<schmars[m]>
scripts/ipkg-build: line 150: 16108 Segmentation fault mkdir "$tmp_dir" -- i swear i've seen this before at least twice and knew the cause, but what was it? anyone remember?
minimal has quit [Quit: Leaving]
bluew has quit [Remote host closed the connection]
<bluecmd[m]>
Hello, forgive me for nagging, but I'd be really happy for some maintainer review feedback on https://github.com/openwrt/openwrt/pull/13039 - and even happier for a merge ;)
<bluecmd[m]>
(It's Cisco vEdge 1000 support)
danitool has quit [Quit: Cubum autem in duos cubos, aut quadratoquadratum in duos quadratoquadratos]
<KanjiMonster>
bluecmd[m]: gave you a few superficial comments
<Mangix>
rmilecki: ah ok. Didn't know it was currently read only.
<bluecmd[m]>
KanjiMonster: Much appreciated!
Daanct12 has quit [Remote host closed the connection]
<bluecmd[m]>
KanjiMonster: "this doesn't match the compatible of your dts, does sysupgrade even accept the image?" yep it does, the only place BOARD_NAME seems to be used is in the filename for the upgrade files *shrug*
<bluecmd[m]>
As for the "vendor,model" - would you go with "imt,n821" for "Interface Master - Niagara 821", "viptela,vedge1000" for the original product, or "cisco,vedge1000" for the latest product?
<bluecmd[m]>
n821 is what the board inside is called - printed in the silk screen
<Slimey>
oh while im thinking about it https://github.com/openwrt/openwrt/pull/12850 how do what hauke wants me to? "The follow up commits are all just fixing problems in the initial commit, please squash them all together and force push them into this pull request. While at it please add line breaks after 72 characters into the commit description of the initial commit."
<bluecmd[m]>
That fits with the other theme as well of being "ubnt" and e.g. "edgerouter"
<PaulFertser>
Slimey: it means what it says, which part raises your questions?
<Slimey>
squashing them all together
<KanjiMonster>
Slimey: "git rebase -i <base>" allows you merging (squashing) commits and updating their commmit messages. <base> is the commit after which you want to edit commits, so usually origin/main
<Znevna>
"Installing the sysupgrade. Do not unpower device!" :P
<Slimey>
ah thanks KanjiMonster, what about the u-boot commands that are longer
<Slimey>
Znevna never
<Znevna>
I think better wording can be found
<Znevna>
lol
<KanjiMonster>
Slimey: commands, urls and (serial) output / error messages are usually fine to exceed the 72 characters
<Slimey>
k
goliath has quit [Quit: SIGSEGV]
<bluecmd[m]>
KanjiMonster: I am refactoring to use the `octeon_3xxx.dtsi` now, and one of the reasons I didn't do that before is that it has not seen so much love over the years - DTC complains a lot (e.g. the SOC should be `soc {` these days not `soc@0 {`). I assume you're fine with those warnings? Also, I will have to use a fair amount of `/delete-node/` and `/delete-property/`. I see those are used in OpenWrt in other places so should be good.
<bluecmd[m]>
Right - but fixing the dtsi might break all the other DTSes all over the world including them!
<bluecmd[m]>
So it's like mission impossible to fix. And if I fix them and break e.g. openwrt people will be pissed I broke something that has no ABI guarantee anyway :P
<KanjiMonster>
octeon was one of the first devicetree users in mips, so it's no wonder that there is a lot of cruft in there. most of the things it did weren't formalized yet