00:06.53 | jonwil | should any of the following packages be added to cssu-devel? |
00:06.54 | jonwil | hald-addon-bme-dev,hildon-im-vkbrenderer3,hildon-input-method-plugins,hildon-status-bar-usb,libcal,libconnui-dev,libgtkadi,location-supl-fix,omap3430-dsp,osso-systemui-modechange,rtcom-accounts-plugin-jabber,sharing-service-facebook |
00:07.06 | jonwil | libconnui-dev definatly should be |
00:07.15 | jonwil | osos-systemui-modechange probably |
00:07.20 | jonwil | no idea on the others though |
00:07.35 | jonwil | wait hald-addon-bme-dev no, its in the bme repo already |
00:09.39 | jonwil | and these are the packages that are out-of-date in -devel relative to git |
00:09.41 | jonwil | clock-ui,community-ssu-enabler,connui-internet,curl,feedservice-plugin-fb-common,ke-recv,libdevlock-dev,libhildonfm,matchbox-window-manager-2,osso-systemui-alarm,osso-systemui-dev,osso-systemui-powerkeymenu,osso-systemui-tklock,profiled,status-menu-applet-profiles,tracker,vte |
00:10.04 | jonwil | I shall now let the maintainers of cssu-devel figure out which of those need to go into -devel and which dont :) |
01:04.10 | merlin1991 | -devel has no "maintainer" it's just some devs with push access |
01:04.41 | merlin1991 | I can add you too |
01:49.19 | jonwil | well I dont know for sure which packages should be in -devel and which shouldn't be |
01:49.27 | jonwil | so i dont want to put things in there that dont belong |
01:54.40 | *** join/#maemo-ssu xes_ (~xes@unaffiliated/xes) |
02:10.25 | jonwil | my god openssl takes forever to compile... |
02:29.11 | jonwil | bah, its STILL compiling :( |
02:29.29 | jonwil | most likely its all those testcases and things that take so long |
02:38.04 | *** join/#maemo-ssu ruskie (ruskie@sourcemage/mage/ruskie) |
05:59.10 | jonwil | http://talk.maemo.org/showthread.php?p=1486997#post1486997 |
06:24.55 | freemangordon | merlin1991: the files still hang in cssu-devel incomming directory |
06:38.46 | *** join/#maemo-ssu _rd (~rd@p5B2C7F08.dip0.t-ipconnect.de) |
06:38.47 | *** join/#maemo-ssu freemangordon (~ivo@46.249.74.23) |
06:47.36 | *** join/#maemo-ssu amiconn (~amiconn@rockbox/developer/amiconn) |
07:22.28 | freemangordon | jonwil: why is eapd so big? any idea? |
07:22.54 | jonwil | probably inefficient (all that C++) and also its doing a lot of different WiFi security things |
07:24.35 | freemangordon | can't we replace it with wpa_supplicant? |
07:25.15 | jonwil | Its not that simple, it talks to a bunch of stuff and a bunch of stuff talks to it |
07:25.25 | jonwil | I never got far enough in reverse engineering it to find out everything it does |
07:25.27 | freemangordon | vid dbus? |
07:25.35 | freemangordon | *via |
07:25.53 | jonwil | via dbus yes but also other things like gconf keys |
07:26.15 | freemangordon | that should be relatively easy to recover and implement |
07:29.59 | freemangordon | jonwil: hmm, there is no need to reinvent the wheel https://www.google.bg/url?sa=t&rct=j&q=&esrc=s&source=web&cd=1&ved=0CCcQFjAAahUKEwiDmrml0uTIAhWK8RQKHcADAPk&url=http%3A%2F%2Fwww.cncmods.net%2Ffiles%2Fdbus%2Feap-dbus.h&usg=AFQjCNE7VHd09OZdNhBuXU-3F9TSUVcIrQ |
07:32.07 | freemangordon | I'd bet we should be able to wrap wpa_supplicant (if there is no exported API) and to reimplement eapd |
07:34.01 | jonwil | yeah the dbus stuff I figured out most of, its the gconf keys that are the problem |
07:34.16 | jonwil | eapd touches gconf keys and then other stuff needs to get the stuff that eapd put there |
07:34.37 | jonwil | oh and I created eap-dbus.h so I know all about it |
07:34.50 | jonwil | well I didn't create it but I edited it and put it up there |
07:34.59 | freemangordon | hmm? "@author Janne Ylalehto <janne.ylalehto@nokia.com>" |
07:35.05 | freemangordon | oh :) |
07:35.26 | jonwil | I forgot where it came from until I opened it and saw :) |
07:45.08 | freemangordon | ok, wpa_supplicant build in SB |
