| inky | true. Wizzup, you cannot imagine, i have so many things to package right now. things that i already tested. I'll do that and open several requests on bugtracker too. | 02:30 |
|---|---|---|
| inky | by the way, i really want to have librewolf, but if i am not mistaken the maps app was running out of disk space? i dont remember how much firefox fork can take, i guess not more than 15g. | 02:31 |
| inky | shell i try and open bugtracker request for librewolf as well? | 02:32 |
| inky | librewolf disabled lots of functionality, and it is not only privacy oriented but also i feel how much faster is it compared to firefox on pinephone too. | 02:35 |
| Wizzup | you can always open tickets | 02:49 |
| Wizzup | I think we'll probably put out eggs in the jib browser bucket for the foreseeable future, but I am also sad that firefox hasn't made it easier to do embedding | 02:49 |
| Wizzup | inky: well packaging things is a great way to have your efforts reach beyond just you and ensure others can enjoy them too | 02:53 |
| Wizzup | and yes, building firefox with all the rust stuff is probably out of the question on our build servers | 02:56 |
| arno11 | dsc_: rtcom leak fix seems to have a big impact on N900 ram usage :) | 11:21 |
| arno11 | for example, when conversations UI and 4 or 5 chat windows are opened, ram usage is back to normal when i close all of them | 11:25 |
| arno11 | (around 110MB free/available) | 11:27 |
| arno11 | (previously it was around 60-70MB) | 11:27 |
| arno11 | let's see if it's still the same after few days | 11:31 |
| dsc_ | arno11: currenty im working on another commit that will save 30MB memory | 11:31 |
| dsc_ | arno11: btw its interesting that you can observe the memory goes back to normal, because in my VM it did not do that quite as much as on your n900 | 11:32 |
| dsc_ | maybe Qt internal garbage collecting is more aggressive when it notices less RAM is available | 11:32 |
| arno11 | ah maybe that's because of custom vm stuff on N900 (swappiness, cache pressure and others) | 11:37 |
| arno11 | *vm custom stuff == virtual memory (to avoid confusion) | 11:44 |
| dsc_ | i figured :p | 11:45 |
| arno11 | :P | 11:45 |
| arno11 | will be interesting to test the next commit and compare to VM | 11:45 |
| dsc_ | im redoing the overview screen | 11:46 |
| dsc_ | so the first screen | 11:46 |
| arno11 | ah ok | 11:46 |
| dsc_ | same design, but better performance | 11:46 |
| arno11 | cool | 11:46 |
| arno11 | bbl | 11:49 |
| * Wizzup tests razr/xt912 image | 12:37 | |
| dsc_ | https://plak.infrapuin.nl/selif/wthwek6v.png | 13:44 |
| dsc_ | creating this part in QtWidgets is going to be 'fun' | 13:44 |
| dsc_ | a table column with a custom delegate | 13:45 |
| dsc_ | not for the faint of heart | 13:45 |
| dsc_ | setCellWidget() is easier, but not performant | 13:46 |
| Wizzup | is the scrolling speed similar? | 13:54 |
| Wizzup | as in, does qtwidgets also use gles? | 13:54 |
| Wizzup | or at least some method better than gtk2 scrolling :D | 13:55 |
| dsc_ | im not sure | 13:56 |
| dsc_ | quick is always hw accel'd | 13:56 |
| dsc_ | widgets probably not | 13:56 |
| Wizzup | might be worth checking, because if it's really slow to scroll we shouldn't go to qtwidgets | 13:57 |
| dsc_ | sure | 13:57 |
| dsc_ | scrolling is fine, i tried with 512 rows on d4 | 14:17 |
| Wizzup | ok, great | 14:26 |
| dsc_ | now lets 512 rows on rows with richtext rendering | 15:30 |
| dsc_ | lets try* | 15:30 |
| dsc_ | now lets try 512 rows with richtext rendering* | 15:30 |
| dsc_ | gr8 success | 15:48 |
| uvos__ | qtwidgets is software rasterized, there is (used to be?) a ogl backend to qpainter but it was allmost universally slower so never enabled by default | 15:52 |
| dsc_ | yeah | 15:55 |
