uvos | can some one think of a hildon application with stackable windows thats also fullscreen? | 01:50 |
---|---|---|
Wizzup | freemangordon: I will look at this tomorrow but something seems to bring tp into off state oe so | 02:21 |
Wizzup | when I enable mobile data but set ui presence to offline, tpring appears in sphone | 02:21 |
Wizzup | if I turn off mobile data it is gone | 02:21 |
freemangordon | what ui presence? IIRC tpring should not appear in accounts ui | 07:40 |
* freemangordon checks | 07:40 | |
freemangordon | https://github.com/maemo-leste/rtcom-presence-ui/blob/maemo/chimaera/lib/pui-master.c#L926 | 07:48 |
freemangordon | so, if tpring appears in presence ui, that's a bug | 07:49 |
freemangordon | also, please provide result of mc-tool show ring/tel/account0 | 07:52 |
arno11 | the new release works fine on n900: boost freq are ok, we have overclock ootb now, nice | 08:01 |
freemangordon | arno11: nice | 08:07 |
freemangordon | Wizzup: also, ring supports neither TP_IFACE_CONNECTION_INTERFACE_PRESENCE nor TP_IFACE_CONNECTION_INTERFACE_SIMPLE_PRESENCE | 08:07 |
freemangordon | so you should not try to check the presence on ring account | 08:07 |
freemangordon | it is either enabled or disabled | 08:08 |
freemangordon | see https://github.com/maemo-leste/rtcom-presence-ui/blob/maemo/chimaera/lib/pui-master.c#L188 | 08:14 |
Wizzup | freemangordon: will check | 10:43 |
Wizzup | freemangordon: could be that voicecall program doesn't deal with this properly | 11:28 |
Wizzup | freemangordon: lol so this is a tp-ring special path: | 11:31 |
Wizzup | if (d->shouldForceReconnect()) { | 11:31 |
Wizzup | WARNING_T("Forcing account %s back online immediately", qPrintable(d->account.data()->uniqueIdentifier())); | 11:31 |
Wizzup | d->account->setRequestedPresence(Tp::Presence::available()); | 11:31 |
Wizzup | } | 11:31 |
Wizzup | I think I have an idea what is going on | 11:34 |
Wizzup | freemangordon: looks like tp-ring reports presence on dbus | 11:37 |
Wizzup | it looks like voicecall things that it's not even 'online' | 11:38 |
Wizzup | thinks* | 11:38 |
Wizzup | but why would tp-ring even act on things like connecting to wifi, feels like something else is instructing it | 11:41 |
freemangordon | Wizzup: I see nothing like that here | 11:42 |
freemangordon | do you have empathy or something installed? | 11:43 |
freemangordon | also, you *cannot* request presence for tpring | 11:43 |
freemangordon | it does not implement the required interface | 11:43 |
freemangordon | maybe they have that patched in SF | 11:44 |
Wizzup | what causes the tp accounts to connect when connected to the network | 11:44 |
Wizzup | freemangordon: it seems happen to change presence in our repo | 11:44 |
Wizzup | s/happen/happy/ | 11:44 |
freemangordon | what presence? | 11:45 |
Wizzup | signal time=1702550119.441680 sender=:1.39 -> destination=(null destination) serial=998 path=/org/freedesktop/Telepathy/Account/ring/tel/account0; interface=org.freedesktop.Telepathy.Account; member=AccountPropertyChanged | 11:45 |
Wizzup | string "CurrentPresence" | 11:45 |
Wizzup | variant struct { | 11:45 |
Wizzup | uint32 1 | 11:45 |
Wizzup | string "offline" | 11:45 |
Wizzup | string "" | 11:45 |
Wizzup | } | 11:45 |
freemangordon | no | 11:45 |
freemangordon | wait | 11:45 |
freemangordon | please give me 'mc-tool show ring/tel/account0' | 11:45 |
freemangordon | forget about dbus | 11:46 |
freemangordon | use mc-tool | 11:46 |
Wizzup | https://paste.villavu.com/raw/RIJ4Ju5ITdc0KzTNh0Uhwr5Ai981oOtXApp14fUl/ | 11:46 |
freemangordon | "Connects: automatically" | 11:46 |
freemangordon | thats what brings it online | 11:47 |
freemangordon | we should not request online status | 11:47 |
Wizzup | so you want auto-connect off? | 11:47 |
