IRC log of #maemo-ssu for Wednesday, 2012-09-12

DocScrutinizer05I'd rather provide diff proper, than fix bb-b00:00
DocScrutinizer05straight forward, zero risk00:00
keriohrmpf, why isn't diffutils-gnu optified? :(00:01
DocScrutinizer05downside: probably some 50k more rootfs usage00:01
*** luf has quit IRC00:01
DocScrutinizer05compared to bb-b patch00:01
kerioDocScrutinizer05: 6131600:01
keriooh, ok :)00:01
* DocScrutinizer05 waves, for good now00:02
keriohell, why isn't the whole of /usr/bin/gnu optified?00:02
* kerio waves back00:02
keriobe safe out there, doc00:02
gregoadpkg uses diff. in src/configure.c: "sprintf(cmdbuf, DIFF ...", and in lib/dpkg.h: '#define DIFF "diff"'00:06
*** nox-- has joined #maemo-ssu00:06
merlin1991well sleeping bug it is then00:07
merlin1991DocScrutinizer05: the reason this is such a sleeping bug is, that ham just runs dpkg with flags so that the question what todo does not arrive00:08
*** luf has joined #maemo-ssu00:08
DocScrutinizer05:nod:00:08
*** nox-- is now known as nox-00:09
*** nox- has joined #maemo-ssu00:10
merlin1991and this is also the reason why the community-ssu enabler in stable runs this hack: https://gitorious.org/community-ssu/community-ssu-enabler/blobs/stable/install-community-ssu#line9600:11
merlin1991because otherwise users would never get an updated transistions.ini and probably would be on tmo "where is this option, my transistions.ini does not contain it"00:11
gregoa(which reminds me that a conffile in /usr/share/$foo is weird)00:13
kerioand in the midst of all this, the fact that versions for -thumb are fscked00:14
kerio:(00:14
gregoaseems like the contest for the longest-and-weirdest-versions-ever is still running00:15
merlin1991gregoa: it always is :D00:15
keriomerlin1991: update every package version with a new epoch plskthx00:16
kerio(don't do that)00:16
gregoamerlin1991: I know :) on each upgrade I'm staring at the screen when I see the "replacing foo-$version with foo-$nextversion" messages00:16
merlin1991gregoa: well it's debian policy to have a package version of $upstreamversion-$debianrevision+$whatevverpostfix00:17
merlin1991so we have $upstreamversion which actually is $upstreamversion-$somerevision+$nokiasuffix and then we add our stuff :D00:18
keriomerlin1991: we have some "-"s in some versions, and fmg botched the postfix, made it -thumb00:18
gregoamerlin1991: no, debian policy is $epoch:$upstreamversion-$debianrevision, where $epoch: is optional; everything after the last dash, if any, is the $debianrevision (which is also not needed but then it's a native package)00:20
keriogregoa: oh, the *last* dash?00:20
keriooh shit00:20
keriothat's the problem then00:20
merlin1991yeah doubledash tanks to silly nokia version00:21
gregoamerlin1991: this whole dance with -X+y+z is totally matter of taste and convention, and makes sense for nmus, derivates, whatever, but stacking stuff like +0m5+0cssu0 is just ... aehm ... ugly00:21
keriogregoa: community-thumb has a silly situation now00:21
gregoakerio: yes that's the problem indeed :)00:21
kerio1:2.2.142-21.1-thumb1 is considered greater than 1:2.2.142-2200:22
keriostill not sure why, though00:22
merlin1991yeah t vs 22 :/00:22
keriobut isn't 2.2.142-21 lesser than 2.2.142?00:22
merlin1991it is greater00:22
keriodammit00:22
keriofreemangordon: we (you) are thoroughly fucked00:22
merlin1991kerio: well the simplest fix would be if freemangordon fixes his -thumb packages with epoch and +thumb at the end00:23
keriomerlin1991: but then his packages won't fit between the -testing packages00:23
merlin1991ah yeah I'd have to epoch mine aswell00:23
merlin1991bloody mess00:23
keriowould you actually do it? :300:23
DocScrutinizer05https://bugs.maemo.org/show_bug.cgi?id=1266400:24
povbotBug 12664: dpkg depends on diff, pager. But neither diff nor pager available00:24
merlin1991nah fmg should just fix his versions and tell everyone who was on -thumb to reflash00:24
merlin1991it had huge warning signs on the thread after all00:24
keriomerlin1991: D:00:25
keriomadness00:25
keriobut good pont00:25
kerio*point00:25
kerioi'll just downgrade manually =D00:25
kerio"downgrade", i mean00:25
merlin1991well a simple reflash with --flash-only=rootfs and then a few lines of apt-get install should get your system to the laste state quickly00:25
merlin1991DocScrutinizer05: apt-get resp dpkg dafiq00:27
merlin1991s/dafiq/dafuq?/00:27
infobotmerlin1991 meant: DocScrutinizer05: apt-get resp dpkg dafuq?00:27
keriomerlin1991: less lines of apt-get install foo=version :)00:30
DocScrutinizer05that's the right way00:31
keriomerlin1991: this will screw up my manual/automatic marks though :c00:31
merlin1991hm booting uboot on hw revision 2204 also fails00:32
merlin1991it only works on 2101 for me00:32
DocScrutinizer05merlin1991: I don't really think "apt-get resp dpkg" is a big problem00:33
merlin1991DocScrutinizer05: my main problem was to understand what you mean by it00:33
keriobut on the other hand, they're fucked anyway, because most things are marked as manually installed00:33
lufmerlin1991: how you get the hw revision?00:35
merlin1991flasher-3.5 -i00:35
merlin1991luf: or in /proc/cpuinfo as "Revision"00:36
gregoakerio: I think bug #12645 can be closed, right?01:00
povbotBug https://bugs.maemo.org/12645 operator-name-cbs-widget does not display the operator name correctly01:00
kerioi think so01:01
kerioamiconn had that problem too01:01
merlin1991well untill you guys say it really is fixed I'm not going to fix it01:12
kerioi haven't encountered that problem again01:13
merlin1991s/fix it/ close it/01:13
infobotmerlin1991 meant: well untill you guys say it really is fixed I'm not going to  close it01:13
gregoait works for me and, IIRC, for kerio, so if amiconn can also confirm it weh should be good01:13
amiconnAs already mentioned I cannot confirm right now. Need to travel somewhere near the border so roaming kicks in01:14
amiconnI don't know when that will happen. Could be in a few weeks as well as next year01:19
*** luf has quit IRC01:21
*** M4rtinK has quit IRC02:05
*** DocScrutinizer05 has quit IRC02:23
*** DocScrutinizer05 has joined #maemo-ssu02:24
*** nox- has quit IRC02:59
*** arcean has quit IRC03:38
*** LaoLang_cool has joined #maemo-ssu05:16
*** amiconn has quit IRC05:38
*** amiconn_ has joined #maemo-ssu05:38
*** amiconn_ is now known as amiconn05:38
*** LaoLang_cool has quit IRC06:18
*** DocScrutinizer05 has quit IRC06:36
*** DocScrutinizer05 has joined #maemo-ssu06:36
*** luf has joined #maemo-ssu07:41
*** M13 has joined #maemo-ssu09:06
*** X-Fade has quit IRC09:17
*** X-Fade has joined #maemo-ssu09:17
freemangordonmerlin1991: what was that flashing for -thumb users?09:18
*** andre__ has joined #maemo-ssu09:38
*** andre__ has joined #maemo-ssu09:38
*** BCMM has joined #maemo-ssu12:07
keriofreemangordon: versions are fscked, fix them and tell people to reinstall stuff, or something12:12
*** Pali has joined #maemo-ssu12:38
keriofreemangordon: your other option is to drop the testing/thumb intertwining possibility and keep the versions as they are12:50
keriountil -testing updates the part before the first "-" or something12:50
*** arcean has joined #maemo-ssu12:54
*** LaoLang_cool has joined #maemo-ssu13:04
*** M4rtinK has joined #maemo-ssu13:10
*** lizardo has joined #maemo-ssu13:41
*** LaoLang_cool has quit IRC13:41
merlin1991bah harm got diff13:58
merlin1991after the initial bug report for diff came for diablo13:59
keriohaha14:01
*** LaoLang_cool has joined #maemo-ssu14:09
kerioi vote for optified difftools14:30
kerioalthough to be honest, busybox is supposed to provide "diff"14:31
kerioso idk14:31
*** LaoLang_cool has quit IRC15:00
DocScrutinizer05fsck, bug report for diablo diff?? how could I miss that?15:09
Paliwhy not to pack single diff binary into diff package?15:15
Paliless is already in SDK and extras-devel15:16
merlin1991DocScrutinizer05: even if we bring diff in over a different package, we should remove the provides line from the busybox control file15:28