07:45.17 | freemangordon | the latest one, from git |
07:49.08 | freemangordon | ph, Pali is one of the maintainers of extras-devel wpa_supplicant :) |
07:49.13 | freemangordon | *oh |
07:50.10 | freemangordon | jonwil: ^^^ |
07:50.22 | freemangordon | I guess we're lucky then |
08:05.54 | Sicelo | eapd .. i support idea :) |
08:08.03 | Sicelo | i can be a beta tester for anything eapd/wpa_supplicant related. |
08:10.19 | freemangordon | we'll have to RE icd2 aiui |
08:10.33 | freemangordon | or not? |
08:10.41 | freemangordon | anyway, gotta run |
08:12.13 | Sicelo | i just know when using wpa_supplicant, maemo programs don't "see" the connection. you need to use "dummy network" interface. only pidgin has issues with that. all others are happy with dummy interface |
08:13.27 | bencoh | how comes pidgin isn't happy with "dummy network" ? |
08:14.09 | Sicelo | wish i knew |
08:14.27 | Sicelo | you have to disable accounts first. only then it works |
08:19.31 | Sicelo | some features of wlancond will be needed if we go wpa_supplicant, especially PSM. i'm currently doing it on interface initialization using iwconfig. wlancond changes the power saving dynamically |
08:37.46 | jonwil | in addition to osso-wlan-security (aka eapd) we will also possibly need to worry about libicd-network-eap, and libicd-network-wps |
08:38.35 | jonwil | we may also need to care about connui-conndlgs-wlan and conui-iapsettings-wlan |
08:38.47 | jonwil | ok, that's got that finished (at last) |
08:40.09 | jonwil | I have cleaned up my N900 and removed some crap I didn't need and then reinstalled self-built packages of all the stuff in community-ssu I want to run on my phone including a self-built copy of microb-engine (the building of which wasn't as painful as I thought it would be) |
08:40.22 | jonwil | plus the latest openssl 0.9.8zf |
08:40.42 | jonwil | and the latest maemosec stuff for the latest root certificate sets |
08:45.02 | *** join/#maemo-ssu freemangordon_ (~ivo@213.222.56.174) |
08:48.17 | bencoh | jonwil: hmmm, last time I tried I didnt manage to build microb-engine in my sb vm :/ |
08:48.40 | bencoh | strange errors everywhere |
08:49.42 | jonwil | It build just fine |
08:50.00 | jonwil | and I can log into my bank so everything is working (including all the security stuff) |
08:50.28 | jonwil | I could be wrong but I think GPS locking is now faster too (so something I installed must have affected that) |
08:50.37 | bencoh | hmm |
08:50.52 | bencoh | maybe just a missing certificate? which supl server do you use? |
08:52.21 | freemangordon_ | bencoh: microb-engine builds just fine in SB |
08:53.07 | bencoh | yeah, I guess it's just me |
08:54.14 | bencoh | freemangordon_: would you pastebin the output of dpkg -l in sb (fremantle_armel, not linaro/4.7) when you have some time? |
08:55.22 | jonwil | bencoh: I can do that for my copy of SB if that helps |
08:55.40 | jonwil | which is based on the Maemo dev VM from Nokia |
08:56.29 | bencoh | jonwil: that'd be great :) |
08:57.03 | bencoh | (I'm concerned about "which autoconf/automake variant is installed" and stuff like that) |
08:57.19 | freemangordon_ | bencoh: http://pastebin.com/ktEcEsTN |
08:57.27 | bencoh | thx |
08:57.36 | freemangordon_ | expires in an hour |
08:58.20 | bencoh | gonna curl that :) |
08:59.00 | jonwil | I will compare that pastebin to mine and see whats different (for interests sake) |
09:03.21 | jonwil | ok, looks like its not much interesting, just some packages where fg has cssu version installed and I have stock |
09:03.27 | jonwil | and some -dbg packages I have that fg doesn't have |
09:04.38 | *** join/#maemo-ssu sparetire_ (~sparetire@unaffiliated/sparetire) |
09:06.56 | bencoh | mine differs a bit (autotools things) |
09:09.21 | jonwil | in any case I cant see any packages freemangordon has that I dont have and want |
09:12.16 | jonwil | I have a lot of packages installed that freemangordon doesn't have |
