- Remove PGP signing from the script as its done in GHA
- Also remove generating .current control file as its obsolete in this design
Co-authored-by: Ricardo Pardini <ricardo@pardini.net>
- rootfs: create readonly global `rootfs_cache_id` (type+hash+date/version)
- add "oras-upload" CLI command (dumb, no retries, single target/single source)
- docker: mount-bind for & re-pass envs `GITHUB_OUTPUT` and `GITHUB_STEP_SUMMARY` down to Docker
- introduce `github-actions.sh::github_actions_add_output()`
- during logs cleanup, dump the Markdown log into GITHUB_STEP_SUMMARY if it is available
- killed `[[ $ROOTFS_TYPE != ext4 ]] && display_alert "Assuming ${BOARD} ${BRANCH} kernel supports ${ROOTFS_TYPE}" "" "wrn"`
- which definitely didn't belong in rootfs
- disable usage of run_host_command_logged_long_running for kernel-make
- stop lying about long_running stuff being any different from non-long_running versions
- stop lying about 'set -e' when there's still a bunch of pipes for "pv" and stuff all around
- require ARCH/RELEASE for rootfs cli
- disallow BOARD/LINUXFAMILY/BOARDFAMILY being set for rootfs cli
- split `prep_conf_main_only_rootfs()` instead of `prep_conf_main_build_single()` for rootfs-only stuff
- run both conf and build without stdin for rootfs CLI (no interactive allowed)
- don't blindly set `ARCH=armhf` for no reason in main-config
- allow skipping family stuff in main-config, via `allow_no_family=yes` (used only in `prep_conf_main_only_rootfs()`)
- kill double loading of userpatches family, not needed/old code
- allow skipping kernel stuff via `skip_kernel=yes` in `config_pre_main()` (used only in `prep_conf_main_only_rootfs()`)
- make `tmpfs_estimated_size` and `available_physical_memory_mib` readonly and global and only ever set once
- remove subshell from `get_rootfs_cache_list()`, new version `get_rootfs_cache_list_into_array_variable()` for sanity
- warns if `ROOT_FS_CREATE_ONLY=yes` is set during cached build path
- add a `wait_for_disk_sync()` after restoring rootfs cache
- fix logging sections for `build_rootfs_only()` and `prepare_rootfs_build_params_and_trap()` in `build_rootfs_and_image()`
- remove the logging section from `get_or_create_rootfs_cache_chroot_sdcard()` call in `build_rootfs_only()` (which is now itself logged)
- @TODO: `deploy_qemu_binary_to_chroot()` is leaking a qemu bin in rootfs cache!
- _told ya it would all make sense eventually..._
- introduce new, real, non-alias `rootfs` command in `cli-rootfs` -- this _only_ builds rootfs, nothing else
- no more `main_default_build_single()`; instead `full_build_packages_rootfs_and_image()`
- introduce `do_with_default_build()` wrapper, for things that need host prepared + aggregation to run
- CLI-specific parts of `main_default_build_single()` moved to cli-build
- put prepare/cleanup pair of `prepare_rootfs_build_params_and_trap()` and `trap_handler_cleanup_rootfs_and_image()` into new `trap-rootfs.sh` (this needs further splitting for the image part)
- refactor `get_or_create_rootfs_cache_chroot_sdcard()` into `build_rootfs_only()`, `calculate_rootfs_cache_id()`
- update library with new files
- "| cut" masked errors in multiple places
- include real pip3 version in python deps hash, all will be invalidated
- be less happy about hitting hash/cache, should happen often
- allows using pre-shallow-ed bare trees, specific to the KERNEL_MAJOR_MINOR
- uses shallow by default if storage is "slow" (mmc), or free disk space < 32g (@TODO tweak this)
- if full tree already there keeps using it
- allow forcing decision with `KERNEL_GIT=full` or `KERNEL_GIT=shallow`
- countdown to confirm if decision based on storage type/free space and user is interactive
- this has some of the worst copy ever written
- @TODO refactor the device type / free space code, reuse across codebase (3-4 other spots detect free space)
- @TODO GHA self-hosted (full) vs hosted (shallow) runners decision
- sprinkle with `wait_for_disk_sync()` around clean/download/extract
- sprinkle git with debugs, and an info between fetch and checkout (otherwise it seems "fetch" takes a long time)
- wait 10s for sync call; otherwise warn user to be patient
- wait_for_sync "reason for sync" everywhere
- add @TODO for /run/user/0 etc and XDG_RUNTIME_DIR in chroot
- rpardini: actually use ports keyring for Debian Risc-V
- `risc-v`: correctly include `debian-ports-archive-keyring` in Risc-V CLI packages
- include in debootstrap list, via `config/optional/architectures/riscv64/_config/cli/_all_distributions/main/packages`
- can't have two `--include` in debootstrap call
- riscv64: move `debian-ports-archive-keyring` from main to debootstrap; also only for `sid`
- don't show stats for patches that are bind applied blind (failed to parse/autogen)
- don't show Author, it's too long
- don't dump Python exceptions in log
- extra: expand Python imports
- reorg `display_alert()` so more frequent levels are handled first
- improve some ANSI color usage, still a mess
- don't write failed commands stackstrace as warnings to screen -- just to file (yes it's redundant, except if in test)
- more useful / subdued stacktraces, to make actual error more obvious...
- do not show path to Docker run-host's logfile, unless the Docker run itself didn't produce logs
- introduce `global_final_exit_code` to allow to maintain/preserve Docker's exit code without triggering an error host-side
- avoid using the "last resort, use lazy umount" by killing tee_pid _and descendants_ before hand. Works!
- introduce `get_descendants_of_pid_array()` using recursive trick to get all descendants
- do not `unset CURRENT_LOGFILE` -- during `logging_error_show_log()` (from SHOW_LOG=no days)
- do `unset CURRENT_LOGFILE` immediately before processing logs -- although it magically survives. mysteries of traps in bash...
- `declare -g` instead of `export` for `start_logging_section()`
- don't `check_and_close_fd_13()` in `run_cleanup_handlers()`
- do run `check_and_close_fd_13()`, but only after archiving the old logs (in logging trap!)
- add a default error for unhandled TRAP's -- of which should be done
- fix reset of `global_tee_pid=0` in section-logging
- aggregation.py:
- only include non-removed pkgs in hash (do'h)
- use a sorted unique of desktop+rootfs+debootstrap pkgs for pkg hash
- make space for having apt sources/gpg keys in hash @TODO
- write a hash-text variable for bash with info, for easy comparison later
- refactor out bash_string_multiline()
- create-cache.sh:
- don't armhf unless `ARMHF_ARCH=yes`
- remove the uninstall stuff until we decide @TODO
- don't show "free space of MOUNT" when there's only SDCARD
- better logging for resolv.conf/machine-id shenanigans does in rootfs cache
- drop dead code, squash some quoting issues
- show the size of the zstd ball after compressing
- don't even go to github armbian/cache nor cache.armbian.com for list of caches if SKIP_ARMBIAN_REPO=yes
- remove mentions of BOARD or BRANCH, not relevant in rootfs cache
- call `check_and_close_fd_13()` directly when `exit_with_error()` called, don't wait for trap
- store PID spawned by `do_with_logging()` under `$global_tee_pid`
- `check_and_close_fd_13()`:
- sprinkle with `sync` to give a chance for stuff to die gracefully
- check if `$global_tee_pid` is set and actually a running PID, if so, kill it
- do not blindly "wait" for stuff, instead, kill it and wait for it to die -- 100% more likely to not hang ;-)
- move `fakeroot_dpkg_deb_build()` from runners.sh to new utils-dpkgdeb.sh
- and make it debug show the size of the source
- clean off `.git` directory earlier for full build
- @TODO: get rid of the "hardlinks" funky biz in there and not-copy `.git` to begin with
- check and make sure the fd is closed before processing and/or deleting logs.
