diejuse | I'm going to record a short video to understand each other better. | 00:03 |
---|---|---|
Wizzup | okay :) | 00:09 |
diejuse | Here is: https://youtube.com/shorts/HfiyCjy5sCk?si=586iJoy1OHdaH901 | 00:41 |
Wizzup | diejuse: so many maemo applications 'spawn' by sending a dbus message to maemo-launcher telling it 'start this program' | 00:48 |
Wizzup | or rather, the programs themselves are a 'symlink' to the maemo-invoker (or so) program, which from argv[0] will figure out what to spawn | 00:48 |
diejuse | Hmm. Is there any workaround so they can start without sending the dbus message? | 00:53 |
diejuse | Like Debian apps? | 00:53 |
Wizzup | not without recompiling, but you'll really want dbus to work for maemo to function properly | 00:53 |
Wizzup | is it not possible to run dbus in proot? | 00:54 |
diejuse | It would be the best but I think I won't be able to do it because it doesn't seem possible in recent versions of Android. | 00:54 |
Wizzup | hmm... how so? | 00:55 |
Wizzup | is this because of the buggy script? | 00:55 |
diejuse | However, I can start Devuan with another dbus. I'm confused. | 00:55 |
diejuse | Because I tried installing Devuan via proot and it works! That's why I'm confused. | 00:56 |
diejuse | I don't know if the script is the culprit. | 00:56 |
diejuse | I mean "dbus-launch –exit-with-session startxfce4" works | 00:58 |
diejuse | on Devuan | 00:58 |
diejuse | But ""/etc/init .d/dbus start" no works | 00:58 |
Wizzup | just for the record the line you pasted has a unicode dash in it | 01:00 |
Wizzup | that won't work in the shell | 01:00 |
Wizzup | well, I think for openrc to run in a virtual environment you have to do some things? I don't fully recall | 01:00 |
Wizzup | I'd have to try it... | 01:00 |
Wizzup | diejuse: do you have instructions on how to set this up? | 01:06 |
diejuse | I wanted to add that my ZFlip5 is rooted and via chroot the maemo apps do work and you can also see the time. | 01:06 |
diejuse | Bur via proot don't | 01:06 |
Wizzup | I think this can probably work, but this will take some careful steps to figure it out | 01:07 |
Wizzup | like, maybe the environment is unclean, or something else is up, and do you start everything in x session, or only specific things? | 01:07 |
Wizzup | and what does say env|grep DBUS print | 01:07 |
diejuse | I start everything from an Android terminal called Termux. First run a script with the "proot" command and a few other things. And once inside the distro, run another script with the commands that I already used with chroot (https://raw.githubusercontent.com/diejuse/chroot_Maemo-leste_on_Android/main/launchMaemo.sh) | 01:11 |
Wizzup | so this is likely not your problem, but the ' &' at the end could make it so that dbus isn't started before other maemo programs that need dbus start | 01:13 |
Wizzup | it looks like you're indeed already touching the softlevel file | 01:13 |
diejuse | "env | grep DBUS" => Nothing | 01:14 |
Wizzup | can you check if dbus is running at all, with ps? | 01:14 |
Wizzup | 'ps xua | grep dbus' | 01:14 |
Wizzup | same for 'ps xua | grep maemo' | 01:14 |
Wizzup | see if maemo-launcher is running | 01:14 |
diejuse | Without & no working. The same. | 01:15 |
Wizzup | yes, that won't make all the difference, it will not continue with starting the other tools | 01:16 |
Wizzup | you'll eventually need something more complicated than this, like run-parts on the xsession files or so | 01:17 |
diejuse | 'ps xua | grep dbus' => Error /proc must be mounted. And this is the real problem that I mentioned before. In modern versions of Android /proc /sys is not mounted with proot. However, with older Android versions (like 7) yes. | 01:19 |
diejuse | proc not mounted is the key | 01:19 |
Wizzup | that is a big problem | 01:20 |
diejuse | I will focus on investigating some workaround to mount /proc with proot. | 01:21 |
diejuse | And in the meantime I would like you to find some way to start Maemo apps without calling dbus. As we said before. | 01:21 |
Wizzup | there is none other than recompiling | 01:22 |
Wizzup | (every package) | 01:22 |
