*** spiiroin_ is now known as spiiroin | 06:27 | |
spiiroin | vknecht is gone, but... having wonky system time at bootup is expected if rtc time-of-day is not writable | 06:30 |
---|---|---|
spiiroin | sometimes it depends on device tree / board config setting, but sometimes modifying just can't be enabled | 06:32 |
spiiroin | in those cases sfos uses "rtc tod + offset" to keep track of what system time ought to be | 06:32 |
spiiroin | that is handled by dsme -> system time / statistics for systemd units started before dsme will be off | 06:33 |
spiiroin | the solution could be to handle the same / similar offsetting in something that gets executed earlier - preinit / stte | 06:34 |
vknecht | spiiroin_, thanks, I'll investigate this :) | 08:58 |
maccraft123 | hi | 17:25 |
maccraft123 | how powerful computer i need to port? | 17:25 |
maccraft123 | i want compile times for kernel to be around 10-20min | 17:25 |
maccraft123 | and will thinkpad z60m suffice? 2gb ram, 1 core pentium m, 2ghz | 17:26 |
maccraft123 | (i have also desktop) | 17:26 |
maccraft123 | qno, s4 mini lte is done? | 17:29 |
maccraft123 | i mean usable state at least | 17:29 |
piggz | maccraft123: power doesnt really come into it | 17:48 |
maccraft123 | piggz: doesnt porting involve compilation? | 17:49 |
piggz | maccraft123: well yes, but I use an old i5, and a lot of compilation can be done on obs | 17:57 |
maccraft123 | piggz: ok | 17:59 |
T4 | <eugenio_g7> Most of the things are pulled by Jolla's repositories in binary form | 17:59 |
T4 | <eugenio_g7> I think the most time consuming bit is the kernel | 17:59 |
maccraft123 | And how to actually start working? | 18:00 |
TheKit | maccraft123, but with 2gb RAM you will pretty likely run into OOM on mer-hybris bits compilation | 18:01 |
T4 | <eugenio_g7> maccraft123: https://sailfishos.org/develop/hadk/ | 18:01 |
maccraft123 | TheKit: Oh, ok, my desktop has 4gb and phenom 9650. Would it suffice? | 18:02 |
TheKit | should be better | 18:02 |
maccraft123 | If device works with postmarketOS, porting would be easier? | 18:05 |
TheKit | it is not very related, as postmarketOS generally does not use Android drivers via libhybris | 18:07 |
maccraft123 | TheKit: But kernel is same? | 18:08 |
TheKit | kernel sources are, but enabled set of options might be different | 18:09 |
piggz | r0kk3rz: where is your build server logated again? | 18:14 |
piggz | r0kk3rz: ive added config for latte-ci, what do i need to do to make it build? | 18:50 |
piggz | sorry, got it maybe | 18:51 |
piggz | r0kk3rz[m]: @eugenio_g7, rats, first attempt failed https://gitlab.com/sailfishos-porters-ci/latte-ci/-/jobs/160374540 | 19:00 |
mal | piggz: are you missing some package from OBS? | 19:03 |
*** olesalscheider_ is now known as olesalscheider | 19:04 | |
piggz | mal: i'll try a local build... | 19:04 |
piggz | mal: https://paste.gnome.org/pxrcleexa | 19:11 |
piggz | the pacakge seems ok https://build.merproject.org/package/show/nemo:devel:hw:xiaomi:latte/mce-plugin-libhybris-nondroid | 19:11 |
piggz | hangon... | 19:12 |
piggz | Problem: mce-1.98.2.2-1.16.1.jolla.i486 conflicts with mce-plugin-libhybris < 1.3.0 provided by mce-plugin-libhybris-nondroid-1.12.3+nondroid.20190209185149.4.g6b53138-1.3.1.jolla.i486 | 19:13 |
piggz | Solution 1: do not install mce-plugin-libhybris-nondroid-1.12.3+nondroid.20190209185149.4.g6b53138-1.3.1.jolla.i486 | 19:13 |
T4 | <eugenio_g7> Wut | 19:15 |
piggz | hmm, version seems ok | 19:16 |
piggz | @eugenio_g7 get that error trying to install -nondroid on device, which is probably why the image build fails | 19:16 |
mal | piggz: maybe tag that? | 19:18 |
piggz | one moment, i'll have to fork it to the shared repo | 19:20 |
piggz | mal: tag as just 1.12.3? | 19:21 |
piggz | oh, that tag exists | 19:21 |
mal | piggz: maybe Provides: xxx = yy.y | 19:23 |
mal | piggz: to spec | 19:23 |
mal | instead of just Provides: xxx | 19:23 |
mal | piggz: I think tag is not the problem but that Provides | 19:24 |
T4 | <eugenio_g7> Could be | 19:24 |