| dsc_ | https://plak.infrapuin.nl/selif/lluse1h0.jpeg | 16:35 |
| dsc_ | this looks simple but in QtWidgets its a bit annoying | 16:35 |
| dsc_ | anyway, 512 rows of this works fine | 16:36 |
| dsc_ | user@devuan-droid4:~/conv$ python3 ~/ps_mem.py -p $(pidof hello) Private + Shared = RAM usedProgram | 16:36 |
| dsc_ | 19.2 MiB + 6.9 MiB = 26.0 MiBhello | 16:36 |
| Wizzup | uvos: the mapphone kexecboot config package should maybe check if /boot/boot exists when it runs | 19:20 |
| Wizzup | I think there is an issue during the image creation if /boot/boot/ doesn't exist, then it doesn't do nything | 19:20 |
| Wizzup | I'm verifying this now | 19:20 |
| unic0rn | hi. a question, is it possible to install leste on a droid 4 without using a microsd card? not sure what's the current setup, will have to find the device since I haven't used it much, but it's certainly rooted. | 19:24 |
| unic0rn | just dropping this here, will come back later, have places to go unfortunately | 19:24 |
| Wizzup | yes, it is possible, but it's tricky and not well documented currently | 19:24 |
| Wizzup | in general messing with the android partitions can be a frustrating thing - booting back to android once might have android nuke them | 19:25 |
| unic0rn | ah. great. then I'll have to check how the droid is set up | 19:25 |
| unic0rn | no plans for using android on that thing. it has some android/cyanogen/whatever installed, that's all I know right now. | 19:26 |
| unic0rn | if I'll fail, worst case it'll get bricked, not a big loss | 19:26 |
| Wizzup | Well, you can't nuke android | 19:53 |
| Wizzup | some parts have to stay intact | 19:53 |
| Wizzup | afaik | 19:53 |
| Wizzup | so in this sense microsd cards are much easier | 19:53 |
| uvos | you can install leste to the internal storage of the deivce | 20:10 |
| uvos | doing so by tempoarly installing leste to an sdcard is easy, but doing so entirely without a sdcard is tricky | 20:11 |
| uvos | https://uvos.xyz/maserati/LesteXT894EmmcInstall.txt | 20:11 |
| Wizzup | this misses an /etc/ I think: /mnt/LesteRoot/fstab | 20:18 |
| Wizzup | dsc_: I also have one charger that affects the ts a bit | 20:52 |
| Wizzup | others don't | 20:52 |
| dsc_ | ah | 20:53 |
| uvos | dsc_: if your charger is affecting the ts, this usually means it has a high degree of mains leakage | 20:55 |
| Wizzup | freemangordon: I am adding hildon-input-method as depedency for hildon-im-fkb and hildon-im-keyboard-assistant as their postinst depends in the hildon-im-recache binary | 20:56 |
| uvos | you can often also feal a high leakage by running your finger allong the back of the d4s display, you may have to flip the plug around to feel the mains "hum" there | 20:56 |
| Wizzup | just got this in my image build: | 20:56 |
| Wizzup | /var/lib/dpkg/info/hildon-im-fkb.postinst: 4: hildon-im-recache: not found | 20:56 |
| Wizzup | dpkg: error processing package hildon-im-fkb (--configure): installed hildon-im-fkb package post-installation script subprocess returned error exit status 127 | 20:56 |
| dsc_ | uvos: AKA it is haunted | 20:57 |
| dsc_ | haunted charger. | 20:57 |
| uvos | more like fiesty and willing to wake you up with a stout zap :P | 20:58 |
| dsc_ | risking my life for the Leste project, receiving current leakage through my fingers on a daily basis | 21:00 |
| uvos | nah its not dangerous | 21:01 |
| uvos | usualy anyhow, current is way to low | 21:01 |
| uvos | unless its really defective ofc | 21:01 |
| dsc_ | :P | 21:04 |
| * Wizzup tries again for xt912 image | 21:20 | |
| Wizzup | this time I think everything will work | 21:20 |
| dsc_ | https://plak.infrapuin.nl/selif/p3gxkylr.png | 23:39 |
| dsc_ | ETA desktop app | 23:40 |
Generated by irclog2html.py 2.17.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!
