libera/#maemo-leste/ Tuesday, 2024-06-11

freemangordondsc_: it is removed after restart only if you didn't abuse it before restart by typing something in the chat window00:00
dsc_right00:00
dsc_yes thats wrong00:00
dsc_will fix00:00
dsc_and also do what you said Re: groupchats00:00
freemangordonok, cool00:00
freemangordongreat00:00
freemangordonperhaps still make a new release before that, because of the segfault fix00:01
dsc_reading the diffs00:01
freemangordonok00:01
dsc_well, I said I will do what you said Re: groupchats but I have not thought of all the "situations" (states) it could possibly go into00:04
dsc_so will verify we can actually do that00:04
freemangordonyeah, makes sense00:05
freemangordonI think it should be ok, but I could have missed some corner case00:05
dsc_i think so too00:06
dsc_(that it will work)00:06
freemangordondsc_: one more thing: I am not sure how 'own' scrollback messages are stored in rtcom db, but they appear in addressbook 'recent contacts'00:20
Wizzupthis is probably something for us to figure out00:21
Wizzup(fmg)00:21
freemangordonyeah, maybe00:22
freemangordonI'll trace what log_event does00:22
Wizzupthis depends on the recent contacts query00:22
freemangordonoh,   remote_alias is nullptr for message sent from the device00:23
dsc_sqlite> select * from GroupCache;00:23
dsc_8|3|idle/irc/dscqemu0-##maemotest|8|8|000:23
dsc_interesting00:23
freemangordonWizzup: could that  remote_alias make the difference00:23
* freemangordon checks00:24
dsc_new build btw00:39
freemangordondsc_: wait00:40
freemangordon(if not late :) )00:40
freemangordonI have another fix, just have to make the commit00:41
dsc_more contributions are always good :)00:41
dsc_and yes you are too late00:42
freemangordonok, I have to test it anyways :)00:43
dsc_so we can look into how to get rid of the config stuff in lib/tp.cpp00:47
dsc_its a bit strange to me, pragmatic solution to save persistent data related to TelepathyChannels:00:47
dsc_- auto-join00:48
dsc_- last message received (prevent duplicates)00:48
dsc_- render things in the overview, needed in some situations00:48
dsc_perhaps for all 3 points we can find solutions in favor of getting rid of `AccountChannel*`00:48
dsc_e.g: how to save auto-join in rtcom somehow00:48
freemangordonI don;t think that belongs to rtcom00:49
freemangordonthis is application setting00:49
dsc_true00:49
dsc_what `AccountChannel*` does is create a temp. obj that represents a channel, in the case Tp has not given us this channel yet00:51
dsc_architecturally, it is a bit suspicious00:52
freemangordonwell, I was wondering why it is needed00:52
dsc_as lib/tp.cpp should only involve Tp stuff00:52
freemangordondsc_: https://github.com/maemo-leste/conversations/commit/b55dc99772b7b0e06c95fb2c56103d19d12ba83e00:54
freemangordonI'll appreciate if you make a new release00:54
dsc_nice00:54
freemangordonthanks!00:56
dsc_:)00:56
Wizzupok, I think locally fixed most of the voicecallmgr problems with sphone apart from the randomly no audio issue01:42
Wizzupgoing to try to look into that now01:42
Wizzupuvos__: so I checked amixer and during calls with voicecall manager and the controls seem to be identical in calls where there is audio and calls where there is no audio02:02
Wizzupso userspace seems to be doing what it should be doing02:03
WizzupI did not see any errors when asking pulseaudio to switch UCM or not02:03
WizzupI tried to manually change alsamixer controls and pulseaudio UCM - switch back and forth, and that didn't make it work either02:03
Wizzupso I'm at a loss, I don't think voicecallmanager is doing anything wrong here02:04
Wizzupvoicecall-manager (daemon) debug output is also identical when it works and when it doesn't work02:06
WizzupI guess I'll compare kernel regs next02:08
Wizzupuvos__: /sys/kernel/debug/asoc/Mapphone Audio/cpcap-codec.0 is also identical when it works / doesn't work02:19
Wizzupuvos__: what I can tell right now is that from the relevent regs the CPCAP_REG_TXI (0x0814) seems to be different02:33
Wizzup0814: 00c2 when it's only one side (remote side can't hear me) and 0814: 0cc2 when it's both sides02:33
Wizzupapparently the /sys/kernel/debug/regmap/spi0.0/registers file is no longer writable so I can't test whether that is it for sure but it mist be02:36
Wizzupso CPCAP_BIT_MB_ON1R and CPCAP_BIT_MB_ON1L are somehow not set02:44
Wizzupthis is called the MIC1R Bias and MIC1L Bias in alsamixer02:45
Wizzuptmlind: you might have some ideas too perhaps02:46
Wizzup(this is 6.1)02:47
Wizzupalso since when is the registers file no longer writeable?02:50
Wizzupmy regtool says as much too:03:09
Wizzup python regtool.py cmp regs_both_works.txt regs_onesideonly.txt03:09
WizzupDifference in CPCAP_REG_TXI.CPCAP_BIT_MB_ON1L: 1 != 003:09
WizzupDifference in CPCAP_REG_TXI.CPCAP_BIT_MB_ON1R: 1 != 003:09
Wizzuphm looks like the define is REGMAP_ALLOW_WRITE_DEBUGFS and I don't see a CONFIG_ option for it03:16
Wizzupuvos__: also very unrelated but if we don't display the dialer buttons in sphone we can't send dtmf03:18
Wizzupany reason not to have those buttons at least clickable within a call?03:18
Wizzup(well visible and clickable)03:19