- change from fd 3 to 13, for extra "this is evil". is it? is 13 an evil number? you decide.
- `prepare_host()`: split; do checks earlier and allow them to be interactive
- introduce `exit_if_countdown_not_aborted()` for "Low Disk Space" and other critical conditions
- split `prepare_host()` into interactive & non-interactive parts
- split off `clean_deprecated_mountpoints()` from prepare into `cleaning.sh`
- introduce and use `reset_uid_owner_non_recursive()` for `.tmp` reset in trap, to avoid disasters
- add more logging sections to default-build.sh, avoid unlogged parts
- introduce generic `prepare_tmpfs_for()`, which manages it's own cleanup/dir removal
- use it for `WORKDIR` (which is `TMPDIR`) and `LOGDIR`
- adapt previous cleanup handlers for those, so they delete their contents but not the dir itself (which might be mounted)
- also: make `ARMBIAN_LOG_CLI_ID` readonly together with other superglobals
- set `XDG_RUNTIME_DIR` & `XDG_RUNTIME_DIR` together with `TMPDIR`
- kernel-make.sh: pass `CCACHE_TEMPDIR` down to Kernel make (thanks @the-Going)
- fix a bunch of quoting issues
- use runners for chroot stuff
- don't error out of failed disable
- makes building `sid` again possible
- add a bunch of TODO's
- consolidate at `prepare_python_and_pip()`
- sanity check for Python version 3.9+ regardless of HOSTRELEASE
- TODO: pip vs sudo/root: need pip 22.2+ to curb warning, not doing it
- .config is now compared with previous one byte-wise, diff shown, and mtime preserved directly via `cp -p`
- this does away with the previous mtime-based BS I had written during the fasthash era (and thus remove file-mtime.sh, regen lib)
- split some functions, to better control the interactive piece in the _middle_ of kernel config process
- no more `kernel_config_maybe_interactive()`, now `kernel_config()` that controls its own logging sections
- extract `kernel_determine_toolchain()` out of config into make
- introduce internal hook `armbian_kernel_config()` that runs before custom version (for .config defaults, coming soon)
- default `SHOW_LOG=yes` if user on terminal
- shortcut: `DEBUG=yes` defaults both `SHOW_LOG=yes` & `SHOW_DEBUG=yes`
- much simpler logging format for log file
- skip all display_alert()'s more verbose than DEBUG from logfile
- set `DEBUG=yes` to log everything again
- skip Markdown assets from being included in ANSI log
- use of `ccze`
- introduce internal var for controlling COMMAND-level log
- remove unused `run_on_sdcard()` that somehow was left there
- drop `export_html_logs()` and `EXPORT_HTML_LOG` -- are no more
- remove a lot of hopefully / hardly confirmed, unused dependencies
- don't use crossbuild-essential-xxx; avoid the c++ compiler that comes with it, install gcc-only instead
- hostdeps: use `libc6-dev make dpkg-dev gcc` (without `g++`) instead of `build-essential`
- drop `btrfs-progs` and `f2fs-tools` (@TODO add in extension when/whereused)
- more: drop `cryptsetup` (@TODO add in extension when/whereused)
- don't be too quiet when doing apt-update for hostdeps
- acng is constant source of headaches
- we don't need it anymore, since we have local caching now
- You can also have acng running on a different machine/container and set `APT_PROXY_ADDR=xxx:3142` if you're running a build farm
- probably breaks a lot of stuff; c++ compilers, distcc, possibly others
- split into late_prepare_host_dependencies / early_prepare_host_dependencies calling adaptative_prepare_host_dependencies() full of if's
also:
- calc deps out of configuration
- reorg checks into obtain_and_check_host_release_and_arch()
- if set, the Docker launcher will not install any dependencies
- so every Docker run will reinstall everything from scratch
- do NOT use, **even if you're Oleg**
- completely remove support for building under `buster` -- that's way too old, sorry.
- de-hardcode `python3` invocations, instead use `python3_binary_path` set by `prepare_python3_binary_for_python_tools()`
- juggle `$HOSTRELEASE`: read from actual host, or determined from Docker image name (during Dockerfile build)
- TL;DR: include and use `python3.9` for focal-like host OS's
- introduce `do_with_conditional_logging()` which only starts logging sections if `do_logging=no`
- with this we should get complete logs (ofc except for the interactive sections)
- config might run on a non-Linux, bash-capable, GNU coreutils enabled machine, so avoid `linux-version`
- use `declare -g` instead of `export` for most of `prepare_and_config_main_build_single()`
- Python patching: tune some logging all-around
- Python patching: *FINALLY* set the dates on the patched files to `max(patch_date, root_makefile_date)` thus making lighting-fast rebuilds again
- new EXTRAWIFI patch generator harness; Python patching EXTRA_PATCH_FILES_FIRST
- Python patching: use temp file for patching rejects; clean it up
- new EXTRAWIFI patch generator harness; Python: mark drivers as autogen, don't split or parse them, read as bytes, apply bytes directly
- new EXTRAWIFI patch generator harness; somewhat-works, but patches are per-family
- Python patching: add the `cache/patch` as a mountpoint
- Darwin Docker performance is 20x with namedvolume; the cached patches are very large
- Python patching: hopefully better Markdown: split series dir; detect more status/problems; arche only for missing Subject
- Python patching: archeology only for missing Subject:, not description; clarify CLI pushing
- Python patching: use `{file_base_name}.patch` for archeology as `file_name` might include a dir
- Python patching: fix: don't skip board/target patches for no reason
- Python patching: fix for series.conf patches (sunxi/sunxi64): don't sort; mark as series and show on summary
- Python patching: don't fail if archeology found no commits (but add warning)
- Python patching: CLI command `kernel-patches-to-git` with archeology summary, and git pushing of results & summary
- Python patching: patches-to-git small fixes, auto-push if it's rpardini
- Python patching: add `patch-kernel` CLI command
- Python patching: commit README.md & gh-pages workflow when apply_patches_to_git
- Python patching: hopefully better markdown
- Python patching: `git add` everything all at once, for speed
- armbian-next: fixes to ORAS-related logging
- armbian-next: actually enable ORAS-based kernel git bare tree seeding/bundles; enable cleaning of bundle articfacts after confirmed working
- armbian-next: introduce ORAS-based kernel git bare tree seeding/bundles (although it is a .tar, not a bundle); this is 20x faster than cloning
- armbian-next: introduce `ORAS` tooling; pull and push functions & downloader/launcher
- WiP: Python patching delusion, pt 1: finding & parsing patches; apply & git commit with pygit2; Markdown summaries (also for aggregation); git-to-patches tool
- Python: Markdown aggregation and patching summaries; collapsible; SummarizedMarkdownWriter
- Python: Markdown aggregation and patching summaries
- Python: reorg a bit into common/armbian_utils; define the `ASSET_LOG_BASE` in preparation for Markdown delusion
- Python patching: initial apply patches & initial commit patches to git (using pygit2)
- Python patching: add basic `series.conf` support
- Python patching: force use of utf-8; better error handling; use realpath of dirs
- Python patching: `git-to-patches` initial hack. not proud. half-reused some of the patches-to-git
- Python patching: "tag" the git commits with info for extracting later; introduce REWRITE_PATCHES/rewrite_patches_in_place
- Python patching: commented-out, recover-bad-patches hacks
- Python patching: shorten the signature
- Python patching: allow BASE_GIT_TAG as well as BASE_GIT_REVISION
- Python patching: git-archeology for patches missing descriptions; avoid UTF-8 in header/desc (not diff)
- Python patching: use modern-er email.utils.parsedate_to_datetime to parse commit date