freemangordon | no | 11:47 |
freemangordon | it is ok | 11:47 |
freemangordon | my question is - what exactly the code you pasted ^^^ is supposed to do? | 11:48 |
freemangordon | if (d->shouldForceReconnect()) { | 11:48 |
freemangordon | etc | 11:48 |
freemangordon | tp already does that for us | 11:48 |
Wizzup | that checks if the account is tp | 11:48 |
Wizzup | as | 11:48 |
Wizzup | er | 11:48 |
Wizzup | check if it is tp-ring | 11:48 |
Wizzup | and then it 'forces' it online again | 11:48 |
Wizzup | but that still doesn't have the proper effect | 11:48 |
Wizzup | so, my question is | 11:48 |
Wizzup | what would cause tp-ring to change it's presence or not be online | 11:49 |
Wizzup | what *software* tells it to do that | 11:49 |
freemangordon | tp-ring *cannot* be online, offline, not available, etc | 11:49 |
freemangordon | because it does not support presence :) | 11:49 |
Wizzup | ok, but in dbus it reports that it is 'connecting' and not 'connected' | 11:50 |
freemangordon | connected has nothign to do with being online | 11:50 |
freemangordon | those are orthogonal | 11:50 |
Wizzup | and this tracks having wifi or gprs | 11:50 |
freemangordon | wait | 11:50 |
freemangordon | 'connected' is TpConnection preoperty | 11:50 |
Wizzup | https://telepathy.freedesktop.org/doc/telepathy-qt/a06885.html#ga601f90d95125a95ed0ff9a546fb1363f | 11:50 |
freemangordon | 'online' is TpAccount property | 11:50 |
Wizzup | this yes | 11:50 |
Wizzup | yes | 11:51 |
freemangordon | so, what is the issue? conneciton is not 'connected' or account is not 'online'? | 11:51 |
Wizzup | so I think I know how to 'fix' voicecall, but still I would like to what *what* in maemo triggers presence | 11:51 |
Wizzup | freemangordon: voicecall sees it as not available and unregisters it with sphone | 11:51 |
Wizzup | but this happens only when not on wifi or gprs somehow | 11:52 |
freemangordon | based on the presence? | 11:52 |
freemangordon | where is that code, link please | 11:52 |
Wizzup | ok, but there's still my question ^^ : D | 11:52 |
freemangordon | nothing in maemo triggers presence in general | 11:52 |
freemangordon | each protocol either supports presence or not | 11:53 |
Wizzup | what makes it *act* on wifi/gprs change | 11:53 |
sicelo | layman explanation ... tp-ring is offline when we're in Flight Mode ... maybe it's the answer to the question? | 11:53 |
Wizzup | because they start to connect immediately | 11:53 |
Wizzup | https://github.com/maemo-leste-upstream-forks/voicecall/blob/master/plugins/providers/telepathy/src/telepathyprovider.cpp#L235 | 11:53 |
sicelo | so it does support some 'primitive' presence | 11:53 |
Wizzup | and the status applet also knows immediately | 11:54 |
freemangordon | sicelo: what is 'it'? | 11:54 |
Wizzup | freemangordon: specifically this check https://github.com/maemo-leste-upstream-forks/voicecall/blob/master/plugins/providers/telepathy/src/telepathyprovider.cpp#L240 | 11:54 |
Wizzup | it checks for status connected | 11:54 |
sicelo | i mean, tp-ring has a basic type of presence. it can be offline when we're in flight mode for example | 11:55 |
freemangordon | so, which one is false there? | 11:55 |
Wizzup | freemangordon: from dbus-monitor, the last one | 11:55 |
Wizzup | d->account.data()->connectionStatus() | 11:55 |
Wizzup | but I haven't added any debug logging yet | 11:55 |
freemangordon | this is not presence | 11:55 |
freemangordon | so lets stop talking about 'online', ok? because we're mixing the things and that doesn't make it any easier :) | 11:56 |
freemangordon | sicelo: no, tp-ring does not have any concept of presence | 11:56 |
Wizzup | freemangordon: I'm not talking about 'online' | 11:56 |
