IRC log of #maemo-devel for Saturday, 2010-05-29

*** fiferboy has quit IRC00:02
*** thunderfest is now known as thunderfest_away00:30
*** SpeedEvil has quit IRC00:33
*** t-tan has quit IRC00:36
*** JLP has joined #maemo-devel00:43
*** JLP has quit IRC00:43
*** JLP has joined #maemo-devel00:43
*** DocScrutinizer51 has quit IRC00:44
*** Docscrutemp has joined #maemo-devel00:44
*** Docscrutemp is now known as DocScrutinizer5100:44
*** SpeedEvil has joined #maemo-devel00:45
*** JLP has quit IRC00:50
*** DocScrutinizer51 has quit IRC00:55
*** Docscrutemp has joined #maemo-devel00:55
*** Docscrutemp is now known as DocScrutinizer5100:55
*** JLP has joined #maemo-devel01:03
*** JLP has quit IRC01:03
*** JLP has joined #maemo-devel01:03
*** rcampbell has quit IRC01:29
*** T7g has joined #maemo-devel02:04
*** tealbird has joined #maemo-devel02:07
*** thunderfest_away has quit IRC02:53
*** Mobsan has quit IRC03:17
*** edisson has quit IRC03:18
*** njsf_ has joined #maemo-devel04:06
*** SpeedEvil has quit IRC04:07
*** SpeedEvil has joined #maemo-devel04:07
*** tealbird has quit IRC04:47
*** JLP has quit IRC05:26
*** JLP has joined #maemo-devel05:28
*** JLP has quit IRC05:29
*** JLP has joined #maemo-devel05:29
*** tealbird has joined #maemo-devel05:38
*** njsf_ has quit IRC05:46
*** tealbird has quit IRC06:25
*** tealbird has joined #maemo-devel07:18
*** DocScrutinizer51 has quit IRC07:33
*** DocScrutinizer has quit IRC07:33
*** DocScrutinizer has joined #maemo-devel07:34
*** Docscrutemp has joined #maemo-devel07:34
*** Docscrutemp is now known as DocScrutinizer5107:34
*** swc|666 is now known as swc|hatesSHY07:56
*** swc|hatesSHY is now known as swc|66607:56
*** BT__ has joined #maemo-devel08:47
BT__hi , I am trying to use qemu , looks like n900 is not supported, qemu-system-arm -M n900 -mtdblock meego_qemu_nand.img08:48
BT__any idea?08:49
*** tealbird has quit IRC09:27
*** swc|666 has quit IRC09:35
*** BT__ has quit IRC10:27
*** rsalveti has quit IRC10:28
*** 50UAABWM5 has joined #maemo-devel10:28
*** w00t_ has quit IRC10:32
*** tealbird has joined #maemo-devel10:47
*** Mobsan has joined #maemo-devel10:51
*** tealbird has quit IRC11:01
*** Acedip has joined #maemo-devel11:02
*** polac has quit IRC11:33
*** 50UAABWM5 has quit IRC12:04
*** csaavedra has joined #maemo-devel12:20
*** csaavedra has quit IRC12:40
*** Acedip has quit IRC14:07
*** burchr has joined #maemo-devel14:53
*** burchr has joined #maemo-devel14:53
*** Acedip has joined #maemo-devel14:55
*** joorin has joined #maemo-devel17:58
*** desu has quit IRC17:59
*** desu has joined #maemo-devel17:59
joorinive tried to google and search the wiki for information about application icons. which resource should i have found? i've created a 48x48 png that i want to show up in the application manager.17:59
joorinusr/share/icons/hicolor/48x48/apps has been suggested as the right location.18:00
joorinbut no go for me.18:01
zokieriirc the icon in appmanager needs to be embedded into debian/control file18:02
joorinzokier, oh? is the application amanger reading te control files?18:02
joorin"manager"18:02
zokieryes18:02
joorinzokier, im doing my experiments directly, no package has been built.18:02
joorinzokier, but how come i could add an .desktop file?18:03
joorinzokier, and it showed up immediately.18:03
zokierhuh?18:03
joorinzokier, im talking about the application manager that is used to start new applications, not install or uninstall applications.18:04
zokierah18:04
joorinzokier, my bad.18:04
joorinzokier, the other application manager will without a doubt read control files. ;)18:04
joorinapparently my google karma is low.18:06
