libera/#maemo-leste/ Wednesday, 2024-05-22

Wizzupuvos: I don't know if something changed but I just received a message that I sent to someone on telegram earlier using the web interface00:16
Wizzuplatest tp-haze and conversations00:16
Wizzupit does not seem to have been logged though, I will investigate00:18
Wizzupthere is an issue however with telegram logging when you make it a contact I think00:21
WizzupI see a group_uid like haze/telegram_tdlib/_asdfasf and haze/telegram_tdlib/_asdfasf-ContactName00:21
Wizzupso they do not get grouped well00:21
Wizzupthe self messages do not get logged correctly either, they get logged like it is a new contact00:31
Wizzupmy last message is incorrect, the group uid was ok01:44
freemangordonWizzup: yes, something changed: https://github.com/maemo-leste-upstream-forks/telepathy-haze/commits/maemo/chimaera-devel07:19
freemangordonin particular https://github.com/maemo-leste-upstream-forks/telepathy-haze/commit/686e10be593ad757a26728e505ba4747e0e5749f07:19
freemangordonand yes, self messages are incorrectly logged07:19
freemangordonat least with haze/fb remote messages appear in conversations like if they were from the other party07:21
freemangordondsc_: perhaps valgrind can help07:23
freemangordonor libasan07:24
freemangordonyou just need to find a way to make conversations exit properly07:25
Wizzupfreemangordon: they do appear like that bad the group_uid is ok, so I will investigate11:05
Wizzuplike that but*11:06
freemangordonok, thanks11:08
freemangordonalso, I wonder if we have scrollback support11:08
freemangordonWizzup: shall I have a look to stop notification being issues for the currently active chat on a new message arrival?11:09
freemangordonwe just need a simple 'click' sound or something11:09
Wizzupwhat do you mean with 'scrollback support' ?11:10
freemangordonwhen device is not locked that is ;)11:10
freemangordonfull history of a chat11:10
freemangordonor partial history11:10
Wizzupthat I understand, but this is also telepathy concept so I am trying to understand what you refer to :D11:10
freemangordonyes, that's what I meant - tp concept of scrollback11:11
freemangordonI already added some support for group chats11:11
Wizzupok, so what we would need for this (ideally)11:15
Wizzup1. matching scrollback matches to existing history11:15
Wizzup2. getting the right scrollback that we actually need (optional, requires also protocol specific stuff)11:15
Wizzupright now we get the same scrollback every time you join a channel and we log it11:16
Wizzupso we get ots of double messages11:16
freemangordonumm, how's that (same scrollback)?11:20
freemangordonwhich proto os that?11:21
freemangordonas I fixed haze (with fb at least) for 'seen' support11:21
freemangordonso no more duplicated messages shall arrive11:21
freemangordonhowever, now we miss the 'seen on other devices' history11:22
freemangordonBTW, I think TP concept of matching is by message itself11:24
freemangordonlike, message ids are not really usefull11:24
freemangordonbut I think we can easily implement that by keeping a message text hash in the DB11:24
freemangordonso we can match lists of consecutive hashes, which should be trivial11:25
Wizzup11:24 < freemangordon> BTW, I think TP concept of matching is by message itself11:26
Wizzupthis might be true, but that won't work for some protocols11:27
Wizzupfor MAM11:27
Wizzup(in xmpp)11:27
freemangordonhttps://telepathy.freedesktop.org/doc/telepathy-glib-1/TpMessage.html#tp-message-dup-token11:27
Wizzupbut yes, we should try to match scrollback to existing messages as step 111:27
freemangordonI don;t think we have any other option11:28
freemangordonsee ^^^11:28
Wizzupyeah so in xmpp when done right I think this will be11:28
Wizzupbut for others it might not be11:28
freemangordonsorry, can;t parse11:28
freemangordonxmpp or not, we have TpMessage, period :)11:29
Wizzupthe message token can be unique there11:29
freemangordon*can*11:29
freemangordonbut is not guaranteed by the specs11:29