tmlindWizzup: there's some kernel kconfig debug option for enabling regmap write06:05
tmlindWizzup: i think it's the REGMAP_ALLOW_WRITE_DEBUGFS that you have to define manually06:07
uvos__Wizzup: no reason, its just that the dialer and the call manager are different plugins entirely that thus far share no code08:59
uvos__tmlind: no you have to chainge a define in the code08:59
uvos__tmlind: ah yes exactly09:00
uvos__Wizzup: the registers where never writeable09:01
uvos__Wizzup: at least not in the time span i have been working with this09:01
arno11dsc_: freemangordon: conversations update to 0.7.15 (from 0.7.10): it takes only 1 or 2 sec max to show the UI on N900 now, great !09:02
uvos__i think its pretty uff for a running applcation to take 2 seconds just to show a window09:03
uvos__but you kown :P09:03
arno11:P09:04
uvos__Wizzup: the main reason why i never implemented dtmf is because on d4 dtmf is exposed via alsa, which is wrong place for it to be09:05
uvos__Wizzup: and for me to implement / test dtmf in sphone i need some platform where dtmf works correctly via ofono's interface09:05
sicelouvos__: you likely have "some platform where dtmf works correctly via ofono's interface" :-p09:39
uvos__sicelo: ok "some platform where dtmf works correctly via ofono where i wont die of brodom wating for sphone to compile" :P10:06
uvos__anyhow i will implement the ui bits for dtmf, thats not so hard10:19
uvos__maybe Wizzup can deal with the backend side10:19
uvos__Wizzup: unfortionaly this means that the non-working audio in vcm and the non-wroking audio in 6.6 are unrelated10:21
uvos__i was hopeing those where the same problem :|10:21
freemangordonarno11: which UI? main? I would say the better test would be the time needed to show chat11:06
Wizzupuvos__: they might be related but thye are not the same problem11:29
Wizzupuvos__: so in 6.6 for example 0618 is set correctly11:29
Wizzupuvos__: I wrote a python script to bring the audio regs in sync with certain text dump of the regs, but I guess I just modified the define yeah11:30
Wizzupuvos__: oh you wrote unrelated :)11:31
Wizzupyes11:31
Wizzupwell, 6.6 is the next step11:31
arno11freemangordon: i meant UI: it tooks 10 sec before last update...12:47
arno11chat UI is still around 3-4 sec12:47
arno11*main UI12:47
arno11arghh sorry that's unclear12:48
arno1110 sec for main UI before update12:48
arno11now 1 or 2 sec12:48
arno11chat UI is 3-4 sec12:49
dsc_so regarding chat clearing/removal, i will keep this functionality as-is for a few days so that people can use/test it13:52
dsc_p2p needs fixing there, other than that its good13:53
dsc_if we then still want something else, I would like to start a discussion on Github13:53
freemangordonwill you fix p2p?13:54
dsc_because merging those 2 features into 1 *may* introduce some corner cases that need to be explored, and I would like to discuss them in a more formal way on Github where we have markdown etc.13:54
dsc_freemangordon: yes, today13:54
freemangordondsc_: I think we shall merge and *iff* there are corner cases we shall fix13:55
freemangordonthere is no 'delete chat' functionality in fremantle, still ,I don;t remember anyone saying it is missing13:56
dsc_perhaps, lets see what e.g uvos/arno think13:56
WizzupI don't ever clear chats fwiw so I am not a good tester13:56
freemangordonI do13:56
Wizzupand I agree what fremantle does seems sensible to me13:56
dsc_Wizzup: you have not tried the feature yet at all13:56
dsc_hence I request to first look at it ;)13:56
freemangordonfrom time to time I clear chat history to clean-up db and free some RAM13:56
freemangordondsc_: again, for p2p chat, 'delete chat' makes no sense13:57
freemangordonfor group chats, "delete chat" is "leave group", no?13:58
Wizzupdsc_: from how rtcom works and I have my preference :)13:58
WizzupI did look at the code13:58
Wizzuptmlind: uvos: so any suggestion how to debug the CPCAP_REG_TXI.CPCAP_BIT_MB_ON1L and CPCAP_REG_TXI.CPCAP_BIT_MB_ON1R regs not being set?14:01
Wizzupbut only in some cases14:01
WizzupI didn't see any difference in amixer settings or dapm settings14:02
WizzupI could try to add some printk statements to kernel code or something14:02
dsc_freemangordon: yes I agree but there are some implications that is hard to discuss on IRC; what should happen in the overview when X is done14:02
dsc_X being clearing, or removal, whatever14:02
dsc_we can discuss it on IRC, but I would prefer an issue so I have time to think about presenting some situations/cases14:03
dsc_and we can solve those edge-cases there14:03
freemangordondsc_: ok, understand14:37
freemangordonso, please fix p2p functionality (which includes no "delete chat" for p2p)14:37
freemangordonand then we'll continue on GH issue tracking14:38
arno11dsc_: btw i tried clear/delete functionalities, works fine on my device, very useful in case of unwanted/spam SMSes14:49
arno11and probably useful to free up some RAM, specialy on N900 :P14:51
arno11and again launch time is very good now14:53
Wizzupit can get better still if we add some dbus thing iirc14:53
Wizzupthen h-d will just focus the conversations window14:53
arno11ah cool14:54

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