<f00b4r0>
author & commiter & SoB as "noreply.github.com" ?!
wenger has quit [Ping timeout: 480 seconds]
<f00b4r0>
(and a malformed commit message - what else could go wrong)
fakuivan has joined #openwrt-devel
<robimarko>
Yeah, my fault
<robimarko>
Fix was correct and I did not inspect the description and SoB
<robimarko>
Will definitively pay more attention
<f00b4r0>
but it looks like it was committed by GH. Don't we have CI tests for this?
<robimarko>
Seems we dont as only CI failure was bulding of a mediatek subtarget which was not that PR-s fault
<f00b4r0>
(don't want to throw a stone at anyone, just make sure we catch these efficiently)
<f00b4r0>
it's not terribad since it's not a code commit
<f00b4r0>
robimarko: did you use some script to commit this? Because it seems really odd that the committer is set to Github: if scripted, this should probably be fixed
<dwfreed>
it's the commit directly from the PR
<dwfreed>
the person who made the PR did the commit from the github web interface
<robimarko>
f00b4r0: Yes, I used the github-merge-pr but that one doesnt change the commiter
<f00b4r0>
robimarko: sounds like a bug
<f00b4r0>
dwfreed: i see
<robimarko>
Do you want me to revert the commit?
<f00b4r0>
no no, who am i to suggest this :)
<robimarko>
Asking in general, as I introduced the issue
<f00b4r0>
just I think we want to make sure this cannot happen again, because it would be pretty nasty if it were copyrigthtable code
<f00b4r0>
the commit script should definitely update the committer imho
<robimarko>
Well, it usually does
<robimarko>
Author gets preserved while commiter is the actual commiter
<f00b4r0>
which would make sense :)
<robimarko>
So, this is some GH trickery being doe if such SoB is used
<robimarko>
Which I do not like
<f00b4r0>
heh. Clearly there's a problem with the tools because that particular commit has *all* the things we want to avoid ;)
<robimarko>
Well, it seems that for whatever reason on this PR formal checks were never ran
<stintel>
if I apply a commit locally I'm the committer
<robimarko>
Well, that is the thing I have no clue
<robimarko>
As the github-merge-pr script just pushes it via git
<robimarko>
I have used it so far without issues like this
<stintel>
I guess I can extend the check-commits hook
<robimarko>
I will poke at ansuel to see why formal checks were not ran on that PR
<jow>
I'd revert that commit
<dwfreed>
I mean, reverting it doesn't really change anything, it's still in git history permanently; it's also 1 trivial line, which is very much not copyrightable
<Piraty>
hwdata seems to be the correct term, thanks for hinting this
<Piraty>
both are related
rua has joined #openwrt-devel
<Piraty>
judging from both bootlogs, this device has 4core cpu and 256M RAM ... https://openwrt.org/toh/yuncore/ax820 . i'm about to fix the data, objections?
<stintel>
1.5MB binary size for release build with size optimizations
<colo>
yay for static linking, right?
<stintel>
ah well it is what it is
<stintel>
most modern embedded devices have decent flash sizes anyway
<dhewg>
stintel: hah, nice!
<stintel>
right now 2 relays communicate over WebSocket with the raw mNDS UDP packet JSON serialized, but I plan to replace that with binary format - dunno if that will reduce size
<stintel>
I should do some cleanup, write a README, and publish it on github/codeberg
<stintel>
but later, need to pick up a friend from the airport, might be a couple of days before I get to that