**** BEGIN LOGGING AT Thu Apr 07 02:59:59 2011
Apr 07 03:25:11 <drack> can you bluesnarf with n900?
Apr 07 04:18:31 <drack> hello?
Apr 07 04:46:33 <pupnik_> hi drack i don't know
Apr 07 04:47:22 <gn00b> just did a hard reset on my nokia n900. now i'm trying to install all my apps again and i got a msg saying 'not enough space in application install memory. uninstall some applications to free enough space for this installation.' i haven't installed anything yet.
Apr 07 04:48:23 <pupnik_> gn00b: open terminal, type df and then press enter
Apr 07 04:48:31 <pupnik_> you will see disk free statistics
Apr 07 04:48:59 <pupnik_> the first line after "Filesystem" should be 'rootfs' which shows available space
Apr 07 04:50:21 <gn00b> rootfs used 128104 available 100836
Apr 07 04:51:57 <pupnik_> that's pretty odd
Apr 07 04:52:29 <pupnik> restart device, maybe something just got munged
Apr 07 04:52:45 <gn00b> just did
Apr 07 04:52:52 <pupnik> same error?
Apr 07 04:53:06 <gn00b> three times...same error
Apr 07 04:54:03 <pupnik> hm
Apr 07 04:55:25 <pupnik> if you do a 'du' is /home/opt full?
Apr 07 04:55:28 <pupnik> sorry
Apr 07 04:55:30 <pupnik> "df"
Apr 07 04:56:18 <gn00b> 95%
Apr 07 04:56:40 <gn00b> home 95% and opt 95%
Apr 07 04:56:59 <gn00b> no wait...
Apr 07 04:57:14 <pupnik> it's the same partition
Apr 07 04:57:30 <gn00b> ok...then it is 95%
Apr 07 04:57:37 <pupnik> that should be about 100MB free, so it's not the problem
Apr 07 04:58:05 <gn00b> shouldn't everything be free since i just did a hard reset?
Apr 07 04:59:01 <pupnik> i'm not sure.  i would try installing a package with apt-get install
Apr 07 04:59:03 <pupnik> a small package
Apr 07 05:07:01 <gn00b> i can't 'sudo gainroot' because i don't have rootsh installed. i can't install rootsh because i can't 'sudo gainroot' ... wtf?
Apr 07 05:25:10 <pupnik> gn00b: sorry forgot about that
Apr 07 05:26:05 <pupnik> gn00b: maybe something went wrong during flashing
Apr 07 05:28:49 <irushad> when is the next update releasing ?
Apr 07 05:41:11 <pupnik> wonder if n900s will *ever* drop below 200 eu
Apr 07 05:42:59 <ds3> pupnik: it might in the US if the TMO/ATT merger goes through
Apr 07 05:47:43 <pupnik> :(
Apr 07 05:51:07 <pupnik> what phone lets you take RAW photos?
Apr 07 05:51:35 <pupnik> n900 :D
Apr 07 05:52:11 <robbiethe1st> pupnik: Now, if only it could /read/ them.
Apr 07 05:53:03 <robbiethe1st> pupnik: I mean, F-cam's great and all, but not being able to look back at pictures you took, even roughly... sort of a pain.
Apr 07 05:53:32 <kerio> robbiethe1st: huh?
Apr 07 05:53:35 <kerio> scroll down
Apr 07 05:53:41 <kerio> or up
Apr 07 05:53:41 <robbiethe1st> ?
Apr 07 05:53:53 <kerio> oh, for raws
Apr 07 05:53:57 <robbiethe1st> Yea
Apr 07 05:53:58 <kerio> nvm then
Apr 07 06:02:19 <pupnik> robbiethe1st: what about 'bless' camera app?
Apr 07 06:02:47 <robbiethe1st> No clue; never used it
Apr 07 07:09:49 <pupnik> annoying scratchbox shows $ at the end of every line
Apr 07 07:09:52 <pupnik> in vi
Apr 07 07:10:04 <pupnik> what's that marking called when it shows tabs, CRs etc
Apr 07 07:10:10 <pupnik> makes cut and paste a problem
Apr 07 07:36:47 <Termana> good morning
Apr 07 07:38:52 <DocScrutinizer> gn00b: you should reflash COMBINED and make shure you don't abort first boot which will, during auto-optification, clean the /opt partition
Apr 07 07:52:59 <N9HUNDRED1983> wat is up
Apr 07 07:57:08 <N9HUNDRED1983> just installed cut
Apr 07 07:57:21 <N9HUNDRED1983> cutetube-QML
Apr 07 07:57:30 <N9HUNDRED1983> and its awesome
Apr 07 07:59:48 <MohammadAG> QML is awesome
Apr 07 08:19:46 <DocScrutinizer> I'm pondering to add Q* to my blacklist that has g* for now ;-D
Apr 07 08:20:18 <Venemo_N900> DocScrutinizer: why is that?
Apr 07 08:20:30 <DocScrutinizer> but then I'm in trouble which toolkit to use for gfx
Apr 07 08:20:46 <DocScrutinizer> Venemo_N900: because of QtMob fubar
Apr 07 08:21:21 <DocScrutinizer> and I mean FUBAR
Apr 07 08:21:30 <Venemo_N900> what's wrong with QtMob?
Apr 07 08:23:06 <DocScrutinizer> mompls I got a link for you
Apr 07 08:23:28 <Venemo_N900> ok, gimme the link.
Apr 07 08:24:05 <DocScrutinizer> source_devel/bugs menu in my bookmarks has this: http://qt.gitorious.org/qt-mobility/qt-mobility/blobs/master/plugins/sensors/n900/n900proximitysensor.cpp#line54
Apr 07 08:24:58 <Venemo_N900> what da fu ck is this?????? :O
Apr 07 08:25:38 <DocScrutinizer> to fix you not only need to rewrite from scratch the n900proximitysensor, but you need to weed out that friggin filesensor object and everything else basing on same poor understanding of how to access sysfs and avoid recurrent timer actions
Apr 07 08:26:03 <Venemo_N900> loool
Apr 07 08:26:49 <DocScrutinizer> werner almesberger meant "yeah, decent oversampling avoids aliasing" ;-P
Apr 07 08:27:40 <Venemo_N900> so it polls the sensor 100 times a second... bleh.
Apr 07 08:27:48 <DocScrutinizer> :-((
Apr 07 08:27:55 <Venemo_N900> :(
Apr 07 08:28:55 <DocScrutinizer> filesensor is an object fundamentally incompatible with the intended usecase of QtMob
Apr 07 08:29:12 <Venemo_N900> mhm
Apr 07 08:29:34 <Venemo_N900> sure, but why don't they just use the native Maemo api?
Apr 07 08:29:43 <DocScrutinizer> nfc
Apr 07 08:30:07 <DocScrutinizer> also why don't they use kevents with their nice signal/slot concept?
Apr 07 08:30:34 <Venemo_N900> probably because they don't know anything about kernel events...?
Apr 07 08:30:42 <DocScrutinizer> probably
Apr 07 08:31:03 <Venemo_N900> looking at the code, it seems that some average, underpaid guy wrote it hastily.
Apr 07 08:31:17 <DocScrutinizer> exactly
Apr 07 08:31:23 <Venemo_N900> but it does explain why Mobility sucks the hell out of the cpu.
Apr 07 08:31:30 <DocScrutinizer> yep
Apr 07 08:32:15 <DocScrutinizer> sorry, QtMob, QA NOT passed, back to devel please!
Apr 07 08:32:45 <Venemo_N900> DocScrutinizer ++
Apr 07 08:32:50 <DocScrutinizer> sorry all you app devels who used QtMob
Apr 07 08:33:14 <Venemo_N900> yay
Apr 07 08:33:30 <DocScrutinizer> same applies to you, back to maemo-devel until either you kicked QtMob or QtMob got their shit sorted
Apr 07 08:34:03 <Venemo_N900> although the problem seems to be trivially fixable by reimplementing that class.
Apr 07 08:34:41 <pupnik> workworkwork
Apr 07 08:34:48 <DocScrutinizer> yup, IF that's the only stinking pile of SH** in that lib. I doubt they got only one
Apr 07 08:35:15 <Venemo_N900> hell yeah
Apr 07 08:35:44 <DocScrutinizer> it's a clear indication of devels without a friggin clue, so I expect the whole lib being of same inferior quality
Apr 07 08:35:52 <timeless_w7ip> docscrutinizer: on the subject of Qt-Mob ... consider this sample code: https://bugs.meego.com/attachment.cgi?id=5243
Apr 07 08:36:47 <timeless_w7ip> oh, for people unfamliar w/ bugzilla...
Apr 07 08:36:55 <timeless_w7ip> to go from an attachment to a bug, add &action=edit
Apr 07 08:37:06 <DocScrutinizer> gasp *COUGH*
Apr 07 08:37:17 <timeless_w7ip> you mean *choke*, i think
Apr 07 08:37:38 <timeless_w7ip> anyway, please don't comment on the bgu
Apr 07 08:37:46 <timeless_w7ip> the bug is an exercise in bugzilla ping pong
Apr 07 08:37:58 <Venemo_N900> how so?
Apr 07 08:38:03 <DocScrutinizer> eh?
Apr 07 08:38:08 <timeless_w7ip> the goal being to get people to figure out how to find people w/o needing twenty or more volleys
Apr 07 08:38:37 <Venemo_N900> lool
Apr 07 08:38:54 <timeless_w7ip> if you're looking at the bug (or any bug), click the "History" link under the reporter at the top of the bug
Apr 07 08:39:22 <timeless_w7ip> the ping pong here is the 'Assignee' field (search for it in that page)
Apr 07 08:39:26 <ruskie> hehe...
Apr 07 08:39:34 <ruskie> they seriously have no clue who should actually handle it...
Apr 07 08:39:44 <timeless_w7ip> oh, give them time!
Apr 07 08:39:48 <timeless_w7ip> this is only 5 volleys :)
Apr 07 08:39:55 <timeless_w7ip> i give them another 5-10 :)
Apr 07 08:40:11 <timeless_w7ip> ruskie: but, as i said earlier, it's a learning experience
Apr 07 08:40:16 <ruskie> I mean... why can't the first person to stumble upon it and actually care enough to commit it do it...
Apr 07 08:40:16 <timeless_w7ip> they need to *learn* how to handle bugs
Apr 07 08:40:35 <Venemo_N900> :D
Apr 07 08:40:36 <timeless_w7ip> ruskie: are you familiar w/ NMU?
Apr 07 08:40:43 <timeless_w7ip> it's generally considered rude
Apr 07 08:40:51 <timeless_w7ip> it exists in some domains, but it's considered rude
Apr 07 08:40:59 <ruskie> NMU?
Apr 07 08:41:02 <timeless_w7ip> the other thing is that if you fix someone else's code, how will they learn?