Wizzupwell if we can't do message archive management and such we will need to extend TP to be able to do it11:29
freemangordonno11:29
freemangordonwe just keep hash of the text for every message11:29
Wizzuphttps://xmpp.org/extensions/xep-0313.html11:29
freemangordonand do 'message window' matching upon scrollback message being received11:30
freemangordonalso, I think we have message timestamp11:30
freemangordonhash+ts should be enough to uniwue identify a message11:30
Wizzupok, I think you may want to reconsider this after some reading11:31
freemangordonmhm11:31
Wizzupif you can for example ask for the latest received message on the client side you can inform the server what to send to you11:31
Wizzupfor this some own custom hash won't suffice, you need the protocol specific message id11:32
Wizzupthis is the 'right' solution to getting the messages you want, not just the unread ones as facebook does currently for you11:32
freemangordonno, FB can do the full history11:32
Wizzupand this requires storing at least some message ids locally, potentially more, but that is just how this ought to work11:32
Wizzupok, but I doubt that you want the full history every time, and with message markers as in xmpp I think you have a much cleaner and easier way to get what you need11:33
Wizzupagain we don't need this now and gabble can't even do it11:33
Wizzupjust things to consider before we start filling db fields :D11:33
freemangordonright11:33
Wizzup(gabble can't even do it -now-)11:33
freemangordonre FB: you request the last N messages11:33
freemangordonor full history11:34
freemangordonnot sure how's that supposed to work11:34
freemangordonI guess it is same/similar for xmpp11:34
* freemangordon reads the specs again11:34
freemangordonoh, it seems I misunderstood it11:35
freemangordon"This is not guaranteed to be unique, even within the scope of a single channel or contact: the only guarantee made is that two messages with different non-empty tokens are different messages."11:36
freemangordonI am not sure I can grok the "...he only guarantee made is that two messages with different non-empty tokens are different messages." part11:36
WizzupIt is probably intentionally vague11:41
freemangordonhmm, libpurple does not seem to provide message ids :(11:47
Wizzuphttps://github.com/petermolnar/awesome-pidgin-plugins11:52
Wizzuphttps://issues.imfreedom.org/issue/PIDGIN-15653/Support-XEP-0313-aka-MAM-Message-Archive-Mgmt11:52
Wizzupso there's pidgin fork/changes with message markers for example11:53
Wizzupalso omemo plugins, would be interesting if we could just load those somehow11:54
Wizzupthis would of course be for tp-haze not gabble (these plugins)11:57
freemangordonok, what is the status of pidgin3?11:58
freemangordon#pidgin12:00
freemangordonmaybe join ^^^12:01
Wizzupk12:02
WizzupI wouldn't look at that now12:02
Wizzupwe have all we need to make this work on pidgin2 I think12:02
Wizzupsee my links bove12:02
freemangordonthey say they use timestamps, which is not exactly message ids12:33
freemangordonbuok, lets not focus on that now, we have more issues12:34
arno11weird: on my N900 conversations UI is still very slow to launch from hildon icon (around 10 sec, even after the first time). but if i start it from a fb chat or sms notification, it starts in 2 or 4 sec max.13:22
arno11*2-4 sec max to show UI and chat window13:25
Wizzupthis is because the first one launches conversations for conversations to find out it is already running and activating itself13:26
Wizzupthe second one I believe just lets conversations know on dbus that it needs to show itself13:26
Wizzupwe can probably have the conversations desktop icon also be a similar dbus call13:27
dsc_10sec while its already in the background?13:31
arno11yes13:31
arno11but not from notifications13:31
dsc_right13:31
dsc_its because clicking on a notification will just raise conversations13:31
dsc_while clicking will actually start a 2nd instance, at which point it finds out that its already running, and exit, and raise the original one13:32
Wizzupdsc_: I think we are in agreement ;)13:32
