- Oct 29, 2024
-
-
Harald Sitter authored
-
Justin authored
-
- Oct 27, 2024
-
-
Justin authored
-
Justin authored
-
Lasath Fernando authored
-
- Oct 26, 2024
-
-
Harald Sitter authored
helps not ending up with corrupt data somehow
-
Harald Sitter authored
we might also consider throwing out manpages. half the time I don't find the right page anyway ^^
-
Harald Sitter authored
we currently have a 768M metadata section, it's got plenty of spare space so let's roll with it for now Metadata, single: total=768.00MiB, used=491.25MiB
-
Lasath Fernando authored
-
- Oct 25, 2024
-
-
Nate Graham authored
As of 0e5ad095c37bb8a4bebe387ed48ae2dbc06bb741 in sysadmin-repo-metadata, at least. So it will get pulled in automatically.
-
Nate Graham authored
Okular, Ark, and VLC. These should constitute enough to have a basic set of apps necessary for normal people to use a computer. Once Dragon Player is on Flathub, we can consider shipping that instead of VLC. Fixes #50
-
Nate Graham authored
We try to follow https://community.kde.org/Distributions/Packaging_Recommendations, which says to pre-install optional dependencies. This adds the user-facing ones.
-
Nate Graham authored
Except for Partition Manager, Spectacle, and XWaylandVideoBridge — which are not yet on Flathub.
-
Nate Graham authored
Part of #50
-
Harald Sitter authored
debugging wasn't very conclusive. the metadata is almost full and the data itself is also almost full let's try bumping the space a bit and see what happens
-
Harald Sitter authored
hopefully that will shed some light on ENOSPC issues
-
- Oct 24, 2024
-
-
Harald Sitter authored
we currently use avahi for it by necessity. applications are generally using libavahi so they implicitly force the use of avahi-daemon. I do have a prototype that bridges the avahi-daemon dbus api to systemd so we don't need the separate daemon, but it is far from finished
-
Lasath Fernando authored
-
Nate Graham authored
Right now our image is basically the prototype of the Testing edition, and for this edition, asserts make sense since the goal is to find issues early. So let's turn 'em on! Fixes #61
-
- Oct 21, 2024
-
-
Harald Sitter authored
in an attempt to fix the build. their server appears to be down right now -.-
-
Harald Sitter authored
-
Harald Sitter authored
last I heared glibc is able to work without the precompiled archive. let's see if that is really the case
-
Harald Sitter authored
should be a list and the key name was changed as well
-
Harald Sitter authored
-
Harald Sitter authored
the socket activates the daemon
-
Justin authored
-
Harald Sitter authored
this should be used everywhere but master on the main repo
-
Justin authored
-
- Oct 20, 2024
-
-
Justin authored
-
- Oct 17, 2024
-
-
Harald Sitter authored
this should help us eventually force meson to strip as well even though kde-builder currently doesn't support it
-
Harald Sitter authored
-
Harald Sitter authored
this is slightly hacky because kde-builder doesn't support running arbitrary targets. we first override ninja with a special program that turns install commands into install/strip commands we then set -DCMAKE_STRIP to a special strip command that invokes eu-strip with the right arguments to strip stuff to /tmp/debugroot from there we then pick it up in the archive generator
-
Justin authored
-
Harald Sitter authored
This reverts commit 5d37d0a8. So, while trying to set this up on KDE infra I hit the rather awkward problem that aurto doesn't really want to be automated. It's very much meant to be operated by a human. That is really just a source for failure. If we want to go down this route we'll want to use aurutils directly and write some higher level tooling on top if necessary. Undoing this for now so we can ponder it some more. Realistically I'd rather explore factoring out the rootfs build into its own job though. That way we can isolate all the binary builds into a job that produces a tarball which can then be consumed to render an image without having to build anything. Seems to me that should rock the boat a bit less while moving in the general direction of splitting things out.
-
- Oct 16, 2024
-
-
Lasath Fernando authored
-
Lasath Fernando authored
-
- Oct 15, 2024
-
-
Harald Sitter authored
doesn't work. we'll need to come up with a better idea
-
Harald Sitter authored
on the base image we have no git so install it first, then try to reset the git permissions
-
Harald Sitter authored
tidier set up and hopefully unbreaks the pipeline
-
Harald Sitter authored
It would appear that invent mangles files permissions so that we end up with world writable files. Instead try to hold on to git. This may ultimately cause trouble with docker but we'll have to try to know. Relates to #27
-