Apr 07 08:41:04 <timeless_w7ip> ~nmu
Apr 07 08:41:04 <infobot> from memory, nmu is Non Maintainer Upload, it's when a debian developer which is not the maintainer of a package uploads an update of said package
Apr 07 08:41:39 <ruskie> if you don't actually maintain anything in 2 months after change happens I don't consider you a maintainer
Apr 07 08:41:46 * timeless_w7ip shrugs
Apr 07 08:41:49 <ruskie> that's my down to earth actually get stuff out the door approach
Apr 07 08:42:06 <timeless_w7ip> i haven't found maintainers to be particularly great in general
Apr 07 08:42:21 <timeless_w7ip> so you won't find me arguing with you
Apr 07 08:42:24 <ruskie> and I'm talking this as a distro dev(source mage)
Apr 07 08:42:42 <ruskie> so I'm familiar with I'll maintain this... then nothing happens
Apr 07 08:42:49 <timeless_w7ip> :)
Apr 07 08:42:59 <timeless_w7ip> anyway, what you see here is a different thing
Apr 07 08:43:04 <timeless_w7ip> it's a flavor of "bug hot potato"
Apr 07 08:43:11 <ruskie> so we simply have ping - do you care about this - no repsones in 14 days - ok I'll play with it
Apr 07 08:43:33 <timeless_w7ip> (or Bug Potato for short)
Apr 07 08:43:36 <ruskie> hehe
Apr 07 08:43:37 * timeless_w7ip wonders if people here are familiar w/ Hot Potato
Apr 07 08:44:17 <timeless_w7ip> anyway, your average engineer has 2-3x more things to do in a given day than he can possibly do
Apr 07 08:44:46 <timeless_w7ip> and he's supposed to work on *his* area
Apr 07 08:45:06 <timeless_w7ip> so his goal should be to find someone who is more closely related to the problem than he and pass it along
Apr 07 08:45:14 <ruskie> I consider myself lucky that I tend to not have 2-3x more things to do
Apr 07 08:45:17 <ruskie> as a sysadmin
Apr 07 08:45:19 <timeless_w7ip> the problem here is that as you've noted, they really have no clue which direction to send it
Apr 07 08:45:32 <timeless_w7ip> someone wrote a nokia email address in the early bits of the bug
Apr 07 08:45:39 <ruskie> I've read it yeah ;)
Apr 07 08:45:46 <timeless_w7ip> that person should have said "I'll send an email to <...> asking him to find an owner"
Apr 07 08:46:04 <timeless_w7ip> and he should have left assigned to himself until the emailee replied
Apr 07 08:46:09 <ruskie> if it were the one making the call the last commiter(or the most common commiter) would be my first choices...
Apr 07 08:46:27 <timeless_w7ip> if you look at the bug's initial bits, there's a url
Apr 07 08:46:36 <timeless_w7ip> at the bottom of the page is the list of comitters to the file
Apr 07 08:46:44 <timeless_w7ip> and you can see that the file was buggy from its initial commit
Apr 07 08:46:47 <timeless_w7ip> (iirc)
Apr 07 08:47:17 * timeless_w7ip waits for ruskie to confirm
Apr 07 08:47:28 <ruskie> hehe
Apr 07 08:47:31 <ruskie> not gonna happen
Apr 07 08:47:36 <timeless_w7ip> why?
Apr 07 08:47:38 <ruskie> to tired to read through commits and code etc...
Apr 07 08:47:43 <DocScrutinizer> yeah, but how will we stop QtMob from contaminating all maemo?
Apr 07 08:47:44 <ruskie> and heading off to get some food
Apr 07 08:47:45 <ruskie> lunch time
Apr 07 08:47:52 <timeless_w7ip> oh, enjoy lunch
Apr 07 08:47:57 <timeless_w7ip> you don't need to read through the commits
Apr 07 08:48:05 <timeless_w7ip> the web ui will show you the oldest version of the file
Apr 07 08:48:14 <timeless_w7ip> and you can quickly search for the code blob and confirm it's still present
Apr 07 08:48:26 <timeless_w7ip> ruskie: the bug is there as a learning exercise for everyone
Apr 07 08:48:31 <Venemo_N900> heheh
Apr 07 08:48:49 <timeless_w7ip> for people like you, i'm trying to show how MXR can help answer questions like this
Apr 07 08:49:14 <Venemo_N900> mxr=?
Apr 07 08:49:28 <DocScrutinizer> meego xross reference
Apr 07 08:49:38 <timeless_w7ip> venemo: well, step 1, did you get to the bug page from the attachment link?
Apr 07 08:50:10 <timeless_w7ip> (step 1 is a "advanced tricks of bugzilla" lesson)
Apr 07 08:50:38 <DocScrutinizer> damn, I already closed the ticket page
Apr 07 08:50:46 <DocScrutinizer> :-)
Apr 07 08:50:59 <timeless_w7ip> well, you can get back there :) ... history or repeat the steps
Apr 07 08:51:07 <timeless_w7ip> i've reclosed the page 10-20 times in the last 10 mins :)
Apr 07 08:52:03 <timeless_w7ip> oh, and i'm now looking at the wrong file..
Apr 07 08:53:11 <DocScrutinizer> heh, this time adding &action=edit took me directly to the ticket instead of to a edit window with the patch text
Apr 07 08:53:26 <timeless_w7ip> hrm, it isn't supposed to do that..
Apr 07 08:53:45 <timeless_w7ip> doc: anyway, see if you can find the helpful source link
Apr 07 08:54:41 <DocScrutinizer> mxr.meego.org/... looks promising
Apr 07 08:55:10 <timeless_w7ip> yeah, just be careful, you want the link for the buggy code, not the victim code :)
Apr 07 08:55:26 <voltagex> is there a way to currently get maemo-pan working on the N900?
Apr 07 08:55:36 <timeless_w7ip> as a hint, the most important link should be in the clickable <url> field in the bug metadata section
Apr 07 08:55:50 <timeless_w7ip> (typically at the top of bugs unless you've changed your bugzilla preferences)
Apr 07 08:57:08 <DocScrutinizer> history on mxr gives me OP A.Stanley-Jones Dec 16 2010
Apr 07 08:58:00 <timeless_w7ip> hrm
Apr 07 08:58:06 <timeless_w7ip> not the 14th?
Apr 07 08:58:21 <timeless_w7ip> i'm using  the drop down box at the bottom of the page
Apr 07 08:59:53 <timeless_w7ip> anyway, as you can see, a small bug can be a very valuable learning experience for many people :)
Apr 07 09:01:42 <DocScrutinizer> for me it is for sure, just I'm not realy involved in fixing bugs ;-)
Apr 07 09:02:06 <Venemo_N900> DocScrutinizer: btw, how do the kernel events that you talked about work?
Apr 07 09:02:29 <DocScrutinizer> why t f does "history" link upper right not show the dec 14. versions?
Apr 07 09:02:40 <DocScrutinizer> netlink socket
Apr 07 09:03:37 <crashanddie_> lol, GAN900, some idiot is trying to get you in trouble on TMO...
Apr 07 09:03:48 <crashanddie_> "I'm stunned that no-one brought up GeneralAntilles... after all, in the US, impersonating an officer of the military is a serious crime!"
Apr 07 09:04:08 <DocScrutinizer> basically you open a socket and wait for msgs from kernel. Kernel sends a msg for every event, with details, and you filter for the events you're interested in
Apr 07 09:04:09 <MohammadAG> lol
Apr 07 09:04:12 <timeless_w7ip> doc: hrm
Apr 07 09:05:00 <Venemo_N900> thx DocScrutinizer
Apr 07 09:05:12 <timeless_w7ip> grr, gitorious requires JS!
Apr 07 09:05:13 <DocScrutinizer> timeless_w7ip: hrm?
Apr 07 09:05:20 <crashanddie_> timeless_w7ip, no it doesn't.
Apr 07 09:05:47 <timeless_w7ip> crashanddie_: disable JS and load: http://qt.gitorious.org/qt-mobility/contacts/commit/470d4a57d58ef8f2b2b7a1817714da925fc02340
Apr 07 09:05:50 <crashanddie_> well, older versions didn't, i browse it on my server with lynx and no JS
Apr 07 09:06:35 <crashanddie_> indeed
Apr 07 09:06:44 <timeless_w7ip> that page is technically useful
Apr 07 09:06:48 <DocScrutinizer> Venemo_N900: http://git.freesmartphone.org/?p=cornucopia.git;a=blob;f=libfsoframework/fsoframework/kobjectnotifier.vala;h=b4d6aca64d67e118a58fb11f72aa084b75f50fb5;hb=00c471c86667ad3d86610296a634496d746ec287#l61
Apr 07 09:06:55 <timeless_w7ip> but not practically useful because their design is stupid
Apr 07 09:06:58 <crashanddie_> though the qt.git and maemo.git (and any nokia.git) may be running versions of gitorious I'm not aware of
Apr 07 09:07:00 <timeless_w7ip> there are two *important* links
Apr 07 09:07:16 <crashanddie_> timeless_w7ip, most of gitorious was designed for design, not practicality
Apr 07 09:07:18 <timeless_w7ip> but there's no indication that they're important
Apr 07 09:07:55 <crashanddie_> timeless_w7ip, feel free to email support@gitorious.org or rant in #gitorious, they're actually really open to helpful comments
Apr 07 09:09:02 <timeless_w7ip> crashanddie: atm i'm ranting about doc's date issue
Apr 07 09:09:15 <timeless_w7ip> it sounds like my repo importer got the date wrong
Apr 07 09:09:15 <crashanddie_> k
Apr 07 09:09:18 <voltagex> I really need Bluetooth PAN working.
Apr 07 09:09:23 <timeless_w7ip> and i'm trying to understand how that could happen :(
Apr 07 09:09:44 <timeless_w7ip> doc: so...
Apr 07 09:09:50 <timeless_w7ip> in theory w/ enough bad time zone conversions
Apr 07 09:09:57 <timeless_w7ip> i suppose bad things could happen
Apr 07 09:10:18 <Venemo_N900> DocScrutinizer: about the QtM bug, we should take it into someone like achipa's attention.
Apr 07 09:11:22 <MohammadAG> or see how proximityd does it and do that ourself
Apr 07 09:11:49 <timeless_w7ip> oh strange
Apr 07 09:11:57 <timeless_w7ip> doc: i think i see the problem
Apr 07 09:12:01 <Venemo_N900> MohammadAG: proximityd doesn't use the accelerometer, does it?
Apr 07 09:12:16 <timeless_w7ip> if you look at the datestamp here: http://mxr.meego.com/hgweb.cgi/contacts/rev/f726ca721a91
Apr 07 09:12:18 <DocScrutinizer> prolly not
Apr 07 09:12:19 <timeless_w7ip> it's +1000
Apr 07 09:12:22 <MohammadAG> no, but the same polling is done for the proximity sensor
Apr 07 09:12:31 * timeless_w7ip wonders if that's Chineese Eastern Time or something
Apr 07 09:12:33 <MohammadAG> remember mediabar?
Apr 07 09:13:08 <Venemo_N900> sure
Apr 07 09:14:07 <MohammadAG> Venemo_N900, I installed Gnome shell 3 btw, it doesn't update the window I'm using
Apr 07 09:14:18 <DocScrutinizer> Venemo_N900: but actually accelerometer is a good example. It probably has quite identical issues
Apr 07 09:14:34 <Venemo_N900> MohammadAG: doesn't update what window?