dsc_yes13:32
Wizzupbut we can have dbus start conversations potentially, and then make the icon just do a dbus call13:33
arno11lol indeed you are13:33
arno11makes sense imo13:33
dsc_is there something in dbus that can prevent running something 2 times?13:34
dsc_how does it check if it is already running?13:34
dsc_or do we then just launch a bash script that checks a lockfile13:35
uvosThe problem ist that QApplication::QAppilcation() takes way to long13:43
uvossphone also has this problem, when run with a qt module starting sphone via the icon takes long and i blocks in this call13:43
uvosif no qt module is used with sphone it starts and informs the other instance very fast13:44
dsc_yes, starting conversations just to check if an instance is already active is wasting resources13:44
dsc_it can be done with something simple13:45
uvossure, same with sphone, (well sphone is a lot better, since it loads all the modules after checking if it needs to raise the other instance)13:45
uvosbut still QApplication::QAppilcation() takes an unresonable amount of time regardless13:45
Wizzupthis we could probably really 'fix' with a maemo-launcher  qt booster13:52
Wizzupuvos: so sorry to pester you with this, but what can I do on razr with android to see if we can fix the screen issue?13:52
uvosWizzup: i mean you could look at /diff the DSS registers see page 2192 in the TRM14:27
uvosbut its a shot in the fairly dark14:27
Wizzupok14:27
Wizzupso we think it's likely not a dts issue but rather a driver issue?14:27
uvoscould be either really14:28
Wizzupok14:28
Wizzupand I think for atrix we made some notes on how to make the dts changes, how it was different from the bionic14:34
Wizzupdid you host those somewhere?14:34
Wizzupthere is the signal map I see14:34
kivaI was 2 1/2 hour MS Teams meeting with Pinephone Keeyboard thru Firefox 115.1. Mic keeps muted so had to write chat of it, when I wated answer. Mic worked about half yeas ago...I think it was bug in M$-code, it did not understand that mic was allowed by Firefox.15:26
kivawated=wanted15:26
kivaAfter that meeting there was stll 39% battery left..thanks for two batteries of PPK.15:29
kivaAnyway it might be also good thing that M$ does not know how to listen mic from Maemo Leste device ;)15:31
siceloteams meeting via phone (3g/4g)? or how?15:37
kiva4g15:37
kivaconnecting with Firefox with 4G internet connection.15:38
sicelooh wow. teams web?15:39
kivayes15:40
freemangordonkiva: on leste? you are my hero :)16:03
uvosWizzup: afaik as i know there is no difference between the atrix and the bionic16:04
uvossensors and modems aside16:04
uvosi think it would just boot a bionic image16:04
Wizzupuvos: right, but we need to make a dts for the diff sensors at least16:06
Wizzupiirc we made some list on irc of what was different16:06
WizzupI g2g16:06
kivafreemangordon: ofcourse with Leste.16:07
uvosWizzup: de decoded the dtb and the signal map16:08
uvosboth of those are on my server16:08
uvosyou can diff them with bionic for differences16:08
uvosbut iirc there are none besides the mentioned sensors and different modem16:08
uvossetup16:08
uvosthe real issue is that it needs clownboot which is annoying16:09
uvosyou could add dmesg for mb865 seams we lack it16:10
Wizzupuvos: yeah I checked your server :)16:46
WizzupI think I already got clownboot sorted?16:47
WizzupI have a working atris 2 with leste here16:47
Wizzupatrix16:47
Wizzupiirc it was the same16:47
Wizzup(as bionic)16:47
dsc_clownboot? :P16:52
Wizzupdouble kexec16:58
uvossilly boot method i devised for devices with locked bootloaders that have no known vunerabilities16:59
uvosthat involves kexecing twice16:59
inkykiva: i think with chromium you would not have a mic problem.22:23
inkykiva, also advice from a former corporate employee (and jobless now) - i compiled pidgin's teams plugin and i was able to be always online for them from my droid4 and leste.22:25
Wizzupit would be nice/cool to have that packaged like we packaged the other libraries23:38

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