libera/#maemo-leste/ Saturday, 2024-11-30

Wizzupfreemangordon: can you look at this tomorrow? https://phoenix.maemo.org/job/hildon-input-method-configurator-binaries/architecture=amd64,label=amd64/16/console02:55
Wizzuplooks like we have two hildon-im-protocol.h02:55
WizzupIn file included from /usr/include/hildon-input-method/hildon-im-ui.h:34,02:55
Wizzup                 from /usr/include/hildon-input-method/hildon-im-ui.h:26:02:55
Wizzupoh I suppose this just needs some ifdef guard02:55
WizzupI'll look tomorrow02:57
Wizzuphm the error is not there on chimaera03:04
WizzupI think03:04
freemangordonWizzup: MAEMO_CHANGES is not defined08:28
freemangordonthats not good08:29
freemangordonfails in chimaera VM as well08:33
freemangordonMAEMO_CHANGES is defined by hildon-1.pc08:34
freemangordonbut it is not used08:34
freemangordonhttps://github.com/maemo-leste/hildon-input-method-framework/commit/a38e68ad9933433b9d1baef266572f44ade26e5d08:35
freemangordonperhaps hildon-input-method-configurator was  build before that patch08:36
freemangordonWizzup: fixed09:20
freemangordondsc_: so, on fremantle, it is /usr/share/telepathy/clients/NotificationUI.client that does the authorization requests09:21
freemangordonbut, it takes care of everything, like incoming sms-es, chats, calls, whatnot09:22
freemangordonI am tempted to RE it though09:24
freemangordonhmm, I will have to09:39
freemangordonthis is the one that shows call icon in the status menu09:39
Wizzupfreemangordon: thx10:03
Wizzupfreemangordon: we can show the call icon in the status menu, it's not that hard I think10:03
freemangordonWizzup: who are 'we'?10:06
freemangordonI know we can, but we don;t10:07
freemangordonwho will do that? sphone? conversations?10:07
Wizzupstatus applet?10:07
freemangordonthat's what I am on10:07
freemangordonthe one from fremantle that does it :)10:07
Wizzupcan't it use connui-cellular to get call state, or monitor dbus mce call state change?10:07
freemangordonit can10:07
Wizzupok, it seemed like you said you'd RE the other parts too10:08
freemangordonmaybe I will10:08
freemangordondepends on what they do10:08
freemangordonBTW in fremantle it listens to TP for call state, iiuc10:08
freemangordonwhich makes sense10:09
freemangordonWizzup: also, I think MFA belongs to the same plugin11:22
WizzupI don't know if that is the case12:35
freemangordonwhat do you mean?12:39
freemangordonsomeone shall register with TP for MFA, no?12:39
freemangordonplease elaborate, it is possible taht I am missing something12:40
siceloMFA ... multi factor auth? What authorization requests did you mean? if incoming friend requests, then MFA doesn't sound like a good fit for that12:55
freemangordonno, I mean that the same place we handle auth requests is good place to handle MFA as well12:55
sicelomfa would be in whatever handles adding/registering the account12:55
freemangordonit does not work like that with TP12:56
sicelooh12:56
freemangordonyou should register some kind of agent to handle MFA and friends12:56
freemangordonin fremantle incoming friend requests are handled in rtcom-notifications-ui12:57
freemangordonit also handles call status icon12:57
freemangordonthat's why I think it is also a good place to handle MFA etc12:58
freemangordonare there any issues with that?13:00
sicelosounds unintuitive, but I guess that's due to lack of understanding of TP13:00
sicelohow does TP handle account password? e.g. Facebook13:01
freemangordonsee https://telepathy.freedesktop.org/spec/Channel_Interface_SASL_Authentication.html13:03
freemangordonsimple password is another story13:03
freemangordon"This can be used to connect to protocols that may         require a password, without requiring that the password is saved in         the Account.Parameters."13:03