- Python patching: unify PatchInPatchFile; better git-commiting; re-exporting patches from Git (directly)
- Python patching: switch to GitPython
- GitPython is like 100x slower than pygit2, but actually allows for date & committer
- also allows to remove untracked files before starting
- Python aggregation: fix missing `AGGREGATED_APT_SOURCES_DICT`
- Python patching: add `unidecode` dependency to pip3 install
- Python patching: don't try archeology if SRC is not a Git Repo (eg, in Docker)
- Python patching: don't try archeology if not applying patches to git
- WiP: Python patching delusion, pt2: actually use for u-boot & kernel patching
- Python patching: much better problem handling/logging; lenient with recreations (kernel)
- Python patching: don't force SHOW_LOG for u-boot patching
- Python patching: don't bomb for no reason when there are no patches to apply
- Python patching: fully (?) switch kernel patching to Python
- Python patching: more logging fixups
- Python patching: capture `kernel_git_revision` from `fetch_from_repo()`'s `checked_out_revision`
- Python patching: fully switch u-boot patching to Python
- Python aggregation/patching: colored logging; patching: always reset to git revision
- Python aggregation/patching: better logging; introduce u-boot Python patching
- Python patching pt3: recovers and better Markdown
- Python patching: detect, and rescue, `wrong_strip_level` problem; don't try to export patches that didn't apply, bitch instead
- Python patching: Markdown patching summary table, complete with emoji
- Python patching: include the problem breakdown in Markdown summary
- Python patching: sanity check against half-bare, half-mbox patches
- Python patching: try to recover from 1) bad utf-8 encoded patches; 2) bad unidiff patches; add a few sanity checks
- Python patching: try, and fail, to apply badly utf-8 encoded patches directly as bytes [reverted]
- Python patching: try to recover from patch *parse* failures; show summary; better logging
- set `GIT_ARCHEOLOGY=yes` to do archeology, default not
- armbian-next: Python `pip` dependencies handling, similar to `hostdeps`
- same scheme for Dockerfile caching
- @TODO: still using global/shared environment; should move to a dir under `cache` or some kinda venv
- WiP: add `python3-pip` to hostdeps; remove `python-setuptools`
- remove `python-setuptools` (Python2, no longer exists in Sid) from hostdeps
- add `python3-pip` to hostdeps; part of virtualenv saga
- WiP: split `kernel.sh` a bit, into `kernel-patching.sh`, `kernel-config.sh` and `kernel-make.sh`
- `advanced_patch()`: rename vars for clarity; no real changes
- Python patching: introduce FAST_ARCHEOLOGY; still trying for Markdown links
- share a single `.git` for all u-boots
- all fetches are done against it, maximizing cache hit ratio drastically
- also reduces the size of each working copy by more than 70%
- split uboot's git stuff into uboot-git.sh
- split `uboot_prepare_git()` from `compile_uboot()`
- now we've a single `.git` for all kernels with all branches; a "master" cache
- back to full stable git bundle usage from kernel.org
- bye bye, "gitballs"
- split kernel-related git stuff into `kernel-git.sh`
- introduce `ARMBIAN_MOUNTPOINTS_DEPRECATED` @TODO actually use for cleaning non-Docker
- add `axel` hostdep for multi-conn download of bundle, which comes from a CDN.
- deprecate old mountpoints for gitballs and linux-kernel
- move cleaning of old sources/kernel to mountpoints code
- introduce `if_user_on_terminal_and_not_logging_add()`;
- also the opposite, `if_user_not_on_terminal_or_is_logging_add()`
- unset `CURRENT_LOGGING_SECTION` when exiting a logging section
- introduce `do_with_logging_unless_user_terminal()`
- sharing most UEFI code, will replace the `virtual` one soon
- x86: patch uboot defconfig to use the `q35` machine type, not `i440fx`
- separate x86 bootscript, due to non-uInitrd-ness of it
- hack ramdisk load address both in u-boot source and bootscript
- use 32-bit u-boot, not 64-bit
- grub: introduce `UEFI_GRUB=skip`, does not deploy GRUB (but does the kernel packages, etc)
- auto-enable qcow2 output for these
- works with both distro's and Armbian's kernels
- cli builds won't have PPAs and other external repos
- packages from such repos should be listed in `packages.external` now, so they're not mismatched
- `PACKAGE_LIST_EXTERNAL` (packages.external) only in DESKTOP builds
- this logic may change later, example:
- we might _also_ want pure-Ubuntu desktops, without external repos
- we might _instead_ want external repos in the cli builds
- this way is easy to change, since everything is kept separate and we can change in a single place.
- introduce `SHOW_CCACHE=yes` for detailed ccache statistics and logging
- fix: kernel build is done under "env -i", so pass CCACHE_DIR down to Make if CCACHE_DIR is set...
- split from kernel.sh, show stats also for u-boot targets; show compile time
- add volume definition (under `${SRC}/cache/ccache`); auto-use that in Docker builds via `CCACHE_DIR`
- better logging for `do_with_ccache_statistics()`
- there's some CCACHE_DIR code from before; unify @TODO
- `undecided` now directly sends to `build`, which might send to `docker`, which relaunches `build` in Docker;
- the _sending to Docker_ part probably needs refactor
- DEBUG_EXTENSION_CALLS is gone. use SHOW_EXTENSIONS instead.
- LOG_ALL_HOOK_TRACES is gone. use SHOW_EXTENSIONS_TRACE instead
- LOG_ENABLE_EXTENSION is gone. use SHOW_DEBUG=yes instead
- Changes how Docker Desktop works too, there's a `case` for switching back if needed.
- Now the linux-only CONTAINER_COMPAT hacks are used also for Darwin (`DOCKER_SERVER_REQUIRES_LOOP_HACKS=yes`)
- `/dev/loop*` is spelled out since they don't exist on Darwin (`DOCKER_SERVER_USE_STATIC_LOOPS=yes`)
- armbian-next: introduce `USE_LOCAL_APT_DEB_CACHE` (default `=yes`) as alternative/in addition to `apt-cacher-ng` (eg, in Docker)
- this uses `cache/aptcache/${RELEASE}-${ARCH}` (in the host) for
- apt cache, by bind-mounting it to `${SDCARD}/var/cache/apt` in the `chroot_sdcard_apt_get()` runner and its usages
- debootstrap, by passing it `--cache-dir`
- utility function to help understand what is happening to cache during usage
- apt itself mantains this cache, removing old packages when new ones are installed. apt does this _by default_
- introduce `DONT_MAINTAIN_APT_CACHE=yes` to skip out of automatic apt maintenance of apt cache, eg, during `remove`s
- don't do `apt clean` and such if using local cache, that would clean the cache, not the chroot
- clean up `install_deb_chroot()` a little, find an unrelated bug there
- WiP: the great cli entrypoint (+docker) rewrite, Phase 6: relaunching structure; re-pass ARMBIAN_BUILD_UUID; use ARMBIAN_COMMAND for log filename; fix for output/logs dir perms
- WiP: the great cli entrypoint (+docker) rewrite, Phase 5: cleanups 4/x; better logging, check & force `DEST_LANG`
- WiP: the great cli entrypoint (+docker) rewrite, Phase 5: cleanups 3/x; don't write to stderr in generated Dockerfile
- it's `drastic red` on non-buildx dockers
- WiP: the great cli entrypoint (+docker) rewrite, Phase 5: cleanups 2/x, logging
- WiP: the great cli entrypoint (+docker) rewrite, Phase 5: cleanups 1/x
- source configs in a logging section.
- Docker: silent, fast retries to make sure `docker system df` works
- shut-up `chown` (no `-v`) output related to `SET_OWNER_TO_UID`
- ask user to wait while `DESTIMG` is rsync'ed to `FINALDEST` -- it's potentially very slow
- use green apple for Mac logging, instead of red apple which might imply error...
- WiP: the great cli entrypoint (+docker) rewrite, Phase 4: run as non-root, maybe-with-Docker
- introduce `is_docker_ready_to_go()`; if it is, and we're not root, use Docker instead of sudo. <- GOOD IDEA? BAD IDEA? lol
- introduce `SET_OWNER_TO_UID` var to be passed to Docker/sudo so written files are owned by the launching user, not root.
