| T42 | <dgmarie2> Ok I built it but now I can't install in TWRP. The error is tar: exec bunzip2: No such file or directory | 02:23 |
|---|---|---|
| T42 | karlin5233 was added by: karlin5233 | 06:11 |
| rinigus | I wrote a small script to handle building multiple SPECs as a single project. as we have OBS not updated to aarch64, it is a way to handle corresponding builds. | 07:42 |
| rinigus | script, docs, and example available at https://github.com/rinigus/tbuilder | 07:42 |
| rinigus | it seems like my approach - generate Makefile using SPECs - has a limitation and maybe I have to rewrite it all into smaller build engine. See issue #1 in that repo | 07:42 |
| rinigus | very limited testing so far, mainly RFC right now. | 07:42 |
| rinigus | pinging few who could be interested, but not limited to them - piggz, mal , Thaodan ^ | 07:43 |
| Thaodan | rinigus: Have you looked into mock? That was suggested by #opensuse-buildservice when asking for something similar to the obs build engine but local. | 17:20 |
| Thaodan | We could path it to use mb2 intead of rpmbuild directly. | 17:21 |
| rinigus | Thaodan: not deeply, just a bit. there was something that I didn't like, maybe rpmbuild bit. but let me read about it again as I have learned few bits while writing my own solution | 17:23 |
| Thaodan | rpmbuild used anyway even sfdk | 17:23 |
| Thaodan | The hard but is the dependency resolution. | 17:24 |
| Thaodan | rpmspec is not relay able because of provides that only there after the package is build. | 17:25 |
| Thaodan | It is better to parse the pkg db | 17:25 |
| rinigus | Thaodan: agree, that's essentially the issue that I've got. although it is possible to approach it differently, as I wrote down in the issue as well. | 17:28 |
| rinigus | as for parsing pkg db, we can let zypper search through system provided ones and build dep tree from packages in that shared dir | 17:28 |
| rinigus | where RPMs are dumped | 17:29 |
| Thaodan | I see mock --chain as most valuable since it handles that own its own. | 17:35 |
| rinigus | Thaodan: do you have the reference to some mans on mock? some examples of --chains use? | 17:43 |
| rinigus | and have you looked already into how to plug it in into mb2? | 17:46 |
| Thaodan | https://github.com/rpm-software-management/mock/blob/606bb7459fd6c762de3a27848655d0f731ed900f/mock/docs/mock.1#L93 | 17:47 |
| Thaodan | However they don't mention something about dependency resolution | 17:47 |
| Thaodan | I guess instead of calling rpmbuild it should call mb2 or similar. | 17:48 |
| rinigus | I presume koji and copr do some dep resolution. maybe via mock | 17:48 |
| Thaodan | That could be. | 17:49 |
| Thaodan | rinigus: See my question in #rpm.org | 17:50 |
| Thaodan | Rinigus: https://github.com/rpm-software-management/mock/issues/170 | 17:53 |
| Thaodan | gg | 17:53 |
| Thaodan | Need to fix that/adapt how obs uses libsolv here. | 17:54 |
| rinigus | Thaodan: OK, from the issue it sounds like mock not ready yet ... I'll try for now to fix flatpak builds using my small tool and see if I can use it meanwhile | 18:23 |
| Thaodan | Good | 18:26 |
| Thaodan | If you need a repo ping me, I can host them | 18:26 |
| rinigus | Thaodan: you can't really host for all - we will need some systematic solution. but that is probably known anyway | 18:32 |
| Thaodan | Sure I can't but yeah.. | 18:36 |
| Thaodan | hosting would be the least issue | 18:36 |
| rinigus | yes, it is | 19:21 |
| T42 | <dgmarie2> I got Sailfish to install, but it gets stuck on boot logo. It shows up in dmesg under mer bootloader. What logs should I look at to figure this problem out? | 20:10 |
| T42 | <zinstack625> Try figuring that out in telnet, device should appear as an rndis ethernet | 20:12 |
| T42 | <dgmarie2> how do I connect? | 20:12 |
| mal | telnet 192.168.2.15 2323 (or 23) depending on how far it booted | 20:12 |
| T42 | <dgmarie2> I tried that earlier and it timed out and never connected | 20:13 |
| mal | do you see usb network interface on host computer? | 20:14 |
| T42 | <dgmarie2> where should I look for that | 20:15 |
| T42 | <dgmarie2> also when i boot it shows up as an mtp device called "always" | 20:15 |
| T42 | <dgmarie2> and i see this in dmesg "[168661.850191] rndis_host 1-1:1.0 usb0: register 'rndis_host' at usb-0000:02:00.0-1, RNDIS device, 76:df:f7:56:57:61" not sure if it is important | 20:16 |
| mal | ifconfig -a | 20:16 |
| mal | hmm, as mtp device? strange | 20:17 |
| T42 | <dgmarie2> it goes away after a second though | 20:17 |
| T42 | <dgmarie2> also Im using a custom hybris installer because the regular one wasnt working, maybe thats causing a problem? | 20:18 |
| T42 | <dgmarie2> I do not see usb network interface | 20:18 |
| spiiroin | there could be android side mtp leftovers in scripts? | 20:19 |
| spiiroin | if there are, depending on timing they could be conflicting with what usb-moded is doing | 20:20 |
| T42 | <dgmarie2> any suggestions on what I should do? | 20:22 |
| mal | well tell what you see in host dmesg when the device is booting | 20:23 |
| mal | hopefully it tells how far it boots | 20:23 |
