<f00b4r0>
incidentally, if I wanted to provide both -tls and non-tls variants of the same package, which requires a different configure flag for each case, can this be done in our current system or would that require two separate "packages" in the feeds?
<f00b4r0>
hmm zlib uses the exact same Config.in snippet style and it works there
* f00b4r0
scratches head
KGB-2 has joined #openwrt-devel
<dwfreed>
f00b4r0: build variants; see dnsmasq for an example
<f00b4r0>
dwfreed: thanks. Looks a bit intricate but I guess I can work with that :)
<Nick[m]>
sur5r: thanks, now still thinking about how to backport it to 24.10
mccurly_ has joined #openwrt-devel
<dwfreed>
set VARIANT in package definition; check BUILD_VARIANT when defining configure options
<mccurly_>
Hello everyone!
<f00b4r0>
if I read that Config.in correctly, the intent was to have TLS enabled by default. That's not the case in our current package
<mccurly_>
"Spoiler alert" :) I am noob into version control systems such as git or the like. I am running podman containers to issue local build requests for my routers. One of these routers, (xiaomi ac2350) has wifi issues in the latest stable release. There are already merged fixes for these issues, (I think), planned to go out in the next stable release. I am currently running SNAPSHOT on this xiaomi router, but it still presented the issues that I was r
<mccurly_>
the last stable one.
<mccurly_>
is it possible to use the local asu containers to build those commits?
<mccurly_>
(sorry for this ignorance)
<f00b4r0>
ha, the Makefile is missing the necessary source statement
<dwfreed>
ah, yes, the Package/foo/config block
<f00b4r0>
yep.
<f00b4r0>
let's PR that
AtomiclyCursed has quit [Quit: ZNC 1.10.0 - https://znc.in]
<dwfreed>
f00b4r0: you may also want to add PKG_CONFIG_DEPENDS := CONFIG_RADCLI_TLS
<f00b4r0>
dwfreed: what does that do, for my education?
<dwfreed>
"specifies which config options influence the build configuration and should trigger a rerun of Build/Configure on change"
<f00b4r0>
oic
<sur5r>
Nick[m]: You can just change that DEPENDS line
<Nick[m]>
sur5r: yeah, doing that. thought about maybe backporting the ipq-wifi history however this adds to much. so just adding a commit. thanks for your help :)
minimal has joined #openwrt-devel
minimal has quit [Remote host closed the connection]
mccurly_ has quit [Remote host closed the connection]
<sur5r>
That was quick :)
<f00b4r0>
didn't realize gettext builds non-parallel. That's super slow
valku has joined #openwrt-devel
<robimarko>
gettext spends most time in configure than configuration, doubt that parallel will help anything
<dwfreed>
autohell
<robimarko>
gettext is probably the worst example
<robimarko>
It has a bunch of targets and calls configure on each one and then builds like one file
<f00b4r0>
ah right, that rings a bell
<dwfreed>
a stash of autoconf cache variables would probably improve performance drastically
<dwfreed>
The one good thing is that because the openwrt build system is one ginormous invocation of make, generally you can get pretty decent parallelization across different packages
<f00b4r0>
"generally" :)
<efahl>
mccurly_: the ASU system uses imagebuilders, and thus do not do any compilation
<efahl>
mccurly_: this means you cannot add patches or modified source to the images they produce
<dwfreed>
he's gone
noltari has quit [Ping timeout: 480 seconds]
robimarko has quit [Remote host closed the connection]
noltari has joined #openwrt-devel
guerby has quit [Remote host closed the connection]
guerby has joined #openwrt-devel
n3ph_ has joined #openwrt-devel
goliath has joined #openwrt-devel
lucascastro has joined #openwrt-devel
n3ph has quit [Ping timeout: 480 seconds]
<efahl>
oh well
noltari is now known as Guest18089
noltari has joined #openwrt-devel
Guest18089 has quit [Ping timeout: 480 seconds]
AtomiclyCursed2 has quit [Quit: ZNC 1.10.0 - https://znc.in]
<mirko>
is sth like this supposed to work or prone to failure anyway, because of not being evaluated at times needed?
<mirko>
ifneq ($(CONFIG_QT_VERSION_5),)
<mirko>
PKG_BUILD_DEPENDS:=qt5tools/host
<mirko>
endif
<mirko>
if it's supposed to work, how to debug it if not working? :) qt5tools host build is not being triggered when building the package above lines are part of
decke has joined #openwrt-devel
mccurly has joined #openwrt-devel
Stat_headcrabbed has joined #openwrt-devel
<mccurly>
efahl: Hi, just caught up the irc logs and seen your answer. Ok I roger that it is not possible due to the imagebuilders (blobs?, that are precompilled), so there is no way around it then.
<mccurly>
dwfreed: also thank you for your assistance
noltari is now known as Guest18092
noltari has joined #openwrt-devel
<dwfreed>
mccurly: the very latest snapshot currently available for download should have the changes you referenced
<dwfreed>
this snapshot is new as of about 12 ish hours ago
<mccurly>
hi dwfreed thank you. how to be sure that it contains it? I am on the snapshot version for that xiaomi ac2350 router. (I used the local podman containerized asu server to install that snapshot version) but the no appearing ac wifi happened in there as well
<mccurly>
dwfreed: roger
<mccurly>
yesterday it wasn't (it was the last time I tried)
<mccurly>
:)
lucas__ has joined #openwrt-devel
<mccurly>
dwfreed: I am going to try it then
Stat_headcrabbed has quit [Quit: Stat_headcrabbed]
<mccurly>
Either it works or it doesn't. I will try it and report back. thank you dwfreed !