freemangordon | sicelo: "also, ring supports neither TP_IFACE_CONNECTION_INTERFACE_PRESENCE nor TP_IFACE_CONNECTION_INTERFACE_SIMPLE_PRESENCE" | 11:57 |
Wizzup | I'm just asking what piece of software in maemo makes it so that tp knows when to reconnect, regardless of this issue | 11:57 |
freemangordon | this is mc related | 11:57 |
Wizzup | because it directly correlates with tp-ring saying it's no longer 'connected' | 11:57 |
Wizzup | so tp-mc is icd2 aware? | 11:57 |
freemangordon | maybe | 11:57 |
freemangordon | I dunno | 11:57 |
freemangordon | there might be some remnants | 11:57 |
Wizzup | I always assumed we have something icd2 aware that restarts things | 11:57 |
Wizzup | it could just be netlink for network interfaces | 11:57 |
freemangordon | no | 11:57 |
freemangordon | is there is something, it is over dbus | 11:58 |
freemangordon | or, it could be ofono | 11:58 |
Wizzup | ok | 11:58 |
freemangordon | Wizzup: sorry, my blood sugar level is very low, bbl in 20 minutes | 11:58 |
freemangordon | :) | 11:58 |
Wizzup | take it easy | 11:58 |
Wizzup | so I will check again in a bit (outside atm), but it looks like tp-ring connection status goes into 'connecting' insteaf of 'connected' | 11:58 |
Wizzup | and this causes voicecall to say 'well, you're not ready then' | 11:59 |
Wizzup | god I wish there was a program that takes a dbus-monitor log and makes it easily to filter it | 12:00 |
freemangordon | afaik there is something | 12:00 |
freemangordon | some gnome tool | 12:01 |
Wizzup | I know of the one to explore interfaces | 12:01 |
Wizzup | not one to monitor | 12:01 |
freemangordon | ah, could be | 12:01 |
sicelo | what do you want to filter? | 12:01 |
freemangordon | wait, I'll check what's going on in tp-ring code | 12:01 |
* sicelo uses busctl nowadays ... seems versatile | 12:02 | |
Wizzup | sicelo: like, I want to see all messages related to ring, but grep won't help since it's all multi line and indented | 12:02 |
Wizzup | sicelo: and 'ring' matches 'string' too, so in general it's pretty stupid | 12:02 |
freemangordon | Wizzup: it seems it changes the status to "connected" when a modem is bound | 12:02 |
Wizzup | I can navigate in vi with seaching for 'ring/' but then there is still a lot of clutter | 12:02 |
sicelo | busctl can probably help ... even has json output, so you can use jq and similar stuff | 12:02 |
Wizzup | freemangordon: well the modem doesn't go anywhere | 12:02 |
Wizzup | sicelo: will check | 12:02 |
freemangordon | see ring_connection_bind_modem() | 12:03 |
Wizzup | freemangordon: when does it change to 'connecting' ? | 12:03 |
freemangordon | on the constructor :) | 12:03 |
freemangordon | *in | 12:03 |
Wizzup | that would mean somethings sets it offline then | 12:03 |
Wizzup | maybe auto-reconnect is exactly the opposite of what we want | 12:04 |
freemangordon | I'd recommend to start tpring from cmdline with logs enabled | 12:04 |
freemangordon | could be | 12:04 |
Wizzup | to me it seems like tp cm just turns it off | 12:04 |
sicelo | only shortcoming of busctl is that you need to be root in order to monitor. otherwise it's flexible | 12:04 |
Wizzup | (but I didn't see that on dbus) | 12:04 |
Wizzup | sicelo: even for non system bus? | 12:04 |
sicelo | no, only system :-) | 12:04 |
Wizzup | yeah none of the tp stuff is on system bus | 12:05 |
Wizzup | lol yep, tp-ring has a new pid now that I reconnect to wifi | 12:05 |
Wizzup | ok, I will look at this later | 12:13 |
Wizzup | tp-ring logs tp_base_connection_change_status: was 0, now 2, for reason 1' | 12:21 |
Wizzup | 'mcd_account_connection_proceed_with_reason: ring/tel/account0 wants to connect, but we're offline, queuing it up' | 12:26 |
Wizzup | yeah it uses GNetworkMonitor to see if we're 'at least partially online' | 12:26 |
