| *** AndrewX192 has quit IRC | 00:10 | |
| *** AndrewX192 has joined #maemo-ssu | 00:10 | |
| *** AndrewX192 has quit IRC | 00:12 | |
| *** AndrewX192 has joined #maemo-ssu | 00:13 | |
| *** AndrewX192 has joined #maemo-ssu | 00:13 | |
| *** _rd has joined #maemo-ssu | 00:19 | |
| *** int_ua has quit IRC | 00:56 | |
| *** kolp has joined #maemo-ssu | 01:08 | |
| *** NIN101 has quit IRC | 01:25 | |
| *** kolp has quit IRC | 01:27 | |
| *** _rd has quit IRC | 01:45 | |
| *** Martix has quit IRC | 01:57 | |
| *** peetah has quit IRC | 02:42 | |
| *** peetah has joined #maemo-ssu | 02:43 | |
| *** jonwil has joined #maemo-ssu | 03:01 | |
| jonwil | ~seen pali | 03:02 |
|---|---|---|
| infobot | pali <~pali@Maemo/community/contributor/Pali> was last seen on IRC in channel #maemo-ssu, 5d 13h 41m 58s ago, saying: 'freemangordon, do you know state of portrait keyboard?'. | 03:02 |
| *** arcean has quit IRC | 04:10 | |
| *** M4rtinK has quit IRC | 04:46 | |
| *** ChanServ sets mode: -o merlin1991 | 04:48 | |
| merlin1991 | hm not exactly the channel I meant :P | 04:48 |
| *** nox- has quit IRC | 05:00 | |
| * merlin1991 currently sets up a qemu-kvm x86 image to test some theories regarding compiling | 05:39 | |
| merlin1991 | great fun! | 05:39 |
| *** amiconn has quit IRC | 05:49 | |
| *** amiconn_ has joined #maemo-ssu | 05:49 | |
| *** amiconn_ is now known as amiconn | 05:49 | |
| *** DocScrutinizer05 has quit IRC | 06:04 | |
| *** DocScrutinizer06 has joined #maemo-ssu | 06:04 | |
| *** peetah has quit IRC | 06:29 | |
| *** peetah has joined #maemo-ssu | 06:29 | |
| *** DocScrutinizer06 is now known as jOERG_rw | 06:30 | |
| *** jOERG_rw is now known as DocScrutinizer06 | 06:30 | |
| *** DocScrutinizer06 is now known as DocScrutinizer05 | 06:30 | |
| *** joshgillies has joined #maemo-ssu | 07:22 | |
| *** sourav has joined #maemo-ssu | 07:40 | |
| *** sourav has left #maemo-ssu | 07:55 | |
| merlin1991 | freemangordon: how on earth did you build microb-engine | 08:34 |
| merlin1991 | I tried to build it on a clean git clone on x86 and even there it failed | 08:35 |
| freemangordon | merlin1991: by using vmware image | 09:28 |
| *** joshgillies has quit IRC | 09:41 | |
| *** Pali has joined #maemo-ssu | 10:33 | |
| freemangordon | jonwil: ^^^ | 10:34 |
| jonwil | hi pali | 10:34 |
| Pali | jonwil, hi | 10:34 |
| jonwil | I have been working with HEAD of http://gitorious.org/community-ssu/operator-name-cbs-widget/ | 10:35 |
| jonwil | and I have some things to say regarding some changes I made | 10:35 |
| *** inz has joined #maemo-ssu | 10:37 | |
| jonwil | Is there a reason you need to specifically listen to Phone.Net.operator_name_change? | 10:37 |
| jonwil | You should not be listening to that directly as the code behind connui_cell_net_status_register listens for you | 10:38 |
| *** luf has joined #maemo-ssu | 10:40 | |
| *** _rd has joined #maemo-ssu | 10:44 | |
| jonwil | This patch | 10:50 |
| jonwil | http://pastebin.com/YWqAA67N | 10:50 |
| jonwil | contains the changes I have in my local copy of operator-name-cbs-widget | 10:50 |
| jonwil | Without these changes I get an incorrect operator name on my N900 | 10:50 |
| jonwil | With these changes I get the correct operator name | 10:50 |
| jonwil | I would like to discuss these changes so we can find out whether my reasoning as to why they need to be made is correct or if not, what the correct change is so that I get the right operator name but others also get the correct operator name (e.g. those who roam to other operators) | 10:51 |
| jonwil | hmmm, guess no-one wants to talk about my changes... :P | 10:54 |
| freemangordon | jonwil: this is IRC after all, give Pali some time :P | 10:56 |
| jonwil | :) | 10:56 |
| Pali | jonwil, if It fixing your bug, commit changes to git | 10:58 |
| *** _rd has quit IRC | 10:58 | |
| Pali | and push new deb package to cssu-devel for testing | 10:58 |
| merlin1991 | freemangordon: any idea what's different from the vmware image to a regular scratchbox installation? | 10:59 |
| freemangordon | NFC | 10:59 |
| Pali | I added that dbus signal because more times I got "empty string" as operator name | 10:59 |
| merlin1991 | freemangordon: do you still have a copy of the virgin image? | 10:59 |
| freemangordon | yep | 11:00 |
| merlin1991 | can you upload it to the server ? | 11:00 |
| freemangordon | and iirc thedead1441 has it online | 11:00 |
| freemangordon | *1441 | 11:00 |
| freemangordon | the fuck :D | 11:00 |
| merlin1991 | I'll ask him then | 11:00 |
| freemangordon | thedead1440 | 11:00 |
| freemangordon | :) | 11:00 |
| Pali | jonwil, if I get "empty string" as operator name again I will add that dbus signal but name will be readed only if operator name will b empty | 11:00 |
| jonwil | I suspect the reason you saw the empty strings is because of the typo where you had priv->operator_name = g_strdup(state->operator_name); instead of priv->operator_name = g_strdup(operator); | 11:00 |
| jonwil | i.e. it was pulling state->operator_name (which was empty) instead of operator (which was not empty) | 11:01 |
| freemangordon | merlin1991: http://talk.maemo.org/showpost.php?p=1310425&postcount=153 | 11:03 |
| jonwil | As for pushing to git, my Gentoo box is having issues right now and wont boot up so I cant push. I do have a .deb file on my phone that matches HEAD plus that patch though. | 11:04 |
| luf | freemangordon: I have vcs for osso-gnomevfs-extra :) | 11:05 |
| luf | I have to check if it's the latest but I think so based on debian/changelog | 11:05 |
| jonwil | I can make the .deb I have available to anyone who wants it (or to anyone who can upload it to cssu-devel) | 11:07 |
| luf | jonwil: for cssu-devel import is needed _armel.changes and all files mentioned there (with the right checksum) | 11:08 |
| jonwil | hmmm ok | 11:08 |
| jonwil | is there someone here who can help by putting the patches into GIT and building for me since my build env is down? | 11:08 |
| luf | jonwil why such terrible namelog (no function or define)? Why don't log it into syslog? | 11:11 |
| *** joshgillies has joined #maemo-ssu | 11:12 | |
| luf | jonwil: the last thing in patch: namelog after namelog ? Why can't you join this two namelogs? | 11:13 |
| Pali | luf, there is gconf key for enabling logs | 11:13 |
| Pali | luf, for debugging is was easier to have two log files | 11:14 |
| luf | syslog.conf isn't enough? come on. But I see there are namelog used already in the cbms so it should be ok - but I hate this work. Just to copy X times the same instead of define or some function call. | 11:15 |
| merlin1991 | freemangordon: I answered nicolai for now | 11:16 |
| luf | jonwil: please fix the last namelog direclty after namelog and I can compile and put it info cssu-devel. | 11:16 |
| merlin1991 | btw freemangordon on that vmware image your scratchbox there, does it have all the cssu packages installed or only the latest build deps? | 11:19 |
| merlin1991 | because that's one thing I didn't try, building on a virgin target, without newer cssu packages | 11:19 |
| Pali | merlin1991, I have patch for dsme & upstart. when you call sudo poweroff it will tell it to dsme for clean shutdown | 11:27 |
| Pali | similar patch as for reboot | 11:27 |
| Pali | but now I added patch for dsmetool for adding parameter -p for poweroff | 11:27 |
| merlin1991 | hm we should leave poweroff as is | 11:28 |
| jon_y | Pali: hey, I have a 1-line patch to get the kernel building for newer toolchain | 11:28 |
| Pali | merlin1991, patch is for telinit | 11:28 |