mal | piggz: maybe Provides: mce-plugin-libhybris = %{version} | 19:24 |
mal | @eugenio_g7 try that latest line I just gave | 19:25 |
T4 | <eugenio_g7> piggz, I can fix that in ~30 min if you don't fork | 19:25 |
T4 | <eugenio_g7> Don't want to* | 19:25 |
piggz | building now on obs! | 19:26 |
piggz | :D | 19:26 |
piggz | building again with correct url | 19:27 |
mal | piggz: hopefully that helps | 19:30 |
piggz | mal: it does | 19:30 |
mal | piggz: good :) | 19:31 |
piggz | @eugenio_g7 so ...... about the test image build ...... i kinda like my tablet how it is, so how about you try it :D | 19:32 |
T4 | <eugenio_g7> Cool! Will try tonight :) Have to do a TWRP backup first! | 19:47 |
piggz | @eugenio_g7 https://gitlab.com/sailfishos-porters-ci/latte-ci/-/jobs | 19:51 |
piggz | i might just post it to XDA and hope for the best! | 19:51 |
piggz | so, bug wise, we have a BT problem ... it seems sometimes i cant connect stuff and need to reboot | 19:52 |
mal | piggz: the legendary release after it builds :) | 19:52 |
piggz | mal: yup ... if it builds, ship it :D | 19:52 |
T4 | <eugenio_g7> Heh, I'd avoid posting to xda for now | 19:53 |
piggz | yeah, i'll let you check it installs and runs first | 19:53 |
T4 | <eugenio_g7> I'd like to post to XDA an image using LVM rather than stowawaying on the base android rom FWIW | 19:55 |
T4 | <eugenio_g7> Given that we don't actually need it | 19:56 |
piggz | i know what you;re saying, but it would ne nice to share our efforts, and we could do that later couldnt we? | 19:57 |
piggz | mal: anyone using opencv on sfos? | 19:58 |
T4 | <eugenio_g7> Yes absolutely, the only thing that bugs me is that will need a reflash on their part | 19:58 |
piggz | yes, but I would write that as a warning | 19:59 |
T4 | <eugenio_g7> Ok | 19:59 |
maccraft123 | what did i missed? | 20:08 |
maccraft123 | miss* | 20:09 |
mal | piggz: not sure if using, I did test it a bit | 20:13 |
mal | piggz: if you want to try https://build.merproject.org/project/show/home:mal:opencv | 20:13 |
maccraft123 | (163/256) Installing: rpm-4.9.1.2-1.3.9.i486 .....................................................................................................................[error] | 20:17 |
maccraft123 | Installation of rpm-4.9.1.2-1.3.9.i486 failed: | 20:17 |
maccraft123 | Error: Subprocess failed. Error: RPM failed: rpm: error while loading shared libraries: librpm.so.8: cannot open shared object file: No such file or directory | 20:17 |
maccraft123 | cant zypper dup | 20:17 |
mal | maccraft123: in which environment? | 20:17 |
maccraft123 | completely broke it now | 20:18 |
maccraft123 | PlatformSDK bash-3.2$ zypper | 20:18 |
maccraft123 | zypper: error while loading shared libraries: librpm.so.8: cannot open shared object file: No such file or directory | 20:18 |
maccraft123 | redoing it | 20:18 |
piggz | mal: i was thinking of doing facial recognition in harbour-advanced-camera ... tho ... that is maybe getting ahead of myself a bit feature wise for now! | 20:22 |
mal | piggz: android cameras usually have it already | 20:27 |
piggz | mal: yes, but, NIH | 20:30 |
mal | piggz: ? | 20:30 |
piggz | not-invented-here | 20:30 |
piggz | a syndrome | 20:30 |
piggz | where devs always have to roll their own implementation of something | 20:31 |
mal | piggz: I still don't understand, camera hal provides that | 20:32 |
piggz | mal: i mean, where is the fun in engineering it using the camera hal, when you can write it using opencv and qt video filters | 20:34 |
mal | piggz: ok | 20:34 |
piggz | i mean, it will just be an interesting experiment, and, fune | 20:34 |
piggz | s/fun | 20:34 |
piggz | but, as I say, maybe for later... | 20:34 |
piggz | someone offered to write icons | 20:35 |
mal | piggz: opencv examples have something you can easily use | 20:35 |
maccraft123 | what if my device has -common in repo? | 20:39 |
maccraft123 | i mean requiers both -common and -$DEVICE github repo | 20:40 |
piggz | maccraft123: they all do...if you mean obs repo | 20:42 |
maccraft123 | piggz: and how to fix it? | 20:43 |
