Werner__ changed the topic of #armbian to: armbian - Linux for ARM development boards | www.armbian.com | Github: github.com/armbian | Commits: #armbian-commits | This channel is logged -> http://irc.armbian.com
xBADDA555 has quit [Ping timeout: 240 seconds]
drobo_00 has quit [Remote host closed the connection]
<TRS-80>
sputnik: Now that I think about it, I seem to vaguely recall some thread about Stretch being "gone" but there were reasons for this. Let me see if I can find it again.
drobo_00 has joined #armbian
drobo_00 has quit [Client Quit]
<sputnik>
TRS-80: i see. wonder why that happened... thanks for lookining up, m8
<TRS-80>
no wait, maybe it was Jessie actually
<sputnik>
ah, jessie is still there, in the archives
<TRS-80>
hang on, I may be spreading bullshit
<sputnik>
mhm
* TRS-80
is being called to dinner
<TRS-80>
maybe someone that knows wtf they are talking about will appear
<TRS-80>
XD
<sputnik>
ehehe, sure, i'll hang out here :)
<c0rnelius>
ha
<therion23>
here i thought finding out you are wrong was the new right
<c0rnelius>
whats wrong Buster? Just cuz he got laid out in less than 10 seconds against Mike Tyson, don't mean he's all bad.
<sputnik>
doesn't boot on C1+, and since i ain't got no serial cable, i can't find out why
<c0rnelius>
good answer
<sputnik>
an honest one
<c0rnelius>
indeed
<sputnik>
jessie boots, though. an old timer, but boots and works. the problem is, it works for two or three days :)
<sputnik>
nah, i just dd'ied the official builds, jessie and then buster, from armbian.com downloads, didn't try anything else. but i guess that's an idea, i could try it out
<c0rnelius>
Just today I tried building a kernel for an amlogic s905x and it failed like no ones business. i then built a vanilla kernel from kernel.org and worked right off the bat. Of course missing some key armbian patches, but it does work.
<c0rnelius>
just saying... worth a go.
<sputnik>
i see. well, i guess i'mma give it a shot, then
wi has joined #armbian
torv has quit [Remote host closed the connection]
torv has joined #armbian
<TRS-80>
isn't there some option to change kernel in armbian-config?
<sputnik>
hmm, i didn't notice, but i wasn't really paying attension, to be frank...
<sputnik>
well, i'll try and find out tomorrow
<c0rnelius>
TRS-80: yeah i forgot about that. there is an option in the armbian-config for using diff kernels.
NeuroScr has joined #armbian
Tony_mac32 has joined #armbian
<TRS-80>
now it's a party! :D
_whitelogger has joined #armbian
dddddd has quit [Ping timeout: 256 seconds]
NeuroScr has quit [Quit: NeuroScr]
<archetech>
quietest friday party i've been at ;p
<TRS-80>
IKR?
* TRS-80
been reading back up on ZFS again, since it is becoming more relevant given some recent developments
xec has joined #armbian
xecuter has quit [Ping timeout: 256 seconds]
<Tony_mac32>
I was playing armor on a space man game
<TRS-80>
With Assault? :)
<Tony_mac32>
No, with DOT
<TRS-80>
Nice. :)
<Tony_mac32>
Besiege might join up, still waiting to see it happen
<Tony_mac32>
there is a huge update coming
<TRS-80>
You guys on Discord? TS?
wi has quit [Ping timeout: 240 seconds]
<Tony_mac32>
Discord, I was trying to get them to do an IRC bridge
<Tony_mac32>
;-)
<TRS-80>
I looked it up on Lutris and Wine DBs, looks like a no go from Gahnoo Leenux
<Tony_mac32>
there is a TS, but it's not the old one
<TRS-80>
Yeah I found.... I forget his name.... older guy who plays that free tank game... idling in there, holding down the fort all by himself one night. :)
<Tony_mac32>
lol I keep meaning to fire up war thunder, I really liked it
<Tony_mac32>
it looks like they're adding battlefield 2142 Titan mode to PS2
<TRS-80>
I saw that carrier or whatever, but didn't get too far into that video
wi has joined #armbian
archetech has quit [Quit: Konversation terminated!]
TRS-80 has quit [Quit: WeeChat 2.3]
sunshavi has quit [Read error: Connection reset by peer]
sunshavi has joined #armbian
archetech has joined #armbian
xBADDA555 has joined #armbian
<xBADDA555>
Hi guys. I accidentally installed Armbian_19.11.6_Bananapipro_bionic_current_5.4.8 (PRO) on my "Banana Pi BPI-M1" (M1). It is working like a charm, but recently I found my mistake, and tried to find the differences between Armbian_19.11.6_Bananapipro_bionic_current_5.4.8 (PRO) and Armbian_19.11.6_Bananapi_bionic_current_5.4.8 (M1) and understand if this may affect how my old banana operates. Running diff on files revealed that
<xBADDA555>
Bananapro_defconfig and u-boot-sunxi-with-spl.bin differs in both versions. I decided to extract u-boot-sunxi-with-spl.bin from Armbian_19.11.6_Bananapi_bionic_current_5.4.8 (M1) and install it on my M1 which is running on armbian dedidcated to PRO. No luck here - usb keyboard stopped working. I reverted back bootloader to PRO version, and now my questions are: Can I leave PRO version code running on M1 machine, without any
<xBADDA555>
consequences? What about the differences that I found here in Bananapro_defconfig
<xBADDA555>
# CONFIG_AXP_ALDO3_VOLT_SLOPE_16 is not set
<xBADDA555>
# CONFIG_AXP_ALDO3_VOLT_SLOPE_08 is not set
<xBADDA555>
# CONFIG_AXP_ALDO3_INRUSH_QUIRK is not set
<xBADDA555>
-CONFIG_AXP_ALDO4_VOLT=2500
<xBADDA555>
+CONFIG_AXP_ALDO4_VOLT=0
<xBADDA555>
# CONFIG_DM_PWM is not set
<xBADDA555>
# CONFIG_PWM_IMX is not set
<xBADDA555>
# CONFIG_PWM_SANDBOX is not set
<c0rnelius>
hey pastebin.com works pretty well
<xBADDA555>
Ah sorry for that.