| Pali | same as for rebooting | 11:29 |
| Pali | it is funny that busybox poweroff on maemo calling telinit 0 | 11:29 |
| merlin1991 | well put the patch somewhere so I can look at it more closely, atm I'm not really sure I'm not talking utter bs | 11:29 |
| merlin1991 | btw any idea what the shut down button in the power key menu does? | 11:30 |
| Pali | calling some dsme dbus method | 11:31 |
| jon_y | Pali: http://paste.debian.net/221282/ | 11:31 |
| jon_y | binutils more strict now | 11:31 |
| merlin1991 | does that still work with the old toolchain aswell? | 11:31 |
| jon_y | I think binutils ignores unknown pseudo ops | 11:32 |
| Pali | merlin1991, here is first patch for dsme: http://pastebin.com/zZXcYPTS | 11:32 |
| jon_y | let me post the whole file, you can compile it to test | 11:32 |
| Pali | merlin1991, and here is second patch: http://pastebin.com/Fg9D0BCB | 11:33 |
| jon_y | merlin1991: http://paste.debian.net/221284/ | 11:34 |
| *** kolp has joined #maemo-ssu | 11:34 | |
| freemangordon | merlin1991: powerkeymenu does mce callback via dbus | 11:35 |
| freemangordon | (for poweroff) | 11:35 |
| Pali | jon_y, why is that patch needed? | 11:35 |
| freemangordon | luf: check zlib on gitorious, "your" patch is the way to go, toldya I was lucky with libpng :D | 11:35 |
| jon_y | binutils was erroring out on the smc line | 11:36 |
| jon_y | well, gas was | 11:36 |
| Pali | jon_y, ok, it is upstream patch? | 11:36 |
| Pali | or did you write it? | 11:36 |
| freemangordon | luf: I backported another NEON patch in zlib, that one is decreasing compression time by ~18% ;) | 11:36 |
| jon_y | my patch | 11:36 |
| jon_y | gas says "this processor does not support smc extensions" | 11:36 |
| freemangordon | ~gas | 11:37 |
| infobot | i guess gas is so much cheaper in the US anyway it wont make any difference | 11:37 |
| jon_y | gnu assembler | 11:37 |
| Pali | GNU assembler | 11:37 |
| freemangordon | aah, ok | 11:37 |
| luf | I told you that your patch will help. | 11:37 |
| luf | freemangordon: How you measure such big speedup? | 11:38 |
| luf | *do | 11:38 |
| freemangordon | jon_y: IIRC you can pass some command line option instead of modifying the code | 11:38 |
| jon_y | is it? | 11:39 |
| jon_y | ok, I just checked with older binutils, bad :( | 11:39 |
| freemangordon | luf: with zpipe < $BIGFILE > /dev/null | 11:39 |
| jon_y | omap3_rng_call_rom_asm: | 11:39 |
| freemangordon | jon_y: AFAIK that modification is upstreamed | 11:39 |
| luf | freemangordon: ok. | 11:39 |
| jon_y | oops | 11:39 |
| jon_y | unknown pseudo-op: `.arch_extension' | 11:40 |
| jon_y | freemangordon: oh, how do they deal with old binutils? | 11:40 |
| freemangordon | jon_y: not using smc | 11:40 |
| jon_y | they encode .dword directly? | 11:41 |
| freemangordon | yep | 11:41 |
| jon_y | ok, that's fine too, I have no idea what the opcode is | 11:41 |
| jon_y | Pali: fix it in kp53 pls :) | 11:41 |
| freemangordon | something "0xEEADBEEF", check in uboot or kernel-power | 11:42 |
| Pali | jon_y, kp52 was not released :-) | 11:42 |
| Pali | I will look at it | 11:42 |
| jon_y | Pali: thanks | 11:42 |
| Pali | now I'm fixing ipv6 support :-) | 11:42 |
| merlin1991 | luf and freemangordon if you guys keep backporting real *speed* patches maemo will get too fast! | 11:42 |
| jon_y | Pali: can you also enable the vlan module? | 11:43 |
| luf | merlin1991: feel free to put there some sleep into the code :D | 11:43 |
| jonwil | luf: try this new patch | 11:43 |
| jonwil | http://pastebin.com/XaWvPzfy | 11:43 |
| freemangordon | merlin1991: don;t worry, we'll integrate something eating 50% of the CPU all the time | 11:43 |
| Pali | seems that ipv6 support for wifi over icd2 is working without problem | 11:43 |
| jonwil | should be the right now one | 11:43 |
| freemangordon | to not break "the user experience" :D | 11:43 |
| merlin1991 | sleep("over 9000"); //this stuff is way to fast we can't pass the speed of light on arm | 11:43 |
| Pali | seems that dhcpv6 and radvd detection working | 11:43 |
| jon_y | merlin1991: speed up loops :) | 11:43 |
| merlin1991 | s/to/too/ | 11:44 |
| infobot | merlin1991 meant: sleep("over 9000"); //this stuff is way too fast we can't pass the speed of light on arm | 11:44 |
| jonwil | if someone can help me get that last patch into Git and compile a package for cssu-devel that would be great | 11:44 |
| jon_y | Pali: also, any plans to include fcam drivers as part of the kp build? | 11:44 |
| luf | jonwil: I'll do it (hopefully in next hour or today) | 11:44 |
| * jon_y found out new custom kernels will bootloop on cssu-thumb | 11:45 | |
| jonwil | ok | 11:45 |
| jonwil | thanks | 11:45 |
| Pali | jon_y, no time & I do not using fcam... | 11:45 |
| luf | jonwil: please give me also patch for debian/changelog | 11:45 |
| jon_y | Pali: ok, do at least the vlan module, thanks | 11:45 |
| merlin1991 | jon_y: well ou do need the thumb patches, otherwise you're in for a bazillion sigills :D | 11:45 |
| Pali | jon_y is it CONFIG_VLAN_8021Q? | 11:45 |
| jon_y | yes that | 11:45 |
| Pali | ok, going to enable it | 11:45 |
| jon_y | merlin1991: well, it was KP51r1 and some newer patches | 11:46 |
| merlin1991 | btw freemangordon did you seperate the thumb patches cleanly into git commits? | 11:46 |
| jonwil | I cant edit debian/changelog myself because my linux box is down | 11:46 |
| freemangordon | merlin1991: yep, I had upstreaming in mind when I did them | 11:47 |
| luf | jonwil: so what should I put there? | 11:47 |
| jonwil | otherwise I would edit the changelog and push to Git myself | 11:47 |
| merlin1991 | jon_y: so you'll probably will need just this: http://gitorious.org/community-ssu/kernel-cssu/commit/53b42852b56c7460034f651a5806eba25061bfa8 | 11:47 |
| *** xmlich02_ has quit IRC | 11:48 | |
| jonwil | Changelog items should be: | 11:48 |
| freemangordon | jon_y: https://gitorious.org/community-ssu/kernel-cssu/commit/ce27033560ddf253f25738513bcdab5f059440c6 | 11:48 |
| freemangordon | https://gitorious.org/community-ssu/kernel-cssu/commit/6c5e10ed0a648f4fa93cd622d21e65f71799c813 | 11:48 |
| jon_y | freemangordon: it's not in Pali's git repo? | 11:49 |
| freemangordon | jon_y: it is | 11:49 |
| freemangordon | but I made the orginal patches in kernel-cssu | 11:49 |
| Pali | I will merge kp52 into cssu-kernel when kp52 will be in extras-devel | 11:49 |
| jon_y | ok, I'm using pali's git tree in garage | 11:49 |
| jonwil | * Add logging to update_widget to log how it determines which names to use | 11:49 |
| Pali | btw, is extras-devel still in read-only mode? | 11:50 |
| Pali | and when will be wiki working???? | 11:50 |
| merlin1991 | Pali: extras devel works | 11:50 |
| merlin1991 | but only via ssh/dput | 11:50 |
| freemangordon | jon_y: depending on the kernel version you may also need https://gitorious.org/community-ssu/kernel-cssu/commit/07002d818c0c8b539d85da142e3caa39eec4e560 | 11:50 |
| Pali | merlin1991, I using dput | 11:50 |
| merlin1991 | wiki is technically ready, but nokia didn't swap the dns entries yet | 11:50 |
| Pali | so extras-devel was migrated? | 11:50 |
| Pali | extras dns was updated? | 11:51 |
| jonwil | * do not listen to operator_name_change as connui_cell_net_status_register does that for you | 11:51 |
| merlin1991 | Pali: nope to both, but afaik extras-* never went read only | 11:51 |
| jonwil | * Correct typo in calculation of operator name | 11:51 |
| merlin1991 | I could be wrong though | 11:51 |
| jon_y | freemangordon: actually, I already have all 3 patches | 11:52 |
| Pali | e.g I do not want to see that some my packages was dissapear | 11:52 |
