IRC log of #maemo-ssu for Sunday, 2013-01-06

kerioluf: what's that?00:00
lufIt's even in meego git ...00:03
lufIt's some dbus call to connect to BT device.00:03
lufUsed in Diablo (only) in osso-gwconnect00:04
*** arcean has quit IRC00:09
*** arcean has joined #maemo-ssu00:10
*** _rd has quit IRC01:06
*** kolp has quit IRC01:37
*** kolp has joined #maemo-ssu01:37
*** Estel_ has quit IRC01:39
*** arcean has quit IRC01:52
freemangordonluf: pong01:55
*** Woody14619 has quit IRC02:00
luffreemangordon: I find the answer ;)02:07
freemangordonok :)02:08
luffreemangordon: BTW there is gvs-backends with obexftp in maemo _but_ it needs obex-data-server which isn't in maemo ;)02:08
freemangordonluf:  gvs-backends? what does that do?02:08
freemangordonthe same as obexd plugin?02:08
freemangordonluf: BTW where?02:09
luffreemangordon: osso-gnomevfs-extra can list the devices ok but there is a problem that it uses btcond which exists only in osso-gwconnect (Diablo package only). I have to rewrite it for bluez.02:09
luffreemangordon: I'm sorry: gvfs-backends02:10
lufgvfs is used in newer gnome (as gnomevfs is old and not used)02:10
freemangordonaah, ok02:15
luffreemangordon: do you have any idea how the icons are choosed?02:15
freemangordonyes02:15
freemangordon:D02:15
lufWhen I use names insted of BT MAC I see no icons :(02:15
freemangordonit takes the icon from device properties :D02:15
freemangordonok, well02:15
freemangordonit is using some gconf value02:15
freemangordonjust a minute02:15
lufWhen I have list of devices it's able to display phone/computer icons based on the name (just BT address)02:16
lufHowever BT address is unreadable for humans :D02:16
*** Milhouse has quit IRC02:17
freemangordonluf: are you sure you setup everything correctly02:19
freemangordon?02:19
freemangordonas it seems there are 2 values, name and address02:19
luffreemangordon: where do you see 2 values?02:19
lufI need to fill GnomeVFSFileInfo *info;02:20
freemangordonI know, gimme some time, i'll provide the links02:20
lufThere is only name and symlink_name. When name is [BT MAC] it shows the icons. otherwise it doesn't.02:20
freemangordonre icon:https://gitorious.org/community-ssu/libhildonfm/blobs/master/hildon-fm/hildon-file-system-obex.c#line31202:21
lufYes. If the name is [BT MAC] it shows the icon right ...02:22
freemangordonluf: for the name: https://gitorious.org/community-ssu/libhildonfm/blobs/master/hildon-fm/hildon-file-system-obex.c#line34002:22
lufThere is no possibility to have readable name and icon in that case (or patch the file you sent).02:22
freemangordonshht, check the code ^^^ :P02:23
lufI see ...02:23
lufThe ossofilemanager is using _get_icon_from_uri02:23
freemangordonluf: no02:24
freemangordonit receives everything ready from libhildonfm02:24
freemangordonand just uses it02:24
freemangordonfortunately we have libhildonfm code02:24
freemangordon;)02:24
freemangordonluf: I am not sure dbus_message_new_method_call ("org.bluez", "/org/bluez/hci0",02:25
freemangordon"org.bluez.Adapter", "GetRemoteName");02:25
freemangordonis a valid for BT402:25
*** Milhouse has joined #maemo-ssu02:25
freemangordonthat's why you don;t have device name02:25
lufNo it's not.02:26
freemangordonyou' rather have to use org.bluez.Device02:26
freemangordonafaik02:26
lufI think it's not valid even for the stock bluez.02:26
freemangordonit is still 402:26
freemangordonbluez4 that is02:27
freemangordonthe code we have is legacy from diablo, which(judging by the code) has bluez302:27
lufOk so it's ok to put there just BT MAC and fix the hildon ...02:27
freemangordonseems so02:27
lufOk.02:28
freemangordonwhere is the code of libhildonfm you use?02:28
freemangordonI can help a bit and re-write that function to use bluez4 dbus api02:29
lufI don't change libhildonfm (I just uncomment the #if 0 you sent me.02:29
freemangordonhmm, ok02:29
luffreemangordon: don't worry. I'll do it.02:29
freemangordonwell, ok :)02:29
lufNow I knwo what to change ;)02:29
freemangordonI am just afraid there are too much places to be changed for one men. but ok :)02:29
lufAnd I know how to obtain the name ;)02:30
freemangordonok, great02:30
lufI see two places (gnomevfs-extras and now also libhildonfm)02:30
lufThanks. It's big help.02:30
freemangordonluf: and missind daemon02:30
freemangordon*missing02:30
luffreemangordon: no. I'll use bluez ;)02:31
freemangordonluf: BTW once i wrote a code to connect rfcomm, lemme check it02:31
freemangordoniirc it is Qt, but still02:31
luforg.bluez.Serial.Connect/Disconnect ;)02:32
lufI need just to pass/determine the path for the device ;)02:32
freemangordonright http://pastebin.com/KHFWygg002:33
lufdbus-send --system --print-reply --type=method_call --dest=org.bluez /org/bluez/`pidof bluetoothd`/hci0/dev_00_00_00_00_00_00 org.bluez.Serial.Disconnect string:FTP02:33
lufYeah I know that ;)02:34
freemangordonluf: where from do you know it, this is mine unpublished code :P02:34
freemangordon(though maybe I should publish it, if not for anything else, then just for reference)02:35
lufqdbusviewer on my PC (also is using bluez 4.99) :D02:35
freemangordon:D02:35
lufAnd of course I know a little bit the bluez sources ;)02:35
freemangordonhehe02:35
lufhttp://maemo.org/api_refs/5.0/5.0-final/bluez/serial-api.txt02:35
lufAnd :)02:36
lufTo be honest I dislike the gnomevfs-extras code ...02:36
freemangordonluf: BTW isn't it better to use UUIDs?02:36
freemangordonthe fuck, this is lots of code. and working. defffinitely i should publish that02:38
freemangordonluf: http://pastebin.com/v0Fy8B4B02:39
freemangordonjust in case you don;t full list of have UUIDs02:40
freemangordons/don;t full list of have/don't have the full list/02:40
freemangordon~ping02:40
infobot~pong02:40
*** jonwil has joined #maemo-ssu02:41
*** NIN101 has quit IRC02:51
luffreemangordon: I'm not sure if it's so precisely defined on all devices.02:52
lufI check FTP support with UUID starting 00001106- :)02:53
freemangordonluf: this file is part of my SW I talked about02:53
freemangordonand it runs on n900.02:53
luffreemangordon: just curious. What does the sw do?02:54
freemangordonyep, the same #define FTP_UUID            "00001106" UUID_COMMON02:54
freemangordonluf: setting applet02:54
freemangordondoes whatnot with BT devices02:54
freemangordonwith vary fanci interface02:54
freemangordon*very02:54
luf:)02:54
freemangordonthe fuck, maybe I should take some rest, my typing skills are degrading every minute :D:D:D02:55
freemangordonluf: I started to write it back in the days when I was playing with extkbd/extmou02:55
freemangordonstock menu BT applet don;t have "connect" button for HID devices02:56
freemangordoneventually I found a way to write my own plugin (for HID), so all this stuff become useless02:57
*** BCMM has quit IRC02:58
*** kolp has quit IRC03:01
*** kolp has joined #maemo-ssu03:01
luffreemangordon: Ahh ok.03:04
lufHowever I don't like BT applet/settings. It's mess of something ...03:04
lufNot able to connect whatever you want (audio, network, ...)03:05
*** dhbiker has quit IRC03:06
freemangordonluf: yeah, I know. my "replacement" does that too, but it is: 1. Qt; 2. Unfinished; :D03:06
luffreemangordon: It's nothing you can't fix ;)03:07
freemangordonhehe03:07
freemangordonluf: one of the UI functions is to connect to BT serial profile, so you can use /dev/rfcommN03:09
* freemangordon will try to put himself together and finish tklock tomorrow :)03:09
lufThere are several of Connect ... as I wrote. rfcomm is just one of them.03:09
freemangordonyep03:09
luf:D03:10
freemangordonluf: because of that project(consisting of ~60-70 files) I feel comfortable with bluez03:11
luffreemangordon; I pushed my last dev on osso-gnomevfs-extras into git. There is only missing connecting (still btcond).03:11
freemangordonhmm, can we find that daemon from somewhere?03:12
lufHowever listing and checking FTP is working.03:12
freemangordonwhat about names?03:12
luffreemangordon: I don't like another daemon when bluez can do that.03:12
lufI'll fix names in libhildonfm later.03:12
freemangordonluf: there might be a reason for Nokia to write a daemon03:12
lufmaybe bluez3 :D03:13
lufI'll try at least. With bluez 4.99 I also superseed bluetooth-dun and maybe more what is natively supported by bluez.03:13
freemangordonok03:14
lufAs few daemons as possible is also good for cpu, memory, ...03:14
freemangordonsure03:14
freemangordonBTW http://maemo.org/packages/package_instance/view/diablo_sdk_free_i386/osso-gwconnect/1.0.12/03:15
lufI know. but Diablo ...03:15
freemangordonnut it is sdk/free, so I suppose there is source code too. going to check03:15
