aiyion has quit [Remote host closed the connection]
aiyion has joined #openwrt-devel
Acinonyx_ has quit [Ping timeout: 480 seconds]
Acinonyx has joined #openwrt-devel
<Namidairo>
weird, this stock dtb has "llllw" for a 4 port switch, and a 1gb memory node for a 256mb device
<stintel>
stock dtb of firebox m200 or m300 has a reference to a vitesse switch while it's a broadcom
<stintel>
don't blindly trust stock dt{b,s}
<Namidairo>
I guess I'll see what port 5 is when I boot the image
Acinonyx_ has joined #openwrt-devel
Acinonyx has quit [Ping timeout: 480 seconds]
<Slimey>
hehe
victhor has quit [Remote host closed the connection]
goliath has quit [Quit: SIGSEGV]
Larhzu has quit [Remote host closed the connection]
Larhzu has joined #openwrt-devel
valku has joined #openwrt-devel
Grommish has quit [Ping timeout: 480 seconds]
spacewrench has quit [Remote host closed the connection]
spacewrench has joined #openwrt-devel
danitool has quit [Ping timeout: 480 seconds]
Acinonyx has joined #openwrt-devel
Acinonyx_ has quit [Ping timeout: 480 seconds]
rmilecki has joined #openwrt-devel
<russm>
with the SDK, is there a way to install .ipk packages for deps of something I'm porting, or is building from source the only (suported) way?
<russm>
specifically, I'm porting something that depends on openssl. as far as I can tell the only way to get openssl in the unpacked SDK is to build it there
<russm>
which also builds all the dependencies, including the kernel
<aparcar[m]>
mangix do you compile on a intel or a M1 Mac?
floof58_ has joined #openwrt-devel
floof58 has quit [Ping timeout: 480 seconds]
danitool has joined #openwrt-devel
<Namidairo>
it took me 3 tries to get the ground pin to adhere properly without the solder joint breaking :(
<Namidairo>
now to never ever remove the cables out of fear it will break again
pmelange has joined #openwrt-devel
<mangix>
aparcar[m]: AMD
pmelange has left #openwrt-devel [#openwrt-devel]
<aparcar[m]>
mangix: and Mac?
rua has quit [Quit: Leaving.]
<mangix>
no
<mangix>
I only have an iPhone. No other Apple hardware
<PaulFertser>
Namidairo: ground pin is usually trickier than the others due to heat dissipation. Probably you should just ramp up the temperature a bit, what was it set to?
<PaulFertser>
Namidairo: flux is important too. E.g. fake rma-223 seems to be only good for very clean (out of the factory) parts.
<Namidairo>
implying I have a temperature controlled iron
<Namidairo>
i just sat it on the pad for longer this time around
<PaulFertser>
The tips are compatible with TS100 so you can get original miniware tips from banggood or wherever if you need other shapes.
<PaulFertser>
TS-BC2 is adequate for most purposes (including drag soldering and 0402 components), and I also got TS-I for space-constrained places, happy with both.
<PaulFertser>
(I have TS100, not Pinecil)
<aparcar[m]>
mangix: oh okay, I thought you're compiling on Macs from time to time
goliath has joined #openwrt-devel
aparcar has joined #openwrt-devel
<aparcar>
nbd: do you use an intel Mac too?
<PaulFertser>
aparcar: (curiosity) did you share the failure log somewhere
<nbd>
aparcar: yes
<nbd>
still waiting for my m1
<aparcar>
nbd: any interest on using this arm64 at some point?
<aparcar>
oh okay
<aparcar>
I could have brought you one from Hawaii ;)
<aparcar>
PaulFertser: one sec I'll post them in a sec
<PaulFertser>
aparcar: I see some other packages fix what seems to be a similar issue, e.g. utils/bluez/patches/202-fix-endianness.patch
<aparcar>
PaulFertser: I'll have a look thanks
mangix has quit [Ping timeout: 480 seconds]
mangix has joined #openwrt-devel
Grommish has joined #openwrt-devel
<karlp>
apk-tools and android 1 are both 2008, wouldn't really say "poorly chosen by android" jsut unfortunate overlap
<aparcar>
karlp: oh didn't know it's such a close overlap
<aparcar>
well unfortunate indeed
victhor has joined #openwrt-devel
minimal has joined #openwrt-devel
<rmilecki>
hauke: thanks for that bcm4908 compilation fix!
<Namidairo>
does uboot-envtools not work in initramfs images
<Namidairo>
oh wait it's just called fw_printenv/fw_printsys
<stintel>
anyone familiar with this message: Wed Dec 1 12:31:44 2021 kern.err kernel: [ 1975.463527] mv88e6085 0x0000000ffe4fc000:10: ATU full violation for a0:42:3f:34:3c:d1 portvec 0 spid 0
rua has joined #openwrt-devel
rua1 has joined #openwrt-devel
rua1 has quit []
rua has quit [Ping timeout: 480 seconds]
rua has joined #openwrt-devel
rua has quit [Quit: Leaving.]
