inky_ | >since its totaly impossible to detemine if something is a mult or a pointer based on just one source file | 00:29 |
---|---|---|
inky_ | uvos, can you explain or link to some page? what is mult? | 00:29 |
inky_ | ah, mult is multiplication. | 00:29 |
uvos | indent cant possibly know if SomeToken is a typedef or a global varaible declared in some header | 00:32 |
inky_ | Wizzup: tor+wireguard, sailfish guy. | 00:33 |
uvos | since it parses one file at a time | 00:33 |
inky_ | i believe my friend added gui wireguard config to sailfish, may be you are talking about him? i can ping him. | 00:33 |
uvos | sicelo: good question, i tried phosh on xt1602 it was not so great at the time (buggy incompleat) but that was quite some time ago | 00:34 |
uvos | sicelo: i should try on a mapphone at some point | 00:34 |
inky_ | uvos: ok i understood completely by the following examples. you mean not possible to understand if it is a dereference or pointer. | 00:34 |
uvos | right | 00:34 |
uvos | sicelo: its beside the point anyhow i dont want some ui to work that way, i want all mobile linux uis to work that way, so that they can share a result in a unified platform, just like desktop linux is one unified platform where you can develop applications for that work with any de and even integrate at least resonably well by following xdg specs | 00:36 |
Wizzup | inky_: nah, audio | 02:07 |
Wizzup | uvos: you want to bump mce version to 1.9? | 12:34 |
uvos | Wizzup: yeah | 13:10 |
uvos | Wizzup: i changed enough and should not interfer with sfos | 13:11 |
uvos | sfos incremented the version to 1.100 and started from there as far as i can tell | 13:11 |
uvos | (its at 1.106.7 rn( | 13:11 |
Wizzup | ok | 13:12 |
uvos | i would also argue that the major version bump is required because i totaly broke all compatability with fremantle | 13:13 |
uvos | (esp. its vendor kernel) | 13:13 |
Wizzup | what else besides the vendor kernel? | 13:14 |
uvos | well mce dose everything different now so it no longer uses the same interfaces to do things out side of the kernel | 13:14 |
uvos | eg hal or bme | 13:14 |
Wizzup | isn't the dbus interface the same? (we rely on that) | 13:15 |
uvos | sure mces dbus interfaces are the same | 13:15 |
uvos | the way it perfomes the actions is different | 13:15 |
uvos | ie it uses different interfaces to enact what is told to do via dbus | 13:16 |
Wizzup | ok | 13:16 |
uvos | like geting the battery state from upower instead of bme that sort of thing | 13:17 |
Wizzup | pleas don't write 'totally broke all compatibility with fremantle' in the future as it will give me and fmg heart attacks ;) | 13:17 |
uvos | heh | 13:17 |
uvos | im just saying you can drop mce 1.9 into fremantle and expect anything to work :) | 13:18 |
uvos | so what i mean by that it breaks compataibility | 13:18 |
uvos | is this | 13:18 |
uvos | *cant drop mce 1.9 | 13:19 |
Wizzup | I get that now | 13:19 |
Wizzup | for me maemo leste is fremantle, ported to newer sw/interfaces, but with many of the same dbus interfaces, which I also see as 'fremantle' | 13:19 |
uvos | ok i allways use fremantle only in the sense of the particular setup used on n900 by nokia | 13:20 |
uvos | the other stuff is maemo and maemo interfaces for me | 13:20 |
uvos | (to avoid confusion in the future) | 13:21 |
Wizzup | yeah, makes sense | 13:21 |
freemangordon | :) | 15:30 |
Wizzup | freemangordon: ping | 15:51 |
Wizzup | freemangordon: in icd2 gconf there is stuff like this: | 15:52 |
Wizzup | network_modules = [libicd_network_wpasupplicant.so,libicd_network_ipv4.so] | 15:52 |
Wizzup | if we want libicd-network-tor and/or libicd-network-wireguard to work, we will need to add them after _ipv4 | 15:52 |
Wizzup | They can be a no-op in most cases | 15:52 |
Wizzup | (note: this is not about libicd-provider-tor) | 15:52 |
Wizzup | freemangordon: I don't see another way to do it, but then we might have to change this in gconf based on what is installed | 15:53 |