Wizzup | ok, so the problem is in mission control | 12:26 |
Wizzup | not in voicecall manager or ring | 12:26 |
freemangordon | yep, I was about to say that maybe tp/mc oversmarts itself | 12:29 |
freemangordon | we shal check what SF are doing | 12:29 |
* Wizzup debugged all of that on the screen+keyboard of the d4 btw | 12:29 | |
freemangordon | why? | 12:29 |
Wizzup | because it was sunny out :) | 12:30 |
freemangordon | heh | 12:30 |
Wizzup | also having to debug a problem that only shows up when ssh doesn't work is tricky | 12:30 |
freemangordon | ah | 12:30 |
freemangordon | yeah | 12:30 |
freemangordon | Wizzup: which file is that? | 12:32 |
Wizzup | https://bugs.freedesktop.org/show_bug.cgi?id=68712 | 12:32 |
Wizzup | ./src/mcd-account-connection.c | 12:33 |
Wizzup | function mcd_account_connection_proceed_with_reason | 12:33 |
freemangordon | ok | 12:33 |
Wizzup | my bet is that this code was never in nokia ;) | 12:34 |
freemangordon | "...use GNetworkMonitor (which is also pluggable)," | 12:34 |
Wizzup | we don't want to fake a network connection for cm | 12:34 |
Wizzup | then it will try to online others | 12:34 |
Wizzup | we want a property to let some cm's connect anyway | 12:34 |
freemangordon | we shall check how do they overcome that in SF | 12:34 |
Wizzup | I don't think they use tp-ring, but ok :) | 12:35 |
Wizzup | voicecall manager has an ofono backend | 12:35 |
freemangordon | ah | 12:35 |
Wizzup | see 957795838715b3b95f2bf8bc116b2328bf6a2881, | 12:35 |
Wizzup | A possible refinement in future Tp versions would be a way for a CM | 12:35 |
Wizzup | to tell MC "this connection does its own connectivity-monitoring, | 12:35 |
Wizzup | so don't kick it when we get disconnected". For now, we assume that | 12:35 |
Wizzup | everything needs the Internet, unless it's one of the accounts that | 12:35 |
Wizzup | wouldn't previously have been bound to a connection (i.e. | 12:35 |
Wizzup | _mcd_account_needs_dispatch() returns FALSE). | 12:35 |
freemangordon | yeah, I was about to look in the code | 12:36 |
freemangordon | to see if we can workaround that | 12:36 |
Wizzup | in ring maybe? | 12:36 |
freemangordon | lemme see | 12:36 |
Wizzup | seems to be 'always_dispatch' | 12:37 |
Wizzup | /* In addition to affecting dispatching, this flag also makes this | 12:37 |
Wizzup | * account bypass connectivity checks. */ | 12:37 |
Wizzup | gboolean always_dispatch; | 12:37 |
Wizzup | ./src/mcd-account-config.h:#define MC_ACCOUNTS_KEY_ALWAYS_DISPATCH "always_dispatch" | 12:37 |
freemangordon | is that some account property? | 12:37 |
Wizzup | trying now | 12:37 |
Wizzup | and yes | 12:37 |
Wizzup | there is also an 'always on' property | 12:38 |
Wizzup | not sure if that's also helpful | 12:38 |
Wizzup | looks like now | 12:39 |
Wizzup | not | 12:39 |
Wizzup | either I did something wrong or it did not work | 12:41 |
Wizzup | I set always_dispatch=false | 12:41 |
freemangordon | no, it should be true | 12:41 |
Wizzup | but then needs_dispatch will be true | 12:42 |
freemangordon | ah | 12:42 |
freemangordon | /* special case flag (for ring accounts, so far) */ | 12:42 |
Wizzup | which one is this | 12:42 |
freemangordon | p, li { white-space: pre-wrap; } always_dispatch | 12:43 |
Wizzup | right | 12:43 |
Wizzup | let me try again | 12:43 |
freemangordon | but you shall set it to true | 12:44 |
freemangordon | it defaults to false | 12:44 |
freemangordon | see p, li { white-space: pre-wrap; } mcd_connection_need_dispatch | 12:44 |
Wizzup | ah | 12:45 |
Wizzup | if (mcd_connectivity_monitor_is_online ( | 12:45 |
Wizzup | mcd_account_get_connectivity_monitor (account)) || _mcd_account_needs_dispatch (account)) | 12:45 |
Wizzup | ok | 12:45 |
freemangordon | mhm | 12:45 |