<mrkiko>
On a non-clean build environment, to build successfully I need to revert the commit that updates ca-certificates, otherwise I get a Python error. Is this known?
castiel652 has joined #openwrt-devel
rua has joined #openwrt-devel
<mrkiko>
TypeError: argument 'data': 'bytearray' object cannot be converted to 'PyBytes'
Gaspare has joined #openwrt-devel
<jow>
sigh, Python's exceptional compatibility strikes again
<jow>
probably a python2 vs. python3 thing
<jow>
mrkiko: there's an open flyspray bug report
<mrkiko>
jow: File "/usr/lib/python3.9/site-packages/cryptography/x509/base.py", line 443, in load_der_x509_certificate
<mrkiko>
jow: thanks!
<jow>
... or so I thought. I swear I've seen a ca-certificates related error rport scrolling by today but can't find it anymore
<jow>
could we add a new `-vxxx` feature indicator and simply guard those usages with `if hostapd -vxxx; then ...; fi` ?
<jow>
might also help with testing vanilla versions
<nbd>
if they click upgrade, why doesn't it upgrade wpad as well?
<jow>
we do not support mass updates
<jow>
only package by package
<nbd>
oh
<jow>
Arguably this could be changed (though I think the [extended] package ecosystemis still not mature enough for that)
<jow>
but it won't solve the issue for existing 21.02.1 installations
<jow>
neither would adding adding version dependency constraints (since the yet installed wpad-whatever does not have thme)
<jow>
and we can't depend on specific wpad versions from hostapd-common easily due to the many different falvors
<nbd>
running a mix of old+new packages definitely isn't a good idea. even if the config options are processed, it will likely break in other subtle ways if netifd + hostapd isn't upgraded as well
<jow>
only the inverse would be possible (wpad-* depending on hostapd-common >= ...)
<jow>
and that would also properly prevent such situations
<nbd>
because the update solves some long standing race conditions with adding wlan devices to bridges
<nbd>
not sure how to encode that into version dependencies that prevent all stupid cases of incomplete updates
<jow>
opkg will either refuse the upgrade or upgrade the other packagesd before upgrading hostapd-common
<nbd>
but wpad-* depends on hostapd-common
<nbd>
we'd have circular dependencies then
<jow>
then at least for the future, wpad-* should depend on a specific minimum version of hostapd-common
snh has quit [Ping timeout: 480 seconds]
snh has joined #openwrt-devel
<jow>
using something like EXTRA_DEPENDS:=hostapd-common (>= $(PKG_SOURCE_DATE))
<nbd>
i think i'd probably use = instead of >=
<nbd>
does that cause them to be updated together if you click upgrade in luci?
<jow>
either that or it is refused, not sure offhand what opkg will do
<jow>
but either is fine
<nbd>
and is it fine if the same dependency appears multiple times (one time with version one time without)?
<jow>
but as I said, it will nto fix the issue at hand but will prevent similar instances in the future
<jow>
I used to do the same in luci (having both DEPENDS:=+libuci-lua and EXTRA_DEPENDS:=libuci-lua (>= 2018-01-01)) and it worked fine
<nbd>
ok
<jow>
since dependencies are ANDed
<jow>
I agree that a mixture of new hostapd and old netifd is not ideal but at least it should work long enough that people upgrade all packages eventually
<jow>
and if there's really show stupper dependencies one could still add netifd (>= ...) to EXTRA_DEPENDS for the rare cases where we absolutely require a new version (e.g. due to a new api)
<jow>
*show stopper
snh has quit [Read error: Connection reset by peer]
<jow>
(still verify the CONTROL file in the final .ipk archives, but it looks correct to me)
<nbd>
looks fine in control too
<jow>
great
<jow>
we really should consider adding versioned libc depends too eventually (Depends: ..., libc (>= xxx))
<jow>
then we can bump the minimum version whenever we do incompatible musl updates (like the recent one that introduced new time64 symbols that binaries will start to rely upon)
<jow>
or rather started to rely upon immediately, being unable to run using older libc.so versions
<jow>
anyhow, one thing at a time
rua has quit [Quit: Leaving.]
rua has joined #openwrt-devel
swalker has quit [Ping timeout: 480 seconds]
swalker has joined #openwrt-devel
<ldir>
" TypeError: argument 'data': 'bytearray' object cannot be converted to 'PyBytes'" hmmmm ca-certificates bump not great. Solved the missing python crypto package - still explodes.