merlin1991it's a plain lie15:28
DocScrutinizer05sure15:29
merlin1991actually that would be the best solution15:30
merlin1991remove the provides line from busybox and push a proper optified package to extras15:30
DocScrutinizer05all for it, full support here15:31
merlin1991hm there even is such a file around already15:31
merlin1991http://maemo.org/packages/package_instance/view/fremantle_extras_free_armel/diffutils-gnu/2.8.1-18maemo3/15:32
DocScrutinizer05yep, I know, I got that ;-D15:32
DocScrutinizer05just no "proper" diff in that it's not in /usr/bin but /usr/bin/gnu/15:32
merlin1991yeah the package needs a slight update so that it goes to /usr/bin15:33
DocScrutinizer05I.E. not in $PATH15:33
merlin1991I'll have look at the other binaries provided and see if that would pose any conflict15:33
merlin1991if not, I'd suggest pushing an updated version to extras that goes into /usr/bin15:34
merlin1991if there is a conflict we can drop the package in the cssu repository, this way we make sure users have the cleaned busybox installed already15:34
DocScrutinizer05how about a 'pkg* that creates alternative to /usr/bin/gnu/diff15:34
merlin1991you mean a pkg that basically depends on diffutils-gnu and just contains a bunch of symlinks?15:35
merlin1991actually scratch the idea that we ship this over extras, cssu should be self containing, if we fix a dpkg problem the fix should live in the cssu repo15:36
DocScrutinizer05yep15:36
DocScrutinizer05(for first line)15:37
DocScrutinizer05can we depend on diffutils-gnu?15:37
merlin1991nope because it is in extras15:39
merlin1991user possibly has extras disabled15:39
merlin1991and has every right todo so15:39
*** DocScrutinizer05 has quit IRC15:41
*** DocScrutinizer05 has joined #maemo-ssu15:41
DocScrutinizer05merlin1991: what's that bullshit that happened to marble with their last update, on cssu-t?15:45
DocScrutinizer05could any of the experts here pretty please have a look at it and help fix it?15:46
merlin1991I'll have a look after lunch15:46
DocScrutinizer05[2012-09-07 15:37:24] <DocScrutinizer05> FUUUCK! new marble has black-on-black-text issue in route-planning screen15:47
DocScrutinizer05[2012-09-07 15:38:06] <DocScrutinizer05> I thouhgt this been a Qt bug fixed like 18 months ago?! >:-(15:47
DocScrutinizer05[2012-09-08 03:00:30] <DocScrutinizer05> [2012-09-07 20:15:11] <Earthwings> DocScrutinizer05: works fine here without cssu15:47
DocScrutinizer05[2012-09-08 03:00:30] <DocScrutinizer05> [2012-09-07 20:15:36] <Earthwings> DocScrutinizer05: the reason could be that we subclass QLineEdit to include the inline clear icon15:47
merlin1991what was their channel again? (and the name of the dev)15:47
*** ZogG has quit IRC15:48
RST38hMarble is second of two programs guaranteed to bring n900 toits knees15:48
DocScrutinizer05[2012-09-08 03:00:30] <DocScrutinizer05> [2012-09-07 20:18:27] <Earthwings> DocScrutinizer05: according to http://talk.maemo.org/showpost.php?p=1256169&postcount=28015:48
merlin1991btw DocScrutinizer05where did you install it from?15:48
RST38hso it is rather irrelevant what font color it is using15:48
DocScrutinizer05#marble  earthwings15:48
DocScrutinizer05IroN900:~# apt-cache policy marble      Installed: 1.4.0-2            500 http://repository.maemo.org fremantle-1.3/free Packages15:50
*** ZogG has joined #maemo-ssu15:50
*** ZogG has joined #maemo-ssu15:50
*** LaoLang_cool has joined #maemo-ssu15:56
*** M13 has quit IRC16:00
*** luf has quit IRC16:12
*** dafox has joined #maemo-ssu16:19
* merlin1991 install marble on 3 devices16:28
merlin1991DocScrutinizer05:  what exactly do I want to reproduce now (and how)?16:29
DocScrutinizer05http://talk.maemo.org/showpost.php?p=1256169&postcount=28016:30
DocScrutinizer05I just ran update via ham to most recent marble version like last friday16:30
DocScrutinizer05I.E. *before* T516:30
merlin1991does the main screen fail you already?16:31
DocScrutinizer05wait, it "fixed itself" for me o.O16:33
DocScrutinizer05obviously last cssu-t update fixed it, since I didn't do a thing to marble itself16:33
DocScrutinizer05(sorry I forgot to check after T5 update)16:34
merlin1991ECANNOTREPRODUCE16:34
merlin1991does marble have turn by turn navigation?16:35
DocScrutinizer05(does main screen...) nope, it's actually been the route planning screen like shown in above tmo post that looked exactly same way for me16:35
DocScrutinizer05yes16:35
kerioneat16:35
keriohave you also tried modrana?16:35
merlin1991hm gotta try it one day, I always used mappero16:35
DocScrutinizer05nope16:35
DocScrutinizer05I got both16:36
DocScrutinizer05and I'm undecided which is better16:36
keriomodrana has voice instructions16:36
DocScrutinizer05mappero and marble too16:36
DocScrutinizer05marbe using on-board routing16:37
DocScrutinizer05monav16:37
DocScrutinizer05weird shit with this black-on-black text issue16:38
merlin1991what was the google supl dns for the agps again?16:42
keriosupl.google.com16:45
merlin1991thanks kerio16:45
keriomerlin1991: have you found a good way to fix the -thumb versions apart from reflashing?16:46
DocScrutinizer05merlin1991: could it be we got some fsckdup Qt build in T4.1, which got fixed in T5(.1) with >> qt-x11-maemo Improved performance of partial updates in raster window surface on X11.<< ?16:47
merlin1991DocScrutinizer05: highly unlikely the change in qt with T5 is only a single patch that got backported by fmg16:48
DocScrutinizer05I'm not suspecting some explicit fix, but a borked build in T416:49
merlin1991while that is possible I highly doubt it16:49
DocScrutinizer05you know, autobuilder :-/. Local builds :-//16:50
merlin1991last qt build we had was in 16.6 almost a year ago16:50
merlin1991did you have a problem with marble that long?16:50
DocScrutinizer05ooh I see16:50
DocScrutinizer05nope16:51
DocScrutinizer05just with last update that obviously changed sth regarding the broken widgets16:51
DocScrutinizer05[2012-09-12 14:47:45] <DocScrutinizer05> [2012-09-08 03:00:30] <DocScrutinizer05> [2012-09-07 20:15:36] <Earthwings> DocScrutinizer05: the reason could be that we subclass QLineEdit to include the inline clear icon16:51
DocScrutinizer05aiui that's a change introduced with last marble update 5 days ago, that also brought the problems16:52
merlin1991well the full change we did in qt is here: http://gitorious.org/community-ssu/qt-x11-maemo/commit/81fb472ad4c97d1ee43ec5550c8d8ce2aeb2c7b716:52
DocScrutinizer05so it might have awakened a bug that slept in qt since 16.616:52
merlin1991possibly, but it is also possible that some *evil* pkg messed with your system qt and only the cssu update of qt threw a clean version on top of the mess again16:53
merlin1991hm seems like the summer is over in vienna, currently there is one of the typical heavy autumn rains16:57
DocScrutinizer05quite possible16:58
DocScrutinizer05(summer) same here. But on sunday it's getting better again16:58
DocScrutinizer05merlin1991: (some evil pkg) though I had to think hard to recall anything I installed during last 12 or so months, that _might_ have touched qt. If anything comes to mind, that's all cmdline tools17:00
DocScrutinizer05but obviously that's still the best story we could come up with, so far17:01
DocScrutinizer05hey! I got a complete BM tarball of pre-T5, where rogue file has to be around still17:02
DocScrutinizer05would it help to pastebin a tar --list of that somewhere?17:04
merlin1991don't think so17:05
* DocScrutinizer05 idly wonders WTF does libqtm* on his IroN90017:23
* DocScrutinizer05 hates qtm17:23
kerioDocScrutinizer05: remove it!17:23
*** LaoLang_cool has quit IRC17:24
DocScrutinizer05~ # du -sh /var/lib/dpkg/17:26
DocScrutinizer0532.1M   /var/lib/dpkg17:26
DocScrutinizer05:-S17:26
DocScrutinizer05~ # df -h /var/lib/dpkg17:27
DocScrutinizer05Filesystem                Size      Used Available Use% Mounted on17:27
DocScrutinizer05rootfs                  227.9M    166.3M     57.4M  74% /17:27
merlin1991DocScrutinizer05: it's over 115 mb on my pc17:27
