fun respawning too rapidly |
18:11.35 | TimRiker | I changed it. removed & added exec |
18:11.40 | kergoth | k |
18:12.09 | TimRiker | wait... |
18:12.12 | kergoth | k |
18:12.25 | TimRiker | should ln -sf /dev/vc/0 /dev/tty0 really be ln -sf /dev/vc/5 /dev/tty0 ? |
18:12.33 | TimRiker | er should ln -sf /dev/vc/0 /dev/tty0 really be ln -sf /dev/vc/2 /dev/tty0 ? |
18:12.39 | Neo|Work | is that ln really needed at all? |
18:12.45 | kergoth | tty0 is the master. the foreground |
18:12.53 | TimRiker | don't know. I'll check... |
18:12.54 | Neo|Work | QWS_KEYBOARD should take care of that or is it for some othe reason? |
18:12.56 | kergoth | so no, vc/0 -> tty0 is correct |
18:12.59 | kergoth | but dunno if its needed |
18:13.47 | Neo|Work | If it's only for Qtopia, then it shouldn't be needed as long as QWS_KEYBOARD is set |
18:14.05 | kergoth | well, it may be opening it for the vt handling in qte |
18:14.08 | kergoth | not just for keyboard |
18:14.17 | kergoth | are you guys buliding a patched qte for devfs yet? |
18:14.22 | kergoth | bleh, building |
18:15.02 | Neo|Work | yes |
18:15.09 | kergoth | bmw_3: did you have the guys linking devdisc add -lnet to their link lines? |
18:15.10 | Neo|Work | using your patch which is ... not perfect. :) |
18:15.20 | Neo|Work | it had hardcoded keyboard to /dev/vc/1 |
18:15.21 | kergoth | hehe, yeah, it was quick and dirty, but did the job |
18:15.22 | TimRiker | seems to work without the link on my pet |
18:15.28 | kergoth | cool |
18:15.48 | TimRiker | keyboard works, touchscreen works. trying virtual keyboard... |
18:16.03 | TimRiker | yep. |
18:16.15 | TimRiker | and buttons. looks safe to remove. |
18:16.29 | Neo|Work | changing vt works as expected too? |
18:16.47 | kergoth | yeah, try switching to 1 and then back to 5/2 |
18:18.21 | TimRiker | eep |
18:18.41 | TimRiker | blinking cursor is there without the link... |
18:19.28 | kergoth | confirming its happy with the link? |
18:20.31 | TimRiker | yep.. adding back the link removes the blinking cursor. leave it in for now. |
18:20.38 | kergoth | k |
18:20.44 | Neo|Work | funky |
18:21.00 | kergoth | probably a missing devfs change in teh vt handling |
18:21.07 | Neo|Work | Looking for any /dev/tty references in qtopia.. |
18:21.18 | bryanw | xrharris: hows the httpd.conf ? |
18:21.21 | TimRiker | will be tty0 explicitly |
18:22.35 | Neo|Work | haha |
18:22.44 | Neo|Work | <PROTECTED> |
18:22.44 | Neo|Work | <PROTECTED> |
18:22.44 | Neo|Work | <PROTECTED> |
18:22.52 | kergoth | hehe |
18:22.52 | Neo|Work | <PROTECTED> |
18:22.52 | Neo|Work | <PROTECTED> |
18:22.52 | Neo|Work | <PROTECTED> |
18:22.52 | Neo|Work | int tty=::open(*dev,O_WRONLY); |
18:22.52 | Neo|Work | if ( tty>=0 ) { |
18:22.53 | Neo|Work | <PROTECTED> |
18:22.57 | Neo|Work | <PROTECTED> |
18:22.59 | Neo|Work | } |
18:23.01 | Neo|Work | dev++; |
18:23.03 | Neo|Work | <PROTECTED> |
18:23.10 | kergoth | errr, why's it looking at vc/1 instead of vc/0? |
18:23.11 | Neo|Work | the second code is writeTerm |
18:23.14 | kergoth | thatd busted |
18:23.20 | Neo|Work | because that's what you sent... |
18:23.28 | kergoth | heh, oops |
18:23.37 | Neo|Work | all /dev/vc stuff used /dev/vc/1 :) |
18:23.41 | Neo|Work | well, the two places |
18:23.48 | kergoth | s/vc\/1/vc\/0/g |
18:24.30 | Neo|Work | that would be why it turns of the cursor all over the place too |
18:24.36 | kergoth | Neo|Work: or, just kill the /dev/vc/1.. we do have a /dev/tty iirc. |
18:24.42 | Neo|Work | and notjust in the term it should kill it in |
18:25.00 | kergoth | yeah, its killing it in the master, not in the current vt |
18:25.02 | kergoth | wonder why that is |