- introduce `mkdir_recursive_and_set_uid_owner()` and `reset_uid_owner()` to reset owner based on `SET_OWNER_TO_UID`
- use it for userpatches files created, logs, and output files, including images and debs.
- @TODOs ref. `$SUDO_USER` which I think the old version of this?
- add a lot of @TODOs, ref being able to relaunch something that's not `build` inside Docker, also add/change params and configs and command.
- initially add `ARMBIAN_DOCKER_RELAUNCH_EXTRA_ARGS`
- WiP: the great cli entrypoint (+docker) rewrite, Phase 3: rpardini is demented, v3
- WiP: the great cli entrypoint (+docker) rewrite, Phase 2: rpardini is demented
- WiP: the great cli entrypoint (+docker) rewrite, Phase 1
- armbian-next: WiP: Docker: actually use the GHA-image as base; pull it every 24hs.
- using image in my private repo.
- this has significant speedup to "start building time" on the 1st run
- move some Linux specific stuff to its own if
- add comments and todo
- armbian-next: WiP: Docker, high-WiP, beginnings of Armbian mount dict, with linux/darwin preferences
- armbian-next: WiP: Docker, configure `BUILDKIT_COLORS`
- armbian-next: WiP: Docker, make docker image from Dockerfile more compact by flattening layers
- armbian-next: `logging`: add whale indicator if build running under Docker
- armbian-next: WiP: `docker`: working with `bookworm`, `sid`, and `jammy` on Darwin & Linux; works with `bullseye` on Linux only
- armbian-next: WiP: `docker`: force ARMBIAN_RUNNING_IN_CONTAINER both in Dockerfile and passed as `--env`; apt update and install in same layer; back to jammy
- armbian-next: introduce `armbian_is_running_in_container()` and `armbian_is_host_running_systemd()`, replacing `systemd-detect-virt` in multiple spots
- WiP: try with debian:bullseye -- can't detect docker at all
- armbian-next: WiP: 2nd stab at new Docker support; Darwin still works; Linux `docker.io` working
- gen .dockerignore together with Dockerfile
- split in funcs
- hacks for Linux and `/dev/loop` stuff, CONTAINER_COMPAT=yes
- mac still works, Linux stuff would break it but I if'fed
- armbian-next: the secrets of `CONTAINER_COMPAT` revealed; add size checking to check_loop_device() and avoid retry when `mknod`ing
- this fails for the right reasons now, causing retries, which are then retried and work ;-)
- this is related to building under Docker on Linux, using docker.io package (not docker-ce)
- armbian-next: remove `.dockerignore` and add it to `.gitignore`; it's going to be auto-generated
- armbian-next: `.dockerignore`: Docker context should only have minimal files and folders, to speed up Dockerfile build
- IMPORTANT: `.dockerignore` is going to be generated from now on: so this is the last commit with changes before removal
- armbian-next: WiP: initial stab at new Docker support; really run the passed cmdline; add Dockerfile to gitignore
- armbian-next: WiP: initial stab at new Docker support; generate Dockerfile; introduce REQUIREMENTS_DEFS_ONLY
- uses REQUIREMENTS_DEFS_ONLY
- works on Docker Desktop on Mac;
- linux TBA
- armbian-next: don't error out if `.git` not present; other small fixes
- armbian-next: general "work or at least don't misbehave when run on a very bare ubuntu:latest instance"
- can't assume things, for example:
- that `sudo` will be available; it might not, and might be already root, no reason to fail
- that `/etc/timezone` will exist
- that `systemd-detect-virt` will be available
- that `git` will be available
- that `locale-gen` will be available
- remove grub's --verbose, it's really too verbose
- trust in TMPDIR, do not use trap at all in compile_plymouth-theme-armbian - settings traps adhoc is forbidden in armbian-next as well
- be more verbose in umount_chroot(), do not try to unmount tmpfs-mounted dir itself, only xx/tmp
- try to be smarter about /tmp being mounted in rootfs-to-image -- in the end the recursive saves us
- run post_debootstrap_tweaks when the SDCARD is still mounted, not after
- don't try to download anything from any mirrors if SKIP_ARMBIAN_REPO==yes
- extract loop_over_uboot_targets_and_do(), and handle `target_make/target_files` parsing there, so `compile_uboot_target` needs no parsing magic
- extract deploy_built_uboot_bins_for_one_target_to_packaging_area() (meant to be called via loop_over_uboot_targets_and_do)
- add hook `post_uboot_custom_postprocess` -- literally run after `uboot_custom_postprocess`
- @TODO: possibly squash with a previous u-boot hook was added
- `VER` gets used in the final *image* name produced, and it is very confusing to get `5.19.0` instead of `5.19.0-rc3` there.
- added debug's so we can find when we're bit, there's 2 variations
- also compile.sh
- shellfmt lib
- split off shell and python tools under lib
- revert removal of stuff a-n no longer uses (ref. compilation): general packaging, mkdeb etc
- editoconfig split off
- extension changes split off
- sources and sources/families split off
- some undue stuff removed or split
armbian-next: manual merge (30) of lib changes between 882f995e21 and 31ac6383e1
armbian-next: manual merge (30) of family/board changes between 882f995e21 and 31ac6383e1
armbian-next: manual merge (29) of family/board changes between 3435c46367 and 882f995e21 (A LOT!)
armbian-next: manual merge (29) of lib changes between 3435c46367 and 882f995e21 (A LOT!)
armbian-next: manual merge (28) of lib changes between revisions af6ceee6c5 and 38df56fbf3
armbian-next: manual merge (28) of sources/families changes between revisions af6ceee6c5 and 38df56fbf3
armbian-next: manual merge (27) of `lib` changes between revisions 9c52562176 and af6ceee6c5
armbian-next: manual merge (27) of `sources/families` changes between revisions 9c52562176 and af6ceee6c5
armbian-next: move `ROOTFSCACHE_VERSION` resolution from GitHub from `main-config` down to `create-cache`
- this way config does not depend on remote...
armbian-next: move `ARMBIAN_MIRROR` selection (network) from `main-config` to `prepare-host`
- this way CONFIG_DEFS_ONLY can run without touching the network
armbian-next: manual merge (26) of MD5-checking via debsums (3955) re-imagined
- @TODO make sure
armbian-next: manual merge (26) of sources/families changes between revisions 20ee8c5450 and 9c52562176
armbian-next: manual merge (26) of lib changes between revisions 20ee8c5450 and 9c52562176
- @TODO NOT including the md5/debsums check, that needs further rewrite
armbian-next: manual merge (25) of lib changes between revisions fe972621c6 and 20ee8c5450
- @TODO hmm Igor is now going out to the network for rootfs cache version during configuration phase!!! BAD BAD BAD
armbian-next: manual merge (25) of family changes between revisions fe972621c6 and 20ee8c5450
armbian-next: manual merge (24) of families changes between revisions 9ca9120420 and 560531a635
armbian-next: manual merge (24) of lib changes between revisions 9ca9120420 and 560531a635
armbian-next: manual merge (23) of all changes between revisions 17b4fb913c and 9ca9120420
armbian-next: manual merge (22) of all changes between revisions 0eb8fe7497 and 1dddf78cd0
- @TODO EXCEPT the insanity about locales/eval/VERYSILENT in #3850, requires deep review
armbian-next: manual merge (21) of all changes between revisions e7d7dab1bb and 0eb8fe7497
armbian-next: fix: patching CREATE_PATCHES=yes
- needed to create output dir
armbian-next: add `python2-dev` dep for old uboots
- cleanup some comments
armbian-next: manual merge (20) of all changes between revisions 6b72ae3c86 and 247c4c45fd
armbian-next: fix: pass `TERM` to kernel's make, so `make menuconfig` can work
armbian-next: fix: git: read commit UNIX timestamp/local date correctly
- `checked_out_revision_ts` was correct; git outputs `%ct` as a UNIX timestamp, UTC-based
- `checked_out_revision_mtime` was incorrect: git output it without converting to local time
- manually convert using `date @xx` so it has correct local time, whatever it is.