| jonwil | * Use service provider name when service_provider_name_type is != 1) | 11:52 |
| jon_y | Pali: since you're working on ipv6, any ideas if the n900 know about 6-4 tunnels? | 11:52 |
| luf | jonwil: I also don't know your credentials (name/mail). | 11:53 |
| Pali | jon_y, I reused diablo packages (+ one closed source for icd2) | 11:53 |
| Pali | and one shell script creating some 6-4 tunnel | 11:53 |
| Pali | interface | 11:53 |
| luf | Pali: Is version for operator-name-cbs-widget 3.2.1-0maemo1+0cssu1 ok? | 11:54 |
| jon_y | nice, will it be part of the kp install? | 11:54 |
| Pali | but I do not know if that closed icd module using it or not | 11:54 |
| Pali | luf, why both maemo and cssu? | 11:54 |
| luf | Just cssu? | 11:55 |
| Pali | I think that we had some decision that cssu suffix is not for our upstream packages... | 11:55 |
| luf | Pali: Ah ok. So 3.2.1 is ok? | 11:56 |
| Pali | (e.g packages which have upstream on gitorious) | 11:56 |
| Pali | 3.2 is last version, so 3.2.1 should be ok | 11:56 |
| Pali | merlin1991, where is info about cssu packages versioning? | 11:57 |
| merlin1991 | wiki | 11:57 |
| merlin1991 | and you are right | 11:57 |
| Pali | some packages have cssu suffix | 11:57 |
| Pali | some not | 11:57 |
| merlin1991 | when our gitorious is upstream we have no cssu suffix | 11:57 |
| Pali | some has unstable distro | 11:57 |
| Pali | some has stable | 11:57 |
| Pali | some has fremantle... | 11:57 |
| merlin1991 | well that is just due to the inconsistance in the source we're working with :/ | 11:57 |
| Pali | some has debian version suffix (version-1) | 11:57 |
| merlin1991 | nobody ever bothered to move that properly around | 11:58 |
| luf | It was my mistake. I never wrote new package for cssu ;) | 11:58 |
| Pali | maybe someboty should write proper (wiki) page about versioning... | 11:58 |
| Pali | jonwil, btw, in diablo there is (open) binary rtsol which ask for M and O bits and then icd decide if dhcpv6 client is needed or not | 11:59 |
| jonwil | yes what about this binary? | 12:00 |
| *** DocScrutinizer51 has quit IRC | 12:00 | |
| Pali | but there is bug (in icd maybe?) that if only O bit is set it will not update dhcpv6 config file | 12:00 |
| Pali | and it trying to get also address (not only dns) | 12:00 |
| Pali | and then dhcp client will timeout... | 12:01 |
| Pali | I will fix it | 12:01 |
| *** DocScrutinizer51 has joined #maemo-ssu | 12:01 | |
| jonwil | Is there a reason you are mentioning this to me specifically? Something you want me to look at? | 12:01 |
| Pali | you asked about ipv6 | 12:02 |
| freemangordon | the fuck, somehow I managed to break my SB :( | 12:04 |
| *** _rd has joined #maemo-ssu | 12:04 | |
| luf | freemangordon: zlib? | 12:04 |
| jonwil | I dont remember asking about ipv6 but maybe it was a while back | 12:04 |
| jonwil | :) | 12:04 |
| freemangordon | luf: no | 12:04 |
| freemangordon | install: `debian/libnss3-dev/DEBIAN' exists but is not a directory | 12:04 |
| freemangordon | dh_gencontrol: command returned error code 256 | 12:04 |
| freemangordon | I got this error with some packages | 12:04 |
| luf | Pali: may I insert into changelog as maintainter you or jonwil? | 12:05 |
| jonwil | what does changelog currently say as maintainer? | 12:05 |
| luf | Pali | 12:05 |
| luf | jonwil: ^^^ | 12:06 |
| jonwil | Go ahead and put me if you like, I dont mind either way | 12:08 |
| Pali | luf, just add entry to changelog | 12:08 |
| jonwil | leave maintainer as-is I guess | 12:09 |
| Pali | merlin1991, did you looked at dsme+upstart patches? | 12:09 |
| luf | Pali: sure I mess something in my head. :D | 12:09 |
| merlin1991 | Pali: yep, they look fine | 12:10 |
| Pali | merlin1991, ok can you create dsme git repo on gitorious cssu? | 12:12 |
| luf | I push it to gitorious ... | 12:12 |
| Pali | I used upstream dsme repo and cut it in fremantle version | 12:12 |
| merlin1991 | where's upstream dmse? | 12:12 |
| jonwil | ok, so does anyone have anything they want me to reverse engineer? Would this closed-source bit being used for IPv6 be worth reverse engineering? | 12:13 |
| Pali | merlin1991: gitorious.org/meego-middleware/dsme | 12:13 |
| Pali | jonwil, I needed to RE libicd-network-ipv6, but now I hacked it with shell scripts... | 12:14 |
| Pali | so now it is not needed | 12:14 |
| jonwil | ok | 12:14 |
| Pali | it working | 12:14 |
| jonwil | libicd-network-ipv6 comes from where? | 12:14 |
| merlin1991 | Pali: https://gitorious.org/community-ssu/dsme | 12:14 |
| Pali | merlin1991, here is upstream meego repo: https://meego.gitorious.org/meego-middleware/dsme | 12:14 |
| Pali | I found here last fremantle version and I commited some fremantle changes | 12:15 |
| Pali | ok | 12:15 |
| kerio | ohai Pali | 12:15 |
| Pali | jonwil, diablo SSU repository and was pushed to extras-devel by me | 12:15 |
| kerio | jonwil: libicd-network-ipv6 is in extras-devel non-free | 12:15 |
| Pali | qgil allowed it | 12:15 |
| Pali | (there was meeting about it) | 12:16 |
| jonwil | ok, so it was a closed-binary in diablo and was originally uploaded to extras-devel for use with Fremantle but is now no longer necessary because of whats been done? | 12:16 |
| kerio | Pali: is there a way to reliably detect if there's a kernel attached to the flashed uboot? | 12:16 |
| Pali | jonwil, yes but it is still needed | 12:16 |
| Pali | there was some problem with connections in that library | 12:17 |
| jonwil | ok | 12:17 |
| Pali | but I fixed that problem with shell scripts | 12:17 |
| jonwil | ok | 12:17 |
| Pali | so RE modification is not needed | 12:17 |
| Pali | because seems it working | 12:17 |
| Pali | kerio, yes | 12:17 |
| Pali | kerio, attached kernel must be at specified address | 12:17 |
| Pali | so look at that offset | 12:17 |
| kerio | Pali: could you add that check to u-boot-update-bootmenu, or would you add it if someone gave you the patch? | 12:18 |
| Pali | and check if there is something or not (only 0x00 or 0xFF) | 12:18 |
| jonwil | one of these days I hope we end up with clones of various ICD and related closed blobs | 12:18 |
| jonwil | or replacements | 12:18 |
| Pali | jonwil, when I will have time I will look at netcfg | 12:19 |
| jonwil | netcfg is? | 12:19 |
| Pali | it sounds like good replacement | 12:19 |
| kerio | Pali: right now, u-boot-update-bootmenu only tries to find the kernel version string, and if that fails, it just adds a generic "boot from attached kernel" entry | 12:19 |
| Pali | netcfg is configuration tool for arch linux | 12:19 |
| jonwil | main problem with icd is not finding replacement, its understanding what the closed blobs expose to the world | 12:19 |
| Pali | jonwil, it is set of bash scripts which configuring networks | 12:19 |
| luf | Pali, jonwil: where can I get libconnui-dev ? (Unmet build dependencies: libconnui-dev and apt-get install can't find it) | 12:19 |
| Pali | luf, in cssu gitorious should be | 12:20 |
| freemangordon | luf: in CSSU repo | 12:20 |
| luf | Ugh. ok. | 12:20 |
| Pali | jonwil, that arch netcfg is modular bash system and writing new module is easy | 12:20 |
| jonwil | http://gitorious.org/community-ssu/libconnui-dev | 12:20 |
| jonwil | Main problem with replacing icd is understanding how other parts of the system talk to icd and its related bits | 12:21 |
| jonwil | lots of bits of code (including lots of bits of closed code) talk to the ICD things in various ways | 12:21 |
| jonwil | and finding out how those bits talk is something I have been working on as much as possible | 12:22 |
