| Age | Commit message (Collapse) | Author |
|
The previous change enabled cross-compilation for the early variant,
this duplicates the change for the standard variant.
Signed-off-by: Ophestra <cat@gensokyo.uk>
|
|
Running this via qemu requires extensive patching, so disable them
for now.
Signed-off-by: Ophestra <cat@gensokyo.uk>
|
|
Useful during testing and debugging to leave / writable.
Signed-off-by: Ophestra <cat@gensokyo.uk>
|
|
Rosa OS does not have the concept of unix user accounts.
Signed-off-by: Ophestra <cat@gensokyo.uk>
|
|
This prevents the kernel from attempting to write to the device, even
though squashfs itself is read-only.
Signed-off-by: Ophestra <cat@gensokyo.uk>
|
|
Useful for sharing a directory during integration testing.
Signed-off-by: Ophestra <cat@gensokyo.uk>
|
|
This change also adds a missing doc comment.
Signed-off-by: Ophestra <cat@gensokyo.uk>
|
|
This directory is not used by Rosa OS, but may be useful during testing.
Signed-off-by: Ophestra <cat@gensokyo.uk>
|
|
Useful for integration testing without rebuilding the system image.
Signed-off-by: Ophestra <cat@gensokyo.uk>
|
|
This optionally halts the boot process to run the specified program.
Signed-off-by: Ophestra <cat@gensokyo.uk>
|
|
WirePlumber 0.5.19 contains a commit fixing the permission management
vulnerability; it did not get treated as such upstream, hence the late
update and lack of CVE or even any meaningful announcement.
Signed-off-by: Ophestra <cat@gensokyo.uk>
|
|
PipeWire WirePlumber has finally updated, backport patch has been
removed. Kernel headers remain unchanged. A libarchive test flake was
encountered during rebuild, but it did not show up after multiple
retries, so keeping it enabled for now.
Signed-off-by: Ophestra <cat@gensokyo.uk>
|
|
The hardcoded name is not correct for externally registered files.
Signed-off-by: Ophestra <cat@gensokyo.uk>
|
|
This uses QEMU to run checks when supported.
Signed-off-by: Ophestra <cat@gensokyo.uk>
|
|
QEMU does not support most things worth testing, and supported tests
that invoke a binary still requires binfmt to work correctly.
Signed-off-by: Ophestra <cat@gensokyo.uk>
|
|
This is quite difficult to figure out just from reading source code.
This change documents as much of azalea as possible, excluding some
fields intended for internal use only.
Signed-off-by: Ophestra <cat@gensokyo.uk>
|
|
This defaults to nil, however it did not accept the untyped Azalea nil.
Signed-off-by: Ophestra <cat@gensokyo.uk>
|
|
This is included in the evaluation error message.
Signed-off-by: Ophestra <cat@gensokyo.uk>
|
|
This is significantly more readable and helpful than simply using
the %#v verb.
Signed-off-by: Ophestra <cat@gensokyo.uk>
|
|
This significantly helps with debugging: the %#v text is incredibly
unhelpful to read, and just ends up filling the terminal with useless
information.
Signed-off-by: Ophestra <cat@gensokyo.uk>
|
|
This was missed during migration to the new split convention.
Signed-off-by: Ophestra <cat@gensokyo.uk>
|
|
It does not seem possible to have a native llvm-config that provides
correct information.
Signed-off-by: Ophestra <cat@gensokyo.uk>
|
|
This points to the headers installed to the cross sysroot.
Signed-off-by: Ophestra <cat@gensokyo.uk>
|
|
This change also adds the --sysroot argument. CMake handles this already
with other flags, but it is a nice to have, anyway, to hopefully handle
some edge cases.
Signed-off-by: Ophestra <cat@gensokyo.uk>
|
|
Meson looks for it in host root, so move it there.
Signed-off-by: Ophestra <cat@gensokyo.uk>
|
|
The make helper overrides CC when cross-compiling.
Signed-off-by: Ophestra <cat@gensokyo.uk>
|
|
This builds the package CGO_ENABLED=0 via go:generate.