- add debugging to `get_file_modification_time()` too
armbian-next: abstract `$QEMU_BINARY` to `qemu-static.sh`: `deploy_qemu_binary_to_chroot()`/`undeploy_qemu_binary_from_chroot()`
- add hackish logic to avoid removing binary that would be needed if image actually contains `qemu-user-static` package
armbian-next: fix `uuidgen` basic dep check; use fake bash `$RANDOM` if uuidgen not available
- not good: we need uuidgen to begin logging, but it may not be installed yet. workaround.
armbian-next: retry 3 times download-only also for `PACKAGE_LIST_BOARD`
- acng is really not helping
armbian-next: allow customizing UBUNTU_MIRROR (ports mirror) with `CUSTOM_UBUNTU_MIRROR_ARM64=host/path`
armbian-next: WiP: kernel make via `env -i` for clean env; show produced /boot tree
armbian-next: manual merge (19) of all changes between revisions b23498b949 and e621d25adc
- the ssh firstrun revert stuff mostly
armbian-next: *breaking change* remove `LIB_TAG` and `.ignore_changes` completely
- one day should be replaced with an "update checker" extension, or even "update-enforcer"
- for now this just causes chaos
armbian-next: `python2` is required for some u-boot builds
- would be "use `python-is-python2` so /usr/bin/python exists and points to Python 2.x" but Jammy does not have that anymore
- python2 is required for some u-boot builds.
- that said, python 2.x is deprecated for a while and needs work thus @TODO
armbian-next: bump Python info gatherer to RELEASE=jammy too
armbian-next: add `KERNEL_MAJOR_MINOR` info to `media` kernel (@balbes150)
- 5.18 is not yet released so might be a problem here
armbian-next: allow to skip submodules during `fetch_from_repo`; introduce hook `fetch_custom_uboot`
- via GIT_SKIP_SUBMODULES=yes, which disables all submodules everywhere
- via UBOOT_GIT_SKIP_SUBMODULES=yes, which disables fetching of submodules during uboot fetch (hidden rkbins anyone?)
- extension hook `fetch_custom_uboot` so we can fetch our own stuff if needed
armbian-next: `initrd` caching fixes (always enable hook; if cache hit, convert to uImage too)
armbian-next: introduce `initramfs`/`initrd` caching
- using hashes of (hopefully) all involved files
- cache hits are rewarded with sprinkly hearts.
- why? this proves we got a reproducible kernel modules build!
- also, you just saved yourself 2-10 minutes of pain
armbian-next: manual merge (18) of changes between revisions 08cf31de73 and c8855aa08d
- heh; most bash code changes are for things already done in -next, or no longer used
- some version bumps, etc
armbian-next: cleanup entrypoint and shuffle `prepare_host_basic()` into logging section
armbian-next: *breaking change* add global extlinux killswitch `ALLOW_EXTLINUX`
- unless you set `ALLOW_EXTLINUX=yes`, then `SRC_EXTLINUX` will be disabled globally.
- add a bunch of logging regarding extlinux, armbianEnv and bootscripts for clarity during build
- this is due to nand-sata-install problems with extlinux
- some boards _only work_ with extlinux; we'll have to handle it later
armbian-next: extensions: `image-output-{qcow2|ovf}`: virtual output formats
- which use `qemu-utils` for `qemu-img` conversion of the .img
armbian-next: extension: `kernel-localmodconfig`: faster/slimmer kernel builds with `make localmodconfig`
armbian-next: extension: `cleanup-space-final-image`: zerofree, slim down firmware, show used space
armbian-next: introduce `do_with_ccache_statistics` and use it for kernel compile
- some TODOs
- better logging for .config copying
armbian-next: *breaking change* really disable apt sources for non-desktop builds
armbian-next: fix: don't manage apt-cacher-ng if told NOT to, not the other way around
armbian-next: `JUST_UBOOT=yes` + hooks `build_custom_uboot()`/`post_write_uboot_platform()`
- post_write_uboot_platform()
- only runs during build, for now (not on device)
- build_custom_uboot()
- allow fully custom, extension driven, building of u-boot
- also partial preparation of uboot source combined with default Armbian build
- HACK: u-boot: downgrade some errors to warnings via KCFLAGS
- fix copy of atf bins to uboot, don't do it if atf's not there
armbian-next: fix: no use testing the host for resolvconf if we're manipulating the SDCARD
armbian-next: sunxi_common: avoid shortcircuit error on family_tweaks_bsp when family_tweaks_bsp_s is not defined
armbian-next: fix: add `zstd` and `parallel` to hostdeps
armbian-next: manual merge (17) of all changes between revisions 64410fb74b and 08cf31de73
- changes about `git safe dir` ignored, I've done the same in a different way
- hash calculation changes ignored, fasthash is completely different
armbian-next: add `crossbuild-essential-armel` so `arm-linux-gnueabi-gcc` is available with system toolchains
- need to for some ATF builds, at least.
armbian-next: rockchip64_common: lotsa logging and debugging
- supposedly no practical changes
armbian-next: grub: better logging
armbian-next: fix for chaos caused by git's fix of CVE-2022-24765 otherwise "fatal: unsafe repository"
- might not be the best solution, but it's the only one I found
partitioning: fix: don't try fixing a bootscript that's not there
- this fixes a bug when "rootpart=2" without rootpart 1 being /boot
armbian-next: cleanups: umount tmpfs-based $SDCARD during cleanup too
armbian-next: indented heredoc, no functional changes
armbian-next: fix shortcircuit as last statement in case of extlinux
- yes, I wasted 3 hours on this tiny bit, so *you* don't have to!