| T42 | <dgmarie2> https://pastebin.ubuntu.com/p/JtPbYwVkFW/ | 20:24 |
| T42 | <elros34> are you sure that dmesg is from your boot time not same time later? idProduct should be 0xD001 for init-script. Also check ifconfig on you host whether you have usb0 interface | 20:34 |
| T42 | <dgmarie2> I do see usb0 interface | 20:36 |
| T42 | <elros34> does it have inet entry set? Something like: inet 192.168.2.1 netmask 255.255.255.255 broadcast 0.0.0.0 | 20:37 |
| T42 | <dgmarie2> it says command inet not found | 20:38 |
| T42 | <elros34> I meant check in ifconfig output | 20:39 |
| T42 | <dgmarie2> oh i misread what you said sorry | 20:39 |
| T42 | <elros34> if no then try something like: ip address add 192.168.2.1 dev usb0 && ip route add 192.168.2.15 dev usb0 | 20:39 |
| T42 | <dgmarie2> @elros34 [does it have inet entry set? Something like: i …], yeah i do see that | 20:41 |
| T42 | <elros34> so telnet 192.168.2.15 2323 works now? | 20:42 |
| T42 | <dgmarie2> @elros34 [so telnet 192.168.2.15 2323 works now?], no I didnt make any changes | 20:43 |
| T42 | <elros34> hm.. what about my previous question about dmesg and idProduct? You should have in dmesg Product: init-debug in real rootfs | 20:45 |
| T42 | <dgmarie2> @elros34 [hm.. what about my previous question about dme …], oh yes I do | 20:46 |
| T42 | <dgmarie2> I also get a notification saying activation of network connection failed on my host system | 20:46 |
| T42 | <elros34> yes it's "normal". So at least device pass initramfs and boot further. I guess it's hybris-17 right? DO you left selinux enabled or disabled? | 20:49 |
| T42 | <dgmarie2> @elros34 [yes it's "normal". So at least device pass ini …], selinux is disabled through the cmdline and im using hybris 16 because 17.1 rebooted directly to fastboot | 20:51 |
| T42 | <elros34> ah so should be easier, do you have hybris-16 changes in droid-config mentioned in hadk-faq? | 20:51 |
| T42 | <dgmarie2> @elros34 [ah so should be easier, do you have hybris-16 …], yes | 20:57 |
| T42 | <dgmarie2> @elros34 [ah so should be easier, do you have hybris-16 …], Do I need the ones for "When you get telnet in the SFOS rootfs " | 20:58 |
| T42 | <elros34> depends on device, you will know when you telnet and check logcat | 20:59 |
| T42 | <dgmarie2> @elros34 [depends on device, you will know when you teln …], Is there anything else I can try to get into telnet? | 21:00 |
| T42 | <elros34> the easiest way is stop booting. You would have to run in recovery: touch /data/.stowaways/sailfishos/init_enter_debug2 then telnet will work provided ifconfig shows inet address | 21:05 |
| T42 | <dgmarie2> I got in! | 21:28 |
| T42 | <dgmarie2> so what do I do now? | 21:31 |
| T42 | <elros34> shortly after reboot and telnet check in dmesg whether selinux is disabled at boot time, run journalctl --no-tail -f and continue booting according to instruction you should see after you have telnet. In journal make sure droid mounts for /system and other partitions works then check droid-hal-init logs and so on | 21:35 |
| T42 | <dgmarie2> it says no journal files were found | 21:38 |
| T42 | <zinstack625> Is systemd even running at this stage? | 21:39 |
| T42 | <elros34> no until you run echo continue ... according to instruction on screen | 21:39 |
| T42 | <elros34> on terminal* | 21:40 |
| T42 | <dgmarie2> I just disabled config_security_selinux in the defconfig. Maybe that was the issue? | 21:53 |
| T42 | <elros34> that must be enabled as in android | 21:53 |
| T42 | <dgmarie2> So I should re-enable it | 21:54 |
| T42 | <elros34> yes | 21:54 |
| T42 | <dgmarie2> ok | 21:54 |
| T42 | <elros34> about selinux, just add CONFIG_SECURITY_SELINUX_BOOTPARAM=y CONFIG_SECURITY_SELINUX_BOOTPARAM_VALUE=0 and that is enough | 21:55 |
| T42 | <dgmarie2> ok I've already done that | 21:57 |
| T42 | <dgmarie2> what about config_audit? Should I disable that | 21:57 |
| T42 | <elros34> no, must be enabled. Its part of selinux dependencies | 21:58 |
| T42 | <dgmarie2> Im not trying to doubt you, but are you sure I can't build without selinux? what if I turn off encryption? Just to see if it helps | 22:04 |
| T42 | <dgmarie2> idk im not really sure what the issue is preventing the boot. I can look through the logs more | 22:04 |
| T42 | <dgmarie2> theres a lot of messages like this "Feb 09 20:23:50 UNKNOWN kernel: binder: 3537:3537 transaction failed 29189/-22, size 32-0 line 3012" could this be the error? | 22:06 |
| T42 | <elros34> maybe you can with selinux disabled and some modification in droid-hal-init but I have never tried. Everybody disable it at boot time and that works. About error hard to tell without context. /usr/libexec/droid-hybris/system/bin/logcat might gives you some additional ideas. | 22:09 |
| T42 | <dgmarie2> Actually I think the issue is I forgot to include the anbox binder in drivers but I have it in the defconfig | 22:38 |
| T42 | <dgmarie2> Im building with a lot of birdzhangs sources cause he doesnt have davinci anymore | 22:39 |
Generated by irclog2html.py 2.17.1 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!