Wizzup | ok, just called myself with wifi off, we're good, but we'll need to set this | 12:46 |
freemangordon | this should be part of ring account creation, no? | 12:46 |
Wizzup | yeah | 12:46 |
freemangordon | like, who created ring/tel/account0? | 12:46 |
Wizzup | we don't have a 'ring account creation' process atm :) | 12:46 |
freemangordon | mhm | 12:46 |
Wizzup | well, this is progress, I'm going to take a break for a moment | 12:47 |
freemangordon | mhm | 12:47 |
Wizzup | so where would we want to make the account? | 13:14 |
uvos | would make sense if the sphone plugin did this OR? | 13:36 |
freemangordon | uvos: no, I don;t think sphone should deal with tp accounts | 13:59 |
freemangordon | Wizzup: who will depend on tp-ring? | 14:00 |
freemangordon | BTW, IIRC, back then tp-ring was creating the account on installation | 14:01 |
freemangordon | Wizzup: wait... | 14:02 |
freemangordon | look at mcp_account_manager_ring_init() | 14:03 |
freemangordon | why is that not called? | 14:10 |
sicelo | arno11 - when you have a chance, please checkout https://github.com/sicelo/libcmtspeechdata/commits/compile-fixes/ and see if it works fine. if so, i will make an MR | 14:13 |
sicelo | basically tries to fix the numerous (annoying) compile errors for cmtspeech | 14:13 |
sicelo | and of course, anyone else is free to look at the commits. tbf, that code can do with some serious refactoring :p | 14:25 |
freemangordon | Wizzup: I think I found what's going on | 14:51 |
freemangordon | ring mc plugin is not installed, IIUC | 14:51 |
freemangordon | that's why we shall create the account by hand | 14:51 |
freemangordon | while rpm installs %{_libdir}/mission-control-plugins.0/mcp-account-manager-ring.so :) | 14:52 |
freemangordon | dh_missing: warning: usr/lib/x86_64-linux-gnu/mission-control-plugins.0/mcp-account-manager-ring.so exists in debian/tmp but is not installed to anywhere | 14:55 |
freemangordon | dh_missing: warning: usr/share/man/man8/telepathy-ring.8 exists in debian/tmp but is not installed to anywhere | 14:55 |
freemangordon | :) | 14:55 |
freemangordon | shall I fork? | 14:59 |
uvos | freemangordon: i agree sphone should not deal with tp accounts | 15:01 |
uvos | but ring is special since its not a "real" account, and simply has to exist any where sphone runs | 15:02 |
freemangordon | uvos: anyway, I found what the issue is | 15:02 |
freemangordon | yes, but tp-ring debian packaging is broken | 15:02 |
uvos | ok | 15:02 |
freemangordon | that's why no account is created | 15:02 |
uvos | ok | 15:02 |
freemangordon | going to fix that | 15:02 |
uvos | thats fine then | 15:03 |
uvos | i just want to ensure it works if only tp sphone (and vcm) is installe | 15:03 |
uvos | d | 15:03 |
uvos | since imo it should | 15:03 |
freemangordon | it will | 15:03 |
freemangordon | well, you need telepathy as well :) | 15:03 |
freemangordon | ah, 'tp' | 15:04 |
freemangordon | yeah | 15:04 |
uvos | you need ofono too | 15:04 |
uvos | but you know what i mean | 15:04 |
freemangordon | well | 15:04 |
freemangordon | yeah | 15:04 |
freemangordon | I just installed empathy(and friends) from my d4 | 15:04 |
freemangordon | *uninstalleds | 15:04 |
arno11 | sicelo: currently looking @cmtspeech | 15:14 |
arno11 | wooo, indeed lot of errors fixed | 15:15 |
arno11 | let me try it | 15:17 |
Wizzup | freemangordon: what does this mc ring plugin do? set this property? | 15:18 |
Wizzup | freemangordon: but yes, we can fork if we have to | 15:18 |
freemangordon | I already did | 15:19 |
freemangordon | but it fails to build :) | 15:19 |
freemangordon | see ci | 15:19 |
freemangordon | I don;t get what's wrong with it | 15:19 |
Wizzup | let me take a quick look | 15:19 |
freemangordon | fails to find the tag, but it is there | 15:19 |