freemangordon*but03:15
lufDon't waste your time ;)03:16
luftklock is waiting for you :D03:16
freemangordonaah, not today03:16
freemangordon:)03:16
freemangordonyep http://repository.maemo.org/pool/diablo/free/o/osso-gwconnect/03:16
freemangordonluf: well, that could be used as reference at least03:20
luf:)03:20
*** kolp has quit IRC03:29
*** kolp has joined #maemo-ssu03:30
freemangordonluf: seems you miss the call to free_device_properties(devprops); on success exit path03:38
freemangordonor I am missing something03:38
freemangordonscratch ^^^, found it03:39
luffreemangordon: :)03:48
freemangordonluf: yeah, I tought that you introduced get_device_properties function03:48
luffreemangordon: you're right I introduced that function.03:49
freemangordonyep, but in a different commit, so I didn;t have the whole picture when I checked the last one. that's why i though there is a bug :)03:50
lufAahhh ok.03:51
lufBTW it's funny we have osso-gwconnect-dev in maemo 5 SDK when there is no osso-gwconnect :D03:52
freemangordon:D03:53
*** dhbiker has joined #maemo-ssu03:53
freemangordonanyway, I am off to bed, night03:54
jonwilhmmm, it looks like http://repository.maemo.org/pool/fremantle/free/o/osso-gwconnect/ conatins osso-gwconnect-dev deb file plus tarball that builds osso-gwconnect and osso-gwconnect-dev04:07
*** M4rtinK has quit IRC04:16
*** kolp has quit IRC04:45
*** amiconn_ has joined #maemo-ssu05:45
*** amiconn has quit IRC05:45
*** amiconn_ is now known as amiconn05:45
*** DocScrutinizer05 has quit IRC06:02
*** DocScrutinizer05 has joined #maemo-ssu06:02
*** nox- has quit IRC06:16
*** dhbiker has quit IRC06:41
*** dhbiker has joined #maemo-ssu06:56
luffreemangordon: coool it works :D07:19
*** Milhouse has quit IRC07:37
*** Milhouse has joined #maemo-ssu07:37
*** Milhouse has quit IRC09:05
*** Milhouse has joined #maemo-ssu09:08
*** freemangordon has quit IRC09:15
*** Milhouse has quit IRC09:15
*** Milhouse has joined #maemo-ssu09:17
*** Pali has joined #maemo-ssu10:00
*** _rd has joined #maemo-ssu10:24
*** jonwil has quit IRC10:36
*** Milhouse has quit IRC10:41
*** Milhouse has joined #maemo-ssu10:42
*** freemangordon has joined #maemo-ssu10:43
freemangordonmerlin1991: seems like you and chemist should issue a new update ASAP, in case you've missed it:http://talk.maemo.org/showthread.php?t=8853510:45
* merlin1991 is currently reading the backlog10:46
freemangordonluf: cool, what works?10:48
*** Milhouse has quit IRC10:53
*** Milhouse has joined #maemo-ssu10:54
*** jonwil has joined #maemo-ssu10:55
freemangordonjonwil: that one is the diablo one, NFC what it does in fremantle repo10:59
*** LaoLang_cool has joined #maemo-ssu11:02
Palimerlin1991, freemangordon: What about adding Maemo upgrade description to CSSU metapackage?11:09
Palihttps://wiki.maemo.org/Packaging#Maemo_upgrade_description11:09
*** Milhouse has quit IRC11:10
Palithat string will be displayed in HAM when you click on package update details11:10
Palithere can be changelong11:10
freemangordonwhy not11:10
Palionly small problem11:10
Palithat string with changelog must be changed every time when updating cssu metapackage11:11
Paliso only "last" update changelog is possible11:11
Paliand problem with missing changelogs on wiki can be fixed11:11
freemangordonPali: sure, but we could have control.in, and replace $SOME_STRING_TO_MATCH in debian/rules11:11
Palias it will display directly in HAM11:11
Paliok11:12
freemangordonPali: BTW I am still not sure how exactly to blacklist a certificate. is move .pem to /blacklist enough?11:13
PaliI do not know11:13
Palitry to ask certman mainatainer11:13
Palihe blacklisted also other certificates after pr1.3.111:13
freemangordonme? nah, i'll leave that to merlin1991 :)11:13
Pali(for cssu)11:13
*** LaoLang_cool has quit IRC11:14
Paliwe should ask him if he will continue to maintain certman or not11:14
Palifreemangordon, we can use this for geting changelog from last debian version: "dpkg-parsechangelog | sed -n 's/^  //p'"11:20
freemangordonPali: yep11:20
PaliI will write patch for debian/rules11:21
freemangordonPali: there are a couple of packages in CSSU right now using a debian/bontrol template, I'll try to find one11:21
freemangordondebian/control11:22
Paliwe do not need template11:22
freemangordonPali: what?11:22
Palijust call that script in debian/rules and update debian/<package>/DEBIAN/control11:22
*** Milhouse has joined #maemo-ssu11:22
Paliit is easier11:22
freemangordonPali: what will happen if you build it twice?11:22
Paliafter dh_gencontrol11:22
Palinothing11:22
freemangordonaah, ok, got it11:23
Pali debian/<package>/DEBIAN/control is cleaned11:23
freemangordonyes, yes11:23
merlin1991we have that maemo upgarde description already in -stable11:26
merlin1991I started that ;)11:26
freemangordonmerlin1991: sure, but it should be altered by hand :)11:27
merlin1991something I did aswell :D11:27
freemangordonnot for the last -stable :P11:27
merlin1991well that's chemist boat now11:28
freemangordonhehe11:28
freemangordonmerlin1991: what about broken CA?11:28
*** Milhouse has quit IRC11:28
merlin1991we're going to blacklist the root ca11:28
freemangordonmerlin1991: do you know how?11:29
PaliI will use this for getting major version string: dpkg-parsechangelog | sed -n 's/Version: \(.*maemo.*\)\..*/\1/p'11:29
freemangordonPali: will that work for -thumb?11:29
Palifreemangordon, what is version string for -thumb?11:29
PaliI see this: 21.2011.38-1Tmaemo7.1 for testing11:30
Pali(major versin will be 21.2011.38-1Tmaemo7)11:30
Paliwithout ".1"11:30
merlin1991freemangordon: https://gitorious.org/maemo-5-certificate-manager/maemo-security-certman/commit/fc21ae95be7f44189be2360d445f4fe3f10ab71211:30
PaliI think it is good to include changelong since major version11:30
merlin1991basically we throw the pem into /etc/certs/blacklist11:30
freemangordonPali: 21.2011.38-1Tmaemo7.1+thumb011:30
*** Milhouse has joined #maemo-ssu11:30
freemangordonmerlin1991: I was reading that commit for the last hour, still not convinced it does just moving a .pem to /blackilst11:31
Paliit cut '.1+thumb0'11:31
Paliso changelog will be from 21.2011.38-1Tmaemo7 to 21.2011.38-1Tmaemo7.1+thumb011:31
merlin1991if you don't get the commit just checkout twice and diff -r11:32
merlin1991then you'll see ;)11:32
freemangordonmerlin1991: ok, i trust you11:32
freemangordonPali: ok, sounds sane, lets see how it will work in practice :D11:33
freemangordonPali: what about from  21.2011.38-1Tmaemo7.1+thumb0 to  21.2011.38-1Tmaemo8+thumb011:34
Palifreemangordon, problem is how to get "from" version string11:35
*** jon_y_ has joined #maemo-ssu11:36
freemangordonPali: wouldn't it be easier to just grep for Changes in dpkg-parsechangelog output and use the lines up to the end?11:39
*** jon_y has quit IRC11:39
freemangordonthough I am not sure how to tell grep to do that11:39
PaliI'm doing it11:40
freemangordonhmm, seems like "dpkg-parsechangelog | grep -A999 Changes:" does the job11:40
Palibut you need to tell start and end version11:40
Paliif you do not specify version, then dpkg-parsechangelog show only *last*11:40
freemangordonPali: exactly11:40
Paliand this is better: dpkg-parsechangelog | sed -n 's/^  //p'11:40
*** jon_y_ has quit IRC11:40
freemangordonwhy do we need some "base" version11:41
Palibecause changelog starts with 3 empty spaces11:41
*** Milhouse has quit IRC11:41
Paliand debian/control needs one space at stary11:41
Pali*start11:41
freemangordonyou either puth the whole changelog, or only the last one11:41
freemangordon*put11:41
Palionly last is bad, because we released more times minor update 1 hour after major11:42
Pali(e.g. Recompiled thumb binaries as non-thumb)11:42
keriowhy don't we just put a link to the wiki for the complete changelog?11:43
freemangordonI see. though I think that won;t work for -thumb, as it does not have -testing's version in debian/changelog11:43
kerioare links clickable in HAM?11:43
*** Milhouse has joined #maemo-ssu11:45
Palifreemangordon, it working also for thumb11:46
PaliI tested it11:46
freemangordonok, great11:46
Paliit generate changelog from '21.2011.38-1Tmaemo7' to '21.2011.38-1Tmaemo7.1+thumb0'11:46
Palithumb has also 'T' in version11:47
freemangordonaah, yes, -thumb has mp-fremantle-community-pr (21.2011.38-1Tmaemo7) unstable; urgency=low in changelog11:49
*** Milhouse has quit IRC11:52
*** kolp has joined #maemo-ssu11:55
Palithis will work if we use version scheme: <something>maemo<something>.<something>12:01
PaliI pushed patch to master12:01