Signed-off-by: Ophestra <cat@gensokyo.uk>
|
|
This makes wayland-scanner available in the cross environment.
Signed-off-by: Ophestra <cat@gensokyo.uk>
|
|
This works around the buggy build system which tries to link against
just-built libraries not available to the linker with sysroot specified.
Signed-off-by: Ophestra <cat@gensokyo.uk>
|
|
Of course, meson fails to add standard flags like these, despite knowing
the sysroot via the mfile. This fixes libraries like libdrm.
Signed-off-by: Ophestra <cat@gensokyo.uk>
|
|
This adds a bunch of autoconf stuff. Autoconf does not support sysroot.
Signed-off-by: Ophestra <cat@gensokyo.uk>
|
|
The make helper overrides CC when cross-compiling.
Signed-off-by: Ophestra <cat@gensokyo.uk>
|
|
This is used by host autotools, so it should not be installed into the
cross sysroot.
Signed-off-by: Ophestra <cat@gensokyo.uk>
|
|
CMake is updated, causing a large amount of rebuilds. PipeWire
WirePlumber still remains unfixed, so the backport patch remains.
Signed-off-by: Ophestra <cat@gensokyo.uk>
|
|
These patches are for tests only, which are no longer run for the
bootstrap stages. This change also removes dynamic linker path override
for the final stage, and works around the test by making musl available
at the expected path in the container.
Signed-off-by: Ophestra <cat@gensokyo.uk>
|
|
Release candidates should not notify an available update.
Signed-off-by: Ophestra <cat@gensokyo.uk>
|
|
This allows the prefix to be changed, at runtime if needed.
Signed-off-by: Ophestra <cat@gensokyo.uk>
|
|
This allows expressions like string concatenation for the KV key part.
Signed-off-by: Ophestra <cat@gensokyo.uk>
|
|
Kernel headers remain unchanged. Kernel configuration was updated via
the new oldconfig helper script.
The new python release broke test_profiling. The complete lack of output
makes it difficult to troubleshoot, so it has been disabled for now.
Signed-off-by: Ophestra <cat@gensokyo.uk>
|
|
This replaces SyscallError with PathError when pathnames are involved.
Signed-off-by: Ophestra <cat@gensokyo.uk>
|
|
This package was the first somewhat large piece of code that was
primarily written using the built-in display of my laptop. The
experience had been deeply unpleasant due to the awkward 3:2 aspect
ratio not working well with my setup, which almost always involved
having the implementation and tests side-by-side. This experience is
what provoked me to re-instate the 80-column limit that I had long
resented and rejected. This package however predates the formal
re-instating of this limit, and was never updated.
Signed-off-by: Ophestra <cat@gensokyo.uk>
|
|
This makes updating kernel configuration much easier, removing the need
to manually edit the generated script.
Signed-off-by: Ophestra <cat@gensokyo.uk>
|
|
This produces linux/version.h that would have been generated for
kernel-headers before the build system overwrote it. Useful for bumping
kernel headers.
Signed-off-by: Ophestra <cat@gensokyo.uk>
|
|
Signed-off-by: Ophestra <cat@gensokyo.uk>
|
|
Signed-off-by: Ophestra <cat@gensokyo.uk>
|
|
The ppc64le target is currently cross-only due to the lack of test
constants. A future change may downgrade riscv64 to the same status.
Signed-off-by: Ophestra <cat@gensokyo.uk>
|
|
This enables ppc64le container and seccomp support.
Signed-off-by: Ophestra <cat@gensokyo.uk>
|
|
This removes the broken test preventing this target from being used.
This change also fixes the meson helper and uses linuxArch for the qemu
suffix.
Signed-off-by: Ophestra <cat@gensokyo.uk>
|
|
This enables the ppc64le target.
Signed-off-by: Ophestra <cat@gensokyo.uk>
|
|
This is different from linuxArch for PowerPC. While the alternative
usually works fine, compiling compiler-rt builtins produces no output
with linuxArch. The musl installation also uses this name.
Signed-off-by: Ophestra <cat@gensokyo.uk>
|