18:25.16 | Neo|Work | it should probably use /proc/self/fd/0 |
18:25.22 | kergoth | yep |
18:25.23 | Neo|Work | both ther eand for default "keyboard tty" |
18:25.49 | kergoth | silly trolls. |
18:26.53 | kergoth | hmm, we have the hid drivers in monolithic, and the mousedev input layer driver as a module.. should work. unless qte doesnt support /dev/input/mouse or somethingm |
18:26.56 | Neo|Work | in the apps? did you export QWS_MOUSE_PROTO=USB and modprobe mousedev? |
18:27.13 | Neo|Work | you have to specifically tell it to use USB mouse |
18:27.51 | Neo|Work | there would have to be support specifically added to support plugging a mouse in and enabling it |
18:27.58 | Neo|Work | and enable the mouse cursor |
18:28.05 | TimRiker | QWS_MOUSE_PROTO=USB gets usb mouse as well as ts then? |
18:28.07 | Neo|Work | likewise disable the mouse cursor when you unplug it |
18:28.10 | Neo|Work | No |
18:28.18 | Neo|Work | there's no way of having both without adding code for it |
18:28.20 | Neo|Work | whcih is not done |
18:28.29 | bryanw | kergoth: i heard you were starting a new build. would you hold on and let me go ask xrharris if he is finished with this httpd.conf changes |
18:28.30 | TimRiker | ah. bummer. reall bummer. |
18:28.35 | Neo|Work | Qt supports multiple devices but there's no "non-code" way of adding more |
18:28.43 | kergoth | bryanw: k |
18:28.49 | Neo|Work | ie. it's an app-level coding task, not qt |
18:28.52 | TimRiker | I thought you had both active? |
18:28.58 | Neo|Work | nope |
18:29.00 | Neo|Work | never have |
18:29.32 | Neo|Work | the api's are odd |
18:29.57 | Neo|Work | I can add an input device, which puts it on the first of a list |
18:30.07 | Neo|Work | there si no way to "pop" i toff the list however |
18:30.27 | Neo|Work | I believe you have to clear the list completely and readd the touchscreen on an "disconnect mouse" event |
18:31.10 | Neo|Work | TimRiker: check the iteration plan to see if this is in there. if not, get it in if you want it :) |
18:34.43 | kergoth | bryanw: i'll start the build, we can at least confirm the permissions changes and stuff are happy |
18:34.48 | kergoth | bryanw: then we'll just do another later |
18:34.50 | bryanw | k |
18:43.41 | mindfunk | kergoth: john just checked in his keypad fix |
18:43.49 | kergoth | yup, he told me |
18:43.52 | kergoth | itll be in this build |
18:43.56 | mindfunk | ya building? |
18:43.58 | kergoth | (hasnt gotten to the kernel yet) |
18:44.00 | kergoth | yeah |
18:44.08 | mindfunk | thansk |
18:44.10 | mindfunk | thanks |
18:45.12 | kergoth | udev 014 out. man its improving fast. and greg k-h got his patches into 2.6 to add things like the input layer, sound, misc devices, etc to the driver model :) |
18:45.16 | kergoth | most spiffy. |
18:45.22 | kergoth | not that it benefits us, obviously, but spiffy nonetheless |
18:45.47 | kergoth | jbot: thursday |
18:45.51 | | Thursday sucks, because it isnt friday (See friday). |
18:55.30 | kergoth | why, oh why, do people insist on mailing toolchain,etc issues to l-a-k instead of l-a |
18:56.52 | TimRiker | what keypad fix is that? |
18:57.14 | TimRiker | kergoth: cause they have not been flamed by rmk lately. |
18:57.37 | kergoth | just talked to him. he says suspend/resume still isnt reliable on handhelds. the fix he committed was a change to the interrupt handling |
18:57.47 | TimRiker | ahh. |
18:58.17 | kergoth | suggested adding debug info to DPM to see the suspend/resume call order |
18:58.49 | kergoth | food is good |
18:59.32 | TimRiker | kergoth`lunch: is it there? "ctrl-a f 7" should show the msgs again. |
19:08.26 | bryanw | ok.. |