| Pali | jonwil, 3rd apps should talk to icd only via libconic and that library is open source | 12:22 |
| freemangordon | guys, what about taking a break from RE stuff and finish/polish what is already in CSSU? | 12:22 |
| Pali | so 3rd apps should not be problem | 12:23 |
| freemangordon | just a proposition ofc | 12:24 |
| *** lizardo has joined #maemo-ssu | 12:25 | |
| jonwil | Plenty of closed source blobs talk to ICD via ways not fully understood | 12:25 |
| jonwil | but yeah I think I might go back and continue to reverse engineer that stuff | 12:25 |
| Pali | better for RE is clockd or iphbd | 12:26 |
| Pali | or system-ui daemon | 12:26 |
| Pali | hm, somewhere on irc is list of all closed daemons | 12:27 |
| freemangordon | yeah, clockd or systemui should be relatively easy to RE | 12:27 |
| Pali | and iphbd should be too | 12:27 |
| freemangordon | both have x86 binaries and are some 20-30 KB | 12:27 |
| Pali | because iphbd daemon was merged to meego dsme | 12:27 |
| Pali | and that is open source | 12:27 |
| Pali | so I think that they moved full source code to dsme module or file | 12:27 |
| Pali | need to look into dsme git history | 12:28 |
| jonwil | We dont need to reverse engineer entire icd daemon, we just need to reverse engineer the dbus/gconf/etc interfaces it exposes to the outside world that other parts of the system rely on | 12:28 |
| jonwil | so that a replacement of icd2 will not break those closed blobs | 12:28 |
| Pali | they (nokia) also added conflicts: and replaces: into debian/control to dsme package | 12:28 |
| luf | jonwil: please install and check prior I import it into cssu-devel http://merlin1991.at/~luf/operator-name-cbs-widget/connui-home-cellular_3.2.1_armel.deb | 12:30 |
| freemangordon | luf: http://merlin1991.at/~freemangordon/zlib/ | 12:32 |
| freemangordon | feel free to benchmark/test :D | 12:32 |
| luf | freemangordon: next round of libpng recompilation is waiting for us :) | 12:33 |
| Pali | merlin1991, dsme and upstart patches commited | 12:33 |
| freemangordon | luf: wait, why recompile? | 12:33 |
| freemangordon | isn;t it using system zlib? | 12:34 |
| luf | freemangordon: due to static linking? | 12:34 |
| freemangordon | hmm, is it statically linked? I doubt, but lemme check | 12:34 |
| jonwil | ok, testing that .deb now to make sure it displays the correct name | 12:34 |
| luf | freemangordon: In -dev it is. | 12:34 |
| freemangordon | luf: we'd better fix it to use zlib .so | 12:35 |
| jonwil | luf: seems to work | 12:35 |
| jonwil | so go ahead and push to cssu-devel | 12:35 |
| luf | freemangordon: And someone can use statically linked libpng (e.g. browser) | 12:35 |
| freemangordon | luf: aah, yes, APNG | 12:35 |
| freemangordon | luf: I will add APNG patch to libpng | 12:36 |
| luf | freemangordon: So I think we want to recompile libpng just in case ;) | 12:36 |
| luf | And apng is good reason ;) | 12:36 |
| freemangordon | luf: why do you think libpng is statically linked to zlib? | 12:36 |
| luf | jonwil: done. Please announce it on TMO | 12:38 |
| luf | freemangordon: Because of libpng12-dev.install: libpng12.a | 12:39 |
| Pali | luf, if you have scratchbox up, can you compile dsme & upstart from git? | 12:40 |
| luf | Pali: I'll leave for a lunch and need to do something prior lunch. So not now. I'll see if I have time after lunch. | 12:41 |
| Pali | ok | 12:41 |
| jonwil | I dont know right tmo thread to post in... | 12:42 |
| jonwil | Pali, can you see if running the new 3.2.1 connui-home-cellular gives you any specific issues? | 12:43 |
| Pali | jonwil, I will test if it working, but I had problems in Prague (and I'm not here yet) | 12:44 |
| jonwil | ok | 12:44 |
| jonwil | If you get issues, don't just go making changes, please talk to me first (via IRC or via email) with some logs and together we can figure out just whats going on :) | 12:45 |
| luf | jonwil: http://talk.maemo.org/showthread.php?p=1310019#post1310019 | 12:45 |
| jonwil | luf: posted | 12:48 |
| luf | jonwil: thanks | 12:48 |
| Pali | merlin1991, I'm preparing OSS replacement packages for initrd-tools and getbootstate | 12:49 |
| Pali | what do you think, can it be hosted on gitorious in cssu? | 12:50 |
| Pali | now getbootstate is in bzr (not in git!) on launchpad | 12:50 |
| Pali | but I can convert repo to git | 12:50 |
| Pali | ~ping | 12:52 |
| infobot | ~pong | 12:52 |
| Pali | jonwil, testing and now operator name is ok | 12:53 |
| jonwil | ok, thats good so far | 12:53 |
| jonwil | if you get problems later, lets get some logs and see whats up | 12:53 |
| Pali | ok | 12:54 |
| jonwil | btw, it looks like the Fremantle version of dsme isn't in the dsme git repo... | 12:55 |
| Pali | jonwil, look at cssu git | 12:59 |
| Pali | I found last fremantle version in git tree | 12:59 |
| Pali | and then I commited versions from dsc packages | 12:59 |
| jonwil | btw, I suspect this commit (or commits near it) are where iphb stuff was merged with dsme | 13:01 |
| jonwil | https://meego.gitorious.org/meego-middleware/dsme/commit/735a5f058e082ff63e3cf65579def48b157b6705 | 13:01 |
| jonwil | one would need to reverse engineer libiphb0 and iphbd as of 0.0.17+0m5 possibly using later stuff from dsme repo as a guide | 13:05 |
| jonwil | if one wanted open copies of Fremantle iphb stuff | 13:05 |
| jonwil | iphbd does not come as x86 binary | 13:06 |
| *** Pali has quit IRC | 13:07 | |
| *** Pali has joined #maemo-ssu | 13:08 | |
| luf | freemangordon: can you please test http://pastebin.com/GTG5uveY ? | 13:10 |
| freemangordon | jonwil: sure, but it is 16k binary | 13:10 |
| luf | freemangordon: don't enable fake devices | 13:11 |
| freemangordon | yeah, saw it | 13:11 |
| freemangordon | can't test now, when I am back home | 13:12 |
| freemangordon | BTW what do you expect ti hapen? | 13:12 |
| freemangordon | *to | 13:12 |
| luf | freemangordon: Just in case you're as curious as me :) I don't test in directly (just through dbus-send). | 13:12 |
| jonwil | it sounds like what Nokia did is that at some point between Fremantle and that commit mentioned above, they merged iphbd into DSME | 13:12 |
| freemangordon | luf: BTW what stops you from testing it yourself? | 13:13 |
| luf | freemangordon: it should list your paired devices (all not just that which supports obexFTP) and maybe it can work at all. No idea more ... | 13:13 |
| luf | freemangordon: time nad hildon change ... (if you can share I'll be glad) | 13:13 |
| luf | Now is the lunch time ... | 13:13 |
| freemangordon | hehe. it is lunch time here too, bbl ;) | 13:14 |
| Pali | jonwil, libiphb is open, see: c7d7345279a0da1f16805510cbbb4281fadf2a3f | 13:15 |
| Pali | https://meego.gitorious.org/meego-middleware/libiphb | 13:15 |
| jonwil | so that matches Fremantle libiphb? | 13:15 |
| jonwil | if so, great | 13:15 |
| jonwil | thats one less thing to RE | 13:15 |
| kerio | jonwil: testing connui-home-cellular | 13:18 |
| jonwil | ok, great | 13:18 |
| jonwil | the libiphb I see in that repo seems to be newer version than Fremantle | 13:18 |
| jonwil | I dont see Fremantle libiphb anywhere | 13:19 |
| kerio | jonwil: installed, i'll try to figure out if everything works | 13:20 |
| kerio | so far, everything good | 13:20 |
| jonwil | ok, so there is no code for ipbhd or libiphb out there that matches Fremantle that I can see. Which means we need to reverse engineer the library and daemon if we want to make changes to them :) | 13:21 |
| jonwil | Although I dont know if there is any reason to need to change that daemon | 13:21 |