Apr 07 09:14:35 <MohammadAG> like, if I'm typing on IRC, I don't see anything new on the window
Apr 07 09:14:37 <MohammadAG> DocScrutinizer, it does
Apr 07 09:14:55 <MohammadAG> DocScrutinizer, you can assume the same for all sensors in Qt Mobility
Apr 07 09:15:05 <MohammadAG> Venemo_N900, the window isn't repainted
Apr 07 09:15:07 <DocScrutinizer> yep, what I suspected
Apr 07 09:15:11 <Venemo_N900> MohammadAG: well it works for me. what distro is that?
Apr 07 09:15:22 <MohammadAG> ubuntu
Apr 07 09:15:29 <DocScrutinizer> that made me utter "QtMob is *FUBAR*""
Apr 07 09:15:42 <Venemo_N900> DocScrutinizer: yeah, unfortunately :(
Apr 07 09:15:56 <DocScrutinizer> ~fubar
Apr 07 09:15:57 <infobot> fubar is F*cked Up Beyond Any Recognition, e.g. "This whole operation is fubar, soldier" (gay lisp included), or a bar addon like Titan Panel and Telo's InfoBar. and everything.
Apr 07 09:16:09 <Venemo_N900> MohammadAG: it's ot here, let's talk about it in pm
Apr 07 09:16:12 <DocScrutinizer> I meant Beyonf All Repair
Apr 07 09:16:24 <timeless_w7ip> doc: ok, grr
Apr 07 09:16:30 <timeless_w7ip> i think gitorious is actually buggy here
Apr 07 09:16:32 <timeless_w7ip> if you load http://qt.gitorious.org/qt-mobility/contacts/commit/470d4a57d58ef8f2b2b7a1817714da925fc02340?format=patch
Apr 07 09:16:36 <timeless_w7ip> Date: Tue, 14 Dec 2010 17:08:51 +1000
Apr 07 09:16:45 <timeless_w7ip> that sure doesn't look like the 16th to me
Apr 07 09:18:41 <DocScrutinizer> timeless_w7ip: that's all hopelessly above my head, I'm stubling thru git and am happy to find some useful thing on accident. No clue to tell what's right or wrong with git(orious)
Apr 07 09:18:42 <timeless_w7ip> doc: sorry... mxr is innocent, gitorious is buggy
Apr 07 09:19:04 <timeless_w7ip> doc: yeah, don't worry about the gitorious
Apr 07 09:19:15 <timeless_w7ip> i think you'll find that mxr is friendlier
Apr 07 09:19:29 <timeless_w7ip> i'm sorry that you ran into a date bug in gitorious, but it isn't my fault (and i'm very happy about that)
Apr 07 09:19:37 <DocScrutinizer> hehe
Apr 07 09:20:18 <timeless_w7ip> doc: in the long run, i'd imagine the gitorious people should thank you for helping them find and hopefully fix their date bug :)
Apr 07 09:20:56 <timeless_w7ip> sp3000: food planning?
Apr 07 09:21:19 <DocScrutinizer> well, without you checking it, I'd probably add it to my long list of "points I don't understand regarding git" and forgot about it
Apr 07 09:21:26 <mece> is there a method for updating icon cache that actually does something?
Apr 07 09:21:31 <mece> in maemo 5
Apr 07 09:22:24 <timeless_w7ip> mece: <rm $path_to_icon_cache>
Apr 07 09:22:25 <mece> gtk-update-icon-cache works about as good as 'these goggles'(tm)
Apr 07 09:22:33 <timeless_w7ip> <kill $pid_of_icon_cache_process>
Apr 07 09:22:41 <DocScrutinizer> mece: prolly not, as I still frequently have that issue of icons of a fresh installed app being just a blue square, until next boot
Apr 07 09:22:54 <mece> timeless_w7ip, serious?
Apr 07 09:22:57 <timeless_w7ip> mece: basically if you kill enough files and processes
Apr 07 09:23:02 <timeless_w7ip> it should "fix" the "problem"
Apr 07 09:23:10 <timeless_w7ip> but i don't know offhand which they are :)
Apr 07 09:23:16 <mece> timeless_w7ip, well it's a problem, not a "problem"
Apr 07 09:23:22 <achipa> Venemo: which bug ?
Apr 07 09:23:22 <timeless_w7ip> whichever :)
Apr 07 09:23:26 <voltagex> bluetooth PAN server anyone?
Apr 07 09:23:45 <achipa> or rather Venemo_N900 ^^
Apr 07 09:24:00 <Venemo_N900> achipa: DocScrutinizer noticed that the way QtM polls the sensors is totally wrong.
Apr 07 09:24:32 <DocScrutinizer> achipa: http://qt.gitorious.org/qt-mobility/qt-mobility/blobs/master/plugins/sensors/n900/n900proximitysensor.cpp#line54
Apr 07 09:24:33 <achipa> Venemo_N900: is there something already filed on the issue ?
Apr 07 09:24:50 <Venemo_N900> achipa: just look at the link
Apr 07 09:24:55 <Venemo_N900> achipa: I'm not sure
Apr 07 09:26:29 <DocScrutinizer> achipa: for a context: you MUST NOT poll on /sys for detecting events on sensors. This rapes your idle state of CPU and thus your battery
Apr 07 09:26:37 <DocScrutinizer> ~2119
Apr 07 09:26:37 <infobot> The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", "SHOULD", "SHOULD NOT", "RECOMMENDED",  "MAY", and "OPTIONAL" in this document are to be interpreted as described in RFC 2119.
Apr 07 09:26:50 <timeless_w7ip> doc: hey, you should switch to using mxr for such urls :)
Apr 07 09:27:28 <DocScrutinizer> timeless_w7ip: I'd probably not even find my way to same location on mxr
Apr 07 09:27:30 <timeless_w7ip> http://mxr.meego.com/qt.gitorious.org/source/qt-mobility/plugins/sensors/n900/n900proximitysensor.cpp#54
Apr 07 09:27:48 <timeless_w7ip> doc: could you do me a favor and try?
Apr 07 09:28:00 <timeless_w7ip> i think you'll find that mxr is at least as easy to use if not considerably easier
Apr 07 09:28:09 <DocScrutinizer> hmm, ok. but first...
Apr 07 09:28:12 <timeless_w7ip> but if it isn't, i need to know what's going wrong so i can improve it
Apr 07 09:28:12 <DocScrutinizer> time for 12648430   ☕
Apr 07 09:28:47 <DocScrutinizer> so bear with me, I need some minutes...
Apr 07 09:29:30 <chem|st> morning
Apr 07 09:29:50 <timeless_w7ip> doc: sure, i have flights to book :)
Apr 07 09:30:41 <timeless_w7ip> wow
Apr 07 09:30:53 * timeless_w7ip has dozens of flights that will cost the same amount
Apr 07 09:35:34 <DocScrutinizer> moi n chem|st
Apr 07 09:37:09 <DocScrutinizer> Venemo_N900: http://www.linuxfoundation.org/collaborate/workgroups/networking/kevent
Apr 07 09:37:24 <Venemo_N900> thx DocScrutinizer
Apr 07 09:37:54 <timeless_w7ip> doc: coffee done?
Apr 07 09:39:15 <timeless_w7ip> hrm...
Apr 07 09:39:24 <timeless_w7ip> is O'Hare oh no! a key point?
Apr 07 09:39:36 * timeless_w7ip wonders if it's worth an extra couple of hundreds of dollars to avoid it
Apr 07 09:39:49 <timeless_w7ip> i can fly HEL=>LHR=>ORD=>BWI
Apr 07 09:41:32 * timeless_w7ip asks kayak to avoid ORD
Apr 07 09:42:15 <timeless_w7ip> cool
Apr 07 09:42:21 <timeless_w7ip> oh
Apr 07 09:42:28 <timeless_w7ip> when's hurricane season? :)
Apr 07 09:42:40 <timeless_w7ip> ah, right, june 1st :)
Apr 07 09:44:37 <drack> does anyone know if the super bluetooth hack works on n900?
Apr 07 09:47:16 <MohammadAG> no, it doesn't
Apr 07 09:47:35 <achipa> metspace interruption... anyway
Apr 07 09:47:45 <drack> i read that it only requires java to run superbluetooth hack
Apr 07 09:48:54 <DocScrutinizer> timeless_w7ip: acomplished I'd say: http://mxr.meego.com/qt.gitorious.org/search?find=/&string=addDataRate
Apr 07 09:49:48 <chem|st> drack: _only_
Apr 07 09:49:49 <DocScrutinizer> even more, it's harrowing
Apr 07 09:52:25 <drack> what? you need more than just java?
Apr 07 09:54:24 <chem|st> drack: what is *superBT* anyway
Apr 07 09:54:50 <chem|st> drack: java is a bi7ch
Apr 07 09:55:41 <drack> it is used to hack blutooth onsony ericsission
Apr 07 09:57:46 <achipa> DocScrutinizer: so the bug is... ? I mean, the real question is do you have event driven notifications from the kernel level - if that's a no, you poll, and if your app said it wants to poll... well, it polls...
Apr 07 09:57:58 <achipa> just to be clear on inefficient != bug...
Apr 07 09:59:21 <alterego> Gawd, not this again
Apr 07 09:59:34 <achipa> hey, I'm new on this one ;)
Apr 07 10:01:26 <achipa> but the thing is, unless there is way to read the sensor change through an event or callback, mobility does it the only way it can -> it's not a *mobility* bug. Whether the kernel driver should provide a more sane interface in addition to just /sys is a different story
Apr 07 10:01:32 <DocScrutinizer> achipa: the hw supports IRQ, the kernel SHALL support IRQ, so polling SHOULD not get used. If anything (kernel sysfs implementation, driver implementation) has a roadblock that defeats use of an event driven architecture, then ther bug is there. If QtMob is using polling despite kevents (or any other means to deal with interrupts) is supported all the way down to API, then the bug clearly is in QtMob
Apr 07 10:03:31 <achipa> well, I'm pretty sure there will be no kernel update for Maemo5, so that part is the way it is - except if somebody cares enough to make a patch for the power kernel or sorts
Apr 07 10:03:58 <achipa> if that happens, I'm more than happy to fix mobility to check and use a non-polling interface
Apr 07 10:03:59 <alterego> cssu
Apr 07 10:04:03 <ArkanoiD-> who needs stock kernel anymore?
Apr 07 10:04:05 <DocScrutinizer> achipa: so you suppose the kernel in fact does NOT support kevents??
Apr 07 10:04:42 <DocScrutinizer> achipa: citation needed
Apr 07 10:04:45 <achipa> DocScrutinizer: It's about the proximity device drivers - I have no idea what/how they work on the inside
Apr 07 10:05:05 <DocScrutinizer> achipa: it's not
Apr 07 10:05:43 <achipa> so you're saying the required kernel level stuff IS there and IS working, it's just mobility not using it ?
Apr 07 10:05:50 <DocScrutinizer> it's about the inplementation of filesensor being incompatible with the purpose of QtMOBILITY
Apr 07 10:06:32 <alterego> O_o
Apr 07 10:06:52 <DocScrutinizer> achipa: unless somebody proves me wrong on that, I'm temped to assume the kernel supports kevents (or another method to get callback invoked on change of a sensor sysnode)
Apr 07 10:06:58 <alterego> Does the kernel support the an irq event driven interface DocScrutinizer?
