<schmars[m]>
there's something old in your tree. maybe clean up build artifacts
<mirko>
everything builds fine - it's just the openwrt dependency checks among the package's metadata in the end
<mirko>
schmars[m]: what artifacts, except old package build trees and some data in tmp?
<schmars[m]>
if you wanna debug it, grep everything for "libubox20230523" to figure out where in the build process the outdated dependency is coming from. otherwise i'd do `git clean -fdx` and start fresh
<schmars[m]>
(something's outdated/old and not being cleared up automatically, as evidenced by opkg still being in use)
<mirko>
i see those old strings e.g. for libubox in build_dir/target-aarch64_cortex-a53_musl/root-qualcommax/usr/lib/opkg/status and build_dir/target-aarch64_cortex-a53_musl/root-qualcommax/usr/lib/opkg/info/libuci20130104.control
<mirko>
point is, they're obviously being created - as is the whole root-$ARCH dir for each run
<robimarko>
Yeah, only if OPKG is being used
<mirko>
purgin bin/ helped - didn't know/expect metadata from their is being reused
<schmars[m]>
aren't packages installed from there? :-)
<robimarko>
Yeah, its like OPKG is still being used
<mirko>
it definitely is, as - as said - I disabled APK support the time being
<robimarko>
Ok, then it just looks like a dirty build
* mirko
wonders why packages are not being re-created despite obviously outdated
<mirko>
packages as in ipkg-archives under bin/
<schmars[m]>
ah got it. anyway it's good to start fresh often - i know it's annoying to full build regularly, but if you want fast builds there's also the sdk+imagebuilder (i use them a lot, daily). full build is just so big that weird heisen-issues pop up fast. every day someone in chat, myself included, with some issue that's fortunately fixed by a fresh start
<robimarko>
BTW, I am wondering what issue did you see with APK to make you default to OPKG?
<schmars[m]>
and main branch right after branching off a new release is especially prone for heisen-issues, much churn with big changes such as opkg->apk
rsalvaterra_ has joined #openwrt-devel
rsalvaterra_ is now known as rsalvaterra
<schmars[m]>
here's someone 1.5y ago with the same libubox20230523 issue, did you not clean your checkout for >1y? :D
sharpasacrayon has quit [Ping timeout: 480 seconds]
<soxrok2212>
yeah ive been following it, maybe ill post there too
sharpasacrayon has joined #openwrt-devel
sharpasacrayon has quit [Remote host closed the connection]
Stat_headcrabbed has joined #openwrt-devel
Stat_headcrabbed has quit [Remote host closed the connection]
dick has joined #openwrt-devel
<totkeks>
anyone knows where fw4 gets the valid values for icmp_types for? i'm getting Section allow_mld (Allow-MLD) option 'icmp_type' specifies invalid value 'mld-listener-query' from the script, even though nftables has no issue with it