| Pali | jonwil we have header files of fremantle version | 13:23 |
| Pali | look if header files are same | 13:23 |
| Pali | and if yes, then new harmattan version could be compatible | 13:24 |
| jonwil | well if we decided we cared, cloning libiphb and iphbd wouldn't be that hard for someone with ARM skills :) | 13:25 |
| jonwil | or x86 skills in the case of libiphb which has an x86 version | 13:25 |
| jonwil | ok, oldest libiphb.h in meego libiphb repo is different to maemo libiphb.h | 13:26 |
| jonwil | maemo libiphb.h defines some dbus interfaces (makes sense, maemo iphbd is separate daemon with dbus interface, somewhere along the line before making code available in meego repos nokia makes it part of dsme) | 13:27 |
| jonwil | hence we need to reverse engineer both lib and daemon if we want to change it | 13:28 |
| jonwil | not that I know much about what iphbd does or why one might care about changing it... | 13:29 |
| jonwil | so yeah if I do any more reverse engineering it will be on the things needed by icd and its plugins :) | 13:31 |
| jonwil | i.e. so that icd and friends can be replaced... | 13:32 |
| Pali | jonwil, or you can look at clockd daemon | 13:33 |
| Pali | I really want to see NTP support in clockd | 13:33 |
| Pali | clockd has very good documentation | 13:33 |
| Pali | with all dbus interfaces | 13:34 |
| jonwil | yeah I know | 13:34 |
| Pali | If somebody rewrite/clone/RE clockd I will add missing NTP support | 13:34 |
| kerio | Pali: do we know how big the nolo header is? | 13:38 |
| *** xes has joined #maemo-ssu | 13:38 | |
| Pali | kerio, what do you mean under "nolo header"? | 13:38 |
| kerio | /dev/mtd3 doesn't just contain u-boot | 13:39 |
| kerio | there's a header, afaik | 13:39 |
| kerio | the offsets are slightly different | 13:39 |
| Pali | kerio, some code about nolo header is in 0xFFFF in dump supprot | 13:39 |
| Pali | *support | 13:39 |
| Pali | (reverse flash) | 13:39 |
| Pali | code which deleting nolo header | 13:39 |
| Pali | so you can look at it | 13:40 |
| Pali | kerio, http://talk.maemo.org/showthread.php?t=87996 | 13:40 |
| Pali | kerio, see src/local.c | 13:41 |
| Pali | there is table nanddump_n900[] | 13:42 |
| Pali | with nolo offsets | 13:42 |
| Pali | for each mtd part there is offset and length | 13:43 |
| jonwil | btw, I think reason clockd/libtime needs to exist (and needs to be closed) is because clockd interacts with the cellular services daemon to handle receiving time from cellular network | 13:43 |
| *** NIN101 has joined #maemo-ssu | 13:43 | |
| kerio | yep, 2048 is also what i'm seeing from od's output | 13:43 |
| kerio | 0 75 3 and then zeros until offset 2048 | 13:43 |
| Pali | jonwil, in one of your dbus header I found that dbus interface for time from cellular | 13:43 |
| Pali | kerio, can you test if that table for n900 is correct? | 13:44 |
| Pali | also if you have time you can test 0xFFFF if working :-) | 13:44 |
| jonwil | yeah its Phone.Net.network_time_info_change' | 13:44 |
| jonwil | and Phone.Net is documented (although from what I have seen, I dont think it was intentional that the documentation was made public) | 13:45 |
| jonwil | but since it IS public, I intend to use it :)( | 13:46 |
| jonwil | :) | 13:46 |
| kerio | Pali: the xloader/nolo split seems to be correct | 13:47 |
| kerio | the initfs split is meaningless | 13:48 |
| kerio | the kernel one also seems to be correct | 13:48 |
| kerio | and ubifs don't care, ubifs don't give a fuck, so 0 offset is correct there | 13:48 |
| Pali | ok | 13:48 |
| Pali | you can test also local dump to one fiasco image (to MyDocs) | 13:49 |
| Pali | or get/change R&D flags via USB | 13:49 |
| kerio | i'm trying to write a sensible check for the existance of an attached kernel in a flashed uboot | 13:50 |
| jonwil | so yeah there is clockd and also libtime | 13:50 |
| jonwil | but is libclockcore any relation to all this? | 13:50 |
| jonwil | hmmm, no, clockcore is library used by clock app | 13:51 |
| *** dhbiker has quit IRC | 13:52 | |
| jonwil | so by clock-ui | 13:52 |
| jonwil | and is therefore not related to clockd | 13:52 |
| jonwil | and not a target if one is reverse engineering clockd functionality | 13:52 |
| jonwil | ooh good, libcityinfo has a -dev package :) | 13:53 |
| jonwil | so basically the targets for reverse engineering would be /usr/bin/clockd, /usr/bin/rclockd and /usr/lib/libtime0.so.0 | 13:55 |
| jonwil | if I had the x86 binaries for those 3 things handy I would take a look | 13:55 |
| jonwil | but I dont because of my dev box being down | 13:55 |
| kerio | Pali: can you check that you have a whole lot of nothing before address 264192 and then something, in your /dev/mtd3? | 13:56 |
| *** joshgillies has quit IRC | 13:59 | |
| jonwil | ok, I found the x86 binaries I need | 14:00 |
| kerio | Pali: ok, i think i got it - for uboot to recognize an attached image, it has to start at address 0x40800 in /dev/mtd3 and it has to start with 0x27051956 | 14:02 |
| *** arcean has joined #maemo-ssu | 14:03 | |
| kerio | so if that string is not there, we can be sure that there's no kernel attached - even if it were, uboot wouldn't be able to use it with a simple "run attachboot" | 14:03 |
| kerio | now... how do i check that in busyboxsh? | 14:04 |
| *** M4rtinK has joined #maemo-ssu | 14:08 | |
| kerio | Pali: ah crap, i would need root privs for it | 14:09 |
| *** xmlich02 has joined #maemo-ssu | 14:10 | |
| *** arcean_ has joined #maemo-ssu | 14:12 | |
| *** arcean has quit IRC | 14:15 | |
| *** arcean_ is now known as arcean | 14:15 | |
| kerio | Pali: would it be ok to ship a sudoers file with a very specific command or two? | 14:18 |
| kerio | i need "/usr/bin/dd if=/dev/mtd3ro bs=1 skip=262144 count=4" | 14:18 |
| Pali | kerio, what about to create bootmenu.src which will add attached entry only if there is kernel image? | 14:19 |
| kerio | yep | 14:19 |
| Pali | e.g. not do that in shell, but in uboot | 14:19 |
| kerio | no, wait | 14:19 |
| kerio | what? | 14:19 |
| Pali | bootmenu.src is uboot script | 14:19 |
| Pali | (regular uboot script) | 14:20 |
| *** tg has quit IRC | 14:20 | |
| Pali | so what about wrting uboot script which check that in uboot? | 14:20 |
| Pali | not in linux? | 14:20 |
| kerio | hm, i don't know much about uboot scripting | 14:20 |
| Pali | uboot using hush | 14:20 |
| Pali | there is also simple "test" command | 14:21 |
| kerio | Pali: how do you do the fancy stuff with osso-product-info and zimage-get-version, though? | 14:21 |
| *** tg has joined #maemo-ssu | 14:21 | |
| Pali | osso-product-info working as normal user | 14:21 |
| Pali | and zimage-get-version only as root | 14:21 |
| kerio | Pali: you should ship a sudoers for that, too | 14:22 |
| Pali | so if you call u-boot-udate-bootmenu as root it will read kernel version | 14:22 |
| Pali | and if not then there will not be kernel version string in name | 14:22 |
| Pali | but what do you trying to do? hide attached kernel when uboot is without kernel? | 14:22 |
| kerio | yes | 14:22 |
| kerio | i know, it's silly | 14:23 |
| Pali | so as I wrote that should be done in uboot | 14:23 |
| Pali | so as uboot script | 14:23 |
| Pali | problem is that you can load via 0xFFFF or flasher your own uboot | 14:23 |
| Pali | without flashing | 14:23 |
| Pali | and hiding attached kernel is bad then | 14:23 |
| kerio | is the bootmenu_0 entry always executed by default? | 14:24 |
| kerio | how would we know what should be ran as default? | 14:25 |
| kerio | if you flash a different uboot, you'll have to rerun u-boot-update-bootmenu | 14:26 |