- better logging for rootfs `mkfs` et al
- introduce `PRESERVE_SDCARD_MOUNT=yes` to preserve SDCARD, MOUNT, and LOOP for debugging
armbian-next: kernel-headers: less verbose, trimmed down tools a bit (perf and testing)
khadas-vim3l: add asound.state for Khadas VIM3L
armbian-next: introduce hook `extension_finish_config()` - late hook for ext configuration
- `extension_finish_config()` is the last thing done in config phase
- use it for determining stuff based on kernel version details, package names, etc
- also tune down some logging which was too verbose
- CI logs with no ANSI escape codes
armbian-next: shuffle around code and logic of `add_desktop_package_sources()`
- @TODO: still needs proper asset logging for sources.list(.d)
- @TODO: tunes down adding of sources/packages to CLI builds, check with Igor
armbian-next: 4.x can't build objtool in kernel-headers; allow for handling that later
- 4.x has a lot more obtuse dependencies
- introduce KERNEL_HAS_WORKING_HEADERS calculated based on KERNEL_MAJOR_MINOR
armbian-next: downgrade `error=misleading-indentation` to warning
- some 4.x kernels patches are really messy
- newer gcc's make that an error now
armbian-next: *allow cross compilation*, even the so-called "reverse cross-compile" (amd64 on arm64)
armbian-next: add `zfs` extension, which installs headers and builds ZFS via DKMS in chroot
- similar to how `nvidia` extension does it
armbian-next: x86: enable `nvidia` extension for all releases (only desktop)
armbian-next: `headers-debian-byteshift.patch` is dead; long-live cross-compiled source-only kernel-headers
- kernel-headers package now only includes _sources_
- postinst does the compilation and preparation for DKMS compatibility
- `tools` dir is included now, which includes the byteshift utilities
- handle special scripts/module.lds case after 5.10
- tested on a 6 combinations of `x86` / `arm64` / `armhf` (3x targets, 2x hosts)
- @TODO: we might be able to reduce the size of tools a bit (perf/tests/etc)
- @TODO: still missing ARCH vs ARCHITECTURE vs SRC_ARCH clarity elsewhere
armbian-next: allow `use_clean_environment=yes` for `chroot_sdcard_apt_get()` and descendants
- this causes command to be run under `env -i`, for a clean environment
armbian-next: manual merge (16) of all changes between revisions be9b5156a4 and 2a8e1ecac1
- many `traps` ignored: we don't use them anymore
armbian-next: fix logging for apt sources/gpg keys
armbian-next: don't leak `if_error_xxx` vars across runner helper invocations; always clean then (even if no error)
- also: fix wireguard-tools install, had a double parameter there
bcm2711: rpi4b: add `pi-bluetooth` which provides working Bluetooth
armbian-next: fixes for (non-)logging when interactively configuring kernel (`KERNEL_CONFIGURE=yes`)
armbian-next: move `lz4` rootfs caches to `zstd`, multithreaded
armbian-next: customize.sh: error handling, do not mount overlay if it doesn't exist
armbian-next: extra info for runners; `if_error_detail_message` and `if_error_find_files_sdcard` globals
- those are unset after running any command
- if error occur, message and/or found files will be included in log, for clarity
armbian-next: manual merge (15) of all changes between revisions 0f7200c793 and 101eaec907
armbian-next: better logging for `rsync` calls everywhere
- make rsync verbose
armbian-next: downloads: skip download if no `ARMBIAN_MIRROR` nor `DOWNLOAD_MIRROR`; less logs
armbian-next: update rockchip.conf from master and use runners
armbian-next: update mvebu64.conf from master and use functions
armbian-next: git: fix `fetch_from_repo` with actual submodules usage
armbian-next: `armbian-next`ify the `nvidia` extension after rebase from master
- driver version is configurable via `NVIDIA_DRIVER_VERSION`
- use runner function to log/error-handle/use apt cache/etc
rpi4b: there's no legacy branch anymore, remove it from KERNEL_TARGET
armbian-next: `download_and_verify` non-error handled; logging is messy [WiP] [HACK]
armbian-next: logging: let ANSI colors pass to logfile; CALLER_PID instead of BASHPID in subshell
armbian-next: enable HTTPS CONNECT in Armbian-managed apt-cacher-ng configuration
- PPAs require it
armbian-next: don't loop forever if we can't obtain ARMBIAN_MIRROR from redirector
- also, don't even try to do it if `SKIP_ARMBIAN_REPO=yes`
armbian-next: manual merge (14) of all changes between revisions 13469fd8a9 and 09e416e31c
- also editorconfig and compile.sh (root) changes
armbian-next: *much* improved logging to HTML; log archiving; consistency
- keep only current logfile
- log to LOGFILE also if SHOW_LOG=yes
- log cmd runtime and success/error directly in runner
armbian-next: *breaking change* use `MemAvailable` (not including swap) and up requirements for tmpfs
- of course add debugging logs
- rename vars
- should really only use this if we've really a lot of completely free RAM to spare
- otherwise OOM killer comes knocking
- or swapping to disk, that is counter-productive
armbian-next: *breaking change* `DEB_COMPRESS=none` by default if not running in CI/GHA
armbian-next: *breaking change* `CLEAN_LEVEL=make` is no more; new `make-kernel`, `make-atf`, `make-uboot`
- allows individual control of what to clean
- this effectively disables `make clean` by default
- rebuilds work and timestamping works for patching, so no reason to clean everytime by default.
armbian-next: refactor `prepare_host`, give `apt-cacher-ng` some much needed attention
- library dir for host-related stuff, pull it out of "general" finally
armbian-next: hostdeps: all toolchains via `crossbuild-essential-arm64`/`armhf`/`amd64`
- trying to sort out hostdeps for Jammy [WiP]
armbian-next: remove `eatmydata` usage, leftover from failed tries to make git faster
armbian-next: fix git origin check, recreate working copy if origin does not match
- fix cold bundle https download progress reporting
armbian-next: finally consolidating logs into output/logs; colorized HTML logs
armbian-next: introduce `do_with_retries()` and use it for apt remote operations during image build
armbian-next: another round of logging tuning/fixes; log assets; git logging
- introduce `do_with_log_asset()` and `LOG_ASSET=xxx`
- separate "git" logging level
- add `trap_handler_cleanup_destimg()` to cleanup DESTIMG
armbian-next: kernel: use parallel compressors; reproducible kernel builds
- also remove leftover mkdebian/builddeb parameters in make invocation
- add pbzip2 to hostdeps
armbian-next: tuning logging for timestamp/fasthash related stuff which is very verbose
- idea is to not completely overwhelm `SHOW_DEBUG=yes` case
- make patching quieter and use file instead of stdin
- set checked_out_revision_ts during git checkout (timestamp version of _mtime)
- timestamp | fasthash logging level (via `SHOW_FASTHASH=yes`)
armbian-next: completely remove mkdebian/builddeb/general-packaging kernel packaging stuff
armbian-next: manual merge (12) of all changes between revisions 34d4be6b7b and 5fe0f36aa8
armbian-next: introduce `PRESERVE_WORKDIR=yes` for heavy debugging
armbian-next: packaging linux-headers again
- do NOT use any output from `make headers_install` - that's for libc headers
- grabs "headers" (and binary tools) directly from the kernel build tree, not install target
- does not produce headers if cross compiling, for now
- produces kernel-headers package for the architecture on which it was built
- doing a single make invocation with build and install for packaging
- using 'make all' in place of vmlinuz/bzImage/image/zImage
armbian-next: apt download-only retried 3 times before installing main packages
armbian-next: fix `VER=` read from kernel-image package, also add `linux` default
armbian-next: some logging for atf compilation
armbian-next: rewrite hostdeps as array, add armhf toolchains
armbian-next: distro-agnostic: cleanups
armbian-next: armbianEnv hooks/debugs (bsp / image)
armbian-next: rpi: completely remove dtb hacks, allowing flash-kernel to work again
armbian-next: refactor new kernel packaging; add linux-dtb package back in finally, upgrades untested
armbian-next: refactor new kernel packaging; extract hook helper, fix kernel symlink
armbian-next: refactor new kernel packaging; add linux-dtb package back in finally, all hooks untested
flash-kernel: fix short-circuits as last statement in functions
armbian-next: do not force `SHOW_LOG=yes` if `CI=true`; let's _trust_ logging and error handling works
armbian-next: back out of setting mtime to the revision time during git checkout.
- of course this causes huge recompiles when wanted revision moves forward
armbian-next: sync 'config' dir from master revision ed589b248a
- this is _also_ getting out of hand... gotta merge soon
armbian-next: sync 'packages' dir from master revision ed589b248a
armbian-next: manual merge (11) of all lib/*.sh changes between revisions 3305d45b81 and ed589b248a
armbian-next: more refactorings, general logging; fixes; split image stuff
- logging flowing correct to LOGDIR, still needs packaging
armbian-next: complete removal of usages of `LOG_SUBPATH`; 100% error handled
- loose ends, use new LOGDIR
- remove the last shortcircuit in extensions execution, now it's 100% error handled
- many logging fixes
- still missing: final log consolidation/cleanup
logging: blue groups if `SHOW_DEBUG=yes` or `SHOW_GROUPS=yes` (console equivalent of CI's avocado)
armbian-next: shut down some too-verbose logging: logo building and update-initramfs
armbian-next: git/patching, kernel: use date from git as mtime minimum for patched files
- use revision's date from git log as mtime for all fetch_from_repo
- fix patched files date at least checkout date, otherwise some patches never build
armbian-next: first attempt at kernel packaging rework; just linux-image pkg, no dtbs yet
- correctly predict KERNELRELEASE, put image-dtbs in the right spot for flash-kernel
- remove dpkg-gencontrol, do it all directly
armbian-next: rework kernel source packaging, now exporting from git, to .tar.zst
- compress with zstdmt (multi-thread zstd), remove pv indicator, it's much faster anyway
- export from git (soon will have all patches et al too)
- better logging, show pkg name
- much, much faster due to zstdmt and deb with none compression
armbian-next: a bit atrocious, nameref loaded, `get_list_of_all_buildable_boards()`
- in the process, add support for userpatches/config structure mirroring core, for easy upstreaming
armbian-next: make `SKIP_EXTERNAL_TOOLCHAINS=yes` default. lets see what breaks [WiP]
armbian-next: keeping stdout clean, use display_alert() for cleanup logging
armbian-next: library cleanups; remove `build-all.sh` for good; bring `patching.sh` back
armbian-next: `interactive_desktop_main_configuration()` and stderr'ed + error handl'ed dialog
- use redirection to fd 3 for dialog, now cleanly on stderr
- `show_menu()` -> `dialog_menu()` et al
- interactive configuration now works again!