*** Milhouse has joined #maemo-ssu12:04
*** _rd has quit IRC12:06
*** toxaris has joined #maemo-ssu12:07
*** Milhouse has quit IRC12:11
*** Milhouse has joined #maemo-ssu12:17
*** zogg__ has joined #maemo-ssu12:18
*** zogg_ has quit IRC12:20
*** Milhouse has quit IRC12:24
*** _rd has joined #maemo-ssu12:31
freemangordonluf: installing libxml2 led dor maemo unable to boot on my devel device :(12:39
freemangordons/dor/to/12:39
infobotfreemangordon meant: luf: installing libxml2 led to maemo unable to boot on my devel device :(12:39
*** _rd has quit IRC12:40
*** _rd has joined #maemo-ssu12:40
*** M4rtinK has joined #maemo-ssu12:43
keriowhich libxml2?12:49
kerioit boots fine here12:49
keriomaybe it should depend on the latest zlib1g12:50
freemangordonkerio: which zlib version?12:50
kerioalso, what does "unable to boot" mean?12:50
freemangordonit just sits on running dots12:50
kerioouch12:51
keriothat's even worse12:51
freemangordonwell, I guess it is boot loop, but I have the device in RD mode12:52
freemangordonso WD does not kick in12:52
freemangordonor lifeguard rest, or whatever12:52
freemangordon*reset12:52
keriouse the recovery console to configure the early ssh/usbnet12:52
kerioand go from there12:52
freemangordoni already flashed :D12:52
keriothe n900 has some pretty amazing on-device recovery systems12:53
kerio>:c12:53
kerioit's unix! you don't reinstall unix! you fix it!12:53
freemangordonnow i'll have the chance to verify PR1.3 -> cssu-thumb upgrade12:53
keriofreemangordon: http://www.ee.ryerson.ca/~elf/hack/recovery.html12:53
*** _rd has quit IRC12:59
keriofreemangordon: btw, what do you think about community-thumb.merlin1991.at as the repo URL?12:59
freemangordonkerio: should I think anything about it?13:00
kerioit would be a neat URL, wouldn't it13:01
freemangordonsure13:01
freemangordonwhy need to change the current one?13:01
keriobecause 1) it's ugly13:01
kerioand 2) apt is stupid and only shows the hostname in a lot of places13:01
freemangordonok, according to HAM, cssu-thumb will freee 29.7MB of device's memory :D13:02
freemangordonso we gain ~35MB (the difference is because kernel-cssu is way bigger than omap1)13:03
*** jon_y has joined #maemo-ssu13:08
keriofreemangordon: less13:09
keriothe kernel isn't actually stored on the system for long13:10
freemangordonkerio: what?13:10
keriokernel-cssu includes the .fiasco for the kernel, but it gets deleted by kernel-cssu-flasher's postinst13:10
kerioor preinst13:10
kerioor something13:10
freemangordonkerio: sure, but the same happens for modules13:11
keriosame for kernel, and kernel-power13:11
freemangordonand modules in kcssu>>omap113:11
freemangordonso the difference gets substracted when HAM reports how much memory will be freed13:11
keriothe omap1 modules aren't deleted13:12
freemangordonthe real benefit for only userland is > than what HAM reports13:12
freemangordonkerio: check it once again13:12
freemangordonkerio: KCSSU Conflicts: and Replaces: kernek, kernel-modules13:12
keriothe "kernel" and "kernel-modules" packages aren't uninstalled13:12
kerio...and HAM honors that? :o13:13
freemangordonkerio: they are uninstalled, i've just checked it13:13
kerioso what's the problem with making HAM uninstall packages via conflicts? :o13:13
freemangordonnfc, who says there is a problem?13:14
freemangordonkerio: BTW even if kernel,kernel-modules are not uninstalled, freed memory is still > of what HAM reports13:16
kerioso we *can* delete operator-name-cbs-widget now that the actual files are moved back in connui-home-cellular13:16
*** NIN101 has joined #maemo-ssu13:18
freemangordonkerio: hmm, weird. kernel IS uninstalled, but not kernel-modules13:20
freemangordonaah, waaait13:20
freemangordonit is uninstalled, but dpkg -l lists it as provided package ;)13:21
freemangordonor whatever "rc  kernel-modules                                      2.6.28-20103103+0m5 " means13:21
kerioremoved but with configuration13:21
freemangordonkerio: what zlib version you have installed?13:22
freemangordonthe one that works with libxml213:22
keriozlib1g: Installed: 1:1.2.7.dfsg-13maemo1+0cssu113:22
keriolibxml2: Installed: 2.6.32.dfsg-5maemo4+0m5+0cssu113:23
freemangordonaiui libxml2 should work with older zlib, ain;t?13:23
kerioif the new libxml2 doesn't work with the old zlib1g, it MUST be dependent on the newer version13:23
kerio~211913:23
infobotThe 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 2119.13:23
freemangordonI know13:23
freemangordonok, will check it13:23
kerioidk, you tell me13:23
keriohaven't you got a bootloop because of that?13:23
freemangordonI had the new one installed iir13:24
freemangordonc13:24
freemangordonhowever, will install libxml2 now13:24
keriofreemangordon: i'd make kernel-cssu conflict with kernel, actually13:25
freemangordonalready there13:25
kerio...oh right13:25
keriowhy not go the even better route and just call your packages kernel and kernel-modules?13:25
freemangordonwill do13:25
*** Pali has quit IRC13:26
*** Milhouse has joined #maemo-ssu13:26
*** xes has joined #maemo-ssu13:27
freemangordonhmm, bootloop again :(13:28
kerioso... new libxml should depend on new zlib, new zlib should Break the old libxml13:29
xesi hate bootloop.... :(13:29
freemangordonkerio: but why new libxml2 does not work with old zlib?13:29
*** arcean has joined #maemo-ssu13:30
freemangordonmakes no sense13:30
freemangordonluf: ^^^?13:30
freemangordonthough it could be broken because of -thumb13:30
*** xes_ has joined #maemo-ssu13:30
*** Milhouse has quit IRC13:32
freemangordonhighly unlikely, but still13:32
*** xes has quit IRC13:33
*** Milhouse has joined #maemo-ssu13:34
*** Pali has joined #maemo-ssu13:39
*** Milhouse has quit IRC13:41
*** Milhouse has joined #maemo-ssu13:41
*** Pali has quit IRC13:48
*** Pali has joined #maemo-ssu13:48
*** MohammadAG has joined #maemo-ssu13:49
*** Milhouse has quit IRC13:53
*** Milhouse has joined #maemo-ssu13:55
*** Pali has quit IRC13:58
keriofreemangordon: yay moar thumb14:01
keriohm, what's changed in tvout?14:02
keriooh, nothing14:03
freemangordonkerio: yeah, I was preparing a new update, when I hit libxml bootloop :(14:03
keriojust update both at the same time14:03
kerioand make sure the dependencies are correct14:03
*** Milhouse has quit IRC14:04
keriomaybe it's libpng that caused the bootloop14:04
freemangordonkerio: aah, wait. it could be that i have newr zlib-dev installed in SB14:04
freemangordonkerio: no, I installed only libxml14:04
freemangordonI have libpng on my primary,it is ok14:04
*** _rd has joined #maemo-ssu14:05
*** Pali has joined #maemo-ssu14:06
keriowell, now xterm goes to portrait14:06
kerio...how useful14:06
* kerio updates his blacklist14:07
*** Milhouse has joined #maemo-ssu14:07
keriooh god, it's already blocked D:14:07
keriohow do i block it more?14:07
luffreemangordon: osso-gnomevfs-extras with browsing the device. However there is/are some bug(s) I have find.14:08
luffreemangordon: I'm able to use bluez for rfcomm connecting ;)14:08
luffreemangordon: libxml2 - it's strange. libxml2 depends on zlib1g >= 1.2.3. But you should have such version.14:11
luffreemangordon: I checked lixml2 patched with stock zlib.14:13
*** arcean_ has joined #maemo-ssu14:14
luffreemangordon: what's your zlib version?14:14
luffreemangordon: Isn't problem thumb libxml with non-thumb zlib?14:15
*** Milhouse has quit IRC14:16
*** arcean has quit IRC14:17
*** MohammadAG has quit IRC14:19
freemangordonluf: could be, will check14:19
freemangordonluf: against which zlib-dev you have libxml built?14:20
freemangordonstock or latest?14:20
*** MohammadAG has joined #maemo-ssu14:26
lufI think againist stock.14:29
*** Milhouse has joined #maemo-ssu14:30
luffreemangordon: ii  zlib1g-dev  1.2.3-9.osso9+0m514:30
freemangordonluf: baad14:31
freemangordonseems it breaks when built agains new zlib14:32
freemangordondowngrade and check14:32
freemangordon*will downgrade14:33
lufStrange :(14:33
*** Milhouse has quit IRC14:36
*** Milhouse has joined #maemo-ssu14:40
freemangordonluf: ARM libxml2 build against old zlib works14:48
freemangordongoing to check thumb build agains old zlib14:48
freemangordon*against14:48
*** Milhouse has quit IRC14:49
*** jade has quit IRC14:51
luffreemangordon: what about ARM libxml2 build againist new ARM zlib?14:54
*** Milhouse has joined #maemo-ssu14:54
freemangordonwill check it14:56
freemangordonluf: though it should be the same as ^^^14:56
freemangordon(thumb agains old zlib)14:57
freemangordon*against14:57
*** Jade has joined #maemo-ssu14:57
*** Jade has joined #maemo-ssu14:57
luffreemangordon: the libxml2 patch depends on ZLIB_VERNUM ... maybe that's the difference in build. But failure with old libxml2 and new zlib is after starting hidlon (showing empty desktop).15:00