Apr 07 10:07:39 <alterego> Sys and inotify don't work very well tbh ..
Apr 07 10:07:45 <alterego> sysfs ..
Apr 07 10:08:00 <DocScrutinizer> inotify is BS, afaik. It's not even meant to get used on sysfs
Apr 07 10:08:04 <alterego> Kif that is what you mean by 'filesensor' ..
Apr 07 10:08:36 <alterego> Anyway, proximity under meego is actually handled by an input driver.
Apr 07 10:08:38 <achipa> alterego: filesensor is a rather dumb polling "are we there yet" thingy
Apr 07 10:09:00 <alterego> That driver is obviously even driven and probably requires a different QtMobility backend.
Apr 07 10:09:12 <alterego> Oh, right.
Apr 07 10:09:14 <DocScrutinizer> http://git.freesmartphone.org/?p=cornucopia.git;a=blob;f=libfsoframework/fsoframework/kobjectnotifier.vala;h=b4d6aca64d67e118a58fb11f72aa084b75f50fb5;hb=00c471c86667ad3d86610296a634496d746ec287#l61
Apr 07 10:09:47 <MohammadAG> the fact that proximityd and mce work without raping the CPU proves that some kind of events are working
Apr 07 10:10:18 <qp>  /nick cmantito
Apr 07 10:10:19 <alterego> Not really, polling not hard :P
Apr 07 10:10:29 <cmantito> oops ;p
Apr 07 10:10:30 <DocScrutinizer> HAHAHA
Apr 07 10:10:43 <DocScrutinizer> polling is a NOGO
Apr 07 10:11:53 <DocScrutinizer> it's not only the action to poll sth, it's also the overhead to transition from C5->C0 and back
Apr 07 10:12:35 <DocScrutinizer> you definitely don't want to do this on *several* sensors ONE HUNDRED times per second
Apr 07 10:12:55 <DocScrutinizer> you don't want to poll at all
Apr 07 10:13:02 <achipa> hey, you're preaching to the choir... but the N900 sensors are the way they are - we can get patches in but don't expect Nokia to optimize the hell out of them
Apr 07 10:13:35 <achipa> (though there might be something in Harmattan that could be useful in that regard)
Apr 07 10:13:36 <DocScrutinizer> achipa: that's aplologetic. There ARE better ways
Apr 07 10:14:01 <achipa> with the eternal words of Linus - talk is cheap, show me the code
Apr 07 10:14:06 <MohammadAG> achipa, there's nothing wrong with the hardware
Apr 07 10:14:22 <DocScrutinizer> and unless someone PROVES there's a roadblocker in kernel (which I strongly doubt there is) I claim we CAN USE better method
Apr 07 10:14:26 <achipa> if I have a patch that replaces that sensor with a non-polling implementation, I'll get it in trunk
Apr 07 10:14:36 <alterego> Like I said, the sensor under meego uses the input driver interface and is event driven.
Apr 07 10:14:55 <DocScrutinizer> achipa: (show code) seen my link above ^^^^ ???
Apr 07 10:15:24 <DocScrutinizer> achipa: I'm not going to replace filesensor in QtMob
Apr 07 10:15:33 <alterego> DocScrutinizer: he means a fix, the amount of time you're spent preaching this "bug" and not fixing it is absurd :P
Apr 07 10:15:34 <DocScrutinizer> that's definitely not my job
Apr 07 10:15:47 <SpeedEvil> Also - the prox sensor can't react at >3Hz - so polling at 100hz is a clear bug
Apr 07 10:15:49 <alterego> "not my job"
Apr 07 10:16:01 <SpeedEvil> Even if polling at 6Hz i also a bug.
Apr 07 10:16:03 <achipa> DocScrutinizer: SHOULD work is a different category than DOES works, that's all I'm saying
Apr 07 10:16:09 <achipa> s/works/work/
Apr 07 10:16:10 <infobot> achipa meant: DocScrutinizer: SHOULD work is a different category than DOES work, that's all I'm saying
Apr 07 10:16:38 <DocScrutinizer> alterego: that'S BS. QtMob is using filesensor at maybe 50 places. So you suggest I offer 50 patches (actually new implementations) to prove QtMob guys were too lazy to find a better way than filesensor??
Apr 07 10:16:43 <achipa> SpeedEvil: now that's a different story - and one we can act on
Apr 07 10:16:53 <alterego> Actually 6 is probably about right, without some kind of a has changed counter parameter.
Apr 07 10:17:08 <MohammadAG> 50 places?
Apr 07 10:17:23 <achipa> DocScrutinizer: I'm saying bitching won't change code. Patches will.
Apr 07 10:17:34 <alterego> DocScrutinizer: actually, just write a sample that shows there's a better way.
Apr 07 10:17:38 <MohammadAG> isn't it just two classes that need patches?
Apr 07 10:17:53 <DocScrutinizer> I'm not going to change code, I'm trying to teach devels to produce better code
Apr 07 10:18:06 <MohammadAG> (without breaking the current methods, so you don't have to rewrite lots of crap)
Apr 07 10:18:07 <alterego> He could just drop something in that replaces file sensor, I guess.
Apr 07 10:18:15 <DocScrutinizer> MohammadAG: won't fly
Apr 07 10:18:24 <MohammadAG> how come?
Apr 07 10:18:49 <DocScrutinizer> http://mxr.meego.com/qt.gitorious.org/search?find=/&string=addDataRate
Apr 07 10:18:57 <achipa> DocScrutinizer: sure, they will learn and do it right next time. Maybe. Still doesn't change the N900 implementation.
Apr 07 10:19:03 <chem|st> SpeedEvil: >2.5Hz if you are talking about state change
Apr 07 10:19:23 <SpeedEvil> chem|st: indeed - frequency of 2.5Hz - 5Hz state-changes
Apr 07 10:19:47 <MohammadAG> I only see 3 files related to the N900
Apr 07 10:20:15 <DocScrutinizer> MohammadAG: that's a bug not related to N900, it's a design level bug
Apr 07 10:20:19 <SpeedEvil> I don't know if inotify is supported on the state.
Apr 07 10:20:24 <chem|st> the sensors polling is 5Hz
Apr 07 10:20:51 <DocScrutinizer> hell the thing is called qtMOBILITY, so it better DOES NOT POLL
Apr 07 10:22:03 <chem|st> so state changes may happen with 2.5Hz at max (if we are talking about oscillation
Apr 07 10:22:11 <DocScrutinizer> inotify afaik is for normal filesystems, triggers on change of inodes etc. AFAIK it can not work on sysfs
Apr 07 10:22:52 <DocScrutinizer> I might be wrong
Apr 07 10:23:01 <MohammadAG> it doesn't work on sysfs
Apr 07 10:23:09 <MohammadAG> not sure if it can/can't, but on the N900, it doesn't
Apr 07 10:23:13 <chem|st> I mean closed state is 2.5Hz to be clear
Apr 07 10:24:08 <DocScrutinizer> that's largely irrelevant if this prox sensor has a f(0) of 1Hz or 5Hz
Apr 07 10:24:19 <alterego> I think you'll find it might work on certain nodes and not on others.
Apr 07 10:24:59 <DocScrutinizer> please learn to use sysfs events
Apr 07 10:25:19 <alterego> Have you even submitted a bug?
Apr 07 10:25:45 <DocScrutinizer> I don't care if the method is called NETLINK socket and kevents, or inotify or whatever. But I'm definitely sure there are better methods than polling
Apr 07 10:27:17 <DocScrutinizer> evidently any bug ticket would get closed as wontfix, as even here in a lengthy dialog you're apologetic and think there's no better way and not even any bug
Apr 07 10:28:00 <DocScrutinizer> plus accepting this bug would mean large parts of QtMob need a rework
Apr 07 10:28:05 <alterego> No one has said that, I'm just calling you lazy for not doiing _anything_ about it.
Apr 07 10:28:42 <alterego> Nothing but moan about it here that is. And who's attention do you think that is gonna get?
Apr 07 10:29:40 <DocScrutinizer> obviously yours, Venemo's, RST38h's, achipa's, MohammadAG's... Plus I opened the topic on #meego several times, and nothing happened
Apr 07 10:30:01 <alterego> That's because I don't think the problem exists on meego
Apr 07 10:30:06 <DocScrutinizer> alterego: I'm just doing what I can. I'm not going to submit patches
Apr 07 10:30:14 <DocScrutinizer> haha
Apr 07 10:30:26 <alterego> A bug report is better than nothing ..
Apr 07 10:30:32 <DocScrutinizer> POLLING on meego is OK then?
Apr 07 10:30:41 <alterego> What fucking polling?
Apr 07 10:30:42 <MohammadAG> <DocScrutinizer> evidently any bug ticket would get closed as wontfix, as even here in a lengthy dialog you're apologetic and think there's no better way and not even any bug
Apr 07 10:30:56 <alterego> proximity sensor doesn't even work under meego in mobility
Apr 07 10:30:57 <MohammadAG> Qt bugtrakcer, not maemo's
Apr 07 10:31:04 <alterego> For the exact reasons I gave.
Apr 07 10:31:29 <DocScrutinizer> meh, I'm definitely fed up with it and your bitching
Apr 07 10:31:33 <alterego> Sensors in meego use a different interface.
Apr 07 10:31:59 <alterego> So I'm not even sure mobilities "filepollng" method would work.
Apr 07 10:32:31 <alterego> Now if you're to idle to submit a bloody bug ... You're damn right O'm "bitching"
Apr 07 10:32:37 <alterego> (at you)
Apr 07 10:32:51 <alterego> At least I'm not denying there's an issue here.
Apr 07 10:33:59 <alterego> And all the energy that is wasted on your rants about this could have gone in to a great bug report ...
Apr 07 10:34:36 <alterego> Even if it's something as simple as: "File polling in low power embedded systems, are you kidding?!?!?"
Apr 07 10:37:41 <xkr47> oh man..
Apr 07 10:37:45 <DocScrutinizer> go ahead. You for sure are more familiar with the correct tracker, correct subsystem, whatever. I don't care. I had a nice relaxed chat with Venemo here, and your perceived bitching of me has been proviked by you entirely
Apr 07 10:39:49 <alterego> "I'm just doing what I can, I'm not going to sumbit patches" or make simple bug reports, or actually do anything?
Apr 07 10:40:16 <alterego> Fine, but why continually bring up the fact you're not going to do any of these things?
Apr 07 10:40:43 <alterego> Anyhow, bbl
Apr 07 10:41:29 <DocScrutinizer> because there are guys *interested* in learning about it, not interested in prodding me to participate in development on a level I feel not comfortable with
Apr 07 10:42:06 <alterego> You mean A socially responsible level?
Apr 07 10:42:47 <alterego> Oh there's a bug, it's a quite major bug, I know, I'll just tell a load of users about it and it'll never get resolved.
Apr 07 10:43:03 <DocScrutinizer> alterego: YOUR energy put into bitching at me supposedly not doing what YOU think is the right thing really better goes to filing a ticket :-P
Apr 07 10:46:06 <alterego> No, because it'd take me longer, I'd have to go through all the sources and validate.
