zyga-mbp has quit [Read error: Connection reset by peer]
zyga-mbp has joined #yocto
<RP>
kanavin: I stopped your build since it was clear it was going to break and I need the workers ;-)
fbre has joined #yocto
prabhakarlad has joined #yocto
<kanavin>
RP: no problem :) if there's something more pressing I have a bit of spare cycles
<RP>
kanavin: the "pressing" issues are an rc2 for 3.3.3 (building) and fixing pseudo for glibc 2.34 and the pseudo patch backlog (now testing a v2). Other than that it is general bugs. I'm slowly trying to get through my backlog
zyga-mbp has quit [Ping timeout: 260 seconds]
zyga-mbp has joined #yocto
Vonter has quit [Ping timeout: 260 seconds]
jwillikers has quit [Remote host closed the connection]
argonautx has joined #yocto
xmn has quit [Ping timeout: 260 seconds]
jwillikers has joined #yocto
prabhakarlad has quit [Quit: Client closed]
Vonter has joined #yocto
prabhakarlad has joined #yocto
<wCPO>
What tools can be assumed to be installed on the host? As I read the WIC code it assumes "install" is installed.
<rburton>
see HOSTTOOLS in bitbake.conf
<wCPO>
rburton: thanks, objcopy is installed, so is make what I'm doing a bit easier :)
Guest58 has joined #yocto
argonautx has quit [Quit: Leaving]
cocoJoe has joined #yocto
<RP>
wCPO: do you want objcopy for the host or target though?
<kanavin>
RP: I'm trying to figure out why bitbake won't show deprecation warnings from python. There's a bunch of stuff that has been deprecated since python 3.3 (!) and it's gone in 3.10 - bitbake breaks badly with that.
<kanavin>
given the recent discussion ;) I think we should not silence such things, and display them front and center.
<kanavin>
but now, time for the bicycle!
<kanavin>
tgamblin, I'll try to take a look
kiran has joined #yocto
<tgamblin>
kanavin: awesome, thanks
<kanavin>
RP: e.g "ImportError: cannot import name 'Mapping' from 'collections' (/home/pokybuild/yocto-worker/buildtools/build/build/tmp/work/x86_64-nativesdk-pokysdk-linux/buildtools-extended-tarball/1.0-r0/testimage-sdk/sysroots/x86_64-pokysdk-linux/usr/lib/python3.10/collections/__init__.py)"
<kanavin>
I can fix all that, but first I want to see the warnings :)
<RP>
kanavin: I've worried about the deprecation code hiding things for a while. It may be best to remove some of it from bitbake, not sure
<RP>
kanavin: certainly it is hiding things it shouldn't be
frieder has quit [Remote host closed the connection]
<kanavin>
RP: there's some 2010 commits from a certain Mr.Larsson adding the hidings. I think we should drop that all - I did that, but the warnings still don't show up - something else is hiding them as well.
zyga-mbp has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
zyga-mbp has joined #yocto
sakoman has quit [Read error: Connection reset by peer]
sakoman has joined #yocto
florian has joined #yocto
sakoman has quit [Client Quit]
BobPungartnik has joined #yocto
prabhakarlad has quit [Quit: Client closed]
BobPungartnik has quit [Client Quit]
dmoseley has joined #yocto
prabhakarlad has joined #yocto
kiran_ has quit [Ping timeout: 252 seconds]
kiran_ has joined #yocto
cocoJoe has quit [Ping timeout: 256 seconds]
kiran_ has quit [Ping timeout: 260 seconds]
florian has quit [Ping timeout: 260 seconds]
kiran_ has joined #yocto
mtudan has joined #yocto
kiran_ has quit [Ping timeout: 260 seconds]
mtudan has quit [Ping timeout: 256 seconds]
kiran_ has joined #yocto
kiran_ has quit [Ping timeout: 260 seconds]
<kergoth>
Hmm, yeah, should likely be dropped. I honestly can't recall what the intention was behind it.
<kergoth>
that's *really* old
<kergoth>
maybe intended to avoid seeing them for bitbake itself when not developing bitbake itself? *shrug*
<RP>
kergoth: if you don't remember and agree I think dropping it all and seeing where we are is probably where to start! :)
<kergoth>
agreed.
<RP>
kanavin: I also tried removing that code and it doesn't seem to help. Have to wonder where warnings are going :/
florian has joined #yocto
<kergoth>
RP: ah, that code wasn't intended to suppress them at all, except for particular cases, but to redirect to bitbake logging. apparently it doesn't work anymore, though.
<kergoth>
wonder why it isn't working
<RP>
kergoth: right, there is something else going on too :/
<kergoth>
looks like the one we'd definitely want to drop would be the suppression of <string>, but we'd have to fix those warnings to give real actual useful info about where it occurred.. but clearly there';s a secondary issue with their flow into the loggers
mtudan has joined #yocto
<RP>
kergoth: right, we need a plumber :)
<kanavin>
RP: the warnings need to be enabled *really* early in the code path. Like, literally on the second line of bitbake executable.