*** Milhouse has quit IRC15:00
freemangordonluf: I know about  ZLIB_VERNUM15:00
freemangordonok, thumb build against old zlib boots :(. I was hoping it is my build env broken, but unfortunately it seems ok15:01
freemangordongoing to check ARM build against new zlib15:01
*** Milhouse has joined #maemo-ssu15:02
luffreemangordon: ZLIB_VERNUM seems ok in build env (not checked in zlib1g-dev).15:02
freemangordonluf: yep, it is ok15:02
freemangordonit is something else15:03
freemangordonluf: BTW do you have backup menu or something installed?15:03
lufNot on dev device :)15:03
freemangordonaah, same here, a pity, i'll have to reflash again if it does not work :(15:04
xes_...and the rescue/recovery flash? It doesn't work?15:08
xes_http://206.253.166.96/N900/rescueOS/15:09
*** _rd has quit IRC15:10
*** Milhouse has quit IRC15:12
freemangordonxes_: don;t have that installed15:26
*** _rd has joined #maemo-ssu15:27
xes_freemangordon: you don't need to install that image, you could boot it using flasher15:34
xes_http://206.253.166.96/N900/rescueOS/documentation.txt15:34
freemangordonxes_: thanks15:36
freemangordonbut usually I have -thumb installed ;)15:36
freemangordonluf: I know what is going on15:37
freemangordonzlib has checks for the same version15:38
freemangordon:(15:38
freemangordonluf: https://gitorious.org/community-ssu/zlib/blobs/master/deflate.c#line22615:44
freemangordonhmm, though it checks only major15:45
freemangordonluf: "gconftool-2: /usr/lib/libz.so.1: no version information available (required by ./libxml2.so.2)"15:53
freemangordonthis is libxml2 build with new zlib-dev :(15:54
*** Pali has quit IRC15:58
*** Pali has joined #maemo-ssu15:58
*** Milhouse has joined #maemo-ssu16:01
freemangordonluf: http://pastebin.com/srZgr7rT16:05
luffreemangordon: LD_LIBRARY_PATH=/usr/lib gconftool-2 ?16:06
freemangordonluf: ?16:06
freemangordonno, LD_LIBARY_PATH=./16:07
freemangordonand I have new libxml2 extracted in ./16:07
*** Milhouse has quit IRC16:07
lufI'll check later. Thanks for all the information16:08
freemangordonluf: trying with bot libs new version16:08
freemangordonyep, it boots16:09
freemangordonluf: i'll fix it, it is a matter of debian/control magic16:11
Palimerlin1991, there is something in /srv/apt/cssu-devel/incoming/16:16
Palilibqt4-* packages16:16
Palifrom Jan  2 15:1516:16
freemangordongregoa: ping16:20
gregoafreemangordon: You sent me a contentless ping.  This is a contentless pong.  Please provide a bit of information about what you want and I will respond when I am around.16:20
freemangordonWTF?16:20
keriogregoa: ping16:20
gregoakerio: You sent me a contentless ping.  This is a contentless pong.  Please provide a bit of information about what you want and I will respond when I am around.16:20
keriooh scripting16:20
freemangordonnice :(16:20
gregoaanyway, I'm here. what's up?16:20
kerioomg, this script is really good16:21
keriogregoa: who programmed you? :o16:21
freemangordongregoa: seems we have problem with zlib and libxml2(again). zlib breaks libxml2 that is currently installed on the device.16:21
keriofreemangordon: that's known16:22
freemangordonbut if libxlm2 is built with newer zlib, so has missing symbols if installed on the device with ol zlib16:22
kerioperhaps you mean the other issue, the new libxml doesn't work with the old zlib16:22
freemangordonkerio: may I finish?16:22
keriogo right ahead16:22
freemangordongregoa: the problem comes from th afact, that in order to build new libxml2 against new zlib one should install new libzlib-dev. but that breaks the currently installed libxml, thus cannot be installed16:23
freemangordon(*conflicts)16:24
freemangordonany idea?16:24
freemangordonif libxml2 is build against old zlib-dev everything is fine, but that looks ugly16:25
* gregoa scratches head16:26
keriofreemangordon: hmm, the issue would be with the autobuilder, right?16:26
freemangordonkerio: we are talking CSSU here, no autobuilder16:26
xes_i' m quick & dirtry...i would move the old .pc file and (for the first part) install manually(no deb) libzlib-dev16:27
*** Milhouse has joined #maemo-ssu16:27
keriofreemangordon: build libxml2 against the old zlib-dev, then rebuild it against the new one that you can install16:27
freemangordongregoa: the only think I can think of is a multi-part update16:27
gregoayeah, it guess that needs some round-trip like this16:27
freemangordoni.e. build libxml2 agains old zlib, issue mp- thingie depending on new zlib, build libxml2 against new zlib, issue anothe -mp16:28
keriofreemangordon: ew, why?16:29
keriothis is not a user issue, is it?16:29
gregoawhy the -mp release?16:29
freemangordonhmm, lemme think16:29
freemangordonaah, yes, to be easier for maintainers16:30
*** FlameReaper has joined #maemo-ssu16:30
freemangordonhmm, well, maybe no -mp16:30
keriofreemangordon: what's bad about compiling libxml2 against the old zlib1g anyway?16:31
keriothe new libxml2 is supposed to only use the API16:31
kerioand not any internal structure16:31
freemangordonkerio: hmm, makes sense16:32
freemangordonok. got it16:32
keriobtw, did you or luf eventually manage to make Breaks work?16:32
freemangordonno16:32
luffreemangordon: I think I know ...16:33
luffreemangordon: when gregoa build new zlib he has problem with zlib1g.symbols16:33
*** Milhouse has quit IRC16:34
freemangordonluf: the same here16:34
luffreemangordon: it seems that our environments ignore that. This should be the difference between old and new.16:34
freemangordonluf: ignore what? elaborate please16:34
freemangordonkerio: BTW the problem as I see it, is that one should not be allowed(by debian/control) to build against new zlib and to install on the environment with the old one16:37
freemangordonritgh now it is possible16:37
freemangordon*right16:37
keriowhy does it not work, btw?16:38
keriothe new libxml, i mean16:38
*** Pali has quit IRC16:38
kerioi mean, when compiled against headers that should still be the same thing for all intents and purposes16:38
freemangordonkerio: http://pastebin.com/srZgr7rT16:38
keriowell, for all legal intents and purposes - aka no accessing internal structures16:38
freemangordonit has missing import symbols, see the paste16:39
keriohm16:39
*** Milhouse has joined #maemo-ssu16:40
keriofreemangordon: well, make it depend on a recent-enough zlib1g16:40
kerioso it won't be possible to install it without the newer zlib1g16:40
freemangordonkerio: and then you won;t be able to install zlib-dev, as it will break currently installed libxml216:40
freemangordon(conflicts:)16:40
*** _rd has quit IRC16:41
keriowell, you're about to recompile libxml2 at that point, right?16:41
keriowhy is it installed?16:41
freemangordonluf: why was Conflicts: needed?16:41
keriofreemangordon: it should be a Breaks:, but that has the same effect - the old libxml2 depends on zlib internals that were changed in the new zlib16:42
freemangordonafter all we're going to ship it in one update16:42
kerioit's a bug in the old libxml216:42
freemangordonkerio: I know16:42
freemangordonthe point is that we'll ship both in one update16:43
kerioit's irrelevant, it's a package that exists right now16:43
freemangordon-mp will depend on both, libxml will depend on new zlib16:43
kerioand is broken by a new, different package16:43
keriowhy does the metapackage have anything to do with this?16:44
freemangordonkerio: because I can;t see how to solve the issue16:44
freemangordonif you have any idea, please share it16:45
keriouninstall libxml2, install zlib-dev, compile new libxml216:45
freemangordonkerio: gconftool-2 depends on libxml216:45
*** FlameReaper has quit IRC16:46
freemangordoni guess if I try to uninstall libxml2, I'll be left with libc6 only :D16:46
keriotell dpkg to stfu and remove anyway16:46
freemangordonkerio: I am not asking for help with dpkg, I know how to solve it in my SB16:47
freemangordonthe problem is that packaging is broken16:47
keriowell, you need to solve it in your SB and in merlin's SB16:47
freemangordonyup16:47
freemangordonand in luf's, arcean's, etc16:47
freemangordonthat's why i think it is better to remove Conflicts: from zlib and to add Depends: to libxml216:48
keriono, Depends: to libxml2 has to be there anyway16:48
keriois there anyway, perhaps just needs a version clause to be added16:48
freemangordonyep, version16:48
keriobut the new zlib1g *does* break the old libxml2, and that's a fact16:49
freemangordonI know16:49
kerioand you can't retroactively change the old libxml216:49