armbian-next: logging: `SHOW_PIDS=yes`
armbian-next: refactor and error-handle most of desktop configuration, incl menus/dialog
- `dialog_if_terminal_set_vars()` in place of `dialog_if_terminal()`
[WiP] ddk stuff, allow if not in `$KERNEL_TARGET`
armbian-next: split `compile_kernel()` function into smaller functions (+logging)
- `do_with_logging()` galore, much better error reporting for kernel
- `do_with_hooks()` is for the future, just a placeholder for now
armbian-next: `do_with_hooks()` placeholder for future ideas
armbian-next: logging: small refactor and `do_with_logging` admit it does not do error control
armbian-next: fix: traps: `trap_manager_error_handled` is integer (`-i`) not array (`-a`)
armbian-next: sunxi-tools: fix logging for sunxi-tools compilation
armbian-next: runners now run bash with `-o pipefail` in addition to `-e`
- attention, only affects stuff run through the functions in runners.sh
armbian-next: kernel: reduce logging clutter (CC,LD,AR)
- hide fasthash_debug under `SHOW_FASTHASH`
armbian-next: `armhf` should make `zImage` -- or should it?
armbian-next: show logs through ccze; avoid ANSI escapes in file; `SHOW_xxx` control
- `SHOW_DEBUG` shows the debug level
- `SHOW_COMMAND` shows all invoked commands through the runners
- `SHOW_TRAPS` to show 'cleanup' and 'trap' level
- `SHOW_TIMING` to show $SECONDS but pretty
- replace hardcoded traps/cleanups logging
armbian-next: add `KERNEL_MAJOR_MINOR=x.z` to every family, manually from the `KERNELBRANCH`
armbian-next: cold/warm bundles import/download/export for fetch_from_repo
- warm remote, if present, can be exported shallow
- if warm remote bundle is present, can be imported shallow too
- fallback to cold bundle if warm not present
- export (from cold, if exists + warm) shallow bundle
- use temp clone and DATE (not rev or tag) for shallowing, WORKS!
- info JSON/CSV, include "config_ok" true/false, kernel/uboot info
- include logs for failed configs too
- core reports ARMBIAN_WILL_BUILD_KERNEL and ARMBIAN_WILL_BUILD_UBOOT now with package names
armbian-next: `KERNELDIR` is out, `KERNEL_MAJOR_MINOR` is in for all `meson64`, `rpi4b` and `uefi`
armbian-next: new kernel workdir layout: cache/sources/kernel/<ARCH>-<KERNEL_MAJOR_MINOR>-<LINUXFAMILY>
- `GIT_FIXED_WORKDIR` is used to ignore 2nd param and use a specific dir
- this now REQUIRES `KERNEL_MAJOR_MINOR` to be set.
- prepare some `WARM_REMOTE_NAME` and related, based on it
armbian-next: JUST_KERNEL=yes (opposed to KERNEL_ONLY=yes) is really just the kernel build
armbian-next: fetch_from_repos now done when actually compiling atf/uboot/kernel, not before
- lib regen after removing empty files (sources.sh and errors.sh are now gone)
armbian-next: linux: back to Torvalds bundle, no tags; reminder about export
armbian-next: full cached kernel build; refactor all make's in a single place, even for packaging
- 2nd+ runs build in less than a minute
- kernel: compile and package in a single step, more efficient?
- KERNEL_BUILD_DTBS yes/no to build or not dtbs, replaces KERNEL_EXTRA_TARGETS
- dtbs_install, modules_install and headers_install now called by Armbian, not packaging
- kernel with split, but identical, build and install makes for modules/headers/dtbs
- make mkdebian and builddeb idempotent as possible
- keep a lot more cache, specially 'debian' folder
- filtering logging of install stuff
- might be a few leftovers, revisit gains with packaging later
- keeping the arm64 makefile Image.gz vs Image hack
- fix order of packaging patch byteshift, but still there
- cleaning of scripts tools on cross compile removed (!)
armbian-next: minor stylistic changes that I end up doing while working on other stuff
- I am `OFFLINE_WORK`'ing, I don't wanna waste 3 seconds, thanks
- OCD-fix of double `local` declarations
[giga-wip] rework patching, introducing fasthash
[wip] git: experiment with stable kernel bundle, and all tags. nice, but for what?
- also: never delete working copy, exit with error instead.
[wip] disable make clean during packaging. I wanna rebuild fast, always [NO PR?]
armbian-next: export CHOSEN_KERNEL_WITH_ARCH for reporting info
- fix info gathering, parse all boards first, and stop if some failed
- fix KERNEL_TARGET regex by including optional "export "
- add export from info to CSV, very basic stuff, but works
[squash] remove ddk bullshit from KERNEL_TARGET
armbian-next: remove file logging of aggregation stuff. config phase can't touch disk anymore.
[WiP] git cold bundle; no: shallow clones/fetched; yes: locally packed repo
armbian-next: reorder functions in file, they have a ~logical call-tree order
armbian-next: move `fingerprint_image()` out of `git.sh` into its own file
logging: fix for double stderr redirect during `fakeroot_dpkg_deb_build`
logging: subdued "command" logging for debugging low level cmd invocations ("frog")
armbian-next: when showing log, emit all host-side invocations as `debug` too.
[WiP] trap handler abstraction, somewhat works!
armbian-next: manual merge (10) of all lib/*.sh changes between revisions a4ae3a2270 and 3305d45b81 - but NOT the git unshallow stuff, that will be rewritten
armbian-next: trapmanager pt1, identifying spots for trap manager intervention
armbian-next: `install_pkg_deb` -> `install_host_side_packages` which is completely rewritten version
- much simplified; compare installed packages vs wanted, and only update if some missing
armbian-next: force u-boot and kernel's gcc to output colors to make easy to spot warnings and errors
docker: pass the `CI` env var through Docker invocation, for GitHub Actions
armbian-next: avoid warning if "file" utility not installed
- should not happen, but better safe than sorry
armbian-next: disable long_running `pv` progress bar for custom case too
- will rework later, for now pipe causes subshell and caos
armbian-next: if `CI=true` then `SHOW_LOG=yes` always
docker: add arm64 toolchain to Dockerfile; warn, but don't break, on modprobe failure
armbian-next: docker: use ubuntu:rolling, fix deps, use `$SRC/cache` as container's cache dir
armbian-next: logging fixes (padding, don't show error more than once, don't remove trap)
armbian-next: fixes for early error handling and logging
- split stacktrace-related functions into their own lib file
- simplify the traps
- some stacktrace formatting for readability
armbian-next: fix: don't `trap` `ERR` twice, it causes bash to go bananas regarding `caller`
armbian-next: `UPSTEM_VER` -> `UBOOT_REPO_VERSION` and related fixes
armbian-next: oops, fix some non-lib changes I missed, up to revision ff4346c468
armbian-next: manual merge (9) of all lib/*.sh changes between revisions 3b7f5b1f34 and ff4346c468
armbian-next: more error handling fixes. avoid shortcircuits.