DocScrutinizer05I guess your PC's / is >>256MB ;-)17:27
merlin1991slightly ;)17:28
kerioDocScrutinizer05: i suppose it could be optified17:28
* DocScrutinizer05 still tries to chill out about that qt weirdness17:28
kerioi don't know if dpkg will complain, though17:28
DocScrutinizer05well, I think temporarily "optiying" /var/lib/dpkg is one of the "tricks" SSU (without C) does to allow system upgrades on tight free space in root17:30
DocScrutinizer05or it's been one of the suggested tricks user can do when "please use PC suite" due to too little free space in /17:31
keriowhy can't we permanently do that?17:42
kerioalthough it should be outside of /opt17:43
kerioit should be something like /home/varlibdpkg17:43
kerioDocScrutinizer05: it's not really in CSSU's best interests, but is it possible to properly mount /usr from the eMMC?17:45
DocScrutinizer05possible? sure17:46
DocScrutinizer05you 'just' have to move around a *lot* of pkgs from /usr/(s)bin to /(s)bin17:47
keriowhy?17:47
keriowhy do i need stuff in /usr to mount /usr?17:48
kerioi need mount17:48
DocScrutinizer05because all those are needed during early boot when /usr wouldn't be mounted yet17:48
keriowhich is in /sbin/mount17:48
DocScrutinizer05or you "fix" /sbin/preinit to mount /usr *really* early17:48
keriohell, it could be mounted right at the time when /proc is mounted17:49
DocScrutinizer05anyway you have to mess around with stuff that might get overridden with next SSU (I know will never come), and thus will render your system borked17:49
* kerio ponders about putting /usr in mmcblk0p317:50
DocScrutinizer05might be well worth the effort17:50
DocScrutinizer05:-)17:50
DocScrutinizer05beware of mmcblk0 vs mmcblk1 confusion in kernel, depending on whether uSD is inserted or not17:51
kerio...wat17:51
keriooh, you mean that they could be reversed?17:51
kerioew17:51
DocScrutinizer05it gets "fixed" only later during boot by renaming done by udev iirc17:51
DocScrutinizer05if uSD inserted, it's mmcblk0 during early boot17:52
DocScrutinizer05if no uSD, eMMC is mmcblk017:52
kerio# if user set up mmc booting, try to find the first mmc device17:52
kerio# which contains /sbin/init, it will be used as root17:52
kerio# also fallback to the mtdblock4 (our flash)17:52
keriowhat the FUCK is this17:53
DocScrutinizer05cruft17:53
*** toxaris has joined #maemo-ssu17:53
DocScrutinizer05or legacy, if you like it better17:54
DocScrutinizer05I bet it's from diablo times, when we still had initrd17:54
DocScrutinizer05where from is that? preinit?17:55
DocScrutinizer05:nod:17:56
kerioCHROOT="/usr/sbin/chroot ."17:57
keriooh for fuck's sake17:57
DocScrutinizer05there's a *lot* of cruft in preinit17:57
keriomeh, at least - as far as i know - there's no mention of /usr/ before that17:58
kerioso i just have to mount /usr before the chroot command is executed17:58
kerio...17:58
keriono, wait17:58
keriowait17:58
kerioBACK THE FUCK AWAY FROM A MINUTE17:58
kerio*FOR17:58
DocScrutinizer05the line above that is pivot_root which obviously is as moot as the whole boot() function17:58
keriowhy the hell does it chroot .17:59
keriogoddammit17:59
DocScrutinizer05pivot_root from initrd to real rootfs17:59
DocScrutinizer05diablo stuff17:59
DocScrutinizer05cruft17:59
keriothere's no pivot_root17:59
DocScrutinizer05                pivot_root . mnt/initfs18:00
DocScrutinizer05                CHROOT="/usr/sbin/chroot ."18:00
kerioi know, but where is it defined?18:00
kerioq18:00
DocScrutinizer05pivot_root is a command18:00
keriowhoops, sorry18:00
kerio q18:00
kerioGAAAAH goddammit trackpad18:00
kerioDocScrutinizer05: i can't find where it's defined though18:00
DocScrutinizer05in man pages?18:01
keriooh, it's a binary?18:01
kerioanyway, it's clearly not our case - i hope18:01
kerioif [ $root_mounted -ne 1 ]18:01
DocScrutinizer05clearly18:01
DocScrutinizer05since we got no initfs18:02
DocScrutinizer05no initrd18:02
kerioroot_fstype="jffs2"18:02
keriooh god why18:02
kerioCSSU feature request: a real init18:02
DocScrutinizer05LOL18:02
keriobut then, how does it exec $CHROOT /sbin/init?18:03
kerio...oh you sneaky fucks18:03
kerioif there's no CHROOT defined, then it's just exec /sbin/init18:03
DocScrutinizer05IroN900:~# init --help18:03
keriovery clever, nokia18:03
DocScrutinizer05Usage: init [OPTION]...18:03
DocScrutinizer05Process management daemon.18:03
keriothe thing i'm afraid of is that it will screw up bootmenu if i put an unconditional mount for /usr here in boot()18:05
DocScrutinizer05LOL @18:05
DocScrutinizer05        # let me see then, what thereat is and this mistery explore18:05
DocScrutinizer05        /sbin/fiasco-do-update18:05
DocScrutinizer05        exec $CHROOT /sbin/init $def_runlevel <dev/console >dev/console 2>&118:05
kerioT2S="${initfs_prefix}/usr/bin/text2screen" GODDAMMIT BOOTMENU YOU ARE NOT HELPING ME18:06
DocScrutinizer05WTF?!!18:07
DocScrutinizer05# Change ownership to allow locking for pulseaudio18:07
DocScrutinizer05chown pulse:pulse /sys/power/vdd2_lock18:07
DocScrutinizer05W*T*F?!!!18:07
kerioDocScrutinizer05: indeed18:08
*** BCMM has quit IRC18:08
DocScrutinizer05ls -l /sys/power/18:10
DocScrutinizer05o.O18:10
kerioDocScrutinizer05: yeah, i see18:13
kerioso?18:13
DocScrutinizer05cat /sys/power/idcode18:13
DocScrutinizer05hehe18:13
DocScrutinizer05IDCODE: 4b7ae02f18:14
DocScrutinizer05Production ID: 00000000 00000000 000000cc cafeb7ae18:14
DocScrutinizer05Die ID: 0c004012 040364fb 00000000 44f6000418:14
DocScrutinizer05kerio: so? just look what PA is messing around with18:14
kerioyeah, yeah, PA is awful18:15
kerioold story18:15
DocScrutinizer05this is plain rogue18:15
kerioanyway, Pali confirms that there's no clean way to mount /usr from a separate partition18:15
keriowhich makes me a really sad panda18:15
kerioeasier way to fix this: boot wholly from a eMMC partition18:16
DocScrutinizer05there is a clean way, but it's hard to make this fly in maemo, due to general borkedness18:16
keriobut it's still a fucking stupid way18:16
kerioDocScrutinizer05: the whole boot process expects a single device file to be the rootfs, and nothing else18:16
PaliDocScrutinizer05, I will include rx51_battery driver into kernel-power. Now I'm testing it and I do not see any problem18:17
DocScrutinizer05iirc they introduced optification with PR1.1, and we so much hoped and prayed that they will fix that shit in PR1.2, do a clean cut and get /usr partition properly mounted from eMMC, but NOOOOO18:18
Paliit only export & convert ADCIN channels to sysfs18:18
kerioDocScrutinizer05: i just don't get why we can't have a proper /etc/fstab that mounts shit18:18
DocScrutinizer05Pali: should be safe18:18
kerioand we need these weird scripts that run *before* init, to mount shit18:19
Paliwhat was before pr1.1?18:19
Palino optification?18:19
PaliI do not remember...18:19
DocScrutinizer05no optification, and pretty crammed rootfs18:20
merlin1991before pr1.1 my n900 rebooted all the time and locked itself thanks to that on boot18:20
DocScrutinizer05which caused 100% of rootfs eaten up every now and then -> bootloop18:21
DocScrutinizer05and they didn't get PR1.2 out via SSU due to missing free space on rootfs18:22
DocScrutinizer05so lcuk invented optification18:22
DocScrutinizer05(maybe it's been PR1.1, not 1.2, can't remember)18:23
DocScrutinizer05I *think* optification came with PR1.1 though18:23
merlin1991DocScrutinizer05: nah I just had sw reset all the time18:24
keriooh for fuck's sake, cal-tool is in /usr too18:24
merlin1991and there was this nasty bug, that after a certain amount of sw reset the boot stopped at the moving dots18:24
keriomerlin1991: why?18:25
kerio:o18:25
DocScrutinizer05yeah, there's been some monkey throwing dice whether a pkg goes to /sbin or /usr/sbin18:25
DocScrutinizer05hehehehe yeah, THAT was nasty18:25
merlin1991kerio: the only way to get the system up was to enable r&d mode18:25
keriowhy was there a counter?18:25
merlin1991ask nokia18:26
keriowas the counter removed?18:26