19:08.37 | bryanw | qcop commands aren't working when run through hotplug |
19:08.46 | bryanw | you can run the same script by h and no problem |
19:09.04 | bryanw | printing out umask,whoami,export,etc... shows that the environment when running by hand is the same as run through hotplug |
19:09.09 | bryanw | any suggestions? |
19:10.27 | *** join/#pet frome (~cgiirc@proxyle01.ext.ti.com) |
19:13.56 | *** join/#pet frome (~cgiirc@proxyle01.ext.ti.com) |
19:14.29 | *** join/#pet frome (~cgiirc@proxyle01.ext.ti.com) |
19:14.42 | *** join/#pet frome (~cgiirc@proxyle02.ext.ti.com) |
19:14.45 | *** join/#pet frome (~cgiirc@proxyle02.ext.ti.com) |
19:36.01 | bmw_3 | Is it too late to switch to gtk+ instead of qt/qtopia? |
19:36.57 | kergoth | heh |
19:37.24 | kergoth | bryanw: "aren't working" means what? error when calling hotplug? add 2>/tmp/blah to the qcop line |
19:37.28 | kergoth | er |
19:37.31 | kergoth | error when calling qcop i mean |
19:40.57 | bryanw | kergoth: qcop returns 1 with no error message |
19:41.13 | bryanw | <PROTECTED> |
19:41.16 | kergoth | heh, how helpful |
19:41.37 | kergoth | bmw_3: did you have them try -lnet on the -ldevdisc link line for the apps that use it? should take care of that error |
19:41.57 | bryanw | ah |
19:42.02 | bryanw | got the error (: |
19:42.09 | bryanw | QSocket::writeBlock: Socket is not open |
19:42.09 | bryanw | No Qt/Embedded server appears to be running. |
19:42.09 | bryanw | If you want to run this program as a server, |
19:42.09 | bryanw | add the "-qws" command-line option. |
19:42.21 | bryanw | works fine when i run it by hand though.. |
19:42.31 | bryanw | anyway, i've got to go talk to people about flash drives |
19:42.31 | bryanw | bbl |
19:42.35 | kergoth | heh |
19:55.01 | Neo|Work | bryanw: it needs to run as the same user as qtopia, which is root. it ALSO needs LOGNAME to be root |
19:55.04 | Neo|Work | could that be the issue? |
19:55.23 | Neo|Work | well, LOGNAME to be identical to the LOGNAME used by qtopia, which is root |
20:04.57 | bmw_3 | kergoth: I talked w/ Joel earlier and he said they have it in there. I have asked him to send me his commands and any failure messages + questions to forward to you. |
20:05.14 | bmw_3 | kergoth: So far, they seem to be too busy to send them to me... |
20:22.57 | kergoth | k |
20:33.00 | mindfunk | kergoth: just a friendly reminder: please tag the build with PILOT1_CANDIDATE_date_10000 |
20:33.51 | kergoth | k |
20:33.54 | kergoth | thanks |
20:37.08 | *** join/#pet dez (~dez@proxyle02.ext.ti.com) |
20:37.53 | kergoth | i should really write a script for building outside of nightly |
20:37.57 | kergoth | been doing it manually |
20:38.08 | mindfunk | =) |
20:38.32 | kergoth | bmw_3: k, keep me posted. i'd like to get the sdk happy so i can move on to other stuff |
20:38.33 | mindfunk | I assume another build is coming? |
20:38.43 | kergoth | already up |
20:38.49 | mindfunk | koo |
20:38.50 | kergoth | same as prevoius, but with /mnt/internal actually existing this time |
20:41.01 | bmw_3 | kergoth: What about the problems I sent you the other day when I tried the toolchain/sdk w/ qte/qtopia build? |
20:41.26 | kergoth | .. |
20:41.31 | kergoth | thats what i'm talking about |
20:41.38 | kergoth | adding -lnet will fix the link problem with the libnet symbols |
20:41.50 | kergoth | the tslib ones, not sure whats going on there, need to look into it further |
20:42.32 | bmw_3 | kergoth: I thought you were talking about the device discovery link problems the apps folk were having. Sorry, I'll make a nore to work w/ David tomorrow on this one. |
20:42.46 | kergoth | they're one and the same |
