32 Replies
sup gamers
can i ask you guys something? - how do you all develop readymade?
like, i dont see many commits on the repo then BAM 1 billion things at once
that would be a question for mado but usually with anything we do the brunt of the work all at once for a bunch of things then pick at the rest
itās been nicknamed ADHD driven development
Are you looking to contribute @tulip ?
if yes just pop into our serverās development chat and ask how you can help :3
for sure LOL
Bootc features. I hear no one is working on this yet š
LOL
god i love working with yāall itās always so fun
I just wish I knew rust better to make readymade faster.
rust devs: I just HATE free labour
I mean, we tried to use a faster pure rust implementation of squashfs but turns out it doesnt even read our damn squashfs images
might as well use the kernel to mount the squashfs and then copy
the i/o code is pretty much well-optimized, the bottleneck is actually the external deps it runs on lmfao
especially systemd-repart
i would rewrite the entire stack in Rust and basically kill distinst with it but meh
GitHub
Install from OCI image (bootc) Ā· Issue #44 Ā· FyraLabs/readymade
This is a tracking issue for an OCI copy backend, which would mount an OCI image and copy its contents to the destination, for bootc.
new tracking issueā¢ļø
im not kyle, but ideally on the deck image it would have to support controller navigation to be useful i think. for the desktop images however it would be great to use something like this as oobe
The app is based on GTK4 so I don't know if GTK4 has some kind of SDL navigation support
libmanette
The simple GObject game controller library.
nice so if that and a way to input text without a keyboard is made then it will 100% work on deck images.
atm we have username and password pre-filled onlyu so those handheld systems can get through the installer without a keyboard
actually this might be way out of scope for readymade rn, since it's meant to be a desktop app first and foremost
would be interesting if we used the OEM thing that ultramarine folks have instead
I'm considering refactoring the entire application into multiple libraries
never liked having that kind of stuff on the installer tbh
because rn the code is a monolithic mess of GTK and system code
gonna do it like how distinst does it
readymade is literally distinst but better š§
true, but dualboot support is very important for annoyingly many handheld users so they will need to have a way set that up, cant rely on the touchscreen working correctly on new handhelds we dont support yet.
before you edited that message i thought "ado" was ur partner or somethhing š
no she's a diva
hā¦hai (absolutely late)
that's a merge commit so it combines the previous commits together
yeye
If y'all can do this in Readymade, you'll be miles ahead of anaconda: https://bugzilla.redhat.com/show_bug.cgi?id=2326270
What a stupid limitation.
The problem here is that the /boot partition would be placed at offset larger than 2 TiB which we do not allow in Blivet. I am not sure if this limit still applies with GPT and EFI boot, but we should at least show a better error message (any possibly just a warning in blivet-gui) than the generic "Unable to allocate requested partition scheme" fot his case. yeah that probably won't affect us