DocScrutinizer05there's aconter that stops on reaching 30 reboots, and iirc they even forgot to reset it *ever*18:26
DocScrutinizer05ROTFL18:26
DocScrutinizer05kerio: I'm not sure about that18:27
keriowas that bug removed at least?18:28
merlin1991DocScrutinizer05: a clean poweroff did reset the counter18:28
DocScrutinizer05maybe it's still lingering somewhere, but now they reset it on proper boot18:28
merlin1991so one could disable r&d untill the next time the reboot limit was reached :D18:28
DocScrutinizer05aah poweroff, ok18:28
DocScrutinizer05yeah, the 'early' times of maemo5, fun fun fun18:30
DocScrutinizer05and we still suffer from heritage18:30
merlin1991well the reboot fun was fixed in 1.1 afaik18:31
DocScrutinizer05yup18:31
DocScrutinizer05dunno how, though. Either they neutered or removed the whole thing, or they simply introduced proper reset of counter on power-off18:31
merlin1991well i guess the main reason I did never ever encounter it again, was that my device didn't reboot every 10 mins while using it after I had 1.1 installed18:32
DocScrutinizer05since I never again heard of that "houston, we got a problem" shit, I gather they neutered it completely18:33
merlin1991hm worth a try, artifially make a kernel panic 30 times in a row and see what happens on boot18:33
DocScrutinizer05:nod:18:33
* merlin1991 thinks of cat /dev/kmem18:33
DocScrutinizer05obvious and easy18:34
keriowat18:34
DocScrutinizer05just ponder how to recover18:34
keriowhy is that a problem?18:34
DocScrutinizer05kerio: reads 0x000018:34
keriooh, lol18:34
merlin1991DocScrutinizer05: well the enable r&d recovery path should still work18:35
* merlin1991 goes on a kernel panic quest18:35
DocScrutinizer05not sure, since kernel will still ooops18:35
DocScrutinizer05you better include that line somewhere next to slider-open query for bootmenu18:36
merlin1991err I do it by hand after each boot? :D18:36
DocScrutinizer05or hell, just replace BM by cat /dev/kmem18:36
DocScrutinizer05ooh18:36
DocScrutinizer05might work too18:36
merlin1991at least that resembles what I experienced more closely18:37
DocScrutinizer05hmm, fair enough18:37
DocScrutinizer05(... 5h later ...)18:40
merlin1991#418:40
DocScrutinizer05;-P18:40
DocScrutinizer05Zen or the art to reboot your device 50 times in a row18:40
*** toxaris has quit IRC18:41
merlin1991hehe, I drive a peugeot 106 Zen, I can do this :P18:42
DocScrutinizer05since we had a lot of bootloop reports but not a single "Housten, we got a problem" complaint, I'd guess it's nuked and gone18:42
DocScrutinizer05do you recall where that friggin counter lived? CAL?18:43
kerioanyway, maemo 5 is a fucking mess18:43
DocScrutinizer05prolly since otherwise a hotfix was easy18:43
merlin1991no idea, back then I was a (l)user and almost had my n900 sent to nokia care before I found tmo :D18:43
DocScrutinizer05hehe18:43
keriofreemangordon: *poke*18:44
DocScrutinizer05I found IRC months before I first visited tmo18:44
merlin1991back then my irc usage was limited to quakenet :D18:44
DocScrutinizer05actually I was on #maemo quite some time before I got N90018:44
DocScrutinizer05I visited #maemo first when lindi- reported GTA02 gets hot when charging prototype N900, so I tried to ask about N900 USB & charging, and crash&die and timeless were as nasty as they could get, so I avoided channel for another 3 months18:47
keriohahaha18:47
DocScrutinizer05http://mg.pov.lt/maemo-irclog/%23maemo.2009-09-13.log.html#t2009-09-13T18:46:0818:54
freemangordonkerio: what about /me18:54
keriofreemangordon: fix versions plskthx18:55
*** luf has joined #maemo-ssu18:55
freemangordonkerio: just for the sake of fixing?18:55
keriobut, you know, no pressure :)18:55
keriofreemangordon: for the sake of the new hildon-desktop!18:55
freemangordonkerio: there is no new hildon-desktop. AFAIK18:55
kerio1:2.2.142-2218:56
kerioinstead of 1:2.2.142-21.118:56
freemangordonand why do you think those differ?18:56
kerioit adds the proper blurless thing18:56
freemangordonis that in -testing?18:56
freemangordonOMG18:56
freemangordonmerlin1991: why?18:56
kerioTHERE'S BEEN A RELEASE BRAH18:56
kerioand it includes new magical things!18:57
merlin1991freemangordon: why not?18:57
kerio!changelog18:57
kerio^ we need a bot for that18:57
freemangordonkerio: keep quiet for a while please, I know what is that as it was me, Estel and arcean to invent it and arcean to code it18:57
DocScrutinizer05http://mg.pov.lt/maemo-irclog/%23maemo.2009-09-13.log.html#t2009-09-13T18:55:1718:57
freemangordonmerlin1991: it is not tested at all18:57
freemangordonbesides arcean noone has ever run it18:58
merlin1991when I asked arcean if I should include his mg on hildon-desktop he said something along the lines of "yes if you find the time"18:58
freemangordonand it is missing a major functionality, from the consistency POV18:58
merlin1991which would be?18:58
freemangordonif you press X on application (outside tasknav) and got to tasknav, thumbnail won;t be blurred18:59
freemangordonanyway, it is too late now :)18:59
keriodoes it work fine with blurless?18:59
*** NIN101 has joined #maemo-ssu18:59
freemangordonkerio: NFC, how am I suppsed to know19:00
merlin1991freemangordon: yeah I remember that it missed that19:00
freemangordonmerlin1991: I am not saying it is buggy, just not tested at all19:00
freemangordonkerio: you'll have your .22-thumb0 soon19:01
merlin1991blame arcean ;)19:01
freemangordonwhy him, MR was sitting unmerged for a month or so :P19:01
merlin1991well he's our hildon-desktop vodoo priest and he said go ahead :D19:02
keriofreemangordon: what about the fact that about half of the -thumb package don't "wedge" themselves in the testing versions?19:02
merlin1991I gotta trust some people :D19:02
freemangordonwell, ok then19:02
freemangordon:D19:02
freemangordonkerio: rephrase please, I am not native Wnglish speaker19:02
freemangordon*english19:02
kerioor, alternatively, what about the fact that about half of the -thumb packages do wedge themselves in the testing versions?19:02
keriofreemangordon: 1:2.2.142-21.1-thumb1 is a higher version than 1:2.2.142-2219:02
DocScrutinizer05merlin1991: seems we need to ressurrect the good old inclusion triage meetings here in #maemo-ssu ;-)19:03
freemangordonkerio: ok, you'll have .22 in next -thumb update19:03
freemangordonDocScrutinizer05: we MUST ;)19:03
keriofreemangordon: it's not just a problem of hildon-desktop19:03
freemangordonkerio: I know, so what?19:03
freemangordonyou'll have to wait for the next -thumb update19:04
kerioi'm just wondering if you'll find a way to make things consistent, in a direction or the other19:04
freemangordonwe(me and merlin1991) screwed the versioning system, btu I don;t see any problem because of that19:04
freemangordonand as DocScrutinizer05 said why you want to downgrade?19:05
DocScrutinizer05after all you shouldn't get updates from testing for any package coming from thumb19:05
keriothe way i see it, you either make all the versions higher than the non-thumb ones, or you make them all fit in the middle19:05
freemangordonkerio: they are consistent, though not in the direction we tought it will be :D19:06
keriono they're not, some packages can be upgraded right now19:06
freemangordonkerio: using apt-get?19:06
kerioor HAM with "display all packages", i suppose19:06
kerioyes, there's no problem for lusers19:06
freemangordonhmm, rm -fr / is much easier way if you want to screw your device19:06
keriobut it's a dependency mess :(19:06
freemangordonkerio: I will say that again (and I'll never give up), -pr thung is for (l)users and they are supposed to use HAM19:07
freemangordonperiod19:07
keriofreemangordon: it fucks with the versioning => it fucks with the dependencies19:07
keriosay, modest19:08
keriomy installed version is 3.90.7-11.1-thumb3, the current testing version is 3.90.7-11.219:08
kerioif a package depends on >=3.90.7-11.2, then i can install it19:08
