Wednesday, 2021-04-14

rinigusI think I cracked pkg(xxx) and similar dependencies for small automated build system. now you can define your project consisting of multiple SPECs and let it compile all of them taking into account dependencies.09:17
rinigustested on pure maps and flatpak so far. for port, will probably have to extend a bit, but logic should be mostly there09:18
rinigusas before, available at
T42Tanti %lastname% was added by: Tanti %lastname%11:20
T42<Jame %lastname%> Fwd from Time: When I lost my job!  I thought I lost it all.  With a little $ 1,000 in cash, I got $ 10,000 with @Mason4117 he may not know it, but it really saved my lifeπŸ‘‡πŸ‘‡πŸ‘‡πŸ‘‡πŸ‘‡ @Mason411711:28
T42<Jame %lastname%>
T42<b100dian> @dontatmepls did you manage to fix the error?11:44
T42<b100dian> I see you're on Linux 5. I got the same with 4.14 and it seems to me that CONFIG_PFK is not compatible with CONFIG_AUDIT=n  (basically the build passes if you dont' disable audit). But that may be blocking somewhere in the future11:44
T42<dontatmepls> I wasn’t able to fix it I think it was a long time ago11:45
T42<elros34> @b100dian It's recommended to keep AUDIT=y12:09
T42<b100dian> Yes, I know.. I will need to find how can this be achieved with a new-ish kernel, on the first pass I got E's build error12:10
T42<b100dian> If not, are there any dragons disabling it through kernel cmdline..?12:12
T42<elros34> sure you can add audit=0 to kernel cmdline12:13
T42<elros34> it will only makes your logs free of audit spam12:14
T42<b100dian> It seems a discussion from 2019 surfaces on duckduckgo when searching for CONFIG_PFK
T42<b100dian> Are Mal or GhosalMartin here to tell the end of it? It seems they are referncing a Mix 3 port (and I'm encountering the error on the Mi Note 10 kernel so..)13:18
malwhat error?13:19
T42<b100dian> actually you were just quoted in the link above ^^ ("mal: not sure if someones mentioned, but a nice person emailed me about an issue i had a while back with disabling selinux, and it seems the flag CONFIG_PFK has to be disabled for selinux to be disabled"), I come to think you were not in that conversation actually?13:24
mal@100dian you could try to see which driver causes the kernel build error and disable it13:34
T42<elros34> What do you mean by disable selinux? I has to be disabled via boot param not CONFIG_SECURITY_SELINUX. That way it will not affect your build. What excatly issue you are trying to fix?13:35
T42<b100dian> The issue is a fscrypto header that gets included both with CONFIG_PFK defined and not defined, leading to this redefinition error E had some time ago: . The 'fix' for me now is to set AUDIT=y. The other conversation I found around PFK references SELINUX indeed. I will (when get to that) disable both fr13:42
T42om cmdline13:42
T42<elros34> yeah just don't follow mer-kernel-check advices about selinux, audit, fw helper, they are wrong13:47
malshow a link to the kernel sources13:47
T42<b100dian> Config changes  will go to hybris-17.1 branch not yet pushed14:03
T42<b100dian> Good to know Elros, thanks14:06
malso you get different error with CONFIG_PFK and enabled and disabled?14:08
mal@b100dian I assume you should have CONFIG_PFK enabled and of course selinux as =y in kernel see chapter 9.3.1 in hadk pdf (you need to have CONFIG_SELINUX=y and CONFIG_AUDIT=y in kernel)14:10
T42<b100dian> Not sure what happened what PFK disabled actually, I'll try and post back. But yes, I'll go with the settings you said, thanks mal14:14
T42<b100dian> *with PFK..14:43
mal@b100dian you mentioned earlier that you have disabled CONFIG_PFK because you disabled selinux15:04
T42<b100dian> I may have sent confusing messages. Let me get back to you when I boot the build machine after hours:)16:10
piggzspiiroin: ping19:10
piggzmal: i have a device that emits KEY_WAKEUP .... i can map that to KEY_POWER in mce, but, it still doesnt wake19:13
piggzhow do i know if mce is listeneing on that device19:13
piggzor is ignoring it19:13
T42<rigaslitsas> hello i'm trying to build sailfish for xiaomi mi note pro. I'm following the HADK pdf for v4.19:46
T42<rigaslitsas> When i try to rpm/dhd/helpers/ --droid-hal19:46
T42<rigaslitsas> I have a weird error about img2simg19:46
T42<rigaslitsas> and some .o files not recognised19:46
T42<rigaslitsas> this is the build log19:48
piggzthinking about it, you dont want to map KEY_WAKEUP to KEY_POWER, as every time there is a ring indication, the screen will invert20:00
malpiggz: what kind of device is it?20:04
piggzmal: pinephone20:04
malpiggz: I mean the device in evdev_trace20:04
piggz---====( /dev/input/event2 )====----20:04
piggzName: "gpio-keys"20:05
piggzID: bus 0x19, vendor, 0x1, product 0x1, version 0x10020:05
piggzType 0x00 (EV_SYN)20:05
piggzType 0x01 (EV_KEY)20:05
piggz         KEY_WAKEUP20:05
piggzmal: walking through
piggzits possible such a device could be ignored20:17
malpiggz: so what triggers that?20:24
piggzmal: the modem triggers a gpio as RingIndicator ... the device tree creates an input device for this...20:27
piggzmal: see
T42<b100dian> mal: I have tried switching both AUDIT and PFK to No and the error is the same. I'm continuing the HADK steps with them "on"23:00
T42<b100dian> At "rpm/dhd/helpers/ --configs" I get "community-adaptation is needed by droid-config-tucana-1-202104142231.aarch64" - in which zypper repo are those packages?23:01
T42<b100dian> oh, I get it:( I need another define for ANDROID_ROOT for PlatforSDK and sb2. Because I also bind-mounted the build folders, they are not visible as they are on the host..23:07

Generated by 2.17.1 by Marius Gedminas - find it at!