- store previous error message in `MSG_IF_ERROR` (still to be actually shown somewhere during error)
armbian-next: more error handling fixes. avoid subshells, shortcircuits, and pipes
- add `CFLAGS=-fdiagnostics-color=always` to kernel compile; would need also for u-boot soon
WiP: indexing JSON into OpenSearch, all-hardcoded version
rpi: add DTB symlink in Debian/Ubuntu standard location /lib/firmware/$version/device-tree; remove build-time-only hacks
- this allows us to remove the most horrible hack
- should allow for correctly working DTB upgrades
- should NOT impact other families, although a new symlink will be created, nothing uses it.
rpi: fix: flash-kernel fix to ignore kernel 'flavour' for all raspi's
armbian-next: don't try to remove packages that are not installed to begin with
- much faster
- new chroot_sdcard_with_stdout() runner, without bash or any escaping.
armbian-next: don't try to enable systemd services for units that don't exist
- those might be removed by a bsp extension, so check for existence before trying to enable
armbian-next: don't error/warn on failure to enable bootsplash when systemd units missing
armbian-next: use indented HEREDOCS for all call_extension_method's
armbian-next: manual merge (8) of all lib/*.sh changes between revisions 1d499d9ac2 and 3b7f5b1f34
armbian-next: manual merge (7) of all lib/*.sh changes between revisions d885bfc97d and 1d499d9ac2
armbian-next: manual merge (6) of all lib/*.sh changes between revisions c7f3c239fe and d885bfc97d
armbian-next: avoid writing to disk during configuration; `ANSI_COLOR=none` logging; make CONFIG_DEFS_ONLY=yes runnable without sudo
- when `CONFIG_DEFS_ONLY=yes`, avoid writing the config summary output.log file.
- refactor that into a function as to be easy to if-out-of
- don't write to disk during aggregate_content() if `CONFIG_DEFS_ONLY=yes`
- don't write to disk during show_checklist_variables() if `CONFIG_DEFS_ONLY=yes`
- don't write to disk during write_deboostrap_list_debug_log() if `CONFIG_DEFS_ONLY=yes`
- don't compress and rotate logs if `CONFIG_DEFS_ONLY=yes`
- don't pretend to be handling errors we can't handle during var capture
- I foresee a world we can build all .debs without sudo
- and a some kind of split of codebase entrypoint due to that future feature
- some python info.py enhancements, not ready yet
armbian-next: shellfmt and regen library (after rebase from master n.5)
tools/shellfmt.sh: exclude "cache" and ".tmp" from formatting, for obvious reasons
tools/gen-library.sh: sort function files, so it does not keep changing between runs on different machines.
- order should not be important, since files only contain functions, but avoid git churn
armbian-next: manual merge (5) of all lib/*.sh changes between revisions 1b18df3c83 and e7962bb2b5
- most PKG_PREFIX work was already done
armbian-next: `TMPDIR` for all, many logging fixes, error handling: leave-no-garbage-behind without needing traps.
- set `MOUNT_UUID` and `WORKDIR`/`MOUNT`/`SDCARD`/`EXTENSION_MANAGER_TMP_DIR`/`DESTIMG` early in do_main_configuration()
- but, they're just _set_ there, dirs are not created early, but on demand later
- still @TODO: actually clean those during error trap. (unhappy path leaves garbage still)
- but does not leave garbage behind during "successful" runs at least (happy path works)
- actually export `TMPDIR` (== `WORKDIR`) during start of build (not config!), so all `mktemp` are subject to it
- `runners.sh` has helpers to avoid passing `TMPDIR` to chroot. Use the helpers! don't call `chroot` directly.
- don't trap/cleanup individual `mktemp` temp dirs during .deb packaging's, all is handled at once now.
- kernel packaging, for example, automatically picks up `TMPDIR` too. So now hosts `/tmp` is mostly left alone.
- fix some "infodumps" that are done into `.log` files directly.
- don't use sudo if `CONFIG_DEFS_ONLY=yes`; we'll only be collecting info, not doing anything.
- simpler logging for `rsync` operations (just dump to stdout, logging will handle it!)
- use padded counter for section logfiles, so we know which order things ran. exported as `CURRENT_LOGGING_COUNTER`
- no reason to use `apt-get` with `-yqq` anymore, since all logging is handled, so now `-y` by default
- desktop: using runners helpers for rootfs-desktop.sh, which should help a lot with acng caching and finding of problems
- extensions: correctly cleanup temp stuff, extensions has its own tmp/workdir now, and is always cleaned up at end of build.
armbian-next: bye `PKG_PREFIX`, hello `run_host_x86_binary_logged()` wrapper function; better error handling
- we've x86-only utilities that might need to be run on non-x86 build machines
- previously duplicated logic in PKG_PREFIX variable refactored into logged function
- added centralized debug logging
- replace all PKG_PREFIX usage with the new wrapper function, which already handles logging and errors.
- mostly FIP tooling invocations
- but also the boot_logo image builder
- wrapper function delegates to common `run_host_command_logged`
- wrap other FIP invocations with `run_host_command_logged` too, for tidy logging
- avoid using conditionals when invoking functions; that completely disables error handling inside the called function
- use explicit bash opts instead of shortcuts like `set -e`
- a _lot_ of debug logging added
armbian-next: always use UPPERCASE labels for FAT32 UEFI filesystems (rpi4b, uefi-*)
armbian-next: shellfmt after rebase onto master
armbian-next: manual merge (4) of all lib/*.sh changes between revisions 23afccf56e and e610f00bc7
- plus ooops
atf: fix for `set -e` mode; fix CROSS_COMPILE quoting
Signed-off-by: Ricardo Pardini <ricardo@pardini.net>
logging: predict the future location of .img file
- otherwise it's really unhelpful
Signed-off-by: Ricardo Pardini <ricardo@pardini.net>
uefi: alias `BRANCH=ddk` to `current`'s `DISTRO_GENERIC_KERNEL=yes`
- no real change, just to match rpi4b's BRANCH=style
- opens space for Armbian-built `current` soon
Signed-off-by: Ricardo Pardini <ricardo@pardini.net>
rpi: `legacy`->`ddk` (distro default kernel), remove overclock
- common vars in bcm2711.conf moved to top
- removed overclock/overvolt that was leftover my old setup
- confirmed: works with rpi3b too, should work with CM4/CM3 and others
- use valid UPPERCASE FAT label for RPICFG (in place of `rpicfg`)
Signed-off-by: Ricardo Pardini <ricardo@pardini.net>
armbian-next: shellfmt again after rebase
Signed-off-by: Ricardo Pardini <ricardo@pardini.net>
armbian-next: manual merge (3) of all lib/*.sh changes between revisions 1035905760 and e4e4ab0791
- missed non lib change on Several improvements for RPi builds (#3391)
- I just realized I will have to drop all non lib changes
rockship: fixes for `set -e` mode in rockship armhf family and bsp tweaks
Signed-off-by: Ricardo Pardini <ricardo@pardini.net>
armhf: enable building armhf targets on amd64 using system toolchains
- SKIP_EXTERNAL_TOOLCHAINS=yes on amd64 should use the same system toolchains as an arm64 build
Signed-off-by: Ricardo Pardini <ricardo@pardini.net>
logging: better logging about family_tweaks and family_tweaks_bsp
Signed-off-by: Ricardo Pardini <ricardo@pardini.net>
kernel: unblock cross compilation, warn about headers package
Signed-off-by: Ricardo Pardini <ricardo@pardini.net>
logging: fixes for sunxi/megous stuff with `set -e`
Signed-off-by: Ricardo Pardini <ricardo@pardini.net>
logging: fix shellcheck references generation
Signed-off-by: Ricardo Pardini <