frgo has quit [Read error: Connection reset by peer]
frgo has joined #yocto
savolla has quit [Ping timeout: 265 seconds]
alperak has joined #yocto
florian has quit [Ping timeout: 276 seconds]
Guest61 has joined #yocto
Guest61 has quit [Quit: Client closed]
Guest60 has joined #yocto
savolla has joined #yocto
savolla has quit [Quit: WeeChat 4.6.3]
savolla has joined #yocto
<Guest60>
Hi, I have a question about Yocto best practice. Let's imagine I have a machine and want to write a kernel recipe. I'm thinking of using override(:mymachine) for variables and adding “mymachine” to COMPATIBLE_MACHINE.
<Guest60>
This recipe supports a single machine(mymachine) at the moment, but if a machine from the same family or company comes in the future, I think it can use the same recipe by adding “newmachine” to COMPTABLE_MACHINE and variables override(:newmachine). Is this the right approach or ?
florian has joined #yocto
<rburton>
if you set COMPATIBLE_MACHINE in the recipe and its just one machine then you don't need to use overrides at all
<Guest60>
I was just thinking of using it to prevent the user from using this recipe by mistake or something, and to let them know that the recipe currently supporting this machine.
<rburton>
they can't if COMPATIBLE_MACHINE is set
<Guest60>
I know, what I mean is that when a new machine comes, compatible host will be updated like COMPATIBLE_HOST = “mymachine and newmachine”. I will support it by using overrides for newmachine as well as mymachine in the recipe.
<rburton>
sure
<rburton>
have a look at linux-yocto for an example
<rburton>
or better yet, just use that instead of your own kernel :)
<Guest60>
Is that the wrong or unnecessary approach?
<Guest60>
what I'm trying to do?
<Guest60>
Looking at linux-yocto I see that this isn't wrong or unnecessary :)
<Guest60>
rburton Thanks!
Guest60 has quit [Quit: Client closed]
ablu has quit [Ping timeout: 265 seconds]
ablu has joined #yocto
Guest74 has joined #yocto
<Guest74>
Hi everyone, Can anyone please tell me how can I generate debug sybols within yocto? I have added EXTRA_IMAGE_FEATURES ?= "tools-debug debug-tweaks dbg-pkgs" in local.conf and app-dbg to get installed. Still no success in having images with debug symbols under /packages-split/. I am using scarthgap. I'll appreciate any help
<rburton>
Guest74: the symbols are already split out by the time they get to packages-split: the binaries are stripped and the symbols are in the -dbg packages
<kanavin>
Guest74, packages-split/ does not contain images, it contains content of packages for a recipe
<rburton>
don't worry about the build tree, if you've installed app-dbg in the image then gdb will find the symbols
<Guest74>
rburton kanavin there should be size difference between normal and debugged binary. I see no difference. So don't really comprehend
<kanavin>
Guest74, debug symbols are split into a separate file, so the original binary is the same
<kanavin>
(and they are also packaged separately, into app-dbg package)
<Guest74>
kanavin so what should be injected to gdb so it can work. the main binary doesn't help as gdb says it has no symbols
<rburton>
is that gdb on target?
<Guest74>
rburton yes
<rburton>
triple check that you've installed the yourapp-dbg package into the image
<Guest74>
I have in my package class
<Guest74>
app \
<Guest74>
app-dbg
<rburton>
verify that app-dbg actually contains something like /usr/bin/.debug/app
<rburton>
if it doesn't then your recipe is broken
<Guest74>
rburton ohhhh yes it's there however hidden. Thank you very much
<rburton>
gdb should be following links automatically and it will just work if the file is actually installed
<Guest74>
rburton yes that works. Is there a flag which builds the symbols for all binaries regardless of -dbg package added to image extra
olani_ has joined #yocto
<rburton>
Guest74: debug-pkgs in IMAGE_FEATURES will pull every debug package into the image
<kanavin>
Guest74, there's also support for debuginfod, which is better in several ways
<kanavin>
no need to bloat the image with debug symbols, gdb will fetch them from the build host
<kanavin>
so far both requests have been assaulted by bots, noting 'missing' things in cryptic ways I am not sure how to address :-/
Tosik has joined #yocto
Guest74 has quit [Ping timeout: 272 seconds]
<Tosik>
Hello. I use Yocto "scarthgap" and trying to include protobuf into SDK. I added TOOLCHAIN_HOST_TASK:append = " nativesdk-protobuf" to local.conf. After build and SDK generation there are protobuf libs in target folder, but no protoc in host folder.