*** Pali has joined #maemo-ssu16:49
kerioi mean, you can do that in the repos, but not in the devices16:49
freemangordonI can, by issuing 2 updates16:49
gregoathere is a version in the Depends, which needs to be bumped. that should happen automatically with ${shlib:Depends} and dpkg-whatever during build. unless the shlib system is borken16:49
freemangordongregoa: aiui it takes only major version16:50
keriogregoa: so one would need the new libxml2 to compile the new libxml2 :s16:50
freemangordonexactly16:50
gregoafreemangordon: currently libxml has "zlib1g (>= 1:1.2.1)"16:50
*** Milhouse has quit IRC16:50
keriofreemangordon: well... it's not exactly a big deal i suppose16:50
gregoaI still think that has to be "broken" in a manual roundtrip in one SB by one person, then the 2 packages with fixed Depends/Conflicts can to uploaded and everyone else can use them16:51
keriofreemangordon: you can just release both at the same time, you don't necessarily need two updates16:52
freemangordongregoa: yep, that will work16:52
freemangordonkerio: :nod:16:52
freemangordongoing to do it for -thumb16:52
keriobut without a Conflicts or a Breaks you'll allow broken configurations16:52
freemangordoni'll increase depends: version16:53
kerioand broken as in "won't boot"16:53
*** Milhouse has joined #maemo-ssu17:06
*** Milhouse has quit IRC17:12
*** dafox has joined #maemo-ssu17:16
*** Pali has quit IRC17:40
*** FlameReaper has joined #maemo-ssu17:44
*** dafox has quit IRC17:46
*** Pali has joined #maemo-ssu17:46
arcean_~packages17:47
*** arcean_ is now known as arcean17:47
*** Milhouse has joined #maemo-ssu17:47
arcean~package17:47
infobotpackage is probably simply not installable and a bug report against17:47
*** FlameReaper has quit IRC17:50
arceanthere's a new qt with disabled QMaemo5OrientationManager in CSSU-devel17:55
luffreemangordon: I'm sorry I had to be afk.17:55
freemangordonPali: https://gitorious.org/community-ssu/mp-fremantle-community-pr/commit/3ce14c9564766bcc4f4b86d901d4d145244b9f3517:55
freemangordonluf: NP17:55
luffreemangordon: you can see in history (tomorrow or previous day) that gregoa has problem building new zlib due to missing includes in zlib1g.symobls17:56
luffreemangordon: it's not problem in our SB.17:56
freemangordonluf: however, I think two-pass build is the right thing we should do17:56
freemangordonThis is what I did for -thumb, don;t see a reason why it should ot work for -testing and -stable17:57
freemangordon*should not17:57
*** xes_ has left #maemo-ssu17:57
Palifreemangordon, that is bad17:57
Pali-v "include all changes later than version"17:57
freemangordonPali: yep17:57
freemangordonthe same as yours17:57
Paliwe need "include all changes equal or later than version"17:58
jonwilso I am still no closer to picking something from http://wiki.maemo.org/Fremantle_closed_packages to reverse engineer that I can actually do :P17:58
Palisee how major_version is parsed17:58
Palifreemangordon, major_version can be also current version17:58
Paliand then it will give for you empty changelog17:58
freemangordonPali: dpkg-parsechangelog --from "21.2011.38-1Tmaemo7"| sed -n 's/^  //p' >> debian/mp-fremantle-community-pr/DEBIAN/control17:59
freemangordondpkg-parsechangelog: unknown option `--from'17:59
Pali--from is same as '-f'17:59
Paliseems that I have new version of dpkg-parsechangelog17:59
Palifreemangordon, try: dpkg-parsechangelog -f "21.2011.38-1Tmaemo7"| sed -n 's/^  //p' >> debian/mp-fremantle-community-pr/DEBIAN/control18:00
freemangordonPali: yes, the one I have here does not support --from18:00
kerioarcean: what will that do?18:00
Palifreemangoron, '-v' is not same as '-f' (or --from)18:00
freemangordondpkg-parsechangelog: unknown option `-f'18:00
*** dafox has joined #maemo-ssu18:00
Paliwtf?18:00
luffreemangordon: I don't understand what do you mean with two-pass build.18:00
Palifreemangordon, upload output of dpkg-parsechanelog --help18:01
arceankerio: prepares system for newer hildon-desktop versions18:01
*** Milhouse has quit IRC18:01
arceankerio: new Qt basically fixes rotation/orientation problems18:01
freemangordon[sbox-FREMANTLE_ARMEL_GCC472: ~/workspace/thumb/T-maemo7.1+thumb1/mp-fremantle-community-pr] > dpkg-parsechangelog --help18:01
freemangordonUsage: dpkg-parsechangelog [<option> ...]18:02
freemangordonOptions:18:02
freemangordon-l<changelogfile>        get per-version info from this file.18:02
freemangordon-v<sinceversion>         include all changes later than version.18:02
freemangordon-F<changelogformat>      force change log format.18:02
kerioarcean: by not rotating unless hildon-desktop does something, right?18:02
freemangordon-L<libdir>               look for change log parsers in <libdir>.18:02
freemangordon-h, --help               show this help message.18:02
freemangordon--version            show the version.18:02
arceankerio: yes ;)18:02
kerioarcean: k :c18:02
Palifreemangordon, that is bad :-(18:02
arceankerio: now every Qt application behaves in the same way as the GTK/Hildon ones18:03
PaliI will try to rewrite it, to use -v18:03
kerioarcean: well, not *now*18:03
freemangordonarcean: while you are on Qt, could you blacklist that turkish CA?18:03
keriounless you mean that it'll automagically work18:03
Palifreemangordon, I need version string before "$MAJOR_VERSION" from debian/changelog18:03
arceanfreemangordon: ok18:03
keriofreemangordon: yay thumby things ^_^18:03
freemangordonPali: ok, can I help :D18:03
Paliyou can see that '-v' will not work with last version18:04
freemangordonsure18:04
Pali    --from, -f <version>        include all changes equal or later than version18:04
PaliI have this option ^^^18:04
freemangordonPali: great, I don;t :)18:04
Palion ubuntu (in new debhleper/dpkg)18:04
freemangordonin scratchbox18:05
kerioeverything can be solved with enough sed, grep and awk18:05
freemangordonPali: are you going to do it now?18:06
Palifreemangordon, you can try to use: dpkg-parsechangelog --since 0 | grep '^ mp-fremantle-community-pr '18:07
Palino, do not have time for it now18:07
Palinow going away18:07
freemangordonwell, i'll use -v for this update18:07
Pali'-v' will not work (it will print empty changelog)18:09
freemangordonIt works on -thumb18:09
freemangordonand results in thumb-only changes18:09
freemangordoni.e. since the last -testing18:10
*** xes has joined #maemo-ssu18:10
Palifreemangordon, try to add this line after MAJOR_VERSION := ...18:14
PaliPREVIOUS_VERSION := $(shell dpkg-parsechangelog -f 0 2>/dev/null | sed -n 's/^ mp-fremantle-community-pr (\(.*\)).*/\1/p' | grep -A 1 $(MAJOR_VERSION) | tail -1)18:14
Paliand replace MAJOR_VERSION with PREVIOUS_VERSION in the end of file18:14
Palithis will maybe work18:15
freemangordonPali: new -thumb is out, i'll wait you to fix it in aster and will forllow for the next update18:15
freemangordons/aster/master/18:15
infobotfreemangordon meant: Pali: new -thumb is out, i'll wait you to fix it in master and will forllow for the next update18:15
* freemangordon prays for his primary device to boot18:16
keriofreemangordon: hm, this is the first time you've updated something in cssu-thumb that hasn't been updated in cssu-testing18:16
freemangordonkerio: no18:16
freemangordonmicrob-engine, Qt, etc18:17
keriohm, were those updates?18:17
kerioi thought you just recompiled18:17
freemangordonno18:17
freemangordonCVE fizes in microb came in -thumb first, for example18:17
freemangordon*fixes18:17
freemangordonsame for some backport in Qt18:18
freemangordonok, booted fine18:18
arceanPali: what's new in maemo-applet-tvout?18:18
freemangordonarcean: nothing18:18
arceanahh, I see18:19
freemangordonkerio: I am pretty confiden there won;t be problems with this update18:19
freemangordonat least I hope so :D18:19
freemangordonkerio: i've been runing updated packages on my device for 2 or more weeks, with no problems18:21
keriofreemangordon: Depends: libc6 (>= 2.5.0-1), zlib1g (>= 1:1.2.1), zlib1g (>= 1:1.2.7)18:23
keriou dun goofd18:23
luffreemangordon: you should change the zlib1g (>= 1:1.2.1) in the zlib package (.shlibs)18:24
freemangordonluf: hmm? isn;t that auto-generated?18:24
freemangordonkerio: besides looking ugly, is there any problem with it?18:26
keriohm, i don't think so18:26
keriounless apt gets confused somehow, but i doubt it18:26
keriojust keep in mind to fix it for the next update :)18:26
freemangordonno, update went fine here18:26
freemangordonyep, for sure18:26
kerioarcean: hm, i don't see the updated qt in -devel18:33
freemangordonkerio: seems it is stuck in /incoming18:34
freemangordonPali: already reported it to merlin199118:34
