*** cquence1 is now known as cquence | 00:07 | |
*** cquence1 is now known as cquence | 03:37 | |
*** cquence1 is now known as cquence | 16:55 | |
nephros | Pooh, this was a journey. I hate hate hate DBus. | 20:43 |
---|---|---|
nephros | rinigus: please test Portal-amber again with Kirigami. | 20:44 |
nephros | Ah, sorry, piggz[m] | 20:44 |
piggz[m] | ooo, ok | 20:44 |
nephros | I got sidetracked and implemented Settings interface as well. Which I *think* could be used to pretend being KDE, and making apps retrieve UI colors and fonts from Lipstick. | 20:45 |
piggz[m] | nephros: it seems to return a file:// path, but doesnt work | 20:50 |
piggz[m] | https://paste.mozilla.org/aG2J1GGb | 20:52 |
nephros | Damnit. | 20:54 |
nephros | Thanks for testing, could you show the dbus trace from Desktop once more? The last link expired. | 20:55 |
piggz[m] | yeah,i was just making one! | 20:58 |
flypig | did you forget | 21:02 |
flypig | Haha; oops; please ignore my cut-and-paste error. | 21:03 |
attah | regardless - yes | 21:03 |
flypig | :D | 21:03 |
attah | (i did forget to read the update today) | 21:03 |
piggz[m] | https://paste.mozilla.org/KdxO5OXu | 21:05 |
nephros | Ah, its again the " variant array [ variant string "file:///foo" ]" vs " variant array [string "file:///foo" ]" thing mak pointed out last time. | 21:07 |
nephros | will Try QDBusVariant instead of QVariant maybe that helps. | 21:08 |
piggz[m] | nephros: also, in the second call, the uri is missing completely? | 21:08 |
piggz[m] | not sure why its repeated | 21:08 |
nephros | Yes I noticed. Now I'm sending the reply manually instead of populating an out property. | 21:09 |
nephros | But still, why would an unset out property have the uri string? | 21:09 |
nephros | I'll look into it. | 21:10 |
direc85[m] | I recently had my share of QVariant[Map|List] shenanigans in Whisperfish. It all worked out though and now we're handling both "QVariantList<QVariantMap<QVariant<QString>,QVariant<QString>>>" and "QVariantList<QVariant<int>>" (IIRC) coming from QML to Rust via C++ shims. Fun! And because everything is essentially QVariant or its decendant, everything "worked" but just returned null values until I got the recipe just right. | 21:44 |
direc85[m] | Now that I think about it, there's still some JSON stuff from Rust to QML going on, I'll make a note to get those fixed too. | 21:46 |
nephros | direc85[m]: Thanks, yea I struggled with this as well, especially as Nemo.DBus likes to make everything variants. | 21:49 |
nephros | Actually I struggle with everything dbus all the time :P | 21:49 |
direc85[m] | Me too... :) | 21:49 |
nephros | piggz[m]: new version, new try. Response seems to have the right type for the stringlist now. | 21:50 |
nephros | direc85[m]: Yeah from time to time I just like to JSON.stringify() in QML and send as plain string over dbus to get by all the triple-marshaling. | 21:52 |
* piggz[m] uploaded an image: (416KiB) < https://matrix.org/_matrix/media/v3/download/matrix.org/LIyVMDRAJAaWCNbcQXWlAISZ/Screenshot_20240306_001.png > | 21:53 | |
piggz[m] | \o/ | 21:53 |
nephros | But I learned about QDBusArgument::beginStructure()/endStructure() and that is cool for arbitrary responses. | 21:55 |
nephros | Woohoo!!! | 21:55 |
nephros | :D | 21:55 |
direc85[m] | If you want to have a peek, those are https://gitlab.com/whisperfish/whisperfish/-/blob/main/whisperfish/src/actor/message/methods.rs#L57 and https://gitlab.com/whisperfish/whisperfish/-/blob/main/whisperfish/src/worker/client.rs#L2727 respectively | 21:55 |
direc85[m] | piggz[m]: That looks cool! I love the elephant too! | 21:56 |
piggz[m] | thats just built into the tokodon client :) | 21:57 |
nephros | direc85[m]: I feel I have fought dbus signatures enough for a day. And I don t speak rust. (But thats not an issue, I dont speak cpp as well.) | 21:58 |
direc85[m] | :D | 21:59 |
direc85[m] | If you limit yourself to only those loops, and ignore every try_into() and unwrap(), it's quite readable I think :Ö) | 21:59 |
direc85[m] | :) | 22:00 |
piggz[m] | nephros: well done, this is super useful | 22:01 |
b100dian[m] | <nephros> "Actually I struggle with..." <- As someone who tried to take a grasp at the dbus stuff in hydrogen and failed, I think you're doing fine :P Struggle, but win | 22:02 |
nephros | piggz[m]: Great! | 22:02 |
piggz[m] | file preview thumbnails might be handy ;) | 22:03 |
nephros | Srsly? Polish feature equests already? :D :D | 22:04 |
nephros | b100dian[m]: working version is xdg-desktop-portal-amber-...-1.231.1.jolla.aarch64.rpm. Note the 231 builds on OBS. :D | 22:05 |
piggz[m] | nephros: do the systemd services work correctly? ive not maanged to get it to work without manually starting things at the terminal | 22:06 |
nephros | piggz[m]: The Picker popup needs some work. Doesn't work on landscape for example. | 22:06 |
nephros | piggz[m]: Not sure actually. The flatpak one should be ok, but I think it must be started after the amber one. | 22:07 |
direc85[m] | nephros: You can optionally Finnish feature requests too ;) | 22:07 |
nephros | Also, that XDG_CURRENT_DESKTOP var must be in the (dbus) environment. | 22:08 |
nephros | could be that it isn't until you reboot, or force it via dbus-update-activation-environment | 22:08 |
piggz[m] | you could add a file for that to /var/lib/environment/compositor i think | 22:08 |
nephros | There is one already. | 22:08 |
piggz[m] | ah ok | 22:09 |
nephros | Ah but its at /var/lib/environment/amber/xdg-desktop.conf | 22:09 |
nephros | Possibly the flatpak daemon does not pick it up. | 22:10 |
piggz[m] | yeah if you stick it in compositor, it will get loaded with lipstick | 22:14 |
b100dian[m] | So the XDG_CURRENT_DESKTOP=Lipstick is not a Qt launcher setting, but a dbus one? | 22:25 |
nephros | b100dian[m]: it is read by the xdg-desktop-portal binary to determine which config to use. | 22:32 |
nephros | b100dian[m]: So it must be in dbusses environment. | 22:33 |
nephros | piggz[m]: do you know if the target graphical-session.target does anything on SFOS? | 22:34 |
nephros | Because xdg-desktop-portal makes itself PartOf it in the unit files. | 22:34 |
b100dian[m] | but wouldn't that just be an EnvironmentFile in systemd service file? | 22:34 |
nephros | b100dian[m]: yes, but in the case of dbus-activation the EnvironmentFile might have been placed after dbus has launched. | 22:35 |
nephros | so dbus might not have it in the env. | 22:35 |
nephros | But that is basically only an issue on initial installation. | 22:36 |
piggz[m] | nephros: i dont think that target is used | 22:38 |
nephros | piggz[m]: thought so. so the units must be adjusted. | 22:40 |
nephros | piggz[m]: thumbnails imolemented. | 22:41 |
piggz[m] | alreadY? :D | 22:41 |
nephros | I'm much quicker in qtquick than in qt ;) | 22:42 |
Generated by irclog2html.py 2.17.1 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!