Wizzup | I am pretty sure I knpw | 15:20 |
Wizzup | just a sec | 15:20 |
Wizzup | freemangordon: | 15:20 |
Wizzup | maemo/chimaera != chimaera | 15:20 |
Wizzup | :) | 15:20 |
freemangordon | oh | 15:20 |
freemangordon | :) | 15:21 |
freemangordon | thanks | 15:21 |
freemangordon | will fix it | 15:21 |
freemangordon | Wizzup: I think you shall delete ring account before upgrade | 15:22 |
freemangordon | also, we shall start tp-ring from xsession, mlocked etc. see RING_REALTIME and RING_MEMLOCK | 15:24 |
Wizzup | mhm | 15:25 |
freemangordon | RING_PERSIST=1 as well | 15:25 |
Wizzup | I need to go do some other things first, I'll be back in an hour or so | 15:25 |
Wizzup | feel free to hack away | 15:25 |
freemangordon | well, I don;t see what more I can do now | 15:25 |
Wizzup | maybe write an xsession file, but not sure where it would go | 15:26 |
freemangordon | we shall decide first who depends on tp-ring | 15:26 |
freemangordon | connectivity mp? | 15:26 |
Wizzup | yeah, I don't know if we need another maemo-telephony pkg, or if the existing one works | 15:27 |
freemangordon | to me connectivity is ok | 15:27 |
freemangordon | so xsession file belongs there | 15:28 |
Wizzup | sounds good | 15:29 |
arno11 | sicelo: works fine and commit 51a6f6c fixed the random weird tone issue apparently :) well done man | 15:30 |
freemangordon | the fuck, account is still not created :( | 15:30 |
Wizzup | are you sure that should make the account? | 15:31 |
Wizzup | freemangordon: I think you can just do mc-tool add ring/tel account0 -p always_dispatch=true or so (note I didn't test this and maybe even got the var name wrong) | 15:32 |
Wizzup | actually ignore the '-p' too | 15:32 |
freemangordon | ugh: | 15:32 |
freemangordon | could not load plugins from /usr/lib/mission-control-plugins.0: Error opening directory ?/usr/lib/mission-control-plugins.0?: No such file or directory | 15:32 |
freemangordon | that's wrong dir | 15:33 |
freemangordon | it should be /usr/lib/$arch/mission-control-plugins.0 | 15:33 |
sicelo | arno11, wow! i only compile-tested this. actually i'm surprised there's a positive 'visible' effect. i was only hoping that i didn't break anything or introduce problems. | 16:09 |
sicelo | Wizzup: so there's a PR for cmtspeech - https://github.com/maemo-leste/libcmtspeechdata/pull/3 | 16:10 |
Wizzup | will checl later tonight ty | 16:19 |
freemangordon | Wizzup: tp-ring creation is fixed | 16:26 |
sicelo | Wizzup: please hold off merging. seems there's a compilation error that only shows up on armhf, but not amd64. i want to fix that one too | 21:11 |
Wizzup | okay, sure | 21:11 |
Wizzup | freemangordon: so I delete the account and apt update / upgrade ? | 21:12 |
Wizzup | freemangordon: btw, if tp-ring autostarts in -devel it means calls will no longer work | 21:20 |
Wizzup | (tp-ring will just hang up if there is no tp client, and sphone uses ofono directly in -devel) | 21:21 |
uvos | so new sphone has /usr/share/sphone/sphone.ini.d and should update the combobox when a backend is added or removed without needing to reopen the window | 22:23 |
uvos | the latter is untested | 22:23 |
Wizzup | great, I'll rebase | 22:24 |
sicelo | Wizzup: https://github.com/maemo-leste/libcmtspeechdata/pull/4 ... this one has better fixes for the compilation errors (compiles equally well with both amd64 and armv7) | 23:16 |
freemangordon | Wizzup: well, one shall install tp-ring by hand | 23:46 |
freemangordon | nothing depends on it still | 23:47 |
freemangordon | sicelo: next time you'd better force-push, instead of creating new PR | 23:53 |
freemangordon | it is easier to review that way | 23:53 |
Generated by irclog2html.py 2.17.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!