freemangordonkerio: I know, but as long as there are no problems I don;t see why should I change it. Of course I will try to fix at least some packages with the netxt -thumb, but cannot promise anything19:08
kerioeven though i shouldn't, because i only have 11.1, but thumbified19:08
DocScrutinizer05kerio: hell! disable cssu-testing repo, get updates from cssu-thumb only!19:09
kerioDocScrutinizer05: not every package is thumbified19:09
kerioAND IT STILL SCREWS WITH DEPENDENCIES19:09
freemangordonDocScrutinizer05: won't work, -thumb needs -testing19:09
freemangordononly a subset of packages are in -thumb repo, there is no reason to put translations there (for example)19:10
DocScrutinizer05then all the packages that aren't thumbified should still be available in -thumb and point to -testing19:10
freemangordonkerio: you have a point, will see what can be done. happy?19:10
kerioi'm not entirely happy, because i'm afraid there's nothing that can be done :(19:11
freemangordonmerlin1991: AIUI we screwed versioning by choosing -thumbN instead of +thumbN, correct?19:11
keriofreemangordon: yeah, using +thumbN means you don't have this problem - and it still works as intended19:11
keriofreemangordon: but people who have the weird versions installed right now will have to "downgrade" to the new packages with +thumb manually19:12
DocScrutinizer05kerio: so what? reflash!19:12
kerioDocScrutinizer05: not me, i know how to fix those19:13
keriofreemangordon: unless merlin1991 decides to make DocScrutinizer05 angry and re-epoch the affected packages19:14
freemangordonkerio: bad idea19:14
freemangordoni'll find another way, eve if it is by using apt-get19:14
kerioyeah, i know19:14
DocScrutinizer05kerio: see? you're suggesting merlin1991will make *everybody* angry ;-P19:15
kerioapt-get install foo=bar+thumbX will work fine19:15
freemangordonkerio: I think there is a swith to apt-get to force downgrade, just need to find it19:15
kerioooh, force downgrade to the current higher version in the repos?19:15
kerioi was looking for it, but i can't find it19:15
* DocScrutinizer05 suggests a tiny bit of shell filter magic19:16
keriofreemangordon: ooh, a priority higher than 1000 will make apt willfully downgrade19:16
DocScrutinizer05incl some sed/awk19:16
kerioDocScrutinizer05: nah, we found it19:16
freemangordonkerio: elaborate please19:17
keriofreemangordon: give correct versions to the packages19:17
freemangordonok19:17
DocScrutinizer05freemangordon: I guess he means repo prio19:17
keriothen give the repository a priority of 100119:17
kerioand then apt-get upgrade will work19:17
DocScrutinizer05:nod:19:17
kerioin theory19:17
Palikp52 for testing: http://talk.maemo.org/showpost.php?p=1265427&postcount=14119:17
kerioPali: ^_^19:17
freemangordonhmm, how to change repo priority?19:17
freemangordonPali: great19:18
DocScrutinizer05it's in the config file at very least19:18
DocScrutinizer05at-cache policy lists repo prio19:18
DocScrutinizer05currently 50019:18
freemangordonhmm, I would rather make a shell script containing apt-get install p1=v1 p2=v2 etc.19:18
freemangordonto be downloaded from TMO19:19
* DocScrutinizer05 marks another day in red on his calendar "kerio been helpful" ;)19:19
freemangordonafter all there are about maybe 20-40 users of -thumb and they are supposed to be able to at leas run a shel script19:19
Palikp52 has updated rx51_battery driver19:19
PaliI will update hald-addon bme and other bme code to use it19:20
freemangordonPali: and please revert *thermalobject* to use libbmeipc and fix libbmeipc to read the temp19:20
keriofreemangordon: downgrading is usually kinda dangerous, but this is only a formal downgrading - the packages will either be the same or be an actual upgrade19:20
freemangordonkerio: :nod:19:20
keriofreemangordon: btw, which release are these packages in?19:21
freemangordon"these packages"?19:21
freemangordonwhich ones19:21
Palifreemangordon, ok we do not have to usee patched dsme-thermal package19:21
freemangordonyep19:21
keriothe thing is, i don't know how to identify the repo :s19:22
DocScrutinizer05Pali: please ensure your bat temperature readings are meaningful/defined for ambient temp range -50°C to +90°C at least19:22
freemangordon:nod:19:23
freemangordonPali: did you implement the polinomial approximation?19:23
Palinot yet19:23
freemangordonyou should do it, upper limit of 45 deg is too low19:23
keriofreemangordon: oh my fucking god19:23
DocScrutinizer05then you should extend your LUT by extrapolated values19:23
kerioi have something glorious in mind19:23
freemangordoniirc shutdown temp is 60 deg or something19:23
kerioPackage: *\nPin: version *-thumb*\nPin-Priority: -1019:24
kerioor something19:24
Paliwhere to implement turning charging off?19:25
Palior is this somewhere in maemo implemented?19:25
freemangordonPali: hmm, good wuestion19:25
Palimce, dsme, bme, ...19:25
freemangordonmaybe we'll need some "bmed"19:25
DocScrutinizer05L3_4 mentions -40°C and +85°C as thresholds for certain operation conditions, so we need at least exceed that range19:25
PaliI think hald-addon-bme should be ok19:25
Palihere is also implemented MCE led pattern signals19:26
freemangordonPali: well, could be, but have in mind we should take care for low battery shutdown too19:26
DocScrutinizer05that's why I requested -50 to +90 at least19:26
PaliI do not want other daemon for charging and battery19:26
keriofreemangordon: it could end up being just a file in /etc/apt/preferences.d/19:26
kerioi really hope this works :D19:26
Palifreemangordon, lowbat is handled by systemui or hildon plugin19:26
freemangordonkerio: no, i've already decided on that, will make a script with correct apt-get command and will attach it on TMO19:27
Palibme only send signal battery empty and somebody turn off phone19:27
keriono, seriously, this shit will work a lot better19:27
freemangordonPali: ok19:27
kerioif it works19:27
keriocurrently testing it :)19:27
Palihald-addon-bme already sending all signals like bme19:27
freemangordonPali: yeah, I know, great job :)19:27
Palihttps://garage.maemo.org/forum/forum.php?thread_id=6342&forum_id=581219:28
* DocScrutinizer05 supports kerio's approach19:28
DocScrutinizer05sounds clean19:28
Paliis really dsme handling turn off charging?19:28
arceanguys, the latest h-d fixes blurless desaturation and adds thumbnail's desaturation (disabled by default)19:29
arceanso no worries ;P19:29
DocScrutinizer05Pali: nah, this is a basic security function that *has* to go to lowest level driver possible, directly19:29
freemangordonDocScrutinizer05: how am I supposed to edit a file in /etc/apt/preferences.d/ ?19:29
keriohrmpf, i can't seem to only filter for -thumb19:30
freemangordonarcean: the point is it is not tested by anyone but you :P19:30
PaliDocScrutinizer05, but bq2415x cannot read any temperature...19:30
DocScrutinizer05freemangordon: hrrrmpf, with a sed script?19:30
keriojust append those three lines19:30
freemangordonfrom when?19:30
freemangordonpreinst?19:30
keriofreemangordon: from TMO19:30
freemangordonfrom where?19:30
freemangordonwell, what is the difference then?19:31
DocScrutinizer05Pali: that's the layering problem I've been talking about since somebody suggested implementing charging completely in kernel19:31
freemangordonit is much easier for me to just make the correct apt-get command than playing with sed, awk, and who knows what else19:31
arceanfreemangordon: it was tested on a couple of devices, but yes - it's not available in -devel branch :)19:31
freemangordonarcean: is it?19:32
freemangordonaah, sorry, i thought it was tested on your device only19:32
arceanin devel branch? not yet :)19:32
DocScrutinizer05Pali: anyway going "hmm. *somebody* will take care about over/under-temperature on charging2 won't fly19:32
freemangordonis it tested on couple of devices  :)19:32
DocScrutinizer05Pali: bq24150 also can't read ENUM and D+/- short19:33
freemangordonDocScrutinizer05: i *think* it is bme who does that right now.19:33
*** M13 has joined #maemo-ssu19:33
freemangordon(temp protection)19:33