Wizzup | well, wait, there is a way... | 01:22 |
Wizzup | try maemo-summoner /usr/bin/hildon-status-menu.launch | 01:22 |
Wizzup | or in your case | 01:22 |
Wizzup | try maemo-summoner /usr/bin/osso-xterm.launch | 01:23 |
diejuse | The good thing about proot vs chroot is that all Android phones, new and old, could use Maemo Leste (non-native). It would be good. People don't usually root their smartphones and everyone could easily try it. | 01:24 |
diejuse | maemo-summoner /usr/bin/osso-xterm.launch => "osso_initialize() failed!" | 01:26 |
Wizzup | I think a better approach is to try to get the maemo environment set up properly | 01:27 |
diejuse | maemo-summoner /usr/bin/hildon-status-menu.launch => "Aborted" | 01:28 |
diejuse | Yes. It would be the best. But it is easy that nothing can be done to avoid the new restrictions on Android smartphones of new versions. That's why I'm also interested in some workaround to start those Maemo apps. | 01:29 |
Wizzup | well, you won't be able to run/use maemo without dbus | 01:30 |
Wizzup | almost everything depends on it | 01:30 |
Wizzup | like, if you wanted to have the useful instant messaging features, it won't work | 01:30 |
Wizzup | or the virtual keyboard, won't work | 01:30 |
Wizzup | it's all dbus | 01:30 |
diejuse | I understand. And why does the dbus that starts xfce4 in Devuan work? | 01:31 |
gnarface | most likely they've patched it | 01:32 |
gnarface | well, actually they've absolutely certainly patched it, but my guess is that most likely the patch is why it works | 01:33 |
gnarface | (didn't read the whole scrollback, but dbus is definitely one of the few debian packages that devuan has forked) | 01:33 |
gnarface | (they all have "devuan" in the version string on the package) | 01:34 |
diejuse | Can't the dbus that Maemo uses have the same benefits as the dbus that is used to start other distros like xfce4 or lxde? | 01:36 |
Wizzup | we use the devuan dbus fwiw | 01:36 |
Wizzup | how do you start devuan? | 01:36 |
diejuse | I simply do: | 01:37 |
diejuse | dbus-launch --exit-with-session startxfce4 | 01:38 |
diejuse | And works | 01:38 |
Wizzup | have you tried starting the maemo script like that? | 01:38 |
diejuse | yes | 01:38 |
Wizzup | (without starting maemo's /etc/init.d/dbus script) | 01:39 |
diejuse | dbus-launch --exit-with session ....... what more? | 01:39 |
Wizzup | whatever your script is to start maemo programs | 01:40 |
diejuse | I tried: dbus-launch --exit-with-session maemo-launcher --send-app-died --booster gtk,cpp --daemon | 01:40 |
diejuse | ok, dbus-launch --exit-with-session /usr/bin/osso-xterm.launch ? | 01:41 |
Wizzup | no | 01:41 |
Wizzup | dbus-launch --exit-with-session launchMaemo.sh | 01:41 |
Wizzup | or something like that | 01:41 |
Wizzup | but really like I said, you'd just need something that executes everything maemo session needs, plus some of the init scripts | 01:41 |
diejuse | I just tried "dbus-launch --exit-with-session launchMaemo.sh" but the result is the same. | 01:44 |
diejuse | Yes. It would be the best. But it would also be nice if the Maemo graphical environment could be started with "dbus-launch" in the same way as the most popular desktops (xfce4, lxde, mate...) | 01:46 |
Wizzup | not everything in maemo is part of a session, like icd2 | 01:50 |
Wizzup | also note that your launch maemo starts dbus twice more: | 01:50 |
Wizzup | /etc/init.d/dbus start & | 01:50 |
Wizzup | /usr/bin/dbus-daemon --session --print-address=2 & | 01:50 |
Wizzup | both those must go I think, | 01:51 |
Wizzup | well you would also need a system dbus I suppose | 01:51 |
Wizzup | and after running dbus-session with print address, you also need to evaluate the address it pritns | 01:51 |
freemangordon | Wizzup: it is the system bus taht does not start, session bus is ok | 07:18 |
humanbeta | sorry typo yesterday, because my lapdock does not have model name on it, it is originally made for Motorola Atrix, not Altrix. | 09:25 |
humanbeta | and one hint for that, if touchpad not work, touch shift of keyboard first. | 09:38 |