freemangordonso, someone shall handle Channel.Interface.SASLAuthentication13:04
freemangordonif we want to stay compliant ofc :)13:04
freemangordonotherwise everyone and her dog may implement whatever UI she likes, but I guess that's not teh point13:05
siceloah, I see13:07
sicelolooks like that can also work for e2e (matrix,telegram)13:09
freemangordonmhm13:09
freemangordonand it is better to be in a system-wide service13:10
sicelo. In future, it could also be used to authenticate with secondary services, or even to authenticate end-to-end connections with contacts. As a result, this interface does not REQUIRE ServerAuthentication to allow for a potential future Channel.Type.PeerAuthentication interface.13:10
freemangordonthat way we do not depend on particular UI (call, chat, etc)13:10
siceloso basically with TP, the MFA challenge is like chatting with a first friend ...13:15
freemangordoncould be, I lack the details13:15
Wizzup12:40 < freemangordon> please elaborate, it is possible taht I am missing something13:38
WizzupI am not sure if that should be the status applet of conversations13:38
Wizzupmaybe for login tokens it can be status13:38
Wizzupfor friend auth requests it shouldn't be13:38
WizzupI think?13:38
Wizzupconversations ought to handle those I think13:38
freemangordonWizzup: why should conversations handle it?13:56
freemangordonlike, how it is better if handled in conversations?13:57
freemangordonif we do chat/auth/mfa/etc notifications out of convaersations, this https://github.com/maemo-leste/conversations/tree/master/src/lib/libnotify-qt will be gone13:59
freemangordonI don;t think having libnotify-qt being a part of conversations is the best approach14:00
Wizzupthat won't be hard I think14:25
Wizzupwe can just package it14:25
freemangordonok, but how it is better?14:26
Wizzupbecause conversations is the program that logs/writes chat messages, so it only makes sense to deal with ayth reqs14:26
WizzupI didn't say mfa14:26
freemangordonATM it is abopok that sends auth requests14:26
freemangordon*abook14:26
freemangordonshall we remove that as well?14:26
WizzupI thought the auth reqs were part of the conversations dev tree currently, am I wrong?14:27
freemangordonyou are14:27
Wizzupif it goes through abook then that's ok with me14:27
Wizzupdoing -all- notifications throug the status menu (of chat messages) seems wrong to me14:27
freemangordonit is not only about chat14:27
freemangordonerr...14:27
Wizzupthen let's not make status do chat :)14:27
freemangordonit is not about chat14:28
freemangordonit is about auth14:28
Wizzup>if we do chat/auth/mfa/etc notifications out of convaersation14:28
freemangordonalso, voice mail notification is there14:28
WizzupYou suggest here to do chat notificatios out of conversations14:28
freemangordonyes14:28
Wizzupand I'm saying I think that's not a good idea14:28
freemangordonthat way *all* rtcom notifications will be in one place, IIUC14:29
freemangordonso call ui will not have to take care of missed calls etc14:29
freemangordonnotifying that is14:29
Wizzupbut, why?14:30
Wizzupwe have it working now14:30
Wizzupwhy go through all the effort to make all the business logic in a status applet14:31
freemangordonwe don; at least does not work properly14:31
freemangordonunless I am missing something14:31
freemangordonok, let me gain a better understanding of what this notifications-ui is doing. I will explain in details and will have another discussion then14:33
freemangordonfor now I will implement at least call status icon14:33
freemangordonok?14:36
Wizzupsure14:36
WizzupI'd just hate for us to have re-do all the work we put in conversations and rtcom logging of all the messages + notifications14:37
Wizzuphandling incoming auth requests/2fa is very different14:37
freemangordonright14:38