*** Aranel has joined #maemo-devel18:06
*** Aranel has quit IRC18:06
*** Aranel has joined #maemo-devel18:06
joorinzokier, any suggestions about where to find information about where to put icons to get them to show up when i tickle the More... icon?18:06
zokiersometimes the icons appear only after boot18:07
zokierthe menusystem caches the icons somehow18:07
joorinzokier, ok. i will not reboot to fix this. id rather try to kill processes before rebooting to "fix it".18:07
joorinzokier, and its not blatantly simple to find information about the different icon directories. why have a gazillion of them?18:08
joorin(yes, i know the reason, just a bit unsatisfied with how things work.)18:09
joorinand the OTA upgrade failed in a nice way too.18:09
zokierwelcome to linux systems :)18:09
joorinzokier, well, ive spent ... some time with linux systems already. the maemo way of handling UI isnt impressing me yet.18:09
joorinif this is all about a bug where application icons wont show up, im not impressed at all.18:10
joorinand, again, why wont a simple search for "N900 application icon location" give more than it does?18:11
joorinhildon-desktop has allocated 107MBytes. nice!18:14
joorinany idea about which process to tickle?18:14
joorinhildon-home ?18:15
joorinkilling hildon-desktop rebooted the device. handy feature...18:17
joorinAND i lost my weather forecast widget. bonus!18:18
joorinand i get to see the icon. apparently i placed it in the right place. lucky me.18:19
joorinreboot to show icons. pfff...18:20
joorinand if i ever want to update it, ill need to reboot again. impressive.18:21
SpeedEvilrss!=memory18:24
joorinspeedevil, ?18:26
SpeedEviljoorin: RSS is a very poor indicator of memory use - if that's what you mean by 'allocated'18:27
joorinspeedevil, ah, yes. i was looking at the VSZ column from ps -def.18:28
joorinspeedevil, proper counting of mapped memory pages would be nice, as always.18:29
* joorin sniffles and moans, "Me uptime, me precious uptime!".18:30
SpeedEvilhttp://www.selenic.com/smem/ works OK on device18:30
SpeedEvilThough not the pie graphs - but meh18:31
*** rcampbell has joined #maemo-devel18:44
*** rcampbell has quit IRC18:51
*** swc|666 has joined #maemo-devel19:12
*** rcampbell has joined #maemo-devel19:22
*** simeoni has quit IRC20:21
*** humis has joined #maemo-devel20:22
*** humis is now known as simeoni20:23
*** humbaba has joined #maemo-devel20:38
humbababrr20:38
*** humbaba has left #maemo-devel20:38
*** Acedip has quit IRC20:44
*** DocScrutinizer51 has quit IRC20:50
*** shinkamui has joined #maemo-devel21:25
joorini found an introduction to debian packaging and lintian was suggested for checking package integrity. a quick search for that with apt-cache gives just traces of an answer. what should i use?21:46
joorinon a sidenote, it was easier to build a package for the first time with the debian tools than with the RPM tools. interesting.21:47
*** mindfaq has joined #maemo-devel21:47
joorinsure, it cut some corners with dependencies and changelog but it worked well.21:47
joorin"i cut", the tools did what i told them.21:48
joorinbut i assume you can get just as detailed and picky with debian packages as with RPM ones.21:51
joorinnot having to list all files that the package regards as pertaining to it was nice.21:51
joorinbut it might be a source for grief if the package is a bit more complicated.21:52
SpeedEvilSorry - no idea. :)22:39
SpeedEvilAnyone any thoughts as to what [43955.309539] omapfb omapfb: Unknown ioctl 0x8004721022:39
SpeedEvil[43956.763336] omap-previewer omap-previewer: previewer_ioctl: Invalid Command Value22:39