20:42.48 | kergoth | :) |
20:42.52 | kergoth | in part anyway |
20:43.10 | kergoth | bmw_3: can you reproduce the build failure on your machine, then come get me so i can hack on it first hand? |
20:43.20 | kergoth | bmw_3: its probably a quick fix, but doing it indirectly is slowing us down |
20:47.27 | TimRiker | bryanw: fixed yet? got HOME LOGNAME QTDIR QPEDIR etc? |
20:49.13 | TimRiker | kergoth: oops. yeah, creating the directory would be good. hmm. actually I thought I did change that. did you have to change something I missed? |
20:49.34 | kergoth | just an install -d in base-files |
20:49.48 | kergoth | was still /mnt/usr in my cvs up a few minutes ago |
20:51.14 | Neo|Work | so I tried to connect a usb keyboard to my pet on battery |
20:51.16 | Neo|Work | it hung |
20:51.29 | Neo|Work | I am guessing it was too low on battery but still not very graceful. :) |
21:07.51 | a-atwood | mindfunk: I have an email in to Lorrie to get it reset, but I need you to put every related for the fresh build somewhere that I can access it. I'm locked out of petnet web access again |
21:19.54 | mindfunk | a-atwood: gimme a sec |
21:26.09 | bmw_3 | kergoth: I can reproduce it on my system fairly quickly, but can we do it tomorrow? |
21:26.16 | TimRiker | kergoth: cool. sorry I missed that. I thought I'd changed it. :-/ |
21:27.02 | TimRiker | a-atwood: I think I can add a local password for you on petnet. |
21:27.46 | kergoth | np |
21:29.54 | mindfunk | kergoth: please tag that last build as like PILOT1_....10002 |
21:30.07 | mindfunk | so far... that looks good |
21:32.01 | kergoth | mindfunk: k, im on it |
21:32.03 | kergoth | good good |
21:51.44 | bmw_3 | kergoth: ping |
21:51.51 | kergoth | bmw_3: pong |
21:52.16 | bmw_3 | kergoth: Can we get together in the morning re: toolchain/sdk? |
21:53.01 | kergoth | bmw_3: thats fine |
21:53.16 | bmw_3 | kergoth: Roughly when? |
21:53.40 | kergoth | lately i've been working really late, so havent been getting in till like 10. i can be in around 9 if you'd prefer |
21:54.00 | bmw_3 | 9 or 10 are both OK w/ me |
21:54.13 | kergoth | k |
21:54.21 | kergoth | ill drop by when i get in |
21:54.40 | bmw_3 | I have been getting here around 6:30 +-, so I will be here when you come by |
21:55.08 | bmw_3 | You want me to have the "broke" build at the broken point, or do you want to be here when I kick it off? |
21:55.32 | kergoth | go ahead and have it at the broken point |
21:55.34 | kergoth | will go from there |
21:55.44 | bmw_3 | agreed |
21:55.55 | bmw_3 | nite-nite 8) |
21:55.59 | kergoth | for i in PET/src/os PET/build/oe PET/build/metadata PET/build/rootfs_conf; do cvs rtag PILOT1_CANDIDATE_`date +%Y%m%d`_10002 $i; done |
21:56.03 | kergoth | whee. |
21:56.03 | kergoth | night |
21:58.48 | TimRiker | when's the next os build? |
21:59.04 | a-atwood | by me, tonight at 21:00 |
21:59.07 | TimRiker | kergoth: did you put one up that just finished? does it have apps in it? |
21:59.23 | kergoth | TimRiker: i dont do the app merge, teh os test folks do. they're confirming the permissions fixes as we speak |
21:59.41 | kergoth | i put one up not long ago though, yes |
22:01.23 | TimRiker | who does the merge, and with what apps image will they merge? |
22:01.42 | kergoth | ask mindfunk |
22:08.05 | TimRiker | mindfunk: who does the merge? and with what? and how the heck can you qualify an os image with no apps? ;-) |
22:09.05 | TimRiker | kergoth: did you boot the new build? did richard's changes get in? does "ls /dev" show more than 2 colors? |
22:09.18 | kergoth | i havent booted it, a bunch of the test folks have though |