DocScrutinizer05:nod:19:33
DocScrutinizer05together with dsme afaik19:33
Paliand why dsme needs temp?19:33
freemangordongood question19:33
DocScrutinizer05there's a friggin lot of thermal management you never heard or thought of19:34
freemangordonPali: well, you can return 90 deg from libbmeipc (hardcoded) and we'll see :D19:34
Paliok :-)19:35
DocScrutinizer05good plan19:35
freemangordonthat way we can find shutdown temp too19:35
Paliwe can also hack twlmadc driver which read some num from /sys and then send fake raw temp value19:35
freemangordonincreasing return value every 20 seconds or on every readtemp() call19:35
DocScrutinizer05look L3_4 for shutdown levels19:35
Paliand we will see what will bme do :-)19:35
freemangordonPali: no, without BME19:36
freemangordonthat way we'll check who does thermal shutdown19:36
PaliI think we can get more values into table19:36
freemangordonif nothing happens on 90 deg, then it is not dsme to shutdown on overtemp19:36
DocScrutinizer05you knew cmt is supposed to not do any calls on >60°C (or whatever), *unless* it's a 911 call?19:36
DocScrutinizer05same for *cell* temperatures <-25°C19:37
freemangordonDocScrutinizer05: sure, but all of them use libbmeipc19:37
DocScrutinizer05just saying19:37
freemangordon(cmr,dsme,puleaudio)19:37
freemangordons/cmr/cmt/19:37
infobotfreemangordon meant: (cmt,dsme,puleaudio)19:37
freemangordonthe fuck :D19:38
DocScrutinizer05fsckng PA19:38
DocScrutinizer05I wonder what the hell this shit is about19:38
DocScrutinizer05might just be warning jingle19:38
freemangordonno idea, but seems like yet another protection19:38
freemangordoncould be19:38
DocScrutinizer05might be PA using this for reducing heat from amp when due, by lowering volume19:39
freemangordon:nod:19:39
DocScrutinizer05might even be calibration for AEC19:39
keriooh for fuck's sake pulseaudio19:39
kerioyou're just an audio daemon, stop being silly!19:39
freemangordonanyway, by replacing the underlying lib we'll have all this functionality in place19:39
DocScrutinizer05(you know, speed of sound in air differs for different air temperature)19:40
freemangordonwe only need to find out if BME does anything but reading the pemt19:40
DocScrutinizer05yup19:40
DocScrutinizer05agree19:40
freemangordonhmm, we can ask someone from meego/nemo gurus if they have any clue on that19:41
DocScrutinizer05obviously it does some averaging19:41
freemangordonI hope it is not under NDA :D19:41
DocScrutinizer05they don't19:41
DocScrutinizer05I already asked19:41
freemangordondamn19:41
DocScrutinizer05they use custom built bme courtesy Nokia19:42
DocScrutinizer05of course blob19:42
DocScrutinizer05so basically they do same as we do19:42
freemangordonDocScrutinizer05: we only need to know if BME reacts in any way on temp changes19:42
freemangordonnothing more19:42
DocScrutinizer05just they don't worry about any functions inside bme they might miss19:42
freemangordonhmm, actually Pali is right, we can fake BME from twl adcin driver19:43
DocScrutinizer05as mentioned above, it seems bme does some post-processing of raw ADCIN values19:43
DocScrutinizer05averaging, leaving out extremes, deglitching, PID, whatever19:43
freemangordonDocScrutinizer05: but we don;t care for that PP, the important part is device protection and if there is any implemented in BME or everything is in dsme19:44
DocScrutinizer05we should adopt work thesis that temp can't change by more than 1°C per maybe 10 seconds19:44
freemangordonDocScrutinizer05: implementing PID should not be that hard19:45
freemangordon:)19:45
DocScrutinizer05sure it's not hard to implement, it's my "jb". Calibrating it though...19:45
freemangordona first order lowpass filter should do the job too19:45
DocScrutinizer05:nod:19:45
keriofreemangordon: ok, i have a good way now19:46
keriokinda19:46
freemangordonand I think i have some matlab source code there, first order digital filter19:46
keriostill trying to figure out how to pin by version19:46
DocScrutinizer05or just pick most frequent value out of a sample window of maybe 1619:46
DocScrutinizer05no averaging19:46
freemangordonkerio: forget about it, I told you how I will do it19:46
freemangordonDocScrutinizer05: median filter?19:47
DocScrutinizer05kinda19:47
freemangordonshould be ok19:47
keriook, fine, i'm just trying to figure out how *I* will do it19:47
freemangordonDocScrutinizer05: 5 samples median filter should be pretty ok19:47
DocScrutinizer05nope, from my test row I did it seems we need at very least 819:48
merlin1991fu you debian policy: The package management system will break the version number apart at the last hyphen in the string (if there is one) to determine the upstream_version and debian_revision. The absence of a debian_revision is equivalent to a debian_revision of 0.19:48
kerioDocScrutinizer05: The final conclusion has to be that not only the official documentation of apt, in regard to pinning, sucks— but the whole design and algorithm that determines the candidate version is borked. It is counter-intuitive, ambiguous and not as flexible as it could be.19:48
kerioLOL19:48
*** M13 has quit IRC19:49
merlin1991freemangordon: that's why ie hildon-desktop fails, because in the -testing case the "revision" is after the -22 in h-d for -thumb it is only -thumb19:49
DocScrutinizer05sounds about right19:49
merlin1991so the upstream version is *higher*19:49
DocScrutinizer05so freemangordon's point gets stronger19:49
keriomerlin1991: does cssu always use + like it should?19:49
kerioDocScrutinizer05: it can still be done19:49
merlin1991kerio: I hope so19:50
merlin1991though afaik we have a few -cssu packages19:50
* merlin1991 goes investigate19:50
kerioyou give 2000 priority to every package from maemo.merlin1991.at19:50
kerioand it'll prompt to up/downgrade to the latest version from there19:50
kerioregardless of installed version19:50
freemangordonmerlin1991: i will take care of -thumb users once we have a clear ploicy WTF needs to be done :D19:50
DocScrutinizer05kerio: *nod*19:51
kerionow, if only -devel wasn't on the same domain too... :P19:51
merlin1991kerio: that could be changed19:52
* DocScrutinizer05 calls it a day where real work been done in cssu r&d19:52
freemangordonmerlin1991: if I change -thumb with .thumb?19:52
keriofreemangordon: use +thumb19:53
kerioit's the "standard"19:53
freemangordonkerio: ok19:53
merlin1991freemangordon:  It may contain only alphanumerics and the characters + . ~ (plus, full stop, tilde)19:53
keriomuch like cssu uses +0cssuX19:53
merlin1991also somebody nuke busybox for me please19:54
keriohehe19:54
kerio~nuke busybox19:54
* infobot prepares 100 missle silos, and targets them at busybox ... B☢☢M!19:54
freemangordonmerlin1991: ok, i'll change -thumbN with +MthumbN, ok?19:54
* DocScrutinizer05 goes placing real thermometer into his freezer, to hopefully find it's ~-25°C in there19:54
keriohm, what should the M be?19:54
freemangordonDocScrutinizer05: -1819:54
merlin1991+thumnN should be enough19:54
freemangordonmerlin1991: well, it will be, but M is in case I screw up something19:55
freemangordonok?19:55
keriohmm, i saw some mojbake from infobot19:55
merlin1991freemangordon: dosfstool version is foobar aswell19:55
freemangordonwhy?19:56
freemangordon(it is not in -thumb still :) )19:56
keriomerlin1991: epoch! epoch!19:56
merlin1991it's the only package with -cssu compared to all the others19:56
merlin1991:D19:56
freemangordonkerio: sto trolling :P19:56
keriofreemangordon: no, he should use an epoch19:56
keriothat's like the official debian policy for version mismatches19:57
kerioif he started using +0cssu0 there it would be the same version mismatch as the thumb packages19:57
freemangordonmerlin1991: no choice for dosfstools19:57
freemangordon:D19:57
merlin1991yeah dosfstools is doomed19:57
DocScrutinizer05lol19:57
merlin1991kerio: actually dofstools could swap to +cssu19:57
* DocScrutinizer05 notices kerio's learning how to NOT troll19:57
kerioDocScrutinizer05: i can still troll if you want :319:58