<kanavin>
RP: I have patches
<RP>
kanavin: ah, that might overlap with a patch I just sent :)
<RP>
kanavin: I'm sure you'd approve of the other two though
<kanavin>
RP: I'd like to weed out all fallout from 3.10 though before I send them
<moto-timo>
👏
<RP>
kanavin: fixing them sounds great, fine by me :)
<RP>
kanavin: see my 3/3 bitbake patch, I think we may need the warnings tweak on a few other of our scripts
<kanavin>
RP: the problematic part was that python only shows warnings in interactive sessions, unless you explicitly configure it otherwise. And it needs to happen before any other imports if you want to see deprecation warnings for the imports.
<kanavin>
so kergoth's tweaks were simply too late, and probably not correct with modern python anyway
<RP>
kanavin: right, we should perhaps have some boilerplate sending them to our logger though?
<kanavin>
RP: I left that in
<RP>
kanavin: I'm thinking of bitbake-server and bitbake-worker scripts
<kanavin>
basically anything executable has to have it
<RP>
kanavin: right, so we probably need this in bin/*
<kanavin>
in other news, openssl 3 is not a smooth ride either
<RP>
kanavin: that isn't surprising! :)
<kanavin>
rust needs to be updated, some components use Werror and kaboom on deprecation warnings that become errors
<kanavin>
rust 1.55 that has openssl 3 support came out 6 hours ago :)
<kanavin>
living on the edge
<RP>
kanavin: keeps things interesting I guess :)
<RP>
kanavin: I'll be surprised if those are the biggest issues to be honest
<kanavin>
RP: I ran bitbake -k world for qemux86 (so it wouldn't match the host exactly) - the fail list so far is cargo-native, ruby-native, subversion, cryptodev-tests, python3-pycairo
<RP>
kanavin: could definitely be worse!
Guest56 has joined #yocto
<Guest56>
when I see message like this, does it mean that it the re try was good?
<Guest56>
WARNING: bzip2-native-1.0.6-r5 do_fetch: Checksum mismatch for local file /home/John_Bates/yocto/bin-npm/yocto/downloads/bzip2-1.0.6.tar.gz
<Guest56>
Cleaning and trying again.
<Guest56>
WARNING: bzip2-native-1.0.6-r5 do_fetch: Renaming /home/John_Bates/yocto/bin-npm/yocto/downloads/bzip2-1.0.6.tar.gz to /home/John_Bates/yocto/bin-npm/yocto/downloads/bzip2-1.0.6.tar.gz_bad-checksum_2ae3e6c958d83d2855a37f13fe29fd95
<Guest56>
WARNING: bzip2-native-1.0.6-r5 do_fetch: Checksum failure encountered with download of http://www.bzip.org/1.0.6/bzip2-1.0.6.tar.gz - will attempt other sources if available
<RP>
Guest56: if the build didn;t later fail, probably
<Guest56>
RP that is what I was thinking too. I never see it fail or see it try again, so I didn't know for sure.
<RP>
Guest56: those do fail the build if they don't resolve FWIW, I've seen that
jmiehe has joined #yocto
<RP>
JPEW: around? Just talking with halstead about what we need to do to setup a read-only prserv port
<RP>
JPEW: Looking at the code do you use the same DB with two servers, one rw, one ro?
<RP>
smurray: or you may know the answer?
<JPEW>
If it's like the hashequiv server, that would work, but I didn't look too close at any of the prserv specific parts
<JPEW>
So I don't know for sure
<RP>
JPEW: sorry, I do mean hashequiv
mtudan has quit [Quit: Client closed]
<halstead>
JPEW: Alright. It does need to run as two separate servers though?
<RP>
JPEW: is that how is is designed to work, on DB, two servers, one for each port (for hashequiv)
<JPEW>
Let me check real quick to make sure
<smurray>
RP: AIUI, PR serv is the same, run 2x servers against the same db file, with one given the read-only option
camus1 has joined #yocto
* RP
just wants to make sure he suggests halstead do the right thing!
camus has quit [Read error: Connection reset by peer]
camus1 is now known as camus
<JPEW>
halstead, RP: Yes run 2 servers. One R/W and one R/O
<RP>
JPEW: thanks!
<halstead>
JPEW: I'll give it a go.
<halstead>
RP: I'll update bitbake during the maintenance and get two servers up.
<RP>
halstead: thanks. This would make our sstate usable again so could be a big win
florian has quit [Ping timeout: 260 seconds]
eduardas has quit [Quit: Konversation terminated!]
goliath has joined #yocto
<RP>
khem: there are some hashequiv warnings in your meta-oe build but we're poking at server bits so don't worry about them ;-)
xmn has joined #yocto
leon-anavi has quit [Quit: Leaving]
wwilly has joined #yocto
wwilly has quit [Ping timeout: 252 seconds]
wwilly has joined #yocto
wwilly has quit [Read error: Connection reset by peer]
wwilly has joined #yocto
yates has quit [Remote host closed the connection]
wwilly has quit [Ping timeout: 252 seconds]
jwillikers has quit [Remote host closed the connection]