*** NIN101 has quit IRC18:34
*** NIN102 has joined #maemo-ssu18:35
arceankerio: I've just updated the Qt on my second n900 via cssu-devel :D18:35
arceanhttp://maemo.merlin1991.at/cssu/community-devel/pool/free/q/qt4-x11/18:35
keriofreemangordon: oh, it's because i have -thumb pinned at a slightly higher priority18:35
luffreemangordon: no. gregoa pointed me ;)18:36
freemangordonluf: ok, will keep it in mind for the next update18:36
luffreemangordon: zlib.cssu/debian/zlib1.shlibs: libz 1 zlib1 (>= 1:1.2.1)18:37
freemangordonanyway, it will be soon, i'll put microb-engine build with system libpng and zlib18:37
freemangordonluf: yeah, saw that18:37
luffreemangordon: how you fixed fixed libxml2 build on top of new zlib?18:37
freemangordon1. build libxml2 against old zlib18:38
freemangordon2. install new libxml218:38
freemangordon3. build and install new zlib18:38
dafoxhi all. First off, happy (belated) new year :). Now my question: I was wondering what the status of -thumb is, is it 'stable' now?18:38
freemangordon4. edit libxml2 debian/control to depend on new zlib18:38
keriodafox: hahahahahahahah18:39
keriohahahahahahahahha18:39
kerioyes, it's quite stable actually18:39
freemangordon5. build libxml218:39
freemangordonluf: ^^ is what i did18:39
freemangordonkerio: quite?18:39
dafoxa few days ago I found an announcement post on tmo for 'stable thumb'18:39
dafoxbut what does it mean?18:40
keriodafox: "stable thumb2 executables"18:40
luffreemangordon: So no fix. Ok. I'll put it into my TODO.18:40
kerioas in, running one won't make the cpu blow up because of the silicon errata18:40
dafoxI mean, is it as stable as ssu-stable for example?18:40
freemangordondafox: IIRC what I wrote, it is explained in that very same OP18:40
dafoxaka, 'safe to install'18:40
keriodafox: it works fine for a lot of people18:41
kerioit's about the same as cssu testing, but a bit smaller and faster18:41
dafoxok, I'm willing to give it a go if it is fairly safe. The main benefit seems to be less memory used?18:41
dafoxwhich would be very nice18:41
freemangordondafox: last time we've checked repo stats, there were > 700 people downloaded it18:41
freemangordonthat was some ~months ago18:42
keriodafox: also about 30MB of free space18:42
freemangordon*~3 moths18:42
dafoxcan it be uninstalled, if need be?18:42
freemangordonno18:42
kerionot really, no18:42
dafoxok, so it's not just enable/disable repo?18:42
freemangordonit is SSU18:42
freemangordonthe same as CSSU-testing and CSSU-stable18:43
kerioif you've got some flavour of cssu yes, you just have to enable cssu-testing and cssu-thumb (and disable cssu-stable if you have to)18:43
keriobut there's no way to uninstall it18:43
keriowhat would that even mean? it's a bunch of updates to a lot of system packages18:43
freemangordonkerio: BTW it was for good that I reflashed18:44
keriowhat did you find out?18:44
freemangordonit turned out libpulse dependency is missing from mp-thingie :D18:44
kerio:D18:44
freemangordonit is the same in -testing AIUI18:45
freemangordon(not that i've checked :D)18:45
dafoxfor example, I enable/disable the extras-devel repository on a per-application basis, but it doesn't try to install older versions of those apps when I disable the catalogue again18:45
dafoxis it like that?18:45
kerioyes, but cssu provides system updates18:46
freemangordondafox: this is system upgrade18:46
dafoxI have cssu :)18:46
kerioyou should keep the repos enabled for that18:46
dafoxok, I'll go make some backups then18:46
*** jonwil has quit IRC18:46
dafoxI do keep the cssu-stable (and testing, I think) enabled at all times18:46
* kerio wonders what the big deal is, when the cssu instructions clearly state to make backups *often* with backupmenu18:47
dafoxmaking backups is such a hassle :/18:47
freemangordondafox: disable -stabe if you are on -testing, it jst slows down HAM18:47
freemangordon*-stable18:47
dafoxok, thanks for the tip!18:47
*** _ade_ has joined #maemo-ssu18:48
freemangordondafox: if you install -thumb make sure -testing IS ENABLED, along with nokia repos18:48
dafoxok18:48
_ade_arcean: the replacement clock if now fixed in landscape with the new Qt from cssu-devel18:50
freemangordonnice :D18:50
arceanwow :D18:51
_ade_arcean: other Qt apps I tested look find, no initial rotation to landscape when flat18:51
freemangordon_ade_: are you sure you set Maemo5AutoOriantation?18:51
_ade_I guess so18:51
freemangordonthe fuck :(. Sorry for the typos18:51
arceanremeber, that you need also a new h-d, which is not in cssu-devel yet18:52
_ade_arcean: can you test for yourself?18:52
freemangordonaah, yes18:52
arcean_ade_: sure, give me a link :D18:52
_ade_arcean: You have the replacement clock installed afaik?18:52
freemangordonarcean: are your changes commited on gitorious? in CSSU that is.18:52
dafoxis it the same for extras- btw? If I have extras-testing, can I disable the 'regular' extras repository?18:53
keriodafox: yep18:53
arceanfreemangordon: only in cloned repos18:53
arcean_ade_: yes, on the other device :D18:53
_ade_arcean: so I just have to wait for a new h-d first?18:53
dafoxcool, because ham is really slow, so any speed up is welcome :)18:53
arcean_ade_: I'll push a new version in ~15 minutes18:54
_ade_arcean: I'll wait for that to test further then18:55
*** xes has quit IRC19:01
*** jester` has quit IRC19:03
merlin1991Pali: I see nothing19:17
freemangordonmerlin1991: you broke you glasses? again? :P19:17
DocScrutinizer05everything fine here? (didn't watch since 8h, due to massive overload)19:17
merlin1991freemangordon: what should i see?19:18
freemangordonmerlin1991: NFC. I am just kidding you.19:18
merlin1991damn you ;)19:19
DocScrutinizer05tztztz19:19
freemangordonmerlin1991: on the side not - who do you expect to blacklist that compromised CA?19:19
merlin1991I guess I'll have todo that19:20
DocScrutinizer05uh?19:20
DocScrutinizer05not everything fine so?19:20
freemangordonDocScrutinizer05: http://talk.maemo.org/showthread.php?t=8853519:20
DocScrutinizer05mhm :-/19:21
keriofreemangordon: i suppose it's a maemosec update19:22
kerionot sure who did the last one19:22
freemangordonnokia19:22
kerionope, it was a cssu thing initially19:22
DocScrutinizer05TURKTRUST   omg19:22
freemangordonkerio: no, it was nokia19:22
freemangordonCSSU just distributed it19:22
DocScrutinizer05no19:22
keriofreemangordon: do you trust the turks?19:22
freemangordonkerio: yes, why not19:23
DocScrutinizer05freemangordon: iirc MohammadAG had a fix long before nokia came up with it19:23
freemangordonusually they are nice people. at least those who live in bulgaria19:23
keriobecause they give out intermediate CA certificates instead of SSL certificates, for instance19:23
freemangordonDocScrutinizer05: iirc nokia sent them .debs with the fix19:23
DocScrutinizer05dunno19:23
freemangordons/them/him/19:24
infobotfreemangordon meant: DocScrutinizer05: iirc nokia sent him .debs with the fix19:24
keriofreemangordon: http://gitorious.org/community-ssu/maemo-security-certman19:24
DocScrutinizer05anyway, do we *need* Nokia to fix this silly issue?19:24
freemangordonno19:24
freemangordonwe have the source code19:24
DocScrutinizer05so, go ahead! :-D19:24
freemangordon<freemangordon> merlin1991: on the side not - who do you expect to blacklist that compromised CA?19:25
kerioapparently the fix was done by jum19:25
freemangordon:D19:25
keriojuhani mäkelä19:25
keriosounds finnish19:25
keriosounds nokia19:25
freemangordonyup19:25
freemangordonI wonder why M$ issued 2 updates for that19:26
Palithat was because I published source code of maemo certman aplet :P19:27
freemangordonyep19:27
PaliI got by email source tarball with LGPL license, so I published it on TMO19:28
Paliand that was problem...19:28
Paliso then nokia looked at it and after months released source code19:28
arcean_ade_: I know where's the problem19:28
arcean_ade_: I'll create a merge request19:28
freemangordonarcean: is in adeclock?19:29
Palibtw, you can see some changes in git repo *after* PR.1.3.1: https://gitorious.org/maemo-5-certificate-manager/maemo-security-certman19:29
arceanyes19:29
Pali(that was already included into cssu)19:29
freemangordonarcean: missing autoorientation?19:29
kerio_ade_: why don't you ship .deb packages?19:30
_ade_kerio: see the TMO thread; a bit long to explain here maybe19:31
DocScrutinizer05>>Can certs be revoked manually on N900?<< good question, can't they, at least in CSSU?19:32
kerioDocScrutinizer05: there's no GUI to do that19:32
arcean_ade_: freemangordon: https://gitorious.org/community-ssu/clock-ui/merge_requests/119:32
