Every successful 'make all' now:
1. BEFORE: restores cache from git if needed
2. AFTER: syncs built packages to git-tracked cache
3. AFTER: auto-commits cache to git (with fallback if not in repo)
Flow: build → cache-sync → cache-commit
Cache survives: make clean, make distclean, git clone
This makes the build system fully resilient for a fork/overlay OS.
Red Bear is a fork/overlay on top of Redox. The upstream build
system wasn't designed for forks — it loses all cached stages on
make clean with no recovery path.
This commit adds a git-tracked build cache:
- local/cache/pkgar/{pkg}/stage.pkgar — per-package cache files
- cache-sync.sh: sync built packages → git-tracked cache
- cache-sync.sh --restore: restore cache → recipe targets
- cache-sync.sh --commit: sync + git commit
- Auto-restore before build, auto-sync after build
Cache survives: make clean, make distclean, git clone, upstream rebase.
Recovery from clean: seconds (restore from git) vs hours (full rebuild).
Adds comprehensive build cache snapshot and restore for overlay OS.
Problem: Upstream Redox build system is single-stream — make clean
destroys cached stage.pkgar files permanently. Build can't recover
without full from-scratch rebuild (2-4 hours).
Solution: Red Bear cache system provides:
- snapshot-cache.sh: Save all stage.pkgar to local/cache/
- restore-cache.sh: Restore from snapshot after make clean
- Auto-restore: Makefile auto-restores cache before build
- Essential cache: Pre-built caches for boot packages tracked in git
- Cookbook fixes: Missing deps trigger rebuild instead of crash
With cache restore, make clean recovery is measured in seconds,
not hours.
Gaps fixed in cookbook:
- modified_all_btree: missing dep → UNIX_EPOCH (rebuild trigger)
- sysroot install: missing dep → skip + rebuild
Root cause: modified_all_btree used ? on missing stage.pkgar,
causing cascade failure when make clean destroyed cached builds.
Fixes:
1. dep stage.pkgar missing → UNIX_EPOCH (triggers rebuild, not crash)
2. dep stage.pkgar missing during sysroot install → skip + rebuild
Build system now recovers from make clean by rebuilding deps.
Verified in QEMU: compositor runs, no exceptions, DRM active.
Greeter reaches 'compositor ready, launching greeter UI'.
All canaries present. Boot completes to login prompt.
- apply-patches.sh: add signature-marker checks for build-system patches
to handle cases where reverse-check fails but patch is already applied
- test-baremetal.sh: auto-disable TUI when stdout is not a terminal;
pass CI=1 to make
- test-live-iso-qemu.sh: pass CI=1 via env to prevent repo cook panic
- scripts/run.sh: auto-disable TUI when stdout is not a terminal;
pass CI=1 to qemu launch
- repo.rs: improve TUI initialization error messages (raw mode + alternate
screen) and rustfmt cleanups
- config.rs: auto-detect TTY presence for TUI enablement; use is_terminal()
instead of relying solely on CI env var
Qt6ShaderTools cmake function is not available in our cross-compilation
setup. Added -C preload with no-op stub function to allow cmake
configuration to proceed past shader compilation calls.
Ultraworked with [Sisyphus](https://github.com/code-yeongyu/oh-my-openagent)
Co-authored-by: Sisyphus <clio-agent@sisyphuslabs.ai>
The sed pattern was stale — source v6.10.0 has 'REQUIRED Qml Quick Gui'
but the old pattern only matched the previous format. Fixed to remove
both Qml and Quick from find_package.
Ultraworked with [Sisyphus](https://github.com/code-yeongyu/oh-my-openagent)
Co-authored-by: Sisyphus <clio-agent@sisyphuslabs.ai>
Add three relibc patches (42 total) to close QtNetwork-critical socket gaps:
- P3-inet6-pton-ntop: AF_INET6 address parsing/formatting with RFC 5952
shorthand, IPv4-mapped suffix support
- P3-tcp-sockopt-forward: forward IPPROTO_TCP getsockopt/setsockopt to
scheme daemon instead of hitting todo_skip
- P3-dns-aaaa-getaddrinfo-ipv6: AAAA DNS record queries, lookup_host_v6,
dual-stack getaddrinfo with sockaddr_in6 entries
Also fix P3-tcp-nodelay to use sys_call_wo + from_raw_parts (const) in the
SOL_SOCKET setsockopt fallback — setsockopt sends data to the kernel, not
reads from it.
Production code fixes:
- scheme.rs: replace unwrap() after checked_mul with match binding,
eliminating a latent panic if code is reordered
- main.rs: log request context_id (PID) on request handling failure
instead of silently discarding the error
- drivers/amd/display.rs: split silent EDID read fallback into
separate match arms with log::warn diagnostics for short reads
and read failures, including byte count and connector index
Test coverage:
- gem.rs: add 4 basic tests for GemManager (create+verify,
close+verify removal, double-close error, invalid handle error)
The 17,046-line redox.patch monolith is no longer referenced in the base
recipe. All 27 individual P2 patches are now listed explicitly in
recipe.toml with symlinks to local/patches/base/.