Wizzup | freemangordon: hm, my status menu is not showing the battery and network status icons anymore, just the time | 14:52 |
Wizzup | in my vm | 14:52 |
Wizzup | let me reboot the vm | 14:52 |
Wizzup | uvos: I pushed to the sphone mr, split out the spelling fixes as you asked, and added a commit to change the formatting, if you like the formatting (it is what you asked), then I will merge it into the main commit | 15:34 |
uvos | Wizzup: thank you | 18:58 |
uvos | Wizzup: so whats the status on this on generall, functionality wise | 18:59 |
uvos | otherwise it looks ok now code wise, the extern "C" thing is just unessecary, please tell me if it failed to compile without it, that just means i forgot to do this in one of the headers | 19:00 |
uvos | if you dont have any further concerns i would merge it now, besides the config changes that i would like for you to remove from the pr so that we can first merge it in a default disabled state so that it may be tested by simply loading it | 19:02 |
Wizzup | uvos: I will fix the extern stuff, I'll fix the config comments yeah, and regarding the current state, I need to test it once more, but regular calls should work fine | 19:03 |
Wizzup | sip/xmpp calls I've never had working with audio but i mostly tested on my vm | 19:03 |
uvos | ok | 19:03 |
uvos | well in default disabled state it cant hurt anyone | 19:03 |
uvos | this way we can test it a bit and enable it in leste-config later | 19:04 |
Wizzup | agreed | 19:07 |
Wizzup | I'll make the changes today, I was fighting this usb modem for two hours and gave up, so I will test on my d4 instead | 19:10 |
Wizzup | (fighting to test the new code) | 19:10 |
uvos | ok, no rush | 19:10 |
uvos | i have some other code for sphone i want to put in the same release | 19:10 |
uvos | (instmod and rmmod for running sphone) | 19:10 |
Wizzup | uvos: ok, I made the changes, now testing on my d4 | 19:35 |
Wizzup | uvos: the debian/control entry will need to be modified to get some -dev pkg and perhaps a recommends on the voicecallmgr package | 19:37 |
uvos | one sneaky extern C still lurking in voicecallmanager.h | 19:37 |
Wizzup | ah ok | 19:37 |
Wizzup | I did grep extern *.cpp | 19:37 |
Wizzup | so yeah | 19:37 |
Wizzup | :D | 19:37 |
uvos | yeah i think we need to change how this works generally | 19:38 |
uvos | so rn sphone depends on things it dosent really need at run time only at build time. | 19:38 |
bencoh | hmm ... silly question, but are you removing extern C from headers? if so, why? | 19:39 |
uvos | these things should be required at build time and optional at runtime | 19:39 |
uvos | bencoh: this is a c++ header | 19:39 |
bencoh | ah | 19:39 |
Wizzup | uvos: yeah some things to fix still: | 19:49 |
Wizzup | sphone: faled to load module comm-voicecallmanager: missing symbol sphone_module_init | 19:49 |
uvos | oh right shit im an idiot | 19:49 |
uvos | that extern c was needed since sphone_module_init and exit is looked up by symbol name | 19:49 |
Wizzup | you want the extern "C" stuff back in the header? | 19:49 |
uvos | the others are passed by function pointer | 19:50 |
uvos | no just around the ones that are looked up by symbol name | 19:50 |
uvos | 1sex | 19:50 |
uvos | er sec | 19:50 |
uvos | well either is about the same unit of time | 19:51 |
uvos | put extern c back but only: | 19:52 |
uvos | https://github.com/maemo-leste/sphone/blob/7cbb675d95269949dd73a25da029e3a9973dea99/src/utils/sphone-modules.c#L188 | 19:52 |
uvos | https://github.com/maemo-leste/sphone/blob/7cbb675d95269949dd73a25da029e3a9973dea99/src/utils/sphone-modules.c#L87 | 19:52 |
uvos | https://github.com/maemo-leste/sphone/blob/7cbb675d95269949dd73a25da029e3a9973dea99/src/utils/sphone-modules.c#L36 | 19:52 |
uvos | for these symbols | 19:52 |
uvos | sorry about that | 19:52 |
Wizzup | no worries, dinner first :) | 19:53 |
uvos | the rest should stay mangled | 19:53 |
Wizzup | uvos: incoming and outgoing works, let me fix the extern stuff | 20:51 |
uvos | https://github.com/maemo-leste/sphone/pull/4/files#diff-1e7de1ae2d059d21e1dd75d5812d5a34b0222cef273b7c3a2af62eb747f9d20aR5 | 21:08 |
uvos | whats with this btw | 21:08 |