<efahl>
ah, but are you running from source or from the pre-build images? podman-compose.yml needs to point locally:
<efahl>
- image: "docker.io/openwrt/asu:latest"
<efahl>
+ image: "asu:latest"
<mccurly>
efahl: bit lost here. I am running from prebuilt containers alright. so should I do the trivial solutions (including that fix on podman-compose.yml) or should I pull the latest 1420 and build asu from sources?
<efahl>
yeah, sorry! ok, try using 'owut list' and then firmware selector, I think that still works ok
<mccurly>
also I am running locally but on a laptop and exposing the asu lan ip to the corresponding openwrt routers
<mccurly>
Ok I'll try that owut list command and get back soon (hopefully ;) )
<efahl>
it will show the error message "blah blah packages missing", but I think the list is good to go. make sure to check it, though!
<mccurly>
owut blob?
<efahl>
no, that's the thing that generates the json blob for a request, typically for debug
<efahl>
the 'list' subcommand is just the packages formatted for Firmware Selector
<mccurly>
efahl: using the list --format fs-user command I get the following : libpthread kmod-lib-crc-ccitt getrandom kmod-pppoe kmod-crypto-acompress libmbedtls luci-ssl kmod-ppp libuclient luci owut kmod-lib-lzo kmod-ath -nftables, would there be anything weary in here?
noltari is now known as Guest18099
noltari has joined #openwrt-devel
<mccurly>
I get the -nftables would remove the nftables firewall
<efahl>
Looks right to me, are there any "user installed" packages that appear to be missing to you?
<efahl>
The -nftables is basically a no-op due to it's dependency from firewall4 and other things.
<mccurly>
I guess not. I did not install any I think (from the top of my buzzed memory ;) )
<efahl>
If you do the build, check the stdout and you'll see 'adding nftables' or something in there.
<mccurly>
no-op?
<efahl>
no operation == does nothing
<mccurly>
ah ok
<mccurly>
lost the 'translation' ;)
<efahl>
Old assembly language jargon...
<mccurly>
no worries
<mccurly>
;)
<mccurly>
so I will try it then
<efahl>
Or just leave it off when you paste into FS package fields...
Guest18099 has quit [Ping timeout: 480 seconds]
<mccurly>
FS?
<efahl>
Firmware Selector
<mccurly>
:)
<mccurly>
Then I am old
<mccurly>
;)
<efahl>
More jargon for the dictionary.
<mccurly>
Yep
<mccurly>
I would think 'file system' .... ;)
<efahl>
Hah, I'm usually the oldest one around, my kids are in their 30s and 40s...
<mccurly>
I concur ;)
<mccurly>
But one's as old as one's feels
<mccurly>
;)
<mccurly>
This seems like garbled or what! HA!
jag has joined #openwrt-devel
<jag>
hi
jag is now known as Guest18101
Stat_headcrabbed has joined #openwrt-devel
Stat_headcrabbed has quit []
Stat_headcrabbed has joined #openwrt-devel
Stat_headcrabbed has quit []
<mccurly>
efahl: I think I might have hurt your 'savant' feelings by my jiberishing ;) And if that is the case, I excuse myself, I like people with sense of humour. I tend to think that 'seasoned' people are more likely to weather the challenges presented wherever. So I hope I did not offend you. And even if that is the case, I have offended myself as well... Sorry
<mccurly>
For what I gather Your commit is not yet in place is it?
<efahl>
Oh, not at all, I always assume the best intentions.
<mccurly>
I mean to the ASU repo at git hub
<efahl>
Correct, I'm awaiting aparcar's approval on the PR.
<mccurly>
efahl: And that is something the wise people (I guess) do
<efahl>
👍
<mccurly>
There are so many things to consider that one could be off by a lot if one lets go of those principles ;)
jakllsch_ is now known as jakllsch
<mccurly>
hi again efahl would you kindly inspect this 'before and after' implementing your 'trivial' fix onto asu/main.py 66th line? ---> http://0x0.st/8E00.txt
<mccurly>
does it seem sensible (more sensible I mean)?
<efahl>
The 500 status response means the index.json handling is still broken, must not be picking up the fix.
<efahl>
You may need to do 'podman-compose --verbose --parallel 1 build' before starting the server
Guest18101 has quit [Quit: Page closed]
<mccurly>
I'll do that, ;) you deserve it! :)
<mccurly>
be back in a while
<efahl>
Oh, maybe it's ok. I didn't see that you have two runs of owut there. Do 'owut check --verbose ...' and see what that downgrade is...
plappermaul has joined #openwrt-devel
<mccurly>
ok. just for the records, I have done that container's rebuilding and the log didn't show anything relevant, at least strikingly... --> http://0x0.st/8E0R.txt
n3ph_ has quit [Ping timeout: 480 seconds]
<efahl>
Looks like no errors, but I'm not sure about the 'Successfully tagged docker.io/openwrt/asu:latest' Does that pull the upstream images???
<mccurly>
efahl: perhaps, I believed that it would not, but since I had not changed the podman-compose.yml file to point locally most probably it got it from upstream... Sorry for the noobishness...
<mccurly>
I had many images pulled down
<efahl>
No apologies, it's arcane as hell...
<mccurly>
But I am arcane as hell ... ;)
<mccurly>
You certainly got me there!
<efahl>
I have to keep a cheatsheet for all the weird, undocumented podman stuff. Even reading the source code is often not enough.
<mccurly>
No maybe you have to keep a cheatsheet for dumb arcane guys like myself but not otherwise! :)
<mccurly>
One thing though there are 3 images that the local asu server uses should redis one be built (and used) locally or should I leave it to pull upstream?