<dhewg>
blogic: fresh build for my one from ~5min ago fails to bring up wifi at all, same config as yesterday, haven't investigated at all, but that null mac line catches the eye? https://paste.debian.net/plainh/e3313da7
<dhewg>
ynezz: hey, I just used rz/sz yesterday :P I consider it essential on all my routers/sbcs. There was no network set up at all, and if you have a serial connection anyway that's the most easy way to transfer stuff
<ynezz>
yeah, many bootloaders are common use case as well
<ynezz>
just reading through the code from the '90s is like museum experience :P
<dhewg>
and combined with picocom it's a breeze, it's just C-r or C-s
<ynezz>
yep, using the same combination right now for EFR32MG flashing
<ynezz>
(EFR32MG is that Silabs MCU being widely used for Thread/Matter, home automation over Zigbee)
<dhewg>
and thanks for that PR, you may have noticed my comment of the prior hack, I just can't touch autohell stuff anymore for sanity/nerdrage reasons :D
<ynezz>
yep, its hell, I first tried to fix it cleanly, like using patches from Debian/Gentoo, but it was still failing to build, thus I give up and simply patched it out, should be good for next 20 years or so :P
<dhewg>
hm, I don't think I have any EFR32MG devices, just some cheap zigbee and esp/wifi ones
<dhewg>
comparing the generated supplicant conf with another node: max_oper_chwidth is 1, the other has 2, and beacon_int is in the global scope, the other node has it in the network{} scope
<robimarko>
/external-toolchain/openwrt-toolchain-malta-be_gcc-13.3.0_musl.Linux-x86_64/toolchain-mips_24kc_gcc-13.3.0_musl/bin/../lib/gcc/mips-openwrt-linux-musl/13.3.0/../../../../mips-openwrt-linux-musl/bin/ld.bfd: /__w/openwrt/openwrt/openwrt/tmp/cc8BABDn.ltrans4.ltrans.o: in function `uc_wpa_rkh_derive_key':
<robimarko>
<artificial>:(.text+0x1e7c): undefined reference to `hmac_sha256_kdf'
lucenera has quit [Read error: Connection reset by peer]
lucenera has joined #openwrt-devel
<hurricos_>
stintel: hey!
nixuser has quit [Read error: Connection reset by peer]
nixuser has joined #openwrt-devel
<mrnuke>
hurricos_: I would have responded to #42 if I had been given the time. I disagree with it :(
<hurricos_>
mrnuke: reverted.
<hurricos_>
I'm still very glad I filed and then reversed the commit. It puts on the record that blogic was unhappy and was uninterested in the context.
<mrnuke>
hurricos_: everyone, including blogic, agrees that a uci-defaults script is a good solution. What I don't get is blogic's insistance that this be carried by the realtek-poe upstream instead of the creators of the JSON format.
<hurricos_>
mrnuke: I think the sword to cut the Gordian Knot is probably to place the uci-defaults in openwrt/package.
<mrnuke>
hurricos_: I agree with you on that
<mrnuke>
feel free to tag 2df137e ("realtek-poe/openwrt: Generate config from board.json") if you think that having something rushed for 24.10 may work
<mrnuke>
that way packages/ feed gets the uci-defaults for 24.10, and we figure out the final solution later
<hurricos_>
I'm just gonna submit a PR to openwt/packages instead, and ask for the prior PR to be closed.
<mrnuke>
hurricos_: thank you.
<hurricos_>
I figured I'd just put a knife to the baby and see who came :^)
<mrnuke>
blogic: if there are good arguments to having the uci-defualts script on realtek-poe upstream, I am open to change my view. To me, it seems it's just one preference over the other.
<mrnuke>
hurricos_: BTW, there's a commit in the repo to fix the github pipelines failing: d2d1e3d (".github/workflows: Appease the github gods and use "v4"")
<mrnuke>
Do you want me to PR it?
<hurricos_>
Oh, please do
<hurricos_>
I've been wondering about that one.
<hurricos_>
I spent far longer than I should have staring at the page that the build failure linked, about the v2 deprecation :^)
<hurricos_>
Can we test CI without merging it? I think Github might allow it. PR and test. If the CI succeeds, merge.