Apr 07 10:46:39 <DocScrutinizer> see, I've shared all my knowledge and notions about this with you, and I'm not involved in development. So clearly now you're better qualified than me to act upon it
Apr 07 10:48:56 <DocScrutinizer> so go thru all the sources, lazy bugger! You at least know where they are, I don't. Then will you PRETTY PLEASE file a bug to whatever tracker is appropriate, and stop bitching at me
Apr 07 10:50:54 <DocScrutinizer> I spotted the bug (after MohammadAG kindly pointed me at it), and I even digged out alternative methods that look promising. Now I happily hand it over to your responsibility to carry on with it
Apr 07 10:51:52 <alterego> I'll keep it in mind, if it is still in place in meego I'll submit a bug
Apr 07 10:52:06 <DocScrutinizer> fair enough
Apr 07 10:52:29 <GeneralAntilles> crashanddie_: sounds like sjgadsby. :P
Apr 07 11:23:00 <mece> hey I found what updates the icon cache (which is not called that anymore, since icon cache is not used)
Apr 07 11:23:45 <mece> touch /usr/share/icons
Apr 07 11:23:46 <mece> :D
Apr 07 11:24:05 <mece> I shall install a random program and see if it works.
Apr 07 11:24:18 <mece> (my test were somewhat unorthodox)
Apr 07 11:31:08 <mece> yep, it works.
Apr 07 11:31:09 <mece> nice
Apr 07 11:32:42 <mece> so a touch /usr/share/icons in the postinst would make icons appear as they should. Or adding the command to gtk-update-icon-cache script
Apr 07 12:11:52 <gn00b> DocScrutinizer: thanx. re-flashing it worked
Apr 07 12:17:46 <achipa> DocScrutinizer: alterego: proxymity stuff is kooky - apparently there are sensors that you have to massively supersample (10x) to be able to get reasonable accuracy, not sure if the N900 has such thing or not, though, but still, interesting
Apr 07 12:25:46 <achipa> X-Fade: ping
Apr 07 12:43:20 <homeasvs> anyone know how I should install mud in scratchbox ? doesn't look like it's packaged
Apr 07 13:31:07 <Venemo> DocScrutinizer, I've just read the rest of the discussion - seems to have been heated up after I left :P
Apr 07 13:34:11 <Venemo> DocScrutinizer, honestly, I don't blame the Qt guys for this mess. The Maemo version of Qt Mobility is equally as hasted and messy as the other Qt for Maemo parts.
Apr 07 13:34:41 <Venemo> DocScrutinizer, for example, keyboard input is broken in Qt. I can't enter some capital letters :D
Apr 07 13:51:22 <GeneralAntilles> Ω
Apr 07 14:10:00 <homeasvs> is there a place I can see what the builder is doing ?
Apr 07 14:10:07 <homeasvs> possibly logs as they update live ?
Apr 07 15:16:52 <DocScrutinizer> GeneralAntilles: omega?
Apr 07 15:23:50 <DocScrutinizer> GRRRRRRRRR
Apr 07 15:23:53 <DocScrutinizer> Apr  7 17:22:48 IroN900 cellular: csd[795]: ISI_SMS .022491> incoming_cell_broadcast(): Incoming cell broadcast
Apr 07 15:24:26 <DocScrutinizer> thank you VERY MUCH for that helpful log msg
Apr 07 15:24:48 <GeneralAntilles> DocScrutinizer: tried to cmd-v a clear, ended up being option-z.
Apr 07 15:30:16 <RST38h> Good...whatever, ladies and gentlemen
Apr 07 15:30:49 <pupnik> good mooning!
Apr 07 15:31:36 <pupnik> RST38h: i spent half the afternoon drawing a pitchfork  :)  http://www.daimonin.org/forums/daimonin-project/800x480-widget-client/msg94907/#msg94907
Apr 07 15:34:37 <RST38h> pupnik: Ah, just get yourself an open-sourced RPG icon set...
Apr 07 15:35:23 <RST38h> pupnik: Or ask wazd if he would be interested in some applied megalomania =)
Apr 07 15:41:41 <pupnik> wazd may be helping out :D
Apr 07 15:42:41 <pupnik> but it's strictly "whatever is enjoyable" ... there are over 8800 tiles
Apr 07 15:47:34 <RST38h> pupnik:aaaaaargh
Apr 07 15:48:37 <homeasvs> is there any way to debug the upstart sequence during boot ?
Apr 07 15:48:48 <homeasvs> I can't figure out why the service I'm adding doesn't start
Apr 07 15:49:40 <alterego> homeasvs: does it "start" when you try and start it manually after boot?
Apr 07 15:49:50 <homeasvs> alterego, yep
Apr 07 15:50:59 <alterego> homeasvs: can you pastebin your upstart file
Apr 07 15:56:51 <deno> hi there
Apr 07 16:12:28 <DocScrutinizer> (upstart) depending on a signal that never comes?
Apr 07 16:13:12 * DocScrutinizer opens envelope from HK
Apr 07 16:27:29 <DocScrutinizer> who's been ordering speakers at ownta?
Apr 07 16:30:28 <DocScrutinizer> They sent the ordered 4 correct ones (I pointed them to the fact they have a wrong picture, they answered "don't worry") plus one like the picture they have on their N900-speaker offer :-D
Apr 07 16:31:58 <DocScrutinizer> they still have the wrong picture on their webpage: http://www.ownta.com/index.php?dispatch=products.view&product_id=65385
Apr 07 16:41:27 <timeless_mbp> docscrutinizer: got time for another round of "adventures in [code]wonderland"?
Apr 07 16:42:01 <timeless_mbp> http://mxr.meego.com/repo.meego.com/source/sofia-sip/libsofia-sip-ua/msg/msg.c?mark=239,245-246#235
Apr 07 16:44:25 <DocScrutinizer> timeless_mbp: slightly bored about it, nevertheless looking at it and can't find anything exciting. What you're asking me to do?
Apr 07 16:44:50 <timeless_mbp> docscrutinizer: the question is what does the code ask memset to do
Apr 07 16:45:06 <timeless_mbp> memset takes 3 arguments, and traditionally coders get at least one of them wrong :)
Apr 07 16:45:26 <DocScrutinizer> WTF??? sizeof(&msg->m_addr)
Apr 07 16:46:03 <DocScrutinizer> I'm not sure but that looks fishy
Apr 07 16:46:06 * timeless_mbp gives docscrutinizer a cookie
Apr 07 16:46:09 <khorben_> it does :)
Apr 07 16:46:23 <toggles_> lol
Apr 07 16:46:29 <timeless_mbp> khorben: care to expand "it does"? :)
Apr 07 16:46:38 <khorben_> luckily the first might match, but I doubt the second will
Apr 07 16:47:05 <khorben_> well the first argument needs a pointer to what will be zero'd
Apr 07 16:47:20 <khorben_> and the third wants its length
Apr 07 16:47:37 <khorben_> so usually, it's like: int i; memset(&i, 0, sizeof(i));
Apr 07 16:47:41 <DocScrutinizer> length of an addr?
Apr 07 16:47:59 <khorben_> but here it says the equivalent of sizeof(&i)
Apr 07 16:48:08 <khorben_> which most probably doesn't match
Apr 07 16:48:08 <RST38h> memset(ptr,value,size)
Apr 07 16:48:09 <DocScrutinizer> yup
Apr 07 16:48:23 <RST38h> so, why do people get it wrong again?
Apr 07 16:48:47 <DocScrutinizer> C&P error
Apr 07 16:48:53 <khorben_> either ENOTENOUGHCOFFEE or 0xbadc0ffee or E2MUCHCOFFEE
Apr 07 16:49:19 <timeless_mbp> rst: *shrug*
Apr 07 16:49:19 <DocScrutinizer> time for 12648430   ☕
Apr 07 16:49:25 <timeless_mbp> people seem to be very good at getting this stuff wrong
Apr 07 16:49:30 <timeless_mbp> ooh, i can read that coffee!
Apr 07 16:51:03 <DocScrutinizer> is msg->m_addr malloc'd?
Apr 07 16:51:50 <timeless_mbp> docscrutinizer: thankfully you have a working cross reference at your fingers :)
Apr 07 16:51:53 <DocScrutinizer> I mean, afaik malloc() nowadays makes sure you get zero'ed memory
Apr 07 16:51:59 <timeless_mbp> no, it doesn't
Apr 07 16:52:01 <timeless_mbp> calloc() does
Apr 07 16:52:04 <RST38h> no,not according to the docs
Apr 07 16:52:06 <DocScrutinizer> timeless_mbp: I'm incredibly lazy
Apr 07 16:52:08 <DocScrutinizer> :-)
Apr 07 16:52:23 <timeless_mbp> malloc() will merilly return memory which was previously used and discarded in the process space
Apr 07 16:52:48 <DocScrutinizer> aaah yes, in same process space, that might be correct actually
Apr 07 16:53:02 <DocScrutinizer> as the process got nuttin to hide from itself
Apr 07 16:53:29 <khorben_> some allocators even allow you to fill it with non-zero data
Apr 07 16:53:33 <khorben_> to trigger bugs
Apr 07 16:53:33 <timeless_mbp> note that a clibrary is free to start allocated memory with a value other than 0x0
Apr 07 16:53:42 <timeless_mbp> some tend to use 0xcccccccc
Apr 07 16:53:55 <khorben_> very platform-specific
Apr 07 16:53:57 <RST38h> "some" aka MSVC
Apr 07 16:54:05 <timeless_mbp> khorben: no, very libc specific
Apr 07 16:54:15 <khorben_> indeed
Apr 07 16:54:17 <timeless_mbp> platforms can have dozens of libcs each with their own behaviors :)
Apr 07 16:54:27 <RST38h> IBM xlc uses some 0xbadf00d or 0xdeadcafe, don't remember
Apr 07 16:54:28 <khorben_> or even, programs using their own allocators
Apr 07 16:54:36 <khorben_> talloc, jemalloc... (mozilla...)
Apr 07 16:55:15 <timeless_mbp> mozilla uses jemalloc for most stuff
Apr 07 16:55:15 <khorben_> java has 0xcafebabe
Apr 07 16:55:35 <timeless_mbp> although parts of it use other things (pldhash has its own uninitialized marker)
Apr 07 16:55:38 <jacekowski> DocScrutinizer: malloc sanitizes memory on some secure implementations
Apr 07 16:55:40 <DocScrutinizer> there, I've heard it again: cafe
Apr 07 16:55:48 <jacekowski> DocScrutinizer: like on linux with grsec/pax
Apr 07 16:56:06 <jacekowski> but that's only then
Apr 07 16:56:10 <jacekowski> take a look at pam code
Apr 07 16:56:15 <timeless_mbp> jacekowski: is that malloc or the kernel allocator?