mal | fix what? | 20:45 |
maccraft123 | How to make config? | 20:46 |
maccraft123 | local_manifest i mean | 20:46 |
piggz | ah, earlier in process....you just need to sync up you tree to be valid | 20:48 |
mal | and check dependencies from the repos from lineage.dependencies | 20:50 |
maccraft123 | apps also count to manifest? | 20:54 |
mal | ? | 20:59 |
maccraft123 | https://github.com/LineageOS/android_device_samsung_serrano-common/blob/cm-14.1/lineage.dependencies | 21:00 |
maccraft123 | FlipFlap app is needed afaik | 21:00 |
mal | you probably can skip the app | 21:00 |
maccraft123 | can recovery zip be made with sailfish? | 21:10 |
vknecht | spiiroin_, too bad, enabling qcom,qpnp-rtc-write in dtsi file causes bootloop... will have to try dsme way | 21:44 |
maccraft123 | ok so i did local_manifest | 21:46 |
vknecht | maccraft123, afaik one can flash or fastboot the produced hybris-recovery.img | 21:47 |
vknecht | tho I only ever tried it on xperia x | 21:47 |
maccraft123 | vknecht: i have serranoltexx | 21:48 |
maccraft123 | i mean s4 mini i9195 | 21:49 |
maccraft123 | the danm repo synchronizes _S L O W L Y_ | 21:49 |
vknecht | you're eager :) | 21:50 |
vknecht | maybe I misunderstood your question ; what I meant is that "make hybris-recovery" (included in "make hybris-hal") produces a recovery image, but nothing like twrp... | 21:53 |
vknecht | maccraft123, just curious, how did you skip FlipFlap ? using <remove-project name="..." /> in local manifest, or directly modifying lineage.dependencies ? | 21:55 |
maccraft123 | vknecht: i didnt specify it | 21:56 |
* vknecht is currently adding FlipFlap to his los base, so curious about the reverse operation :) | 21:56 | |
* maccraft123 takes a break, fixing second monitor | 21:57 | |
eugenio | @adampigg: this should be merged otherwise BT doesn't get set-up https://github.com/sailfish-on-latte/droid-config-latte/pull/12 | 22:05 |
eugenio | (haven't tried the build yet, waiting for the pad to charge...) | 22:06 |
spiiroin_ | vknecht: apart from incorrect system time early in the bootup it should be harmless | 22:17 |
spiiroin_ | vknecht: but if you really want to try to do something about, it then something like: test -f /var/tmp/delta-time && date -s @$(($(date +%s)+$(cat /var/tmp/delta-time))) | 22:20 |
spiiroin_ | ... should appy the same fixup as what dsme is going to do -> doing it early enough should make large system time jumps go away | 22:21 |
spiiroin_ | well, "hwclock -s" before that | 22:21 |
spiiroin_ | i.e. set system time to rtc tod, then adjust it by using the offset file written by dsme | 22:22 |
spiiroin_ | (if it is there) | 22:22 |
T4 | <adampigg> @eugenio_g7 ah ok, hope it works well when you get charge! | 22:44 |
T4 | <adampigg> btw, mine seems to work absolutley fine if it has a solid good charge | 22:44 |
T4 | <adampigg> works all day | 22:44 |
T4 | <adampigg> but, sometimes it seems like it doesnt get a full charge, but still says 100%, but then falls really quickly and gets sporadic | 22:45 |
T4 | <adampigg> and, I dont seem able to add you to the porters-ci group, maybe only r0kk3rz can | 22:46 |
eugenio | yup, requested access | 22:47 |
eugenio | anyway yesterday I emptied the battery in the hope to trigger a recalibration, not sure how well it works now | 22:48 |
eugenio | I tried the build and it boots fine, so that's great | 22:48 |
eugenio | sensors doesn't seem to work, haven't looked at them yet | 22:49 |
eugenio | UI, audio, wifi works | 22:50 |
eugenio | some icons seem small, not sure if that's how it should be on tablets as I haven't had the fortune to get a tbj :D | 22:51 |
eugenio | rpm installation via libzypp seem broken | 22:57 |
eugenio | hmmm nothing provides requested libhybris-libGLESv2 | 22:58 |
eugenio | dhi should be masked | 23:08 |
eugenio | I'd refrain for releasing just now but we're close | 23:09 |
eugenio | will hack more on it tomorrow, good night ;) | 23:10 |
Generated by irclog2html.py 2.17.1 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!