22:11.19 | mindfunk | very easily ;) |
22:11.43 | mindfunk | So... I do the os qual... then merge in the apps, then do some sanity checking |
22:11.52 | mindfunk | the apps guys test the apps |
22:12.20 | mindfunk | but... since pilots are so hecktic, we've been merging before os testing |
22:12.37 | mindfunk | s/hecktic/hectic |
22:13.13 | mindfunk | the script that merges the two is mktotopkg |
22:13.23 | mindfunk | in PET/src/os/tools |
22:13.44 | mindfunk | TimRiker:poke |
22:21.55 | bryanw | here |
22:21.57 | bryanw | TimRiker: |
22:21.59 | bryanw | TimRiker: pong |
22:23.43 | bryanw | i was off getting 188 usb flash drives |
22:23.44 | bryanw | (: |
22:28.13 | bryanw | TimRiker: works now |
22:28.20 | bryanw | kergoth: sending you new uflash script |
22:28.25 | bryanw | didn't have HOME and LOGNAME set |
22:28.35 | kergoth | k |
22:35.07 | kergoth | bryanw: k |
22:35.38 | xrharris | bryanw: I just sent you an email with the httpd.conf we need for PET2PET |
22:35.45 | bryanw | k |
22:37.09 | a-atwood | mindfunk: <pester>?</pester> |
22:37.24 | kergoth | bryanw: put the httpd.conf and the uflash in your users area on petdoc and let me know when they're there |
22:37.31 | kergoth | after sanity checking the .conf of course |
22:38.02 | bryanw | will do |
22:38.04 | TimRiker | bryanw: how's the speed? do we get binders now? |
22:38.20 | bryanw | TimRiker: the drives we picked out did that 45 binder speed test in 6.5 seconds |
22:38.29 | bryanw | 64mb for $30 |
22:38.31 | kergoth | TimRiker: woo, holger schurig is OE hacking again, now i'm not the only one improving the core :) |
22:38.42 | bryanw | they fit in the happy hacker keyboards and come with extension cables |
22:38.56 | TimRiker | mindfunk, kergoth: the apps merge should be part of the build. but you guessed that. ;-) |
22:39.10 | bryanw | TimRiker: i'll bring one down and let you see if you want |
22:39.57 | bryanw | kergoth: it'll be a tarball that extracts to $PWD/hotplug and that hotplug dir needs to be whats in /etc/hotplug . i changed a few files so i'm just sending the whole lot |
22:40.05 | bryanw | and i'll test the httpd.conf now |
22:40.06 | kergoth | bryanw: k |
22:40.44 | bryanw | TimRiker: and i won't be doing bzflag for gba. i got the stuff i needed for gamecube development instead (: |
22:41.17 | kergoth | hehe |
22:41.44 | bryanw | kergoth: users/me/hotplug.tar |
22:46.59 | bryanw | TimRiker: what should the server admin email address be ? |
22:50.54 | kergoth | bryanw: uploaded new image with the hotplug updates. |
22:52.18 | bryanw | kergoth: eek need to get this httpd.conf change from xrharris in as well |
22:52.24 | bryanw | im editing/testing it now |
22:52.30 | kergoth | yeah, ill do another when you give me that |
22:52.34 | bryanw | k |
22:57.42 | kergoth | err, crap, i didnt add in the omap-rtc link and autoload of omap-rtc. should do that for this upcoming build |
22:57.51 | kergoth | too much stuff to keep track of |
22:57.51 | kergoth | heh |
22:58.33 | kergoth | TimRiker: you know of a decent linux outliner and/or decent task tracking software? |
22:59.06 | bryanw | kergoth: petdoc/users/me/hotplug.tar petdoc/users/me/httpd.conf |
22:59.11 | bryanw | httpd.conf in /etc/apache/ |
22:59.12 | bryanw | (: |
22:59.19 | kergoth | has hotplug changed since the last one/ |
22:59.25 | kergoth | 15 minutes ago |
23:01.52 | a-atwood | mindfunk: ok, what is the label to this so I can pull the os? |
23:02.28 | bryanw | not since the hotplug.tar no |
23:06.04 | TimRiker | kergoth: never really looked. entering tasks in bugzilla with dependancies has worked, but it leaves a lot to be desired. gforge has some of that integrated now. sill lacking. |