Apr 07 16:56:15 <khorben_> and some implementations use mmap() instead of the heap
Apr 07 16:56:17 <jacekowski> it cleans memory before free
Apr 07 16:56:20 <timeless_mbp> there's a rather significant difference
Apr 07 16:56:21 <jacekowski> timeless_mbp: kernel allocator
Apr 07 16:56:25 <jacekowski> + malloc
Apr 07 16:56:35 <timeless_mbp> cleaning before free otoh is libc
Apr 07 16:56:38 <timeless_mbp> and much more amusing
Apr 07 16:56:43 <khorben_> jacekowski: security code should always clear its memory before freeing it
Apr 07 16:57:10 <khorben_> and actually, as soon as it's not required anymore
Apr 07 16:57:14 <jacekowski> timeless_mbp: as in, kernel does it + hardened glibc ( not used very often, can do that as well )
Apr 07 16:57:17 <timeless_mbp> khobern: so... http://mxr.meego.com/repo.meego.com/source/shadow-utils/libmisc/list.c#124
Apr 07 16:57:20 <alterego> Just had a couple of ideas for small Qml apps.
Apr 07 16:57:23 * alterego starts hacking.
Apr 07 16:57:24 <DocScrutinizer> hah, you bet I need some coffe plus popcorn now. This starts to get amusing...
Apr 07 16:57:24 <jacekowski> timeless_mbp: hardened kernel*
Apr 07 16:57:44 <timeless_mbp> i believe that code is correct, albeit unforunately written
Apr 07 16:57:52 <jacekowski> timeless_mbp: that's why on machines with grsec you will get 0'ed memory
Apr 07 16:57:55 <timeless_mbp> i.e. it would read better to use 'char*'
Apr 07 16:58:06 <timeless_mbp> khorben: can you confirm?
Apr 07 16:58:13 * timeless_mbp is tired
Apr 07 16:58:21 <khorben_> timeless_mbp: if 1. xmalloc exits if it fails 2. there is no integer overflow, then yes
Apr 07 16:58:42 <jacekowski> khorben_: woot?
Apr 07 16:58:43 <timeless_mbp> khorben: overflow is so much fun
Apr 07 16:58:45 <jacekowski> exit on malloc fail?
Apr 07 16:58:47 <timeless_mbp> we'll ignore that for now
Apr 07 16:58:48 <jacekowski> that sucks
Apr 07 16:58:56 <timeless_mbp> jacekowski: mozilla does that
Apr 07 16:58:59 <timeless_mbp> virtually nothing else does
Apr 07 16:58:59 <khorben_> jacekowski: malloc() can fail...
Apr 07 16:58:59 <jacekowski> and another thing, not checking for NULL on malloc
Apr 07 16:59:06 <timeless_mbp> yeah...
Apr 07 16:59:10 <jacekowski> khorben_: yes, and you should check for NULL
Apr 07 16:59:14 <khorben_> aria2/intl/relocatable.c
Apr 07 16:59:14 <khorben_> line 41 -- # define xmalloc malloc
Apr 07 16:59:17 <khorben_> GREAT
Apr 07 16:59:22 <timeless_mbp> heh
Apr 07 16:59:25 <jacekowski> not just kill process allocating memory
Apr 07 16:59:25 <khorben_> ok it's probably not the one
Apr 07 17:00:01 <khorben_> (this cross-reference tool is not helping much)
Apr 07 17:00:05 <timeless_mbp> jacekowski: we aren't talking about kernel murder
Apr 07 17:00:09 <timeless_mbp> we're talking about process suicide
Apr 07 17:00:10 <khorben_> (even ctags is more clever)
Apr 07 17:00:23 <timeless_mbp> khorben: for some things it's helpful
Apr 07 17:00:38 <timeless_mbp> for xmalloc you'd probably want to use 'shadow::xmalloc'
Apr 07 17:00:53 <jacekowski> timeless_mbp: still, if process calls malloc and libc terminates on fail it's more like a murder than suicide
Apr 07 17:00:54 <timeless_mbp> it should give vaguely better answers
Apr 07 17:00:58 <khorben_> of course, but for things like "msg_t" and "xmalloc"...
Apr 07 17:01:01 <khorben_> ok, trying
Apr 07 17:01:19 <timeless_mbp> khorben: actually probably shadow-::xmalloc or something
Apr 07 17:01:28 <khorben_> it was good already
Apr 07 17:01:33 <timeless_mbp> shadow-utils::xmalloc
Apr 07 17:01:36 <timeless_mbp> :)
Apr 07 17:01:44 <khorben_> 54 	ptr = (char *) malloc (size);
Apr 07 17:01:44 <khorben_> 55 	if ((NULL == ptr) && (0 != size)) {
Apr 07 17:01:44 <khorben_> 56 		fprintf (stderr, _("malloc(%d) failed\n"), (int) size);
Apr 07 17:01:44 <khorben_> 57 		exit (13);
Apr 07 17:01:58 <khorben_> casting as int is ridiculous
Apr 07 17:02:05 <khorben_> but not a problem here
Apr 07 17:02:24 <khorben_> still there is a corner case
Apr 07 17:02:36 <khorben_> if size overflows and is 0
Apr 07 17:02:42 <khorben_> then it still can return NULL
Apr 07 17:02:44 * timeless_mbp loads http://mxr.meego.com/repo.meego.com/ident?i=rtp::pa_rtp_context_init_recv
Apr 07 17:02:48 <khorben_> and the program can crash
Apr 07 17:02:48 <timeless_mbp> yeah
Apr 07 17:03:05 <jacekowski> khorben_: strange style
Apr 07 17:03:20 <khorben_> jacekowski: I've been auditing way, way worse code :)))
Apr 07 17:03:22 <jacekowski> khorben_: in most places people write variable==constant
Apr 07 17:03:36 <timeless_mbp> khorben: it sounds like you do the same sort of stuff i do :)
Apr 07 17:03:37 <khorben_> actually no
Apr 07 17:03:42 <jacekowski> and here it's opposite
Apr 07 17:03:46 <khorben_> constant == variable is better
Apr 07 17:03:59 <khorben_> because if you mistype "=" for "==" then it doesn't compile
Apr 07 17:04:02 <jacekowski> yeah, it will give you an error if you mistype
Apr 07 17:04:06 <timeless_mbp> http://mxr.meego.com/repo.meego.com/source/pulseaudio/src/modules/rtp/rtp.c#290
Apr 07 17:04:34 <timeless_mbp> khorben: i agree, but it's impossible to get people to change to that style
Apr 07 17:04:42 <khorben_> I don't even do it myself
Apr 07 17:07:56 <DocScrutinizer> ohnoes, poor japan. 7.4
Apr 07 17:08:14 <pupnik> do they even notice that?
Apr 07 17:08:31 <timeless_mbp> heh
Apr 07 17:09:18 <pupnik> 'waitress, a napkin please, my ramen has spirt'
Apr 07 17:09:31 <jacekowski> DocScrutinizer: again?
Apr 07 17:09:47 <DocScrutinizer> 2:30 ago
Apr 07 17:13:19 <timeless_mbp> https://bugs.meego.com/show_bug.cgi?id=15580 filed
Apr 07 17:13:20 <povbot`> Bug 15580: was not found.
Apr 07 17:13:21 <timeless_mbp> i didn't include the patch
Apr 07 17:13:26 <timeless_mbp> if anyone wants to, they can
Apr 07 17:13:38 <timeless_mbp> as a note, MXR includes a 'raw file' link at the top right of each page
Apr 07 17:13:44 <timeless_mbp> so you can use that to grab the source and make your patch
Apr 07 17:54:24 <pupnik> sdl sleep() isn't buying me much power saving at all :(
Apr 07 18:02:29 <ShadowJK> is that plain sleep()? for how long are you sleeping
Apr 07 18:02:50 <DocScrutinizer> Onagawa NPP on auxiliary Diesel generators :-S
Apr 07 18:08:24 <RST38h> Doc: What? Another one?
Apr 07 18:09:18 <ShadowJK> ah, so that's why my arduino is low on power
Apr 07 18:09:39 <chx> DocScrutinizer: https://twitter.com/#!/JRNHeadlines/status/56028623907471360
Apr 07 18:09:50 <DocScrutinizer> RST38h: quake 7.4 caused power cut on Onagawa NPP, running on emergency local power now
Apr 07 18:10:17 <RST38h> Doc: Ok, but still on diesel, right?
Apr 07 18:10:58 <RST38h> chx: You mean, all these people won't get to run-to-the-hills now?
Apr 07 18:11:27 <DocScrutinizer> RST38h: see chx link
Apr 07 18:11:55 <chx> DocScrutinizer: my understanding is that they recovered
Apr 07 18:12:07 <ShadowJK> "Electricity from Onagawa nuclear plant near Sendai went out after quake." ... "Onagawa nuclear plant in Miyagi loses 2 of 3 external power systems " ... "Power recovered at Onagawa nuclear plant following post-earthquake outage."
Apr 07 18:12:43 * ShadowJK added the <bold>
Apr 07 18:12:53 <ShadowJK> unless it got filtered
Apr 07 18:12:59 <SpeedEvil> You should really have gone with <blink>
Apr 07 18:13:25 <ShadowJK> it doesn't work on anything but BitchX and a proper console
Apr 07 18:13:53 <DocScrutinizer> well, here it cancelled the highlight of your post :-D
Apr 07 18:14:27 <RST38h> ircii here, all your <crap> efforts for nothing=)
Apr 07 18:15:20 <ShadowJK> So the first: plant stopped supplying power. Big surprise, it's supposed to automatically shut down during earthquake. Second post: The plant has 3 redundant links to outside grid. Two of them failed. (what about the UPS and diesels??). Third: everything working okay now
Apr 07 18:15:44 <DocScrutinizer> yup
Apr 07 18:16:12 <RST38h> diesels and ups are not links
Apr 07 18:16:44 <DocScrutinizer> well, our TV news here aren't top up-to-date as they used to be on day-2 of the big boom
Apr 07 18:16:45 <RST38h> so, I guess the diesels worked, until they restored the links, everything is ok, no apocalypse this time, no lulz
Apr 07 18:18:13 <RST38h> There is an aticle about schools sending students to help cleaning up after the earthquake in Miyagi
Apr 07 18:19:07 <DocScrutinizer> hmm, my set of 4 pcs spare N900 speakers is kinda more exciting
Apr 07 18:19:12 <RST38h> With cute schoolgirls in masks shoveling junk
Apr 07 18:19:41 <SpeedEvil> Quadrophony!
Apr 07 18:19:52 <RST38h> Will N900 do Dolby with 4 extra speakers? =)
Apr 07 18:19:59 <pupnik> wonder if there's an oprofile kernel for n900
Apr 07 18:20:12 <ShadowJK> It's amusing how media coverage in western world made it look so much more dramatic with everyone wearing facemasks. I guess they forgot to mention that it's socially acceptable to wear a mask anytime over there :-)
Apr 07 18:20:22 <DocScrutinizer> I guess it will still be able to playback Thomas Dolby, yes ;-P
Apr 07 18:20:30 <RST38h> Shadowjk: that too
Apr 07 18:20:46 <RST38h> Shadowjk: but the girls are also in light full-body suits.
Apr 07 18:21:03 <RST38h> No exactly the bunny suits, but close enough
Apr 07 18:21:54 <ShadowJK> I've heard companies and corporations hve their own assigned areas of tidying up too, and that the all the staff except critical staff go outside to pick up trash a few days a year?
Apr 07 18:22:31 <DocScrutinizer> ShadowJK: never noticed that aspect, as I'm used to that habit - been in Taipei long enough to not even notice it anymore
Apr 07 18:22:32 <RST38h> ShadowJK:Very....Soviet.