freemangordonso at least auth request shall go there IMO14:39
freemangordonincoming auth request handling that is14:39
WizzupI am fine with that, I wonder what dsc thinks14:47
WizzupI ma back to more daedalus today, hopefully make it so that things can boot with network working14:48
freemangordonI think he'll be fine with that14:48
freemangordondsc_: ^^^14:48
siceloWizzup: ah, that's why wlan doesn't want to connect on the L5, is it?14:58
siceloseemed that icd2 segfaults or something (but it was late at night and i was exhausted.)15:05
Wizzupsicelo: yeah15:55
Wizzupprobably15:55
Wizzuplibwpa is an .a file15:55
Wizzupso we need to rebuild it15:55
Wizzupgimme a few hours15:55
Wizzupfreemangordon: I haven't looked at daedalus and gnome tracker, do we want to build our own tracker forks for now? I don't think we have to worry about this now necessarily, just wondering16:44
freemangordonI thinkth it is tracker 3.x there16:53
freemangordon*think16:53
freemangordonwhich means I will have to port mafw stuff16:53
freemangordonand no, I think we shall not build our own tracker, we fixed bugs there, no?16:53
WizzupI am not sure about the last statement16:56
Wizzupwe definitely fixed bugs and/or cherry picked fixes16:56
freemangordonI meant: I don;t think we shall build our tracker fork, as it is 2.x16:58
freemangordonand we just fixed some bugs in it16:58
freemangordondaedalus comes with 3.x (if I saw that correctly yesterday doing dist-upgarde on sicelo's L5)16:59
freemangordonI assume the patches we cherry-picked in our tracer are already in 3.x16:59
freemangordonsounds sane?17:00
Wizzupyeah I think so17:00
freemangordonok17:00
Wizzupand I think this means that mafw-tracker-source should not be build yet since it needs tracker 3+17:00
Wizzupif I got this right?17:01
freemangordonmost-probably you can;t build it17:02
freemangordonso yeah, do not try to17:03
freemangordonLMK when we are at it17:03
freemangordonI will port it17:03
WizzupI can try to build it if you want, it's next up17:06
Wizzupbut I can continue and ignore it for the time being, so it's not a blocker17:06
Wizzupfyi https://github.com/maemo-leste/bugtracker/issues/75117:06
freemangordonnot bad :)17:07
WizzupI will make a lot more progress toda17:10
Wizzupy17:10
freemangordonmafw-upnp-source? did it ever compile?17:13
Wizzupprobably not, I will check it later17:22
WizzupI guess we need to sort out mesa too17:24
Wizzupmaybe our chimaera mesa is newer than daedalus still17:24
Wizzup(since we have our own)17:25
Wizzupah no, we're on 21.x and daedalus is 22.x17:26
sicelo:-)17:29
siceloand that was the blocker for L5 all along17:29
Wizzupmesa?17:33
Wizzupnot xorg?17:33
siceloactually both17:33
siceloin fact we may need an even newer mesa than what's in daedalus. i still had stuff not redrawing, e.g. vkb all black, and letters appear only when i touch them17:34
siceloor status menu showing up all black, until i press the buttons 'blind'17:35
Wizzuphmm, ok17:35
WizzupI was hoping we wouldn't, but we will see :D17:35
Wizzupthen we also need newer libdri, libglvd, libdrm, etc17:36
Wizzupbut it's ok, we did that in the past oto, so17:36
Wizzuptoo, so17:36
siceloyeah i'll test properly during course of the upcoming week.17:36
Wizzupfreemangordon: do I still build libgofono for daedalus?17:43
Wizzupor do we not need it with the upcoming connui-cellular17:43
Wizzupiirc we don't, right?17:43
freemangordonwe don;t, but icd plugin still requires it17:44
Wizzupoh?17:46
Wizzupok17:46
freemangordonhttps://github.com/maemo-leste/libicd-network-ofono/blob/master/debian/control#L517:47
Wizzupyes I see17:50
dsc_<freemangordon> I think he'll be fine with that22:39
dsc_yes, as long as the UX is good, user should be #122:39

Generated by irclog2html.py 2.17.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!