keriothere's, however, a commandline tool19:32
DocScrutinizer05I seem to remember we at least dicussed adding such GUI19:33
Sicelokerio: which tool? openssl?19:33
DocScrutinizer05...-function19:33
kerioSicelo: cmcli19:34
_ade_arcean: just an order issue then. It could affect other Qt apps?19:34
kerioit's in maemosec-certman-tools19:34
Sicelothanks. will have a look :)19:34
freemangordonarcean: hmm, why is the order important?19:34
freemangordonarcean: this will break more application, I can bet19:35
arceansetWindowFlags(Qt::Window) will always clear the windows flags19:36
freemangordonarcean: according to docs? ok then.19:37
freemangordonBTW why is that neede at all?19:37
freemangordon*needed19:37
arceanI don't know :D, _ade_?19:37
_ade_arcean: it is not a part I did19:38
freemangordonafter all it inherits QMainWindow19:39
arceanI would remove it then19:39
freemangordon_ade_: remove that19:39
_ade_I will19:39
_ade_arean: I have not found other Qt apps with this issue for now. And you where talking about a new h-d that was needed?19:41
kerioSicelo: there's a script on TMO that claims to allow the blacklisting of stuff19:42
arcean_ade_: yes, it's needed. Every window which supports portrait mode should have PORTRAIT_SUPPORTED windows flag19:42
Sicelokerio: i'm already in love with cmcli ;)19:42
_ade_acrean: and the latest available h-d lacks that?19:43
arceanno, it was the setWindowFlags(Qt::Window) which cleared that flag :D19:44
DocScrutinizer05freemangordon: nice spotting :-)19:44
_ade_arcean: okay, so I do have the proper h-d?19:44
arcean2.2.146-2? then yes ;)19:45
_ade_arcean: sorry then, I thought you where going to release a new h-d in 15 mins related to this. Misunderstood ;-)19:46
arceanhehe19:46
arceanI'm still going to release a new version :P19:47
_ade_arcean: removing this line did the trick indeed, thanks19:54
_ade_kerio: btw, Merlin1990 said he would package the replacement clock a long time ago. Maybe it helps if you remind him of that every now and then :)20:01
_ade_s/Merlin1990/Merlin1991/g20:02
infobot_ade_ meant: kerio: btw, Merlin1991 said he would package the replacement clock a long time ago. Maybe it helps if you remind him of that every now and then :)20:02
Sicelokerio: seen the cript now. neat.20:06
tadzikoooh, new thumb \o/20:13
tadzikfreemangordon++20:13
dafoxok, so I'm about to install the -thumb stuff. However when I click the link in the tmo post it tells me that the repo is already installed but disabled (which is correct). But when i try to follow the cssu instructions from step 5,  I don't get any more warnings (ok), but at step 6 I don't see the 'Community SSU' icon. Can I skip this step since I already have cssu?20:23
keriodafox: yep20:24
keriojust tell HAM to check for updates20:25
kerioand do the system update20:25
dafoxok, thanks. I hope this works :)20:25
kerio(connect wallcharger and wait, it'll take a bit)20:25
arceannew hildon-desktop (2.2.147.90) @ cssu-devel20:31
*** Martix has quit IRC20:36
*** Martix has joined #maemo-ssu20:37
merlin1991arcean: what's new in it?20:37
arceanmerlin1991: http://talk.maemo.org/showthread.php?p=1311611#post131161120:40
keriomerlin1991: nothing takes priority over the cert issue20:40
merlin1991freemangordon: you're right, there's more to it than moving the file to a different place20:41
merlin1991damn it20:41
merlin1991kerio: guess what I'm looking at atm20:41
kerio<320:41
keriojust look at how it was fixed on the first cert issue20:42
merlin1991well we have a problem20:42
merlin1991the lists of valid certs are signed20:42
merlin1991(blacklist also)20:42
keriough :s20:43
keriowell, at worst we can just delete the cert and pretend it never existed20:45
keriobut putting it in the blacklist would be better20:45
keriomerlin1991: what checks the signature?20:46
merlin1991kerio: can't do that either, because then a cert that should be in the approved list is missing20:46
keriomerlin1991: cmcli can remove certificates though20:47
keriomerlin1991: we could ask the dude who did the first blacklisting!20:49
merlin1991that's what I'm planning todo20:49
merlin1991hm the cmcli code just unlinks the file in quesiton20:54
keriomerlin1991: we could/should add our own public key to the ones that can sign the cert list, i suppose20:54
DocScrutinizer05who T F is signing a revocation list?21:00
DocScrutinizer05that's utter nonsense21:01
merlin1991well i have no idea why it is signed but both the trusted and the blacklist are signed21:01
merlin1991check /etc/secure/s/certman.*21:03
freemangordonmerlin1991: signed? with which key?21:04
merlin1991gotta check but I think it's with the key in /etc/certs/trusted/root.key21:04
freemangordonmerlin1991: hmm, what is the problem then?21:05
merlin1991hm how do I check if a key is pw protected :D21:05
freemangordonmerlin1991: I bet it is not ;).21:06
freemangordonhowever, gimme some time to remember the command21:06
freemangordonmerlin1991: did you try to do dpkg-buildpackage21:07
freemangordonit could be that signing is a part of package build21:07
merlin1991that works, but it never touches the files / signatures21:07
freemangordonwell, then the signature is generated on install time21:08
freemangordonand we have nothing to worry about21:08
merlin1991the lists are presigned21:08
freemangordonoh, wait21:09
freemangordonmerlin1991: check that commit https://gitorious.org/community-ssu/maemo-security-certman/commit/23037012de495c4dfd2ece1e903b7bb1648168ab21:11
freemangordonseems like you should do whatever magic is needed on the device (or PC) and then copy the result21:11
merlin1991hm I could test that21:12
merlin1991let's see if the signature changes21:12
freemangordonyep21:12
freemangordonmerlin1991: you should yse cmcli to move the certificate in question to blacklist domain21:14
freemangordonaiui21:14
freemangordon*use21:14
freemangordon-a to add -r to remove21:15
freemangordonmerlin1991: d937b34e05fdd9cf9f1216aeb6892feb253a881c21:15
*** _ade_ has quit IRC21:17
merlin1991freemangordon: yep seems to resign21:20
merlin1991though I didn't find that code inside the cmcli source when I skimmed it21:21
freemangordonmerlin1991: are you sure the signature is not regenerated on install time?21:21
merlin1991pretty sure21:22
freemangordonhttps://gitorious.org/community-ssu/maemo-security-certman/blobs/master/debian/maemosec-certman-common-ca.postinst21:23
merlin1991ah wait perl /usr/bin/c_rehash /etc/certs/common-ca21:23
freemangordonexactly21:23
merlin1991interestingly my cert manager only shows me the disabled certs, not the valid ones21:25
freemangordonmerlin1991: on the device?21:26
merlin1991yes21:26
merlin1991the one on stable shows me all21:26
*** _rd has joined #maemo-ssu21:26
freemangordonthe one on thumb shows them all, it is something with your device21:27
freemangordonthough I wonder what's this with microb.21:27
freemangordoni.e. that you should remove the certificate by hand21:28
merlin1991libmaemosec-certman0 diverts the microb-engine security lib away21:31
merlin1991so that should be not neccessary21:31
freemangordonyes, that is why I wonder21:31
freemangordonBTW if new microb is installed after the diversion what will happen?21:32
*** _rd has quit IRC21:32
keriofreemangordon: diversions are a dpkg thing21:33
freemangordonsure21:33
keriomicrob will happily install itself in the diverted location21:33
keriothe diverted files, i mean21:33
freemangordonok, thanks21:33
freemangordonmerlin1991: hmm, i removed that cert, but it is still listed in cpl applet21:33
freemangordonoh, wait, it is on my second device :D21:34
freemangordonhmm, still listed too21:35
keriofreemangordon: i read somewhere that blacklisted certs still show up in the UI21:35
freemangordonkerio: it is not blacklisted, just removed21:35
keriohm, weird21:35
freemangordonwell, i'll reboot21:37
*** Milhouse has joined #maemo-ssu21:39
freemangordonthe fuck, it is still there21:40
freemangordonooh, we have 2 certificates from TÜRKTRUST21:46
freemangordonmerlin1991: the second one is c126ef0d847fc578cabfa616229289c42af952e721:49
*** Milhouse has quit IRC21:54
keriofreemangordon: oh, heh21:55
keriofreemangordon: instead of deleting, can't we blacklist?21:56
freemangordonwe shal do both aiui21:56
freemangordonok. microb refuses to use https with turktrust's site21:57
*** Pali has quit IRC21:57
kerio:D21:58
merlin1991kerio: we'll blacklist21:59
*** _rd has joined #maemo-ssu22:02
freemangordonmerlin1991: hmm, I think we should be very careful22:05
freemangordonsee this https://blog.mozilla.org/security/2013/01/03/revoking-trust-in-two-turktrust-certficates/22:05