23:06.41 | TimRiker | there are gui project management apps for linux, but none with windows versions as well that I've found so I've not considered them. |
23:07.16 | TimRiker | the outline mode on openoffice is usable. if that's what you are after. |
23:07.30 | TimRiker | pasting into the impress (slide shows) works well too. |
23:10.41 | a-atwood | kergoth: what is the label you put on the build that just went to release? Need to pull ther kernel sources for Wipor |
23:10.42 | TimRiker | bryanw: bring the cube in so I can see what the kit looks like. |
23:10.50 | a-atwood | kergoth: er Wipro |
23:11.07 | TimRiker | I'm looking for someone with a action replay kit for the xbox so I can hack mine without opening the case. |
23:11.30 | bryanw | TimRiker: it consists of cube controller>keyboard/mouse connector , cube network adapter, and phantasy star online game(has a network exploit for you to run your own code) |
23:11.52 | bryanw | going to pickup the last copy of pso for 15 miles after work today |
23:11.52 | TimRiker | oh cool. so got it booting linux yet? |
23:11.53 | bryanw | (: |
23:12.02 | bryanw | TimRiker: no one has it booting liniux yet |
23:12.05 | bryanw | but there are many dev kits |
23:12.11 | bryanw | you can write you own code for it |
23:12.14 | bryanw | linux is in the works |
23:12.17 | TimRiker | xbox needs a savegame and 007. that's why I need the action replay to load the save game. |
23:13.31 | kergoth | Neo|Work: ping |
23:13.41 | kergoth | a-atwood: PILOT1_CANDIDATE_10002, iirc |
23:13.42 | kergoth | er |
23:13.53 | kergoth | a-atwood: something like that, bu tits datestamped between dandiate and 10002 |
23:13.56 | kergoth | s/dand/cand/ |
23:14.01 | kergoth | gah, cant type |
23:14.05 | Neo|Work | kergoth: pong |
23:14.19 | TimRiker | http://gc-linux.sourceforge.net/pic/linuxpreview2.jpg |
23:14.38 | kergoth | Neo|Work: compared app behavior to ts_test behavior. apps are getting jumpy events. ts_test is solid as a rock. its almost like qte is doing ts_read_raw, not ts_read calls. |
23:15.15 | Neo|Work | kergoth: if it would be reading raw, the events would be way out of bounds |
23:15.28 | Neo|Work | which to the user looks like the touchscreen isn't working |
23:15.39 | kergoth | well i dunno. outside the apps its flawless |
23:15.41 | kergoth | its strange |
23:15.55 | Neo|Work | and you wrote the code :) |
23:16.06 | kergoth | heh, yeah, i know |
23:16.19 | kergoth | Neo|Work: where are hte qte sources ben weatherall is using? |
23:16.27 | Neo|Work | petdoc |
23:16.29 | kergoth | i need a dev environment so i can debug |
23:16.34 | Neo|Work | one sec |
23:16.37 | kergoth | is he applying any patches to it before the build? |
23:16.37 | kergoth | k |
23:17.08 | Neo|Work | qtopia-trolltech-2.3.7-p3-arm-* |
23:17.12 | Neo|Work | that is patched versions |
23:17.14 | kergoth | thats sources? |
23:17.14 | kergoth | k |
23:17.18 | Neo|Work | well, arm-debug and arm-release |
23:17.22 | Neo|Work | the "-sdk" ones have no source |
23:22.58 | kergoth | Neo|Work: is there a paint app or something that i can start -qws just for testing ts? |
23:23.36 | Neo|Work | hmm |
23:23.38 | mindfunk | kergoth: there is a rumor going around that there is another build... |
23:23.49 | Neo|Work | only image editor but I never had it running on the device... |
23:24.11 | Neo|Work | but you can always run tirun -qws tisetup and use the scrollbars |
23:24.19 | kergoth | mindfunk: i uploaded a new one with the new hotplug bits, but i havent built one with the new apache config yet. working on it. |