| kerio | which is still fairly sensible | 14:26 |
| kerio | (or manually type run attachboot) | 14:26 |
| kerio | or/and | 14:26 |
| Pali | hm, really it is not easy, because attached kernel is in bootmenu_0 | 14:27 |
| Pali | and removing it will need to reorder menu... | 14:28 |
| kerio | Pali: unless the user specified a different default! | 14:28 |
| Pali | yes | 14:28 |
| Pali | then you need it reorder again | 14:29 |
| Pali | not easy | 14:29 |
| kerio | Pali: btw, those calls should be with a full path | 14:29 |
| *** luf has quit IRC | 14:29 | |
| *** luf has joined #maemo-ssu | 14:32 | |
| kerio | Pali: i'm still not sure how to manipulate strings of unprintable bytes in ash/busybox | 14:39 |
| Pali | kerio, hard or not possible | 14:40 |
| Pali | because our version of busybox is useless | 14:40 |
| Pali | I used more hacks in locale-resolver-config package | 14:40 |
| Pali | look at it, it is in extras | 14:40 |
| kerio | Pali: can you tell me what's the first line of the output of sudo dd if=/dev/mtd3ro bs=1 skip=$(( 0x40800 )) count=4 | /usr/bin/od -H | 14:42 |
| kerio | please? | 14:42 |
| Pali | not now | 14:43 |
| kerio | :( | 14:43 |
| kerio | jonwil: do you use uboot? | 14:43 |
| Pali | 0000000 56190527 | 14:50 |
| Pali | 0000004 | 14:50 |
| Pali | 4+0 records in | 14:50 |
| Pali | 4+0 records out | 14:50 |
| Pali | kerio ^^^ | 14:50 |
| kerio | hrmpf | 14:50 |
| kerio | Pali: try with -h | 14:51 |
| kerio | it's like backwards | 14:51 |
| kerio | but meh, i'll deal with it | 14:51 |
| Pali | 0000000 0527 5619 | 14:51 |
| Pali | 0000004 | 14:51 |
| Pali | 4+0 records in | 14:51 |
| Pali | 4+0 records out | 14:51 |
| kerio | meh, -H is close enough | 14:52 |
| kerio | Pali: *should* u-boot-update-bootmenu be runnable by the user? | 14:59 |
| Pali | I think yes | 15:00 |
| kerio | then we need both commands in a sudoers file | 15:00 |
| kerio | i'll send you a patch manually if it's ok - i don't even have a gitorious account | 15:00 |
| *** xmlich02 has quit IRC | 15:12 | |
| jonwil | no I dont use uboot | 15:13 |
| jonwil | btw, I am currently trying to identify the names of as many functions in clockd as possible | 15:14 |
| jonwil | based on the info in the docs | 15:14 |
| *** FlameReaper has joined #maemo-ssu | 15:14 | |
| kerio | Pali: it appears to work! :D | 15:15 |
| *** _rd has quit IRC | 15:20 | |
| *** _rd has joined #maemo-ssu | 15:27 | |
| kerio | Pali: http://dl.dropbox.com/u/1405219/u-boot-tools_check-attached-kernel.diff | 15:28 |
| kerio | Pali: you'll also need to /usr/sbin/update-sudoers in postinst and postrm | 15:29 |
| kerio | and find a way to mark /etc/sudoers.d/u-boot-tools.sudoers as *not* a config file | 15:29 |
| kerio | so it gets properly deleted on the normal remove | 15:30 |
| *** Pali has quit IRC | 15:46 | |
| *** Pali has joined #maemo-ssu | 15:48 | |
| freemangordon | APNG patch applies cleanly, trying to build microb with system libpng/zlib | 15:49 |
| *** Pali has quit IRC | 15:53 | |
| *** Pali has joined #maemo-ssu | 15:55 | |
| *** Pali has quit IRC | 15:57 | |
| luf | freemangordon: nice | 16:05 |
| *** luf has quit IRC | 16:06 | |
| *** Martix has joined #maemo-ssu | 16:09 | |
| freemangordon | microb-engine-dbg/usr/lib/debug//usr/lib/microb-engine/libxul.so: No space left on device | 16:16 |
| freemangordon | dh_strip: command returned error code 256 | 16:16 |
| freemangordon | :D:D:D | 16:16 |
| *** int_ua has joined #maemo-ssu | 16:16 | |
| gregoa | jonwil: connui-home-cellular 3.2.1 shows the wrong name for me: https://paste.debian.net/221361/ 3.2.1 shows "A1", 3.1.1 shows "bob". background: "bob" is a budget brand of "A1", so the network is technically from "A1" but still "bob" is the actual provider | 16:33 |
| jonwil | What does the stock Nokia widget display? | 16:34 |
| jonwil | A1 or bob? | 16:34 |
| jonwil | "bob" I assume | 16:34 |
| jonwil | ? | 16:37 |
| merlin1991 | bob | 16:38 |
| jonwil | I was looking for an answer from gregoa :) | 16:38 |
| kerio | :( | 16:38 |
| merlin1991 | well it is an austrian provider and I have it aswell :) | 16:38 |
| jonwil | ok | 16:38 |
| * kerio shakes his fist at MVNOs | 16:39 | |
| jonwil | so does the new 3.2.1 display "A1 | 16:39 |
| jonwil | for you also? | 16:39 |
| merlin1991 | I didn't try the new one yet | 16:39 |
| jonwil | ok | 16:39 |
| gregoa | jonwil: yup, it was always "bob" before this thingie was changed in cssu, since then it changes from version to version :/ | 16:40 |
| jonwil | ok | 16:40 |
| kerio | enable the logging | 16:40 |
| merlin1991 | btw gregoa where in austria do you live? | 16:42 |
| gregoa | merlin1991: innsbruck | 16:42 |
| *** peetah has quit IRC | 16:47 | |
| *** peetah has joined #maemo-ssu | 16:47 | |
| *** xes_ has joined #maemo-ssu | 16:47 | |
| *** xes has quit IRC | 16:50 | |
| jonwil | gregoa (or anyone else with the same operator) please run this command | 16:56 |
| jonwil | dbus-send --system --print-reply --dest="com.nokia.phone.SIM" /com/nokia/phone/SIM Phone.Sim.get_service_provider_name | 16:56 |
| jonwil | and give me the results | 16:56 |
| gregoa | jonwil: https://paste.debian.net/221367/ | 16:57 |
| *** jonwil has quit IRC | 17:01 | |
| *** jonwil has joined #maemo-ssu | 17:02 | |
| jonwil | ok, gregoa, can you run this | 17:02 |
| jonwil | dbus-send --system --print-reply --dest="com.nokia.phone.net" /com/nokia/phone/net Phone.Net.get_registration_status | 17:02 |
| gregoa | jonwil: https://paste.debian.net/221373/ | 17:03 |
| jonwil | what about this one | 17:08 |
| jonwil | dbus-send --system --print-reply --dest="com.nokia.phone.net" /com/nokia/phone/net Phone.Net.get_operator_name byte:0 uint32:1 uint32:232 | 17:08 |
| jonwil | Run it a few times changing the byte:0 to byte:1 and etvc | 17:08 |
| jonwil | so 0 then 1 and on up to 4 | 17:08 |
| *** luf has joined #maemo-ssu | 17:23 | |
| *** Pali has joined #maemo-ssu | 17:23 | |
| luf | freemangordon: do you need some SAN attached to the N900? :D | 17:26 |
| jonwil | I think the only way I can understand whats going on is for someone on this "bob" operator and who is running the stock widget to collect a dbus-monitor dump of what happens when they connect to this operator (e.g. switching from airplane mode into regular mode) | 17:26 |
| kerio | Pali: did you receive the patch? | 17:27 |
| jonwil | That will explain just which dbus bit is resulting in the "bob" string being passed to the widget | 17:27 |
| Pali | kerio, I have link | 17:27 |
| *** arcean has quit IRC | 17:28 | |
| Pali | kerio, "unmarking" as config file for deb package can be done by removing it from debian/<package>/DEBIAN/conffiles after dh_gencontrol | 17:29 |
| *** arcean has joined #maemo-ssu | 17:30 | |
| kerio | oh, it's dh_gencontrol that automatically adds files in /etc/ to conffiles? | 17:30 |
| Pali | s/dh_gencontrol/dh_installdeb/ | 17:32 |
| infobot | Pali meant: kerio, "unmarking" as config file for deb package can be done by removing it from debian/<package>/DEBIAN/conffiles after dh_installdeb | 17:32 |
| Pali | dh_installdeb creating that conffiles | 17:32 |
| kerio | i thought it was just "automatic" in a deeper sense | 17:32 |
| Pali | see manpage | 17:32 |
| Pali | dh_installdeb >>> In v3 compatibility mode and higher, all files in the etc/ directory in a package will automatically be flagged as conffiles by this program, so there is no need to list them manually here. <<< | 17:33 |
| Pali | we using compatibility 5 | 17:33 |
| kerio | but they can be removed at some point, i see | 17:34 |