Apr 07 18:27:13 <RST38h> "Naked Florida man opens fire on SWAT bot with AK-47"
Apr 07 18:30:13 <pupnik> people need better armaments now
Apr 07 18:31:16 <pupnik> http://en.wikipedia.org/wiki/Type_69_RPG
Apr 07 18:31:33 <pupnik> "The origin of the RPG can be traced back to the WWII German Panzerfaust, based on which the Soviets have developed a range of grenade launchers."
Apr 07 18:31:54 <pupnik> das ist gut, ja
Apr 07 18:32:15 <RST38h> Yeaa, but you can just use the geniune RPG-7 part with a variety of different ammunition types
Apr 07 18:35:03 <RST38h> pupnik: Too bad the naked guy did not have any, or he could have use it on the police bot
Apr 07 18:38:02 <pupnik> http://www.operatorchan.org/k/arch/src/k187624_Chinese%20Type%2098%20%28PF-98%29%20120mm%20anti-tank%20rocket%20sys.jpg
Apr 07 18:38:10 <pupnik> "miss, can you stow this for me?"
Apr 07 19:19:35 <ZogG> http://www.youtube.com/watch?v=5ocq6_3-nEw&feature=youtu.be&a
Apr 07 20:15:08 <LjL> so my sister's N810 keeps rebooting in a loop... barely shows the desktop, then blam. anything i can try before reflashing?
Apr 07 20:19:02 <LjL> i think it's very possible that a full internal disk was partly the cause
Apr 07 20:19:42 <ArkanoiD-> ah, my scud batteries just arrived
Apr 07 20:57:18 <gn00b> OK. i've done a hard reset 3x's now with 3 different *.bin files, but i still have the same error...'Not enough space in application install memory. Uninstall some applications to free enough space for this installation.' can anyone please help???
Apr 07 21:00:26 <gn00b> when i look at the available memory, it shows that i have 253.0MB available
Apr 07 21:03:44 <pupnik> i wonder if you can get root without installing a package
Apr 07 21:04:10 <gn00b> can't install rootfs
Apr 07 21:04:21 <gn00b> sudo asks for a password
Apr 07 21:05:19 <pupnik> what does the application manager log say when you try to install an app?
Apr 07 21:05:25 <pupnik> there's a button somewhere for the logfile
Apr 07 21:07:28 <gn00b> gonna have to put it in pastebin
Apr 07 21:08:13 <pupnik> well is there one error message?
Apr 07 21:08:40 <pupnik> also you can get root without installing extra apps  http://wiki.maemo.org/Root_access
Apr 07 21:14:07 <gn00b> http://pastebin.com/juS1GG6r
Apr 07 21:15:39 <MohammadAG> ever dropped the device?
Apr 07 21:16:28 <gn00b> never
Apr 07 21:17:53 <gn00b> bought it used on nov 12th. worked fine up until now
Apr 07 21:19:35 <pupnik> gn00b: http://talk.maemo.org/showthread.php?t=70631
Apr 07 21:19:37 <pupnik> same error
Apr 07 21:20:28 <pupnik> https://bugs.maemo.org/show_bug.cgi?id=10305
Apr 07 21:20:30 <povbot`> Bug 10305: Application Manager in PR1.2 says not enough space even though there is.
Apr 07 21:20:32 <pupnik> bug discussion
Apr 07 21:20:46 <pupnik> you need to free up space gn00b
Apr 07 21:21:59 <gn00b> doesn't it free up space when i do a hard reset?
Apr 07 21:22:49 <pupnik> if you do a df, what partition has apt-worker: freespace(/home/)= 161075200 ?
Apr 07 21:23:06 <gn00b> all my music/pics are stored on my microsd card. the only thing on the phone is apps. and there really weren't that many
Apr 07 21:23:25 <pupnik> i recall your opt was quite full of stuff
Apr 07 21:24:16 <gn00b> df doesn
Apr 07 21:24:28 <gn00b> df doesn't show apt-worker anywhere
Apr 07 21:24:48 <pupnik> yea do you see anything with about 161 mb free?
Apr 07 21:25:53 <gn00b> * /home and /opt both have 157
Apr 07 21:27:07 <pupnik> in one thread a guy says he needs 200MB free
Apr 07 21:27:14 <pupnik> and that that solved it for him
Apr 07 21:28:11 <gn00b> let me get be clear on something...a hard reset is just like a clean install, correct?
Apr 07 21:33:23 <vi__> is it possible to install maemo to eMMC and swap to nand?
Apr 07 21:37:58 <vi__> is it possible to roll our own firmware?
Apr 07 21:39:09 <SpeedEvil> In principle, it's possible to dothat, I've never seen anyone report trying, with success though.
Apr 07 21:39:54 <MohammadAG> yeah sure, just ask Nokia Legal
Apr 07 21:40:51 <vi__> could I use rescue kernel with usb mode and dd my own image on and off the device?
Apr 07 21:40:56 <SpeedEvil> yes
Apr 07 21:41:21 <SpeedEvil> Orsomething similar - I don't know id trhat exacrt way will actually work
Apr 07 21:41:29 <MohammadAG> dd-ing the mtdblock is something I suggest you don't do
Apr 07 21:41:34 <MohammadAG> for your own good :P
Apr 07 21:41:45 <vi__> lol, ok broski
Apr 07 21:42:03 <MohammadAG> for the eMMC, it shouldn't matter
Apr 07 21:42:22 <MohammadAG> but for the rootfs, I had to recover with Ubuntu on the N900
Apr 07 21:42:26 <SpeedEvil> I know someone that's reflashed their OneNand from a meego boot - due to the USB connector being broken.
Apr 07 21:42:55 <vi__> if there was just some way i could rip the current state of my device to my own personal image
Apr 07 21:43:02 <vi__> thatd be sweet
Apr 07 21:43:11 <SpeedEvil> IIRC backupmenu will do that
Apr 07 21:43:21 <MohammadAG> yeah
Apr 07 21:43:26 <vi__> i was reading on the forum a post by egoshin
Apr 07 21:43:37 <vi__> that talked about moving the rootfs
Apr 07 21:43:55 <vi__>  you know the one i am referring to?
Apr 07 21:44:18 * MohammadAG remembers it
Apr 07 21:44:41 <pupnik> gn00b: not of everything
Apr 07 21:45:44 <pupnik> any of you remember what can cause HAM to give disk full error when there's apparently free space on rootfs?
Apr 07 21:46:24 <pupnik> i think his /opt and /home has only 157 MB free which someone reported as causing the error
Apr 07 21:46:37 <gn00b> could i connect to my pc via usb and delete everything? then do a hard reset? or would that brick it?
Apr 07 21:46:51 <pupnik> well there should be things in there you can delete
Apr 07 21:47:00 <pupnik> not necessarily 'everything'
Apr 07 21:47:39 <pupnik> for example i seem to have a bunch of old programs in /opt
Apr 07 21:48:15 <pupnik> if you look in there and see programs you don't have installed, those directories should be safe to remove
Apr 07 21:48:57 <BCMM> pupnik: does dpkg have a way to ask which package owns a specified file?
Apr 07 21:49:05 <BCMM> (like dpgk -L, only backwards)
Apr 07 21:49:08 <pupnik> yes BCMM and i keep forgetting it
Apr 07 21:49:15 <gn00b> ok...just got the truth from my son. he tried to install debian, but it didn't take.
Apr 07 21:49:20 <pupnik> erm maybe not dpkg itself
Apr 07 21:49:44 <pupnik> BCMM: http://en.wikipedia.org/wiki/Apt-file
Apr 07 21:49:45 <vi__> was egoshins script...sucessful?
Apr 07 21:50:03 <gn00b> pupnik, what can i go in and delete?
Apr 07 21:50:06 <vi__> can i install it in paralell with backup menu?
Apr 07 21:50:29 <pupnik> gn00b: to see what takes up space in a directory, i do this
Apr 07 21:50:38 <pupnik> du -s /opt/* |sort -n
Apr 07 21:50:42 <vi__> i.e. kb open = BU menu kb blosed = eMMC boot
Apr 07 21:51:23 <vi__> which directories have to be left on NAND?
Apr 07 21:51:37 <vi__> for this sucker ro keep booting?
Apr 07 21:51:48 <pupnik> heh, how do you do a | on n900 without the terminal hack
Apr 07 21:51:58 <pupnik> well you don't need the |sort -n part
Apr 07 21:52:07 <MohammadAG> blue + ctrl
Apr 07 21:52:12 <pupnik> ty
Apr 07 21:52:20 <vi__> pupnik: you alter your rx-51 file
Apr 07 21:52:31 <gn00b> 34588 /opt/lib     32352 /opt/usr     13020 /opt/var
Apr 07 21:53:05 <pupnik> ah that's not taking up the space then
Apr 07 21:53:14 <pupnik> then it's some junk in home, probably the debian install
Apr 07 21:53:58 <vi__> MohammadAG: does egoshin still hang out on here?
Apr 07 21:54:19 <MohammadAG> not sure if he even got on here
Apr 07 21:54:42 <vi__> so he is another qwerty12
Apr 07 21:54:53 <vi__> a genius that died before his time
Apr 07 21:55:15 <pupnik> qwerty12 died?
Apr 07 21:55:25 <vi__> lol
Apr 07 21:55:31 <vi__> figurativly
Apr 07 21:56:07 <pupnik> oh forum burn-out
Apr 07 21:56:58 <pupnik> gn00b: still with us?
Apr 07 21:57:10 <gn00b> yep
Apr 07 21:57:18 <gn00b> getting desperate
Apr 07 21:57:23 <pupnik> don't
Apr 07 21:57:39 <pupnik> du -s /home/user
Apr 07 21:57:47 <pupnik> du -s /home/user/*
Apr 07 21:57:48 <pupnik> sorry
Apr 07 21:58:07 <MohammadAG> why not use apt-get?
Apr 07 21:58:24 <pupnik> wants to use HAM
Apr 07 21:58:46 <gn00b> can't sudo
Apr 07 21:58:59 <pupnik> you can becomeroot in another way as well
Apr 07 21:59:08 <MohammadAG> flip the switch on R&D mode
Apr 07 21:59:19 <pupnik> but what should he do as root?
Apr 07 21:59:22 <gn00b> 1695420 /home/user/MyDocs     8 /home/user/Xorg.0.log
Apr 07 21:59:26 <pupnik> besides free up the space
Apr 07 21:59:56 <MohammadAG> check if /home is writable, if not, fsck it
Apr 07 21:59:56 <pupnik> don't paste everything it spets out
Apr 07 22:01:09 <pupnik> MohammadAG: do you think it's related to this bug? https://bugs.maemo.org/show_bug.cgi?id=10305
Apr 07 22:01:10 <povbot`> Bug 10305: Application Manager in PR1.2 says not enough space even though there is.
Apr 07 22:01:36 <pupnik> or do you think it might be related to only 157 MB free in /home?