DocScrutinizer05we are all sure about that19:58
kerio(and srsly, you're a bit too trigger-happy with your op powers)19:58
merlin1991because in dosfstools case there is no "revision" but the cssu one so swapping to + would increase the "upstream" version which supercedes any "debian revision"19:58
kerio(or with the threatening of your op powers)19:58
keriomerlin1991: neat19:58
freemangordonyeah19:59
kerioif there's no other "-" then version dependencies will work fine19:59
DocScrutinizer05kerio: I usually just explain my moods19:59
kerioit's still something that you have to remember for the next release, though19:59
DocScrutinizer05doesn't imply I follow them indiscriminately19:59
DocScrutinizer05~lart debian20:00
* infobot DoSes debian20:00
keriofreemangordon: i still think the pinning is better than the apt-get line20:00
kerioespecially because not everyone has every package from -thumb installed20:01
DocScrutinizer05that's why I suggested shell filter magic20:01
kerioi wonder if HAM will respect priorities20:02
Palihald-addon-bme code updated20:02
Palibut not tested yet20:02
DocScrutinizer05hardcoded script with apt-get lines for sure won't pan out20:02
freemangordonkerio: WUT?20:02
keriofreemangordon: I SAID NOT EVERYONE HAS EVERY PACKAGE FROM -THUMB INSTALLED20:02
freemangordonhow they dont have every package from thumb?20:02
DocScrutinizer05don't shout!20:02
freemangordonhear, her20:02
keriofreemangordon: for instance, busybox-power :)20:03
freemangordonkerio: it is not -mp dependency ;)20:03
freemangordonthus not a part of CSSU20:03
DocScrutinizer05ooh, for mp dependencies it's for sure hardcoded fixed set20:04
freemangordonDocScrutinizer05: that is why a script with apt-get is better than any shell script magic ;)20:04
keriofreemangordon: not every package in community-thumb is referred to by the metapackage :)20:06
freemangordonkerio: it is. (at least should be)20:06
keriobusybox-power20:06
freemangordonif you find something missng, i'll be glad to fix it20:06
keriothe weird rotation lock applet20:06
freemangordonkerio: stop that20:06
keriono, you stop that!20:07
kerio:320:07
freemangordoncssu-thumb consist of mp-fremantle-community-pr and its dependencies20:07
merlin1991#maemo-ssu, the place where kerio lives on the edge :D20:07
freemangordonthere is fennec in thumb repo, should we consider it a part of CSSU?20:07
freemangordonmeh20:07
keriofreemangordon: you should still kinda rerelease busybox-power with +thumb0 instead of -thumb020:08
keriobut there's no other "-" in the version, so at least there's that20:08
freemangordonkerio: ask iDont for that20:08
freemangordonhe is the maintainer, I only do dpkg-buildbackage20:08
keriooh crap, fennec has that "problem" too20:08
freemangordonkerio: fennec is in no way related you yoyr edgy place20:08
freemangordons/you/yoyr/to your20:09
kerioi'm just trying not to get into a dependency hell20:09
freemangordonyou are already there :P20:09
keriobecause those version numbers are just /wrong/20:09
keriolike, dead wrong20:09
kerioand you can't reliably depend on package(>=version) if some versions are wrong20:09
freemangordonthat is why I depend on =20:10
freemangordoncheck you -mp thingie20:10
freemangordon*your20:10
keriofreemangordon: yeah, about that...20:10
kerioyour exact dependencies are kinda bad20:11
merlin1991freemangordon: well at least the -thumb repo fixes itself, because I only keeps 1 package in the index20:11
kerioyeah, the repo is fine once you change the versions20:11
kerioand the mirrors inside n900s will be fine once they pin the repo with priority 2000 and do an upgrade :)20:11
freemangordonmerlin1991: you should cleanup it before the new -thumb release. otherwise i wont be able to upload the packages20:12
merlin1991freemangordon: about that, tell me when you're ready to upload and I'll wipe it?20:12
freemangordonsure20:12
freemangordoni will start right now, but it will take a while20:13
freemangordona 2-3 days maybe, I am short on spare time these days20:13
*** dafox has quit IRC20:16
keriofreemangordon: btw, "curl" is also thumb and not in the metapackage20:17
freemangordonkerio: btw you don;t have the latest -thumb20:18
keriomind you20:18
freemangordonhttps://gitorious.org/community-ssu/mp-fremantle-community-pr/commit/302f4af52fb4c872de8d67bed603950350bbc59620:19
kerio"curl" is also thumb and **should not** be in the metapackage20:19
freemangordonwhy?20:19
keriobecause it is not necessary20:19
freemangordonwhere it is supposed to be?20:19
freemangordonyeah, sure20:19
keriono, seriously20:19
kerioremove it - you won't get any dependency problem20:20
keriolibcurl3-0 is the thing stuff uses20:20
freemangordonkerio: it fixes several bugs20:20
kerioi meant "try to uninstall it from your n900"20:20
kerioneat, iphone 5 announcement20:21
kerio112 grams, 326 PPI display, 4 inches screen, 1136x64020:21
kerioANOTHER ROW OF ICONS20:22
keriothis. changes. everything.20:22
DocScrutinizer05suuure20:25
kerio"It certainly looks on-par with current console gameplay graphics."20:27
keriothe current consoles are 7 years old!20:27
DocScrutinizer05who the flying fuck wants CONSOLE GAMEPLAY GRAPHICS on his phone? Morons!20:33
RST38hyeah, morons, all of them20:34
keriobut... i like super mario :(20:34
DocScrutinizer05there's an annoying tendency to turn each and every gadget into a next-gen gameboy nowadays20:34
RST38hThat is pretty low on the list of today's annoying tendencies20:35
DocScrutinizer05yep, but still20:35
kerioyou say that like each and every gadget *shouldn't be* a next-gen gameboy20:35
RST38hI would even say, it is a consequence20:35
DocScrutinizer05of general cerebral decay?20:35
kerioooh, the new iphone has three microphones20:36
kerioi wonder what pulseaudio would do with that20:36
* DocScrutinizer05 grows a third ear20:36
RST38hAttention spans are dangerously low now and it looks like people never quite grow up20:36
RST38hSo, yeah, what you just said20:36
kerioback in my day, we had to pay attention to videogames!20:36
RST38hNow, they had to start making "casual" video games for these morons20:37
RST38hNot enough dedication to play the real thing20:37
kerioooh, the data connector is reversible20:38
kerioand i bet the port doesn't break if you look at it wrong20:39
DocScrutinizer05*burp*20:39
* kerio glances at the µusb port in his n90020:39
kerio"For the record, I'm weeping at the thought of how many adapters I now have to buy..."20:40
kerioJUST DON'T BUY THE NEW IPHONE YOU DOLT20:40
kerioit's not mandatory20:40
DocScrutinizer05kerio: take it to #maemo please20:42
keriok20:42
keriofreemangordon: btw, why do you need to wipe the repo before uploading packages again?20:42
DocScrutinizer05kerio: iirc he said otherwise the repo won't take the new upload, due to some version check20:45
*** toxaris has joined #maemo-ssu20:45
keriooh, makes sense :)20:45
*** toxaris has quit IRC20:50
keriowait, why do i have status-area-orientationlock-applet 1.0+0cssu3.1?20:52
kerio:o20:52
keriowhere did that come from?20:52
keriomerlin1991: modest-l10n-mr0 doesn't have a correct version20:52
*** toxaris has joined #maemo-ssu20:52
keriounless it's intended that the cssu package is below the nokia one20:52
merlin1991that's why we depend on the various l10n packages20:54
DocScrutinizer05kerio: why would you NOT have status-area-orientationlock-applet?20:55
kerioDocScrutinizer05: i have it at a higher version than community-devel20:55
keriomerlin1991: 6.11+0m5 is greater than 6.11+0cssu320:55
DocScrutinizer05I gather that's the version in T then20:55
kerioi assume you wanted 6.11+0m5+0cssu3 :)20:55
kerioDocScrutinizer05: let me rephrase that20:55
kerioDocScrutinizer05: i have it at a higher version than any repo20:55
merlin1991kerio: check ie modest-l19n-nono20:56
merlin1991with dpkg -l20:56
merlin1991it's only the mr0 which has a higher version from nokia20:56
merlin1991all the single packages had various lower versions20:56
keriomerlin1991: hm, why do you still provide a package then?20:57