| kerio | i'm also looking for a better way to do the check for the magic string | 17:36 |
| merlin1991 | jonwil: system or user bus? | 17:38 |
| jonwil | system bus | 17:38 |
| jonwil | which IIRC means you need extra stuff in the dbus configs to make that logging possible | 17:39 |
| Pali | run as root | 17:39 |
| gregoa | jonwil: (sorry, I was shopping): https://paste.debian.net/221379/ | 17:53 |
| *** arcean has quit IRC | 18:15 | |
| merlin1991 | jonwil: http://cdnm.at/~christian/maemo/cssu/bob-system.log | 18:37 |
| luf | freemangordon: ping | 18:45 |
| DocScrutinizer05 | (([2013-01-03 10:28:14] <merlin1991> hm we should leave poweroff as is)) absolutely! don't mess with semantics of basic system functions unless pretty good reason to fix a severe bug, and even then it's often better to keep the bug as is and rather offer an alternative way | 18:45 |
| jonwil | ok, so from reading that log I see that operator_name_change is called with the operator name "bob" and operator_name_type NETWORK_EONS_FULL_OPER_NAME | 18:45 |
| *** NIN101 has quit IRC | 18:45 | |
| jonwil | this signal should result in operator_name_change_cb in libconnui_cell.so being called | 18:46 |
| jonwil | that then fills in the relevant fields in the struct network_state and calls the network status callback | 18:47 |
| luf | freemangordon: BTW maybe I'm not able to get the osso-gnomevfs-extra working :( | 18:48 |
| jonwil | merlin1991, are you using the stock nokia operator name widget? | 18:48 |
| merlin1991 | stock | 18:48 |
| jonwil | ok | 18:48 |
| merlin1991 | that log is from a vanilla pr1.3.1 | 18:49 |
| jonwil | ok | 18:49 |
| jonwil | what I need next is for someone (anyone) with this "bob" operator to install the latest version of the custom operator name widget, then remove any copies of /home/user/opername.log, then reboot (or power off then back on) the phone, then pastebin or send me the full opername.log as it is | 18:50 |
| *** xes_ has left #maemo-ssu | 18:55 | |
| merlin1991 | gregoa: can you do that ^^ ? | 18:58 |
| gregoa | jonwil, merlin1991: will do | 18:59 |
| luf | freemangordon: so just the patch for osso-gnomevfs-extra doesn't work. | 19:04 |
| gregoa | jonwil: cbsms.log: https://paste.debian.net/221413/ and opername.log: https://paste.debian.net/221414/ (with connui-home-cellular 3.2.1 after a reboot) | 19:09 |
| jonwil | ok, thats what I needed to see | 19:10 |
| jonwil | From reading the code and what I have seen through dbus, it looks like state->operator_name should be "bob" | 19:10 |
| jonwil | but its not, its an empty string | 19:11 |
| gregoa | sounds reasonable | 19:13 |
| jonwil | I am stumped as to why its presumably sending the operator_name_changed signal with the value "bob" yet its not passing "bob" into widget_net_status_cb | 19:17 |
| jonwil | aha, I think whats happening is this: | 19:27 |
| *** luf has quit IRC | 19:29 | |
| freemangordon | Pali: ping | 19:32 |
| *** desaparecido has joined #maemo-ssu | 19:34 | |
| merlin1991 | jonwil: I'm still waiting :D | 19:34 |
| *** desaparecido has quit IRC | 19:40 | |
| *** FlameReaper has quit IRC | 19:47 | |
| kerio | hm, i found a minor bug in the portrait messaging CSS | 19:51 |
| jonwil | so what I think is happening is this: | 19:51 |
| kerio | the time is sometimes displayed right at the right of the text instead of being on the far right | 19:51 |
| Pali | freemangordon, pong | 19:53 |
| Pali | kerio, I think I saw it too, but I really do not know how to debug css | 19:54 |
| Pali | kerio, ask some web developer :-) | 19:54 |
| jonwil | 1.libconnui_cell creates the network_state structure and fills in operator_name with blank string | 19:59 |
| jonwil | 2.libconnui_cell recieves one of the signals it has registered for (one that isnt operator_name_change) | 19:59 |
| jonwil | hmmm wait no | 19:59 |
| *** _rd has quit IRC | 20:00 | |
| jonwil | I think I know whats going on, not 100% sure how to fix it though. | 20:06 |
| *** NIN101 has joined #maemo-ssu | 20:07 | |
| *** NIN102 has joined #maemo-ssu | 20:13 | |
| *** NIN102 has quit IRC | 20:14 | |
| *** NIN101 has quit IRC | 20:15 | |
| *** NIN101 has joined #maemo-ssu | 20:16 | |
| *** jonwil has quit IRC | 20:16 | |
| *** arcean has joined #maemo-ssu | 20:22 | |
| freemangordon | Pali: I received an "improved" version of messaging-ui html/css (didn;t check what exactly) | 20:38 |
| freemangordon | do you want to look at it? here http://speedy.sh/KNbww/rtcom-messaging-ui.zip | 20:38 |
| freemangordon | kerio: if you wish test that, to see if it solves what you reported ^^^ | 20:39 |
| kerio | freemangordon: fucking DOS line endings? :S | 20:44 |
| freemangordon | kerio: NFC, toldya, I received that from tmo user knarf via PM | 20:45 |
| kerio | btw, why the hell did we add yet another package that changes the files from another package instead of changing the original package? | 20:50 |
| merlin1991 | because the original css is part of the closed package | 20:54 |
| merlin1991 | we can't do it directly | 20:54 |
| kerio | "closed package"? | 20:54 |
| kerio | i thought we had permission to redistribute stuff | 20:54 |
| Pali | we have full css files in git repo | 20:57 |
| Pali | not patches | 20:57 |
| Pali | kerio, merlin1991, freemangordon: I cannot update that zip file | 21:00 |
| Pali | need PK compat. v6.3 (can do v4.6) | 21:00 |
| kerio | \_o_/ | 21:00 |
| kerio | you also need to convert the line endings | 21:00 |
| kerio | Pali: will give you a tarball, one sec | 21:01 |
| kerio | Pal2013-01-03T21:13:23 <freemangordon> aah, where is luf :( | 21:02 |
| Pali | I'm going to update ipv6 packages to extras-devel... | 21:14 |
| Pali | do you know something special where is ipv6 disabled? | 21:14 |
| kerio | Pali: i think it's disabled by default in microb | 21:14 |
| Pali | now I'm enabling ipv6 support in icd2 and microb | 21:15 |
| kerio | how are you enabling it? | 21:15 |
| Pali | and adding ipv6 lines in /etc/hosts, /etc/resolv.conf | 21:15 |
| Pali | enabling in /etc/avahi/avahi-daemon.conf and mdns4 to mdns in /etc/nsswitch.conf | 21:16 |
| kerio | how will you be able to configure the ipv6 stuff though? | 21:16 |
| *** nox- has joined #maemo-ssu | 21:16 | |
| kerio | the UI is still closed :( | 21:16 |
| Pali | kerio, here is postinst file: https://bazaar.launchpad.net/~pali/+junk/maemo_ipv6-support/view/head:/debian/postinst | 21:17 |
| Pali | kerio, configuring static ipv6 address need gconf :-( | 21:18 |
| kerio | until someone rewrites the configuration UI | 21:18 |
| kerio | hopefully also adding proper support for multiple APNs, and for the configuration of DUMMY and USB | 21:18 |
| Pali | by default it using RA and/or dhcpv6 (wide) client | 21:18 |
| kerio | what's RA? | 21:18 |
| Pali | router advertisment | 21:19 |
| Pali | it will check for M and O bits | 21:19 |
| Pali | if none is set then it using stateless configuration | 21:19 |
| Pali | if O is set then it assign stateless ipv6 address and using dhcpv6 client for dns server | 21:20 |
| Pali | and if did not receive RA or receive M bit, then it spawning dhcpv6 client for getting address and also dns | 21:21 |
| Pali | ping DocScrutinizer05 | 21:45 |
| DocScrutinizer05 | pong | 21:45 |
| Pali | DocScrutinizer05, do you know where is ipv6 support disabled in maemo5? | 21:45 |
| Pali | where could be also enabled? | 21:46 |
| DocScrutinizer05 | sorry, no idea | 21:46 |
| kerio | #maemo-ssu - dragging a kicking and screaming fremantle into the modern age, one step at a time | 21:49 |
| freemangordon | merlin1991: do you know if fixed libxml2 is in cssu-devel? | 22:01 |