uvos | i presume you are mocing by target later and this can go away? | 21:09 |
Wizzup | uvos: that can go I think | 22:02 |
Wizzup | I will remove it too. | 22:02 |
Wizzup | just remnants of me trying to make cmake work with qt | 22:02 |
Wizzup | freemangordon: what do you think, should we use fremantle group_uid for our tp logging or use our own | 22:11 |
Wizzup | huh, actually it looks like it *is* the same... | 22:13 |
Wizzup | I meant for sms specifically | 22:13 |
Wizzup | oh no, it is not the same. | 22:14 |
Wizzup | so on fremantle the group_uid for sms is the last 7 digits of a phone number | 22:14 |
Wizzup | on leste it's ring/tel/account0-[full phone number here] | 22:15 |
Wizzup | I don't know why fremantle make it only the last 7 digits, maybe they know something we don't and it's never going to conflict (doubt it), or they tried to save space? I don't know | 22:15 |
Wizzup | I guess you have the full number anyway in the remote_uid | 22:15 |
Wizzup | in any case, from my POV the *only* reason to make it like fremantle that I can think of is so that when you import the rtcom el db from fremantle and you continue a sms conversation, it will end up in the same group/window | 22:16 |
Wizzup | so, fremantle: | 22:17 |
Wizzup | sqlite> select group_uid from events order by id desc limit 2; 6081360 | 22:17 |
Wizzup | 6081360 | 22:17 |
Wizzup | leste (turned some digits into x to anonimise) | 22:17 |
Wizzup | sqlite> select group_uid from events order by id desc limit 2; | 22:17 |
Wizzup | ring/tel/account0-+31623xxxxxx | 22:17 |
Wizzup | ring/tel/account0-+31623xxxxxx | 22:17 |
Wizzup | maybe this was nokia's way of normalising numbers with and without prefix? | 22:31 |
freemangordon | I don't think it makes sense to keep ring/tel/account0 in the db | 22:35 |
freemangordon | this just gets duplicated, no? | 22:35 |
freemangordon | last 7 digits from the phone number? I guess we shall read a bit | 22:36 |
freemangordon | to me id does not seem unique | 22:36 |
freemangordon | *it | 22:36 |
freemangordon | perhaps it is combined with something else | 22:37 |
Wizzup | this is how *all* other group ids are apart from sms fwiw | 22:37 |
Wizzup | and if you have more than one modem, you probably want to know which one was used? | 22:37 |
freemangordon | no, why do you care? | 22:37 |
freemangordon | it is conversation with the same person, no? | 22:38 |
bencoh | isn't group_uid just used as some kind of index to speed up searches? | 22:38 |
Wizzup | bencoh: it's not just used to speed up, I think | 22:38 |
bencoh | (not that it really makes any sense in a practical way, but ....) | 22:38 |
freemangordon | have to run now, sorry | 22:38 |
Wizzup | freemangordon: I mean I care in the same way that I can have different xmpp accounts messaging the same person | 22:39 |
Wizzup | they're still different conversations I think | 22:39 |
Wizzup | bencoh: I think it is used as a GROUP BY as well, I thought that is why it was called group_uid | 22:40 |
Wizzup | all the code for this is open btw, so we could check, but I think if something has the same group_uid but different say remote_uid, then it will still be grouped together | 22:40 |
bencoh | in convs you mean? | 22:48 |
Wizzup | in anything that uses rtcom-eventlogger-ui, but yes, also in conversations since it uses rtcom-eventlogger | 22:48 |
Wizzup | this is an example ui for rtcom-eventlogger-ui by nokia | 22:49 |
Wizzup | https://github.com/maemo-leste/rtcom-eventlogger-ui/blob/master/example/example.c#L134 | 22:49 |
bencoh | wth is happening there: static_group_by = (1 + static_group_by) % (RTCOM_EL_QUERY_GROUP_BY_GROUP + 1); | 22:51 |
Wizzup | I think just an int for the different group bys for UI purposes | 22:52 |
bencoh | oh, nevermind, it just allows changing the group by method | 22:52 |
bencoh | yeah | 22:52 |
uvos | Wizzup: https://github.com/maemo-leste/sphone/pull/4/files#diff-80b51c8dc924426a6084f06211134017153c134d5b330f3719e0800ce8c71d87R37 memleak | 23:04 |
uvos | you mean strdupa? | 23:05 |
uvos | theres also something wrong with indentation here: https://github.com/maemo-leste/sphone/pull/4/files#diff-85ebdd8b0c290e2ac6d547f92a7036e6152a4a9abb3436bbd00af52636009c58R18 | 23:09 |
Wizzup | uvos: does sphone make it's own copy of the backend name? | 23:10 |
Wizzup | I mean, I don't think toStdString().c_str() makes copies of the data that need to manually be freed, does it? | 23:10 |
uvos | sphone is universally function-call-dosent-change-owership aside from very few (obvious) exceptions | 23:11 |
uvos | so yes it copies itself | 23:11 |
Wizzup | ok, in that case we can just pass along the c_str() | 23:11 |
uvos | yes | 23:11 |
Wizzup | so no need for strdup | 23:11 |
Wizzup | uvos: note that it was freed in the module, so I don't think that's a memleak still | 23:12 |
uvos | i dont see where its freed | 23:12 |
Wizzup | https://github.com/maemo-leste/sphone/pull/4/files#diff-80b51c8dc924426a6084f06211134017153c134d5b330f3719e0800ce8c71d87R47 | 23:13 |
uvos | ah ok, well why carry this anyhow | 23:13 |
Wizzup | yes, I changed the code now locally | 23:14 |
Wizzup | the cmakelists seems to have a tabs/spaces snafu | 23:14 |
uvos | instead of just asking sphone_comm_get_backend | 23:14 |
Wizzup | hm? | 23:15 |
Wizzup | ah, wasn't used anywhere else | 23:15 |
Wizzup | I just force pushed | 23:15 |
uvos | if you need the name elsewhere you can just sphone_comm_get_backend(id)->name | 23:15 |
Wizzup | right | 23:15 |
uvos | you also dont need to pass a is_valid_ch function | 23:17 |
uvos | the default is everything | 23:17 |
uvos | not that it matters mutch | 23:17 |
uvos | see https://github.com/maemo-leste/sphone/blob/7cbb675d95269949dd73a25da029e3a9973dea99/src/utils/comm.c#L81 | 23:18 |
Wizzup | I was planning to eventually change that function perhaps, but good to know | 23:19 |
uvos | yeah makes sense | 23:19 |
uvos | for now it could be the same as ofonos no? | 23:19 |
uvos | maybe not if sip works | 23:19 |
Wizzup | freemangordon: bencoh: maybe I should just follow fremantle group_uid for now, since we've been doing that on fremantle for years anyway, can always change it later? | 23:19 |
Wizzup | sip works, just not with audio yet, but that is not sphone's fault | 23:19 |
uvos | ok | 23:19 |
uvos | sip should really be its own backend | 23:19 |
uvos | ie the module should register 2+ | 23:20 |
Wizzup | it does | 23:20 |
uvos | oh ok | 23:20 |
uvos | but not with different shema | 23:20 |
Wizzup | that might need some work, but sip can also be used to call regular numbers | 23:20 |
uvos | ok | 23:20 |
Wizzup | btw, just stating the obvious, but the ofono and voicecallmanager module should not really be active at the same time (they can be, but telepathy-ring will mess with the plain ofono module) | 23:21 |
Wizzup | in any case I think the pr maybe is ready now? | 23:36 |
uvos | seams so | 23:36 |
Wizzup | freemangordon: also the group-uid not just being the phone number is also because a sip number can also call a regular number, for example | 23:37 |
uvos | Additionally, the module relies on qtvoicecall, which is a (currently | 23:38 |
uvos | maemo specific) addition to qtvoicecall | 23:38 |
uvos | i dont grok this | 23:38 |
uvos | not that this is super important | 23:39 |
Wizzup | I probably missed a word there | 23:42 |
Wizzup | I think I exported some functionality of qtvoicecall because they only exported a qml class to talk to qtvoicecall | 23:42 |
Wizzup | and I didn't want to use qml | 23:42 |
Wizzup | https://github.com/maemo-leste-upstream-forks/voicecall/commit/69314e769ac459d80fb41c3b0f45e07c324d674d | 23:42 |
uvos | ok | 23:44 |
uvos | Wizzup: i found one more issue but i will fix it myself | 23:48 |
uvos | https://github.com/maemo-leste/sphone/blob/b1eecb4199efa2152f178be1f9bab56f0919bfa7/src/modules/voicecallmanager/CMakeLists.txt#L1 | 23:48 |
uvos | you get pkg-config as an optional dependancy but then expect it to be there 2 lines later | 23:49 |
uvos | should be required ofc | 23:49 |
uvos | or guarded by if | 23:49 |
Wizzup | check | 23:52 |
Wizzup | yes please fix if you can | 23:52 |
Generated by irclog2html.py 2.17.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!