23:24.34 | kergoth | Neo|Work: theres something wrong with teh scrollbars above and beyond the ts |
23:24.36 | mindfunk | ok... pls let me know when it is there |
23:24.47 | kergoth | Neo|Work: the position on teh srollbar isnt anywhere near the position of the cursor qte draws |
23:24.59 | kergoth | cursor is correct, bar position is not |
23:25.06 | kergoth | dunno whats up there |
23:26.14 | Neo|Work | well |
23:26.29 | Neo|Work | you don't see the cursor moving as much it seems |
23:26.39 | Neo|Work | the refresh rate of the screen is kind of slow |
23:26.55 | Neo|Work | so it becomes kind of a blur. :P |
23:26.55 | kergoth | i mean, i saw pat release the ts |
23:27.01 | kergoth | left it for seconds |
23:27.11 | kergoth | cursor stayed inteh right place |
23:27.15 | kergoth | scrollbar was still off |
23:27.59 | kergoth | so doesnt seem to be just refresh |
23:28.25 | Neo|Work | it could be what I just experienced perhaps |
23:28.34 | Neo|Work | press, move scrollbar around, release |
23:28.39 | Neo|Work | it jumps back to original position |
23:28.41 | Neo|Work | is that it? |
23:29.03 | kergoth | no, what i saw was that the cursor was right, or clsoe to it, but the position of hte scrollbar was like 30 pixels off of the cursor position |
23:29.18 | kergoth | it varied, but the point is that the cursor didnt match the scrollbar |
23:29.33 | Neo|Work | I have seen the cursor jump like 30 pixels |
23:29.35 | Neo|Work | if not more |
23:29.38 | Neo|Work | and then move back |
23:29.41 | kergoth | yes, the cursor can screw up |
23:29.44 | kergoth | thats not the point |
23:29.44 | Neo|Work | but it's almost impossible to see it doing it |
23:30.15 | kergoth | i'm not saying there isnt a touchscreen problem, i'm saying theres more going on than just that. |
23:30.16 | Neo|Work | it's possible that it got a release event, press event, release event |
23:30.19 | Neo|Work | I've seen that too |
23:30.29 | Neo|Work | I honstly don't think there is |
23:30.42 | Neo|Work | if the events were printed out it probably would become clearer |
23:30.42 | kergoth | thers nothing wrong with gettin ga release, press, release |
23:30.52 | kergoth | yeah, indeed. adding debug messages now |
23:30.55 | Neo|Work | if you pressed, draged, released, sure :) |
23:31.09 | Neo|Work | we can argue or discuss it but it's pointless because we can't see what's going on heh |
23:31.18 | Neo|Work | btw, the best place to debug is readMouseData |
23:31.33 | Neo|Work | all other locations agree with that data (I traced it around when I got seriously odd behavior) |
23:31.49 | Neo|Work | or I should say, nothing happens between readMouseData and the end mouse event that changes it |
23:32.01 | kergoth | k |
23:32.17 | Neo|Work | well, I should say, thave I have seen / noticed |
23:32.29 | Neo|Work | I think all processing of ts events is actually in that file |
23:36.25 | *** join/#pet richardw (~richardw@proxyle02.ext.ti.com) |
23:39.06 | mindfunk | kergoth: build ready? |
23:51.22 | TimRiker | kergoth: busybox needs to be u+s for the apache scripts to use su |
23:51.25 | kergoth | TimRiker: it is |
23:51.29 | TimRiker | where should I do that? |
23:51.35 | TimRiker | oh. since when? |
23:51.35 | kergoth | i did it earlier |
23:51.41 | TimRiker | today? |
23:51.41 | kergoth | this morning |
23:51.43 | kergoth | yeah |
23:59.04 | bryanw | kergoth: *poke* , build? |
23:59.19 | kergoth | uploaded |
23:59.22 | kergoth | this ts thing is weird |
23:59.31 | kergoth | the cursor appears right over the fin ish button, yet the button never gets pressed |
23:59.51 | kergoth | oh, dont mind me, i'm a moron |
23:59.55 | kergoth | button was greyed out ;) |