Coverage gap closed: ixgbed/src/device.rs was the only file not covered
by any individual patch. Added P2-ixgbed-error-handling.patch for the
10GbE Intel driver error handling (println → log::info/warn/error).
Build verified: CI=1 make r.base completes successfully with the new
patch list. The monolith file is preserved as backup but no longer applied.
Analysis shows existing P0/P1 patches cover ~85% of kernel/redox.patch
(2,335 lines). Extract the two uncovered sections as new patches:
P2-redbear-os-branding.patch (65 lines): Redox->RedBear OS branding in
aarch64, riscv64, x86_shared start files + device init logging milestones.
P3-eventfd-kernel.patch (368 lines): Full EventCounter implementation
in event.rs with blocking read/write, semaphore mode, wait conditions,
and EventScheme eventfd path dispatch in scheme/event.rs.
Update desktop status doc with Wave 2 changes.
unassign_device: clear DTE and submit hardware INVALIDATE_DEVTAB_ENTRY
and INVALIDATE_INTERRUPT_TABLE commands with completion wait (was
previously only clearing the software HashMap).
TRANSLATE opcode (0x0012): walk IOMMU page tables for IOVA-to-physical
address resolution.
fstat: return proper MODE_DIR/MODE_FILE and sizes for all handle kinds
(Root, Control, Domain, Device).
Remove #TODO from recipe.toml.
Enforce that every source-tree edit must be mirrored to local/patches/
and wired into recipe.toml in the same session. Apply the policy
retroactively to userutils res/issue and res/motd (Redox → Red Bear).
relibc select_epoll() forced timeout=0 when any FD doesn't support epoll
(e.g. TTY on Redox), causing busy-loop. Poll with 100ms interval instead.
Also add stdint.h to signal/cbindgen.toml sys_includes so signalfd_siginfo
struct types (uint32_t, int32_t) resolve without build errors.
Phase 4 KDE Plasma preparation:
kded6: new recipe at local/recipes/kde/kf6-kded6/ building the KDE
daemon from source. Depends on kf6-kconfig, kf6-kcoreaddons,
kf6-kcrash, kf6-kdbusaddons, kf6-kservice — all already built.
Added to redbear-full.toml package list. D-Bus activation file
already existed; removed TODO now that recipe exists.
kirigami: rewrite from stub to real CMake build. qtdeclarative
(Qt6Quick) is now available, so the real Kirigami can be built
instead of installing dummy cmake configs and a static lib placeholder.
Added qtshadertools and qtsvg as additional dependencies.
Phase 3 input chain wiring:
udev-shim: when scheme:evdev is registered (by evdevd), probe for
event0..event7 devices and create /dev/input/eventN nodes pointing to
scheme:evdev/eventN. This bridges evdevd's evdev devices into the
/dev/input namespace that libinput and compositors expect.
libinput: remove -Dudev=false and add libudev-stub as a dependency.
The libudev-stub recipe provides libudev.so that reads from scheme:udev
(udev-shim), giving libinput a working udev enumeration path instead of
stub functions that return NULL.
Input chain is now: hardware → /scheme/input → evdevd → scheme:evdev →
udev-shim → /dev/input/eventN → libudev-stub → libinput → KWin.
Upstream relibc netdb DNS lookup has two bugs exposed by Rust 2024 edition
strict unsafe handling:
1. packet_data is moved into Box::via into_boxed_slice() but the retry
loop tries to call packet_data.as_ptr() on the moved value. Use the
already-created raw pointer packet_data_ptr instead.
2. close() is a safe function in relibc, so wrapping it in unsafe{}
triggers unused-unsafe (promoted to error by -D unused-unsafe). Remove
the unnecessary unsafe blocks around close() calls.
Patch carries in local/patches/relibc/P3-netdb-lookup-retry-fix.patch and
is applied via the relibc recipe patches list.
The #[cfg(target_os = "redox")] variant of wait_for_shutdown had dead code
after the tokio::select! block. The select already returns Result<(), _>, so
the trailing Ok(()) was unreachable and caused a type mismatch when the compiler
tried to coerce the select result into (). Remove the dead code.
Convert 14 config files from the legacy init script format (plain-text
commands) to the systemd-style TOML .service format. The init daemon
supports both formats; this eliminates the legacy path entirely so that
all services use the richer, more structured TOML unit format.
Key changes per config:
- base.toml: split 00_base into 00_base.service (tmpdir) + 00_sudo.service
(sudo daemon); remove redundant 00_drivers and 10_net (handled by
existing .service files from the base recipe)
- minimal.toml: split 30_console into 29_activate_console.service +
30_console.service + 31_debug_console.service
- desktop-minimal.toml: convert 20_display and 30_console to .service,
add 29_activate_console and 31_debug_console overrides
- x11.toml: convert 10_dbus, 10_xenv, 20_orbital, 30_console
- redoxer.toml: split 10_net into 10_smolnetd.service + 10_dhcpd.service,
convert 30_redoxer
- redbear-legacy-*.toml: update override references to .service paths
- acid.toml, auto-test.toml, os-test.toml, sys-build.toml: direct conversions
vasilito