Apr 07 22:02:08 <MohammadAG> I have no idea how HAM calculates this stuff
Apr 07 22:02:44 <pupnik> there was some discussion from someone who had HAM error-out when /home was < 200MB
Apr 07 22:02:51 <LjL> any hint about an N810 that's suddenly decided to reboot-loop, aside from reflashing? (its internal disk is likely full, i suspect that's at least part of the reason)
Apr 07 22:03:12 <pupnik> i mean > 200MB
Apr 07 22:05:43 <pupnik> gn00b:  du -s /home/user/MyDocs/*    might show you some huge directories as well that you can delete
Apr 07 22:06:02 <pupnik> heh, did for me
Apr 07 22:06:30 <gn00b> yep...there's the debian image files
Apr 07 22:07:35 <pupnik> it's cool your son wants to install debian though :|
Apr 07 22:08:06 <gn00b> yeah, but not on my n900
Apr 07 22:08:11 <pupnik> :)
Apr 07 22:09:51 <gn00b> "1 application successfully installed"
Apr 07 22:09:55 <gn00b> :-)
Apr 07 22:10:09 <pupnik> awesome
Apr 07 22:10:19 <gn00b> thanx for your help
Apr 07 22:10:27 <pupnik> you're very welcome
Apr 07 22:10:36 <gn00b> was already on ebay about to bid on another one
Apr 07 22:13:07 <pupnik> i'll post the workaround on that bugreport
Apr 07 22:13:19 <vi__> mm, i gotta get another n900. i am hoping they will go cheap when the n9/n950/whatever is released
Apr 07 22:13:27 <pupnik> me too
Apr 07 22:13:32 <vi__> unless it is shit
Apr 07 22:13:43 <vi__> in which case price may rise!
Apr 07 22:14:05 <vi__> has anyone compiled hdparm for n900 yet?
Apr 07 22:14:25 <vi__> oh yeah, debian binary
Apr 07 22:14:40 <pupnik> hmm can't log in
Apr 07 22:15:03 <vi__> how much faster is NAND than eMMC??
Apr 07 22:15:19 <SpeedEvil> vi__: It's hard to say.
Apr 07 22:15:27 <SpeedEvil> Raw read and write of large bloks is not.
Apr 07 22:15:35 <SpeedEvil> I have not done small random write tests.
Apr 07 22:15:37 <pupnik> gn00b: do you have an account on https://bugs.maemo.org/?
Apr 07 22:15:43 <gn00b> not yet
Apr 07 22:16:01 <gn00b> guess i should get one, huh?
Apr 07 22:16:15 <vi__> can the NAND be repartitioned?
Apr 07 22:17:20 <vi__> or should i say partitioned?
Apr 07 22:17:45 <SpeedEvil> It is partitioned int 5 (?) at the moment
Apr 07 22:17:48 <SpeedEvil> into
Apr 07 22:18:04 <SpeedEvil> It is_not_ a block device - it's important to realise this.
Apr 07 22:18:13 <SpeedEvil> Trying to treat it as a block device will screw you up.
Apr 07 22:18:19 <vi__> oh?
Apr 07 22:18:34 <pupnik> god damn it, i just changed my pass and i'm not getting on
Apr 07 22:18:41 <vi__> how do you mean?
Apr 07 22:18:50 <SpeedEvil> The raw nand is a mtd device - a raw flash - this has bad blocks.
Apr 07 22:18:59 <pupnik> gn00b: if you'd file that workaround to the bug i'd appreciate it
Apr 07 22:19:05 <SpeedEvil> If you treat it as a block device, you ignore bad blocks. This is bad.
Apr 07 22:19:05 <pupnik> i want to get back to my program
Apr 07 22:19:17 <vi__> right
Apr 07 22:19:49 <vi__> so how do nokia work around this issue?
Apr 07 22:19:56 <pupnik> https://bugs.maemo.org/show_bug.cgi?id=10305
Apr 07 22:19:57 <povbot`> Bug 10305: Application Manager in PR1.2 says not enough space even though there is.
Apr 07 22:20:13 <SpeedEvil> vi__: ubifs understands bad blocks
Apr 07 22:20:32 <SpeedEvil> cat /sys/class/ubi/ubi0/bad_peb_count
Apr 07 22:20:41 <SpeedEvil> This is bad blocks. Mine has one.
Apr 07 22:20:49 <vi__> so i cant really use it as a disk then?
Apr 07 22:20:52 <SpeedEvil> No.
Apr 07 22:21:07 <SpeedEvil> Not unless the firlesystem understands that it's flash.
Apr 07 22:21:17 <vi__> aaah
Apr 07 22:21:41 <SpeedEvil> It is really 2009 (in my case) seperate 128K seperately erasable storage units.
Apr 07 22:21:41 <vi__> ubifs is a file system?
Apr 07 22:21:44 <SpeedEvil> yes
Apr 07 22:22:33 <SpeedEvil> the emmc is too - but it has hardware to hide this.
Apr 07 22:22:36 <SpeedEvil> Also themicroSD
Apr 07 22:23:24 <vi__> could i put a swap file onto it then?
Apr 07 22:23:38 <vi__> using 128k pages
Apr 07 22:23:56 <vi__> or am I just barking up the wrong ass here?
Apr 07 22:26:01 <vi__> was egoshin right when he implied there is much cpu overhead involved in dealing with ubif?
Apr 07 22:26:20 <vi__> SpeedEvil: ?
Apr 07 22:27:43 <vi__> gtg, will have to ask you more questions tomorrow speeeeeeed
Apr 07 22:39:43 <gn00b> pupnik, can i pm you? about to file bug
Apr 07 22:43:11 <SpeedEvil> Vi: swap on 128K pages will basically do hideous things.
Apr 07 22:43:51 <SpeedEvil> Vi: Swap on flash is not really a solved well problem. However, using the native pages is about as bad an approach as you can imagine.
Apr 07 23:08:16 <chx> wind.ca!
Apr 07 23:08:19 <chx> zakkkkm: hi
Apr 07 23:08:29 <zakkkkm> chx: hey!
Apr 07 23:08:40 <chx> zakkkkm: i am also on Wind
Apr 07 23:08:43 <zakkkkm> chx: nice :)
Apr 07 23:08:52 <zakkkkm> i got the $40 unlimited back in december :)
Apr 07 23:08:55 <chx> zakkkkm: what are your experiences?
Apr 07 23:09:04 <zakkkkm> Had it since like August.. perfect
Apr 07 23:09:17 <zakkkkm> i did 38GB bw one month.. no throttle.
Apr 07 23:09:37 <zakkkkm> I live in Toronto, so there isnt signal problems.
Apr 07 23:09:43 <chx> ah Toronto
Apr 07 23:09:54 <chx> I live on the wild west ;)
Apr 07 23:09:57 <zakkkkm> heh
Apr 07 23:10:11 <zakkkkm> out of city have some signal problems on Wind it seems.
Apr 07 23:10:20 <zakkkkm> I love it though.
Apr 07 23:11:31 <chx> zakkkkm: i have a $45 winter wonder plan
Apr 07 23:13:34 <chx> zakkkkm: the miracle plans are better but this is not too bad either
Apr 07 23:20:32 <zakkkkm> chx: sorry, had to go down and get pizza
Apr 07 23:21:12 <chx> zakkkkm: my only grief is that the laptop plans cant be combined with voice so we are stuck with port block
Apr 07 23:21:32 <zakkkkm> i currently use my n900 to tether to my desktop
Apr 07 23:21:40 <zakkkkm> the laptop plans dont have port blocking?!
Apr 07 23:22:44 <chx> no they dont
Apr 07 23:22:51 <chx> it's a different AP
Apr 07 23:22:57 <zakkkkm> http://www2.windmobile.ca/en/Pages/data-plans.aspx is really cheap at $29 , with free data stick
Apr 07 23:22:59 <zakkkkm> thinking of it.
Apr 07 23:23:01 <chx> we use internet.windmobile.ca but it's broadband.windmobile.ca
Apr 07 23:23:14 <chx> yeah they discounted that plan
Apr 07 23:23:21 <zakkkkm> hate always having to tether
Apr 07 23:25:36 <zakkkkm> but yeah, wind is awesome
Apr 07 23:25:53 <chx> i found it a bit flaky
Apr 07 23:26:14 <zakkkkm> got the unlimited everything back in december for $40 a month.. dont have to worry about anything
Apr 07 23:26:19 <zakkkkm> prepaid too
Apr 07 23:27:05 <chx> i found fido a lot more stable for tether ( i have the iphone data plan , 6g for $30 not bad at all) but it's not AWS
Apr 07 23:27:10 <zakkkkm> thanks to Wind, i bought the N900 heh
Apr 07 23:27:23 <zakkkkm> which iphone?
Apr 07 23:27:24 <chx> yeah Wind enabled us that
Apr 07 23:27:31 <chx> just iphone data plan
Apr 07 23:27:35 <zakkkkm> you know you change to 1400 MTU?
Apr 07 23:27:42 <chx> i wouldnt touch the iphone with a ten feet barge pole
Apr 07 23:27:44 <zakkkkm> when tethering
Apr 07 23:27:49 <chx> no
Apr 07 23:27:51 <chx> tell me more
Apr 07 23:27:53 <zakkkkm> heh
Apr 07 23:27:57 <zakkkkm> thats why your network is flaky
Apr 07 23:28:04 <zakkkkm> it stops at like 29mb download eh?
Apr 07 23:28:07 <zakkkkm> cant even finish a download?
Apr 07 23:28:41 <chx> eh
Apr 07 23:28:46 <zakkkkm> right?
Apr 07 23:28:54 <chx> right :(
Apr 07 23:28:57 <zakkkkm> aha
Apr 07 23:29:02 <zakkkkm> you have to change your mtu to 1400
Apr 07 23:29:10 <zakkkkm> by default the packet size is wrong.
Apr 07 23:29:20 <zakkkkm> for some reason Wind changed it.. mobilicity did too
Apr 07 23:29:30 <chx> let me see
Apr 07 23:29:31 <zakkkkm> so eventually the download will end.. result of it being wrong..huge technical reason fo rit
Apr 07 23:29:47 <zakkkkm> simple change .. " ifconfig DEVICE(probably ppp0)  mtu 1400"
Apr 07 23:29:52 <zakkkkm> in my case its ifconfig ppp0 mtu 1400
Apr 07 23:30:03 <zakkkkm> and bam, perfect tether
Apr 07 23:30:11 <chx> Apr  7 00:49:21 veyron pppd[17455]: rcvd [LCP ConfReq id=0x0 <auth pap> <mru 1500> <asyncmap 0xa0000>]
Apr 07 23:30:18 <chx> that's wrong, right?
Apr 07 23:30:22 <zakkkkm> theres your problem :)
Apr 07 23:30:29 <zakkkkm> run ifconfig find the device
Apr 07 23:30:47 <chx> right, next time it'll be fine, i am on Shaw right now, the N900 is just fallback
Apr 07 23:30:50 <zakkkkm> if for instance its ppp0 ... do ifconfig ppp0 mtu 1400 ( as root, so probably sudo  )
Apr 07 23:30:57 <chx> i fixed /etc/ppp/options on the laptop
Apr 07 23:31:04 <chx> no need to change every time, there is an mtu
Apr 07 23:31:38 <zakkkkm> so yeah, should be perfect now :)
Apr 07 23:32:22 <zakkkkm> wow time goes fast, should be getting ready for work soon
**** ENDING LOGGING AT Fri Apr 08 02:59:58 2011