09:21.46 | bencoh | dpkg-checkbuilddeps: Unmet build dependencies: autoconf2.13 automake1.7 |
09:21.52 | bencoh | I wonder how you can build microb-engine |
09:32.59 | freemangordon_ | bencoh: just install those, what is the problem? |
09:34.00 | bencoh | no problem with installing it, it's just that it doesnt show on your dpkg -l :) |
09:50.29 | jonwil | wow, installing the cssu status-area-applet-battery was a good idea, it gives you much more useful info than the stock widget does |
10:06.08 | jonwil | I commend Pali for his work on status-area-applet-battery |
10:29.57 | *** join/#maemo-ssu M4rtinK (~M4rtinK@ip-89-103-133-54.net.upcbroadband.cz) |
10:31.56 | *** join/#maemo-ssu Pali (~pali@Maemo/community/contributor/Pali) |
10:32.38 | freemangordon_ | Pali: hi! what do you think about replacing eapd with wpa_supplicant? |
10:32.58 | Pali | freemangordon_: do you know how to do it? |
10:33.29 | freemangordon_ | we can reimplement eapd dbus and gconf interfaces |
10:33.37 | freemangordon_ | to call wpa_supplicant |
10:33.50 | Pali | it is not easy as you think... |
10:34.00 | Pali | wpa supplicant does not support hooks |
10:34.05 | freemangordon_ | I don;t think it is easy :) |
10:34.12 | freemangordon_ | hooks? what kind of? |
10:34.29 | Pali | so first you need to modify wpa supplicant |
10:34.48 | Pali | get back events, when wpa supplicant change state etc |
10:34.58 | freemangordon_ | could you provide an example of such a hook, so /me to understand |
10:35.11 | Pali | wpa supplicant connected to AP |
10:35.24 | Pali | EAP auth finished |
10:35.31 | freemangordon_ | how does NetworkManager do it? |
10:35.45 | Pali | EAP auth failed because of wrong password |
10:35.55 | Pali | EAP auth failed because of unsupported configuration |
10:36.03 | Pali | and more EAP reasons |
10:36.05 | freemangordon_ | ok, how nm does it? |
10:36.18 | Pali | networmanager map every error message to wrong password |
10:36.24 | Pali | and ask you enter it again |
10:36.40 | Pali | even if error is different as wrong password |
10:37.01 | freemangordon_ | ok, wpa_supplicant has dbus api, I guess it will be easy to signal over the dbus with the correct error code |
10:37.17 | freemangordon_ | if it is not already there |
10:38.46 | freemangordon_ | Pali: https://w1.fi/cgit/hostap/tree/wpa_supplicant/dbus/dbus_old.h |
10:39.10 | Pali | dbus api does not provide all information |
10:39.34 | freemangordon_ | there is also dbus_new.h :) |
10:39.40 | Pali | both |
10:40.06 | freemangordon_ | you think it will be very hard to extend it to provide whatever needed? |
10:40.26 | Pali | two years ago I planned to write my own software as replacement for network manager |
10:40.52 | Pali | and I stopped it because wpa supplicant was not good to write WPA and EAP stuff correctly |
10:42.13 | freemangordon_ | well, I guess it developed for the last 2 years |
10:42.18 | bencoh | oh, by te way |
10:42.23 | bencoh | jonwil: http://abicheck.sourceforge.net/ |
10:42.27 | bencoh | might help you :) |
10:42.54 | freemangordon_ | Pali: do you have any other idea for eapd replacement? |
10:43.09 | Pali | no, reason why I did not finished my SW |
10:43.29 | Pali | wpa supplicant is probably the best solution from all existing |
10:43.31 | freemangordon_ | also, we don't want to replace icd2, but eapd :) |
10:43.42 | Pali | does not matter |
10:43.55 | Pali | icd2 is another piece of broken SW |
10:44.34 | jonwil | replacing eapd with wpa_supplicant isn't that hard if we can deal with the dbus and gconf stuff eapd does |
10:44.39 | freemangordon_ | BTW, I don;t see what is the problem with wpa_supplicant not providing enough information. Who uses that missing info and what kind it is? |
10:45.52 | freemangordon_ | also, I don;t believe wpa_supplicant does not provide enough information to distinguish the error reasonss |
10:46.00 | freemangordon_ | dbus api that is |