SpeedEvilin dmesg might be?22:39
SpeedEvilThe ioctl is FAT_IOCTL_GET_ATTRIBUTES - which confuses me.22:40
*** Docscrutemp has joined #maemo-devel23:01
*** Docscrutemp is now known as DocScrutinizer5123:02
joorinspeedevil, confused by the ioctl call, which ioctl is called or something else?23:13
SpeedEvilI'm wondering where it's coming from - and what's calling the framebuffer with a FAT ioctl23:14
joorinspeedevil, how do you know its that ioctl? asking to try to cut away error factors.23:15
SpeedEvilWell - that's the only thing in the kernel picked up by gre23:16
SpeedEvilp23:16
joorinspeedevil, ok.23:16
joorinspeedevil, might it be a simple "off by one", or typo? in what range are the omapfb ioctls?23:17
joorinor, in which range. i do hate english sometimes.23:17
SpeedEvilhmm - I should check that.23:17
SpeedEvillooking now23:18
SpeedEvilOkaaay. So something is opening the preview and framebuffer drivers, and spraying random ioctls at them.23:37
*** Aranel has quit IRC23:41
joorinspeedevil, random sounds very odd.23:41
SpeedEvilIndeed.23:41
SpeedEvilHowever - I can see nothing obviously buggy in the omapfb reporting of the ioctl number.23:42
SpeedEvilAnd the previewer is also complaining about the same command.23:42
SpeedEvilWell - some incorrect ioctl - don't know if it's actually the same one23:43
joorinspeedevil, reporting of ioctl is typically just moving an int so that being fsck up is not too likely.23:44
lcukSpeedEvil, do some monitoring23:44
lcukie does it happen when you have specific apps on23:45
SpeedEvillcuk: I'm unsure - I need to try to track it down next time.23:45
joorinspeedevil, but, to put it in some perspective, a failed call might be better than a real call with bad values. ;)23:46
lcukyeah23:46
SpeedEvilWell - yeah.23:46
SpeedEvilI'm slowly getting into reading bits of kernel, and understanding them - and this was a good excuse23:47
joorinspeedevil, but i agress on it being weird if the ioctl changes and isnt deterministic.23:47
joorin"agress"? where did that come from? "agree"!23:47
SpeedEvilThere are 6 or 7 of those sequences23:47
SpeedEvilthree omapfb, then a previewer error23:48
joorinspeedevil, just guessing here: might it be some shifting that has gone awry? as in, someone trying to be a bit to smart when creating the ioctl call? that would generate very weird values.23:49
joorinbut, to even make the ioctl call, you need to open the right fd so its not that easy to happen to get it wrong.23:50
SpeedEvilPossibly.23:50
SpeedEvilI'm wondering if it's the camera app doing screwy stuff23:50
SpeedEvilOr maybe one of the many things I have from extras-devel23:50
joorinsince ioctl calls arent "special" in any way, its easy to just muck about and see what happens. extras-devel might very well be a good guess.23:52
joorinlcuk, and when it comes to MilkDrop, i got a bit preoccupied with getting a garage account and meddling with debian packaging.23:57
lcukjoorin, sure get some simple data processing functions together theres a few potential pats to follow23:57
lcukpaths23:58
joorinlcuk, the FFT part compiles, in C++, so im planning the conversion while trying to find easy ways to put it to good use.23:58
lcukit compiles? how fast can it run23:59
lcukit seems reasonable and feasible to do per pixel operations on 320*240 data per frame and still stay around 25fps23:59
joorinlcuk, i havent done any testing. im thinking about connecting some simple OpenGL representation to footify and see what happens.23:59
lcuksure sounds like a good test23:59

Generated by irclog2html.py 2.15.1 by Marius Gedminas - find it at mg.pov.lt!