<xBADDA555>
If it is not safe to keep running PRO code on M1 machine, then what else should be done except switching u-boot-sunxi-with-spl.bin to proper version, to force usb keyboard to work?
Tony_mac32 has quit [Ping timeout: 256 seconds]
lykt has quit [Quit: leaving]
dddddd has joined #armbian
lykt has joined #armbian
patrixl has quit [Quit: Leaving.]
dddddd has quit [Ping timeout: 260 seconds]
xBADDA555 has quit [Quit: Leaving]
sputnik has left #armbian [#armbian]
Strykar has quit [Quit: /quit]
<lanefu>
Yo
Strykar has joined #armbian
<c0rnelius>
Joe
xBADDA555 has joined #armbian
xBADDA555 has quit [Quit: Leaving]
xBADDA555 has joined #armbian
DaRock has quit [Ping timeout: 268 seconds]
<lanefu>
IgorPec: i'm re-enabling CI updates on PRs.. code seems to be working pretty well now
<IgorPec>
not sure if we need that, but perhaps once
<lanefu>
right now it's just doing kernel builds
<lanefu>
but eventually i'll do images
<IgorPec>
yes, just an idea to automatically recreate rootfs cache if number is changed ... but on the other hand
<lanefu>
i'm a little weirded about security with images because any stranger with a PR can effectievly run a shell script as root on my server :P
<IgorPec>
yes :)
<lanefu>
that being said, once we get further along, being able to break down the different build steps like rootfs etc in to jobs, we can do some fun stuff
<IgorPec>
consider this an experiment where anything can happen
<lanefu>
Jenkins finally added websocket support for their runners.. so that will make it easier to integrate other runners now that SSH wont be needed https://jenkins.io/blog/2020/02/02/web-socket/
<IgorPec>
ok, there is one problem with this cache rebuild ... until caches are not made and populated accross network, users will be force to build from packages
<lanefu>
will be forced to build what from packages?
<IgorPec>
yes, because cache with this number will not be done and uploaded
<IgorPec>
we are talking about a window of 12h
keesj has quit [Ping timeout: 240 seconds]
<IgorPec>
now i change id locally, rebuild, upload, wait 12h and then change master branch
<IgorPec>
this way must be done here as well
<lanefu>
so just trying to make sure I understand you correctly
<lanefu>
when you rebuild root fs caches, it makes it outdated for 12 hours.. so users have to build.... What? from packages
<IgorPec>
new caches are not available which means build script runs a deboostrap and make a local cache
<lanefu>
Ohhh okay
<IgorPec>
it doesn't download it
<lanefu>
gotcha... not a showstopper.. just an inconvience
<IgorPec>
yes, we could also solve it in the build script somehow
<IgorPec>
if current doesn not exists, check previous
<IgorPec>
otherwise build
<lanefu>
yeah... maybe some sort of index file on the server where build cache is for pointing to latest and previous.. and build script can just parse that and make the right decision?
<lanefu>
i havent' looked at FS cache code in a few months, so not sure what its doing right now
<IgorPec>
i will fix this problem on the way
<lanefu>
cool
<lanefu>
IgorPec: should i add CI to commits? or focus on expanding functionality of PR testing?
<IgorPec>
IMO testing is more important ... what functionality is now for commits and what is still nice to have?
brittman914 has joined #armbian
<brittman914>
greetings!
<lanefu>
so basically all it does now is attempt to build a kernel based on the files that changed... so either for a specific board if a board config changed, or it will pick a board in a family if a family hanged
<lanefu>
hey brittman914 !
<lanefu>
so i can apply that same thing to commits and just let it try to build an appropriate kernel
<lanefu>
it doesnt really focus on the build engine itself right now
<lanefu>
so i dont know if thats an aspect we want to look at next
<IgorPec>
usually testings are done at the MR so doing that also for commits, not sure if we need tht
<lanefu>
okay cool
<lanefu>
i wont worry about that then
<lanefu>
IMHO i still think we commit directly to master too often, but i understand things just need to get done
<IgorPec>
yes, thats purely optional ... is it worth to add "pass building" while creating one image
<IgorPec>
yeah, i try to do MR as much as possible
<lanefu>
hmmm
<IgorPec>
will do my best :)
<lanefu>
so I guess the next thing to figure out is, when to test building an image based on PR vs just test building a kernel
<IgorPec>
i would just put that to the time based scenario and voila
<lanefu>
oh
<lanefu>
just like a nightly test?
<IgorPec>
yes
<lanefu>
oh okay!
<IgorPec>
one image which will reflect in "building passed/failed"