merlin1991I'll prepare a test package22:05
freemangordonthey are not blacklisting root certs, just compromised ones22:05
freemangordonaiui22:05
merlin1991we could just blacklist the intermediate ones22:06
freemangordonexactly what we should do22:06
merlin1991but ofc the article doesn't link to them22:06
freemangordoni'll check in IE, if I can find them22:06
* freemangordon starts IE for the first time for the last couple of years :D22:07
keriofreemangordon: they should be a CA, and they fucked that up22:08
keriothey should go out of business22:08
freemangordonkerio: well not exactly22:08
freemangordonread their explanation22:08
freemangordonmerlin1991: yep, found them :D22:09
keriothey're "reviewing"22:09
freemangordonmerlin1991: do you have windozw around?22:09
keriootoh, TURKTRUST only admitted the problem after google revealed it22:09
merlin1991freemangordon: it's windoze, and yes22:10
freemangordonyou'll find both sertificates in "untrusted publishers"22:10
freemangordonmerlin1991: if you want I can export them for you22:11
merlin1991that would be great22:11
merlin1991are you sure those are the intermediate and not the root certificates?22:12
freemangordonyep, you'll see22:12
* freemangordon wonders if openssl can read p7b format22:13
merlin1991fmg afaick the windoze cert storage can export pem format22:15
freemangordonwell, it is .cer, but it is x509 base64 encoded, so yes. I got them both exported, lets see of openssl will like them22:17
freemangordonlooks fine22:19
freemangordonmerlin1991: do you know where .pem file names are derived from? md5 of the file?22:19
merlin1991no idea22:20
freemangordon:(22:20
*** nox- has joined #maemo-ssu22:20
merlin1991I don't even understand why they have a derived name22:20
freemangordonmerlin1991: well, I want to name the files correctly before giving them to you22:22
merlin1991we'll have to email Juhani Mäkelä then22:30
merlin1991I have no idea how those file names are generated22:30
freemangordonmerlin1991: gimme some time22:30
freemangordonit is some hash, i just need to find which one22:30
freemangordonmerlin1991: hmm, could be certificate serial number22:31
merlin199172:03:21:05:c5:0c:08:57:3d:8e:a5:30:4e:fe:e8:b0 => 408f39269c4c862399c65109a6e6f2c1fea7f6b7.pem how?22:35
merlin1991:D22:35
freemangordonseems it is not that :D22:35
freemangordonmerlin1991: hmm, i'll install them and cmcli will give me their names :D22:38
freemangordonhehe:22:40
freemangordon64fb1b863db84af24482f9563dea26c0f4e3b334.pem22:40
freemangordonef918e90bbc79150e645852bd7b6fa8145fe0063.pem22:40
*** _rd has quit IRC22:41
*** _rd has joined #maemo-ssu22:42
freemangordonmerlin1991: http://merlin1991.at/~freemangordon/turktrust_blacklist/22:42
freemangordonplease check them, just in case22:42
merlin1991hm how do I see that those are intermediate cas?22:44
freemangordonopenssl x509 -text -in $CERT22:46
freemangordon?22:46
merlin1991well my knowledge of all the fileds amounts to almost 022:49
merlin1991so I can't tell what I'm seeing :D22:49
freemangordonadd those as blacklisted on your device22:50
freemangordonand check in settings applet22:50
freemangordonhmm. maybe i should have exported the whole chain22:52
freemangordonmerlin1991: looks ok there22:56
*** NIN102 has quit IRC22:56
merlin1991yep it does22:57
freemangordonmerlin1991: though I wonder why IE shows one of them as issued to 8.google.com22:59
freemangordon*.google.com23:00
freemangordonbut maemo shows some cryptic turcish name23:00
freemangordon*turkish23:00
freemangordon"e-islem.kktcmerkezbankasi.org"23:00
freemangordonmerlin1991: oh, wait23:01
freemangordoni didn;t export that one23:01
freemangordoni'll have to add it too23:01
freemangordonmerlin1991: we have 3, not 2 certificates to blacklist23:03
merlin1991but everywhere else it states 2 certs to blacklist23:03
freemangordonbut M$ updated 3 - 2 intermediate and the one used for MTIM attack23:04
freemangordongot it?23:04
freemangordonthe others talk abou intermediate certificates only23:04
freemangordonbut there was one issued to *.google.com23:04
freemangordonusing one of the intermediate certs23:05
merlin1991ah k23:06
freemangordonok, uploadad23:06
merlin1991well if we blacklist the intermediate the one issued to *.google.com has no valid chain anyway23:06
freemangordonI know23:06
freemangordonI wonder why M$ did it that way23:06
keriowhy are we still trusting that CA? really23:07
freemangordonkerio: we can only follow M$/Mozilla, etc23:07
freemangordonwe are not to decide on that matter23:08
freemangordonkerio: i read somewhere that lots of turkish banks use this CA23:08
keriolet's follow mozilla pls :323:08
freemangordonif we blacklist them... won;t be good23:08
freemangordonkerio: that is what we are doing23:09
freemangordonmerlin1991: well, it is up to you to decide whether to include *.google.com certificate23:09
freemangordonbut if it was me, i'd rather include it23:09
merlin1991the *.google.com cert has no chain wihout the other one23:10
keriofreemangordon: if your bank uses that CA, you should probably refrain from using your bank's website23:10
merlin1991no point in spamming our cert db23:10
freemangordonmerlin1991: check you blacklist store to see what I mean23:10
keriomerlin1991: +1, the *.google.com is just the only certificate that we know of23:10
keriothere might be a lot more23:10
freemangordonmerlin1991: cmcli -c blacklist -L23:11
freemangordonmerlin1991: though you are most probably right and we should not include that23:11
merlin1991check it before or after I inlcuded the cert?23:12
freemangordondoes not matter23:12
merlin1991so you're talking about the login.yahoo.com certs?23:13
freemangordonmerlin1991: however, I am convinced we should not include *.google.com23:13
freemangordonmerlin1991: yep23:13
freemangordonand addons.mozilla.org23:13
freemangordonbut all of them are signed by root cert23:14
freemangordon*.google.com is not23:14
freemangordonso i agree it should not be included23:14
merlin1991okay :)23:15
merlin1991hm which link to quote in the git commit message?23:15
freemangordonplus we can always include it later if mozilla includes it23:15
freemangordonBTW it will be funny if we are the next after google and M$ to blacklist those23:16
freemangordonbefore mozilla and major linux distributions :D23:16
merlin1991anyway which link to quote?23:17
freemangordonthe one from thread on TMO23:17
freemangordongoogleblogspot or something23:17
freemangordonhttp://googleonlinesecurity.blogspot.com/2013/01/enhancing-digital-certificate-security.html23:17
freemangordonmerlin1991: BTW Pali pushed a commit today to mp- thingie, you may want to revert it as it is briken23:21
freemangordon*broken23:21
merlin1991what's wrong with it?23:23
freemangordondoes not work :D23:23
freemangordonyou can't build -mp23:23
merlin1991rofl23:23
merlin1991okay building new certman stuff now23:25
merlin1991update is in testing-testing23:40
freemangordongoing to test it23:42
freemangordonmerlin1991: what is the repo?23:43
merlin1991the one you enable with http://cdnm.at/~christian/maemo/cssu/testing-testing-enabler_0.1_all.deb ;)23:43
freemangordonhmm. maybe it won;t work on -thumb23:46
merlin1991it's T7.223:46
freemangordonaah, ok23:46
freemangordongood, only certman packages are updated23:48
merlin1991hehe had todo it via x11vnc23:50
merlin1991my wireless does not go into this room :P23:50
freemangordonrebooting, lets see23:51
freemangordonmerlin1991: something is broken, I see only blacklisted certificates23:53
freemangordondid you put hash file too?23:53
merlin1991yes23:53
merlin1991ah wait I have an idea23:54
merlin1991I bet apt is at fault23:54
keriomerlin1991: you said you could only see the blacklisted certificates, right?23:54
kerioeven before, i mean23:54
merlin1991yep, but only on my testing device23:54
freemangordonmerlin1991: doing perl /usr/bin/c_rehash /etc/certs/common-ca don't help23:56
merlin1991freemangordon: check /etc/secure/s/certman.*23:56
merlin1991then you'll see why23:56
kerio...config files huh23:57
kerio.dpkg-old23:57
merlin1991yep23:57
merlin1991nah it's .dpkg-dist23:57
keriooh, i was just guessing :)23:57
keriomerlin1991: it's a bug, remove them from the conffiles list23:57
freemangordononly certman.blacklist.dpkg-dist23:57
merlin1991but anyway move that to get rid of your manually generated hash file and you'll see the certs again23:57
kerioclearly they aren't config files in the usual sense23:58
freemangordonaah, yes23:58
merlin1991kerio: they are in no conffile list, they are conffiles automatically because they sit in /etc23:59

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