merlin1991because it builds from the source20:58
merlin1991and nobody ever cared to make the dependencies proper20:58
keriohrmpf20:58
keriofukken nokia20:58
DocScrutinizer05indeed20:58
keriomerlin1991: it seems like a waste20:59
DocScrutinizer05my mantra: the MP is *EVIL*20:59
DocScrutinizer05and we should try hard to get rid of it, every single dependency less in MP is a win20:59
kerioDocScrutinizer05: it's still needed for HAM upgrades20:59
keriojust having *proper* deps on the other stuff is fine21:00
DocScrutinizer05hmm, don't know if that's a good or even valid rationale21:00
DocScrutinizer05it's just bulk-release centric concept21:00
DocScrutinizer05we *could* update every single shit on a per-opportunity basis, even in HAM (with slight updates/fixes/patches)21:01
DocScrutinizer05maybe get a MP called 'core system'21:02
IronLegendYay i need nokia PC suite to update now xD21:02
DocScrutinizer05hmmpf21:02
DocScrutinizer05where's my [general notice] ?21:02
merlin1991IronLegend: that msg is a lie21:02
DocScrutinizer05[GENERAL NOTICE] if you're using CSSU-T, do NOT use "update all" in HAM, you need to first install/update the opername widget, only *then* go update the rest. Read http://talk.maemo.org/showthread.php?p=1264963#post1264963 ff21:03
merlin1991IronLegend: install the operator name update first21:03
IronLegendah thanks21:03
kerioDocScrutinizer05: just make it so that it's posted every 5 minutes in #maemo and here21:03
DocScrutinizer05friggin konversation has no scripting21:03
* DocScrutinizer05 spanks merlin1991 a little bit21:04
IronLegendDocScrutinizer05: or you could put it in the topic ?21:05
* merlin1991 bitchslaps back21:05
kerioguys, get a room21:05
DocScrutinizer05wtf is wrong with chanserv?21:05
IronLegendit dislikes you21:06
DocScrutinizer05ooh no /op21:06
merlin1991DocScrutinizer05: usinga an inferior irc client atm?21:06
*** ChanServ sets mode: +o DocScrutinizer0521:06
DocScrutinizer05nah, inferior cerebral perfomance21:07
DocScrutinizer05and actually a new installation of whole system that doesn't have all the goodies I got on the old one during 5 years21:07
DocScrutinizer05like /opm21:07
DocScrutinizer05and /dopm21:07
DocScrutinizer05(user defined commands)21:08
DocScrutinizer05so I went for /op #maemo-ssu instead of /cs op #maemo-ssu21:08
*** DocScrutinizer05 changes topic to "do NOT use "update all" in HAM, you need to first install/update the opername widget, only *then* go update the rest. Read http://talk.maemo.org/showthread.php?p=1264963#post1264963 ff || Maemo Community Seamless Software Update "CSSU" channel, http://wiki.maemo.org/Community_SSU | Known bugs: http://j.mp/communityssu-bugs | Channel logs: http://mg.pov.lt/maemo-ssu-irclog/ | Sources: htt"21:09
kerioDocScrutinizer05: /cs topic?21:09
DocScrutinizer05thought as much :-S21:09
merlin1991also you hit topiclen with your update21:09
freemangordonmerlin1991: apt debian/rules is screwed21:10
*** DocScrutinizer05 changes topic to "Maemo Community Seamless Software Update "CSSU" channel, http://wiki.maemo.org/Community_SSU | Known bugs: http://j.mp/communityssu-bugs | Channel logs: http://mg.pov.lt/maemo-ssu-irclog/ | Sources: http://gitorious.org/community-ssu/ | Latest version (testing): 21.2011.38-1Tmaemo5.1; (stable): 21.2011.38-1Smaemo4.1 || do NOT use "update all" in HAM, you need to first install/update the"21:10
DocScrutinizer05:-/21:10
freemangordonatp is compiled with on -O-x flags :(21:10
merlin1991freemangordon: screwed is a bit vague, any additional hints?21:10
freemangordon^^^21:10
kerio-x?21:10
freemangordonkerio: typo21:10
freemangordon-Ox :)21:11
keriooh21:11
keriowhat's -Ox?21:11
*** DocScrutinizer05 changes topic to "Maemo Community Seamless Software Update "CSSU" channel, http://wiki.maemo.org/Community_SSU | Known bugs: http://j.mp/communityssu-bugs | Channel logs: http://mg.pov.lt/maemo-ssu-irclog/ | Sources: http://gitorious.org/community-ssu/ | Latest version (testing): 21.2011.38-1Tmaemo5.1; (stable): 21.2011.38-1Smaemo4.1 || READ!!: http://talk.maemo.org/showthread.php?p=1264963#post1264963"21:11
freemangordonmerlin1991: it exports CXXFLAGS, but actual configure line does "cd build && CXXFLAGS="$(confcxxflags)" ../configure --disable-block-hidden $(confflags)"21:11
*** ChanServ sets mode: -o DocScrutinizer0521:12
merlin1991freemangordon: seems silly21:24
freemangordonmerlin1991: did you look at it (debian/rules)?21:31
merlin1991I'm currently doing dinner preperations21:32
*** _xnt14 has quit IRC21:32
DocScrutinizer05oooh dinner, what a nice idea21:32
DocScrutinizer05((<kerio> DocScrutinizer05: /cs topic?)) nah, the topic editor in Konversation is actually fine21:33
DocScrutinizer05while /cs topic is clumsy21:34
*** _xnt14 has joined #maemo-ssu21:35
freemangordonmerlin1991: I consider that a bug, going to fix it in master, ok?21:41
freemangordonmerlin1991: scratch that, somehow flags are passed21:58
freemangordonthough the way it is passed is ugly :(22:00
*** luf has left #maemo-ssu22:01
keriohuh, this pinning thing actually gives you a neat way of uninstalling CSSU without bothering with weird stuff22:14
keriopin every repo at 2000, remove the cssu repos, update & upgrade22:15
freemangordonkerio: there is -thumb version of apt in the repo, you may want to try it ;)22:15
keriofreemangordon: which version is it?22:15
freemangordon0.7.20.2maemo13.2+0cssu1+thumb022:15
kerioyay ^_^22:16
keriofreemangordon: ZOMG SO FAST22:20
freemangordonyep ;)22:21
kerioexcept that not really22:21
freemangordonit is faster22:21
* kerio pictures a wonderful world where cssu-th is actually for the "fremantle-thumb" release and the versions are the same as cssu22:23
kerioand you just pin it at 600 priority and it magically works fine22:23
*** Spark25x has joined #maemo-ssu22:41
Spark25xheya22:41
Spark25xwhy the new SSU update ask for pc connectivity?Is this mandatory?tnx in advance22:41
merlin1991see /topic22:41
kerioSpark25x: first update the operator name widget, make sure it's updated and that application manager stopped doing its things, then do the upgrade22:42
DocScrutinizer05[GENERAL NOTICE] if you're using CSSU-T, do NOT use "update all" in HAM, you need to first install/update the opername widget, only *then* go update the rest. Read http://talk.maemo.org/showthread.php?p=1264963#post1264963 ff22:43
Spark25xand if ill do a "fresh install" tommorow i will need to follow same process?:P22:43
DocScrutinizer05no22:43
merlin1991Spark25x: nope this is only valid for an update to the current version22:43
Spark25xdoc sorry i cleaned the buffer and didnt had the chance to read the topic22:43
Spark25xmy bad22:43
merlin1991you can get it back with /topic ;)22:44
Spark25xi know now :P22:44
Spark25xi meant i did the mistake to not read it before ask22:44
Spark25xhehe22:44
*** NIN101 has quit IRC22:45
keriofreemangordon: btw, please use inequalities in the metapackage :322:45
Spark25xand another question if you dont mind,where can i see the changelog etc?the wiki and ssu features page arent update it yet (hope this is not a silly question)22:45
keriothey are22:46
keriohttp://wiki.maemo.org/Community_SSU/Changelog#T-maemo522:46
Spark25xkerio okey ill check again lol sorry22:46
DocScrutinizer05been updated yesterday22:47
DocScrutinizer05even prior to rollout of actaul update22:47
DocScrutinizer05(though only minutes)22:47
merlin1991why do people get so demanding lately regarding cssu? people used to be happy when the changelog arrived a week later :D22:48
DocScrutinizer05hehe22:48
Spark25xtommorow ill make a reflash couse i want to get rid of nitroid :P and im happy that i can test the new ssu before this :D22:48
keriodo i need anything in particular to make obexd work?22:50
Spark25xokey tnx for the answers have a good night :)22:51
merlin1991kerio: if it doesn't work then you have no bluethooth :D22:51
*** Spark25x has quit IRC22:54
*** BCMM has joined #maemo-ssu23:49
*** Pali has quit IRC23:59

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