<KanjiMonster>
kanavin: I guess the missing commitref bot is complaning that your backported commits do not contain the original commit's hash in whatever format they expect in the commit message
olani_ has quit [Ping timeout: 244 seconds]
<kanavin>
KanjiMonster, I get that they want some kind of reference, it's just unclear what I should put there.
<RP>
kanavin: That does look confusing :/
<rburton>
Tosik: did you mean to add nativesdk-protobuf-c?
<rburton>
Tosik: ah no you meant nativesdk-protobuf-compiler
vthor has joined #yocto
vthor has quit [Changing host]
vthor has joined #yocto
pidge has joined #yocto
<Tosik>
rburton: I want to have "protoc" in my HOST SDK folder. So, user can generated .cpp files from *.proto files.
<Tosik>
rburton: I found old thread here https://github.com/openembedded/meta-openembedded/issues/628 but this change is already in "meta-openembedded/meta-oe/recipes-devtools/protobuf/protobuf_4.25.3.bb" file. I also tried to remove "-Dprotobuf_BUILD_PROTOC_BINARIES=OFF" in this file, then I got protoc, but in TARGET folder :), not in HOST.
<rburton>
Tosik: which is why i said you want to add nativesdk-protobuf-compiler, because that's the package the compiler is in. (in walnascar, too busy to dig into scarthgap)
<zeddii>
-tiny absolutely is not supposed to be defconfig
<zeddii>
tiny shuts absolutely everything off, and explicity enables what it needs. otherwise, it isn't so tiny. it is just "the default"
bjdooks has joined #yocto
Vonter has quit [Ping timeout: 276 seconds]
<alperak>
I think I misrepresented myself, I'm talking about how the variable is handled, not what it does.
<alperak>
In kernel-yocto.bbclass(line 470), the shell case statement uses the "*alldefconfig" pattern. Here, the asterisk (*) acts as a wildcard and matches any characters before "alldefconfig".
<alperak>
Therefore, any value ending with "alldefconfig" including non standard forms like "--alldefconfig" or "cia-glow-in-the-dark--alldefconfig" will be matched and processed if i'm not wrong.
<alperak>
Wouldn't this cause confusion for the user?
<zeddii>
I don't see any issues with it. There are really only the two valid options. We just have a default if one isnt' recognized, and if someone tries to use such a bizarre value in that variable, there's only so much we can do to help them.
bjdooks has joined #yocto
sudip_ has joined #yocto
bjdooks has quit [Remote host closed the connection]
bjdooks has joined #yocto
<zeddii>
when I wrote the code originally, the flag was passed directly (hence the --), but there was some discussion about having some translation, hence the variants with no dashes. And I tweaked the case statement to handle both ways.
Xagen has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
Xagen has joined #yocto
jpuhlman has quit [Read error: Connection reset by peer]
jpuhlman- has joined #yocto
Xagen has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<fullstop>
is there any way to modify / append the debugfs_script when building an image with wic?
<fullstop>
I have a few directories in there which have content excluded with the rootfs plugin, but the base directory remains.
<fullstop>
The timestamp in a reproducible build is not applied to those directories and it causes more data to be sent than I'd like when doing delta updates.
<fullstop>
For example, /boot is excluded since it contains a kernel, etc, and I don't want that in the root filesystem. This is the end result, though: https://pastebin.com/raw/0nkA48cF
<fullstop>
mtime is set to whatever time it did that operation. It is not set to the fixed value like most of the other directories.
<fullstop>
I mean, I can modify partition.py but that's hardly a long-term fix.
jmd has quit [Remote host closed the connection]
druppy has joined #yocto
eduter has quit [Quit: Client closed]
druppy has quit [Ping timeout: 272 seconds]
alperak has quit [Quit: Connection closed for inactivity]