| freemangordon | (zlib fixed libxml2) | 22:01 |
| freemangordon | merlin1991: I plan to put libpng, zlib and microb-engine in cssu-devel, I need to know if I should put libxml2 too | 22:02 |
| *** kolp has quit IRC | 22:15 | |
| DocScrutinizer05 | http://mg.pov.lt/maemo-ssu-irclog/%23maemo-ssu.2011-11-29.log.html#t2011-11-29T00:04:00 | 22:31 |
| DocScrutinizer05 | ^^^ that's not yet the ultimate check, which more was with a regex checking for [13579bdf] as last digit of an addr | 22:33 |
| DocScrutinizer05 | oops echan | 22:34 |
| kerio | freemangordon: your tklock seems a bit too eager to lock the screen | 22:35 |
| kerio | i have to unlock it to receive a call and i have to unlock it at the end | 22:35 |
| freemangordon | kerio: hmm, I am having about 20-40 calls per day, never seen that | 22:36 |
| kerio | to be fair, i kinda like it | 22:37 |
| DocScrutinizer05 | dafaq | 22:39 |
| freemangordon | kerio: are you sure it is tklock to blame? as it acts on dbus signals | 22:44 |
| *** luf has joined #maemo-ssu | 22:44 | |
| luf | freemangordon: fixed libxml2 is in cssu-devel (also announced on TMO) | 22:48 |
| freemangordon | luf: microb works with libpng+neon+apng+zlib :) | 22:49 |
| luf | freemangordon: good. | 22:49 |
| freemangordon | (latest versions) | 22:49 |
| freemangordon | luf: going to put zlib, libpng and microb-engine too | 22:50 |
| luf | freemangordon: please check in zlib package that it contains Break: libxml (<2.7.7) in debian/control. | 22:50 |
| freemangordon | luf: hmm, ok | 22:50 |
| luf | I mean zlib1g package | 22:50 |
| luf | Breaks: libxml2 (< 2.6.32.dfsg-5maemo4+0m5+0cssu1) | 22:50 |
| freemangordon | Breaks: libxml2 (< 2.6.32.dfsg-5maemo4+0m5+0cssu1) | 22:51 |
| freemangordon | Conflicts: zlib1 (<= 1:1.0.4-7), zlib1g-udeb | 22:51 |
| luf | I'm sorry this line. | 22:51 |
| luf | It's ok. Thanks. I was not sure if I included it. | 22:51 |
| freemangordon | ^^^ is what I have here | 22:51 |
| freemangordon | but it is not 2.7.7 | 22:51 |
| luf | It's ok like it is. 2.7.7 is my mistake. | 22:52 |
| *** xes has joined #maemo-ssu | 22:53 | |
| freemangordon | luf: but will libxml2 in cssu-devel work with older zlib? | 22:54 |
| freemangordon | hmm, wait, it should work, ain't? | 22:54 |
| freemangordon | ok, zlib is in cssu-devel | 22:56 |
| Pali | freemangordon, "Breaks: libxml2 (< 2.6.32.dfsg-5maemo4+0m5+0cssu1)" is not good line | 22:56 |
| Pali | use "<<" or "<=" | 22:56 |
| luf | Pali: why? | 22:56 |
| Pali | there was some change in debian what '<' means | 22:57 |
| Pali | by some version of dpkg/debhelper it is << and by other <= | 22:57 |
| luf | 2.6.32.dfsg-5maemo4+0m5+0cssu1 is the first ok version so in that case << (but I think I used <) | 22:57 |
| luf | Pali: ok, thanks. | 22:57 |
| Pali | luf, I got this problem when I compiled package only on ubuntu | 22:57 |
| Pali | it worked fine | 22:58 |
| Pali | but when I pushed it to extras for maemo it not worked | 22:58 |
| Pali | because maemo understood that '<' not as ubuntu | 22:58 |
| *** lizardo has quit IRC | 22:58 | |
| freemangordon | merlin1991: ping | 23:03 |
| freemangordon | merlin1991: some problem in cssu-devel repo while importing zlib | 23:03 |
| freemangordon | luf: hmm, seems I have to raise libpng version :( | 23:04 |
| freemangordon | ok, libpng is in the repo | 23:10 |
| *** arcean has quit IRC | 23:10 | |
| *** arcean has joined #maemo-ssu | 23:13 | |
| freemangordon | arcean: what is the status of h-d, I saw you commit a lot of stuff in the last few days | 23:15 |
| luf | freemangordon: It's ok to raise libpng version as added APNG ;) | 23:20 |
| freemangordon | luf: already done, package is in cssu-devel | 23:21 |
| freemangordon | feel free to test it :P | 23:21 |
| luf | freemangordon: I'll test it tomorrow. Today I want to go through gnomevfs-extra bt listing ... | 23:27 |
| *** Pali has quit IRC | 23:27 | |
| freemangordon | luf: fingers crossed, though it won;t be an easy task | 23:31 |
| luf | freemangordon: the problem is that I never used dbus message reading with such low level function. | 23:32 |
| luf | freemangordon: it seems easy to list them. | 23:33 |
| freemangordon | well, dbus API is not that complicated, I am sure you'll figure it out | 23:33 |
| luf | freemangordon: I hope so :) But just it takes me longer time. | 23:33 |
| freemangordon | luf: keep in mind that ListDevices returns all devices, not just paired ;) | 23:33 |
| luf | freemangordon: it's not a problem in firs round ;) | 23:34 |
| freemangordon | so you'll have to do "GetProperties" for every device in the list, and check if it is paired | 23:34 |
| freemangordon | ok, but put it in todo | 23:34 |
| luf | Sure I need GetProperties to also check if the device support obexFTP ;) | 23:35 |
| freemangordon | yeah | 23:35 |
| freemangordon | luf BTW are you sure there is no such code somewhere over the inet? | 23:35 |
| freemangordon | did you google enough? :D | 23:36 |
| freemangordon | even if it needs porting, it will be easier than write it from scratch | 23:36 |
| freemangordon | luf: BTW - please, create a project on gitorious (in your repo) and commit in small chunks. I bet you will want a code review at some point, massive commits are harder to be reviewed | 23:38 |
| luf | freemangordon: it's cssu related. | 23:38 |
| luf | Why should I create separate project for it? | 23:39 |
| freemangordon | luf: because merlin1991 is MIA right now | 23:39 |
| freemangordon | and we don;t ave such project in CSSU | 23:39 |
| freemangordon | *have | 23:39 |
| luf | What is MIA? | 23:39 |
| freemangordon | ~mi | 23:39 |
| infobot | extra, extra, read all about it, mi is More Information; saying something "doesn't work" means absolutely nothing, because there are so many millions of ways in which a computer can fail; someone saying "X doesn't work" is really saying nothing, but if they "X says no screens were found", that at least gives the people on #debian a clue as to the area where the problem is. Even better, tell us what hardware you're using, what version of X, what ... | 23:39 |
| freemangordon | ~mia | 23:39 |
| infobot | from memory, mia is missing in action | 23:39 |
| luf | I have local git repo (I imported the nokia'a CVS) ;) | 23:40 |
| luf | I mean version control system ;) | 23:41 |
| freemangordon | luf: my point is that this plugin is so deeply burried in the hierarchy, that it will be virtually impossible to catch a memleak, buffer overflow or invalid pointer dereference. Having the code commited will help us to catch those(if exist) much more easily | 23:42 |
| freemangordon | *commited in small chunks | 23:43 |
| luf | freemangordon: I have list of devices :D | 23:43 |
| freemangordon | luf: good | 23:43 |
| freemangordon | no, great :D | 23:43 |
| luf | Just the names are terrible but step by step ... | 23:43 |
| freemangordon | luf: have in mind that ossofilemanager expects the address in the form of obex://[XX:XX:XX:XX:XX:XX] | 23:44 |
| luf | I know. But I hope it should display something more readable ;) | 23:44 |
| freemangordon | so having BT MAC as name is expected | 23:44 |
| freemangordon | well, I guess device name is easy to get | 23:45 |
| luf | freemangordon: Also in GetProperties | 23:51 |
| freemangordon | :nod: | 23:52 |
| arcean | freemangordon: I'm just improving h-d :P | 23:54 |
| arcean | maybe during the weekend I'll upload new Qt to cssu-devel | 23:54 |
| freemangordon | arcean: so no more bugs? we can put it in cssu-devel then | 23:59 |
Generated by irclog2html.py 2.15.1 by Marius Gedminas - find it at mg.pov.lt!