10:46.24 | Pali | wpa supplicant is also used for choosing and connecting to wifi networks |
10:46.37 | Pali | but this functionality is probably implemented in icd2 |
10:46.54 | Pali | also scanning for wifi networks is in wpa supplicant |
10:47.11 | Pali | so first you need to remove that code from icd2 |
10:47.14 | freemangordon_ | Pali: but I think we can just strip that functionality |
10:47.19 | Pali | and let wpa supplicant do that job |
10:47.21 | kerio | yea fuck icd2 |
10:47.32 | freemangordon_ | we miss the UI replacement |
10:47.42 | freemangordon_ | but yes, we can do that as well |
10:47.55 | Pali | in that case I would suggest to use eapd |
10:48.00 | freemangordon_ | and we'll kill 2 closed blobs at once |
10:48.39 | Pali | basically we should use full wpa supplicant |
10:48.51 | Pali | otherwise I do not see any reason to try hack it into maemo |
10:49.39 | freemangordon_ | what problems do you expect if we choose that? |
10:50.04 | Pali | roaming support |
10:50.15 | freemangordon_ | (whatever it is) |
10:50.29 | Pali | and similar stuff under WPA together with EAP |
10:50.40 | Pali | wpa supplicant should be in scanning state |
10:50.52 | Pali | when it manage wifi network fully |
10:51.07 | jonwil | The packages that should be potential targets for replacement here are: |
10:51.11 | jonwil | libicd-network-eap |
10:51.22 | jonwil | libicd-network-ipv4 (maybe, not sure what that does) |
10:51.29 | jonwil | libicd-network-wlan (maybe) |
10:51.33 | jonwil | libicd-network-wps |
10:51.37 | jonwil | osso-wlan-security |
10:51.46 | jonwil | osso-wlan (maybe, its open source though) |
10:51.50 | Pali | libicd-network-ipv6 too |
10:51.58 | jonwil | that isn't shipped on a stock install |
10:52.17 | Pali | yes, it is in diablo or fremantle extras |
10:52.47 | jonwil | As for UI we have connui-conndlgs-wlan and connui-iapsettings-wlan to deal with |
10:52.55 | jonwil | replace, clone or otherwise deal with |
10:53.31 | jonwil | icd2-network-wlan-config comes into this too (although that's a non-binary package) |
10:53.33 | Pali | now if you are opening question about killing icd2 (+all plugins)... do you want to look at my SW for managing networks? |
10:53.54 | jonwil | I am not suggesting we replace icd2, just all the wlan stuff |
10:54.53 | jonwil | Might hit some of this wlan stuff in IDA to see what its interface is |
11:50.21 | freemangordon_ | doesn't Jolla use wpa_supplicant? |
12:36.50 | Pali | probably yes, together with connman |
12:38.38 | freemangordon_ | I guess we should be able to reuse it on maemo |
13:08.20 | jonwil | FYI, replacing all of icd2 instead of just the WiFi bits will be a lot more work |
13:29.15 | *** join/#maemo-ssu Pali (~pali@Maemo/community/contributor/Pali) |
13:35.52 | *** join/#maemo-ssu Pali (~pali@Maemo/community/contributor/Pali) |
13:53.18 | *** join/#maemo-ssu Pali (~pali@Maemo/community/contributor/Pali) |
13:56.29 | *** join/#maemo-ssu freemangordon_ (~ivo@213.222.56.174) |
14:12.40 | *** join/#maemo-ssu Pali (~pali@Maemo/community/contributor/Pali) |
14:41.51 | *** join/#maemo-ssu freemangordon_ (~ivo@213.222.56.174) |
14:58.57 | *** join/#maemo-ssu DrCode (~DrCode@5.28.134.3) |
17:37.47 | *** join/#maemo-ssu futpib (~futpib@176.104.194.78) |
19:42.48 | freemangordon | merlin1991: please import .debs in cssu-devel/incomming |
20:14.30 | *** join/#maemo-ssu _rd (~rd@p5B2C7F08.dip0.t-ipconnect.de) |
21:05.55 | *** join/#maemo-ssu jonwil (~jonwil@27-33-80-219.tpgi.com.au) |
21:12.06 | jonwil | hi |
21:28.02 | *** join/#maemo-ssu starbuster (~g@194.44.84.246) |
21:31.05 | *** part/#maemo-ssu starbuster (~g@194.44.84.246) |
23:08.30 | merlin1991 | freemangordon: done, sry about the missing autoimport, will fix it on the weekend |
23:25.44 | jonwil | Anyone here know anything about wpa_supplicant (in general, not as it applies to maemo) |
23:30.10 | *** join/#maemo-ssu M4rtinK (~M4rtinK@ip-89-177-124-111.net.upcbroadband.cz) |