Wizzup | unless we have libicd-network-tor installed by default | 15:54 |
Wizzup | freemangordon: I suppose we could have some postinst scripts that add/remove them | 16:03 |
freemangordon | Wizzup: sorry, could you remind me why we need to add them to network_modules? | 16:11 |
freemangordon | they are SPs, no? | 16:11 |
* freemangordon is afk | 16:12 | |
Wizzup | freemangordon: yes, we have a service provider mode, but also a non-service provider mode | 16:13 |
Wizzup | freemangordon: e.g. no matter which network you pick, it should use tor | 16:13 |
Wizzup | service provider is only per iap | 16:13 |
freemangordon | ah, I see | 16:13 |
freemangordon | maybe we shall reowork that | 16:13 |
freemangordon | *rework | 16:14 |
Wizzup | rework what exactly? | 16:14 |
freemangordon | to get modules in a different way | 16:14 |
freemangordon | because I don;t see how you can register/unregister a modules from a postinst/postremove scripe | 16:14 |
freemangordon | *script | 16:15 |
Wizzup | freemangordon: using gconftool | 16:15 |
freemangordon | but this is a string list, is gconftool capable of adding/removing items? | 16:16 |
freemangordon | if it is, then fine | 16:16 |
Wizzup | freemangordon: with some scripting yeah | 16:16 |
freemangordon | if it needs scripting, then we shall add that to icd2-dev | 16:17 |
Wizzup | postinst to icd2-dev? | 16:17 |
Wizzup | I don't think that makes sense' | 16:17 |
freemangordon | hmm, no, sorry | 16:17 |
freemangordon | my bad | 16:17 |
freemangordon | helper script in icd2 that is called by postinst/postrm to edit the list | 16:17 |
freemangordon | postinst/postrm of module | 16:18 |
freemangordon | or, we can create dh_ helper | 16:18 |
freemangordon | have to run, ttyl | 16:19 |
Wizzup | ok | 16:20 |
Wizzup | parazyd: the custom_ui stuff for icons and such for srv_providers are per provider_id, which is really annoying/stupid | 17:30 |
Wizzup | this means I think that we have to write the custom_ui stuff for each tor config for example | 17:31 |
Wizzup | and also clean them out | 17:31 |
parazyd | Where does that go? | 17:31 |
Wizzup | I wonder if we should make the icons per service_type instead | 17:31 |
Wizzup | sec | 17:31 |
Wizzup | https://github.com/maemo-leste/libicd-provider-tor/blob/master/etc/libicd-provider-tor.schemas#L27 | 17:31 |
Wizzup | the 'tor-provider' is the service_id | 17:32 |
Wizzup | I don't really see why we would want that per service_id | 17:33 |
parazyd | So you mean each config has to have it or? | 17:33 |
Wizzup | unless a service would offer widely different services but still have the same service_type | 17:33 |
Wizzup | parazyd: yes | 17:33 |
Wizzup | and I think that's kind of stupid | 17:33 |
parazyd | Same | 17:34 |
parazyd | Especially if they're the same type | 17:34 |
Wizzup | the custom_ui stuff is read from only a few places, so we could just change that | 17:34 |
parazyd | I mean, there is a usecase, but quite an edge | 17:34 |
Wizzup | it's connui/iap-common.c | 17:34 |
Wizzup | connui-common/connui/iap-common.c | 17:34 |
parazyd | Yeah please do | 17:34 |
Wizzup | freemangordon: ^^^ ? | 17:34 |
Wizzup | (do you agree is the question) | 17:35 |
Wizzup | bbiab | 17:35 |
uvos | parazyd: have you removed mce-dev from jenkins yet? | 19:28 |
parazyd | I didn't understand what has to be removed. If you build the new package it should just override the old one | 19:29 |
parazyd | As I wrote in the pull request | 19:29 |
uvos | true, but haveing 2 jobs that result in the same pacakge is wierd and someone could accidentaly trigger the old mce-dev in the future causein the version in the repo to get downgraded no? | 19:30 |
parazyd | Done | 19:31 |
parazyd | I thought you were referring to the .debs | 19:32 |
uvos | no | 19:32 |
uvos | those should stay ofc | 19:32 |
parazyd | Yeah and they just automatically get replaced when you build the new package so no worries there | 19:32 |
uvos | right | 19:32 |
uvos | Wizzup: please build mce for devel or give me the rights to do so myself | 19:33 |
Wizzup | uvos: right I asked this morning when I wanted to tag the new version | 20:00 |
uvos | Wizzup: asked what? | 20:00 |
Wizzup | about 1.9 | 20:01 |
Wizzup | parazyd: we could also keep the custom_ui per provider_id but make a default gconf entry for provider_type and make connui read it | 20:01 |
uvos | right ok, yeah please tag out 1.9 and merge into devel branch | 20:01 |
parazyd | Wizzup: I'm ok with whatever is simpler | 20:02 |
Wizzup | parazyd: I could somehow see the use in being able to customise per config | 20:08 |
Wizzup | just don't want our code to have to write it | 20:08 |
Wizzup | so a default per type seems sensible to me | 20:08 |
parazyd | *nod* | 20:08 |
parazyd | I also wrote above that I see a usecase, but I think it's really an edge-case | 20:09 |
parazyd | ACK @ default | 20:09 |
Wizzup | yeah we at least need a default | 20:09 |
Wizzup | parazyd: ok, I'll add a default | 20:12 |
uvos | "Not using system installed systemui dbus names" | 20:12 |
uvos | hmm | 20:12 |
uvos | not a big deal rn but something is wrong there | 20:12 |
Wizzup | I think that is a big deal because then it disables maemo stuff iiuc | 20:14 |
Wizzup | you probably lack some debian/control stuff | 20:14 |
uvos | that dosent | 20:15 |
uvos | but yeah it dose later | 20:15 |
uvos | SystemUI support disabled | 20:15 |
parazyd | Wizzup: Nice, thanks | 20:15 |
Wizzup | parazyd: looks like it can even add scan results from gconf via services heh | 20:17 |
Wizzup | uvos: yikes, that'll break some stuff | 20:17 |
Wizzup | should we abort the build? or | 20:17 |
uvos | Wizzup: sure cancle it, im on it rn wrt solving it | 20:18 |
Wizzup | uvos: let's just let it build I think | 20:19 |
Wizzup | parazyd: I am going afk in a bit, can you help uvos rebuild or merge fixes? | 20:19 |
uvos | nobody upgrade for a couple min then :P | 20:20 |
parazyd | uvos: Added mce to your packages list | 20:21 |
Wizzup | I think he still can't push | 20:22 |
uvos | its missing osso-systemui-dev as a dependacy | 20:24 |
uvos | not sure how it worked before | 20:25 |
parazyd | uvos: You should be able to push if you accept the invite | 20:25 |
uvos | great should work now | 20:28 |
uvos | hopefully no one upgraded in the last couple of minutes | 20:28 |
Wizzup | :D | 20:29 |
uvos | SystemUI support enabled | 20:30 |
uvos | ok should build all the right modules now | 20:30 |
Wizzup | uvos: breaking -devel isn't the end of the world fwiw | 20:31 |
Wizzup | but yeah | 20:31 |
uvos | still would be kinda nice | 20:31 |
uvos | if you could dry run the build on jenkins | 20:31 |
uvos | and then download the deb too inspect | 20:32 |
uvos | without it ever going live anywhere | 20:32 |
freemangordon | uvos: you're missing pkg-config dependencies too it seems | 20:35 |
freemangordon | you should not count on debian/control to assure the dependencies | 20:35 |
uvos | freemangordon: its totaly modular, it just builds whatever modules it can. | 20:35 |
freemangordon | ah, ok | 20:36 |
freemangordon | Wizzup: not sure I understand the question, and also it was a very long and tiresome week for me, I don't have power for maemo now :( | 20:36 |
freemangordon | lets discuss on Sunday | 20:36 |
freemangordon | sorry | 20:37 |
uvos | repo mce works fine here | 20:41 |
uvos | please upgrade and test at your lesiure | 20:41 |
Wizzup | parazyd: ok I'll fix that default thing tomorrow, there is something weird but I can't figure out what | 20:55 |
Wizzup | maybe we just remove the per service_id stuff, I don't think it makes much sense anyway | 20:55 |
parazyd | Wizzup: What does it do at its current state? | 21:17 |
Wizzup | parazyd: no icons, no nothing | 22:02 |
Wizzup | I'll fix it tomorrow | 22:03 |
parazyd | Wizzup: ok, let's just make sure that most of it can be copypasted also for wg and ovpn :) | 23:07 |
Wizzup | :p | 23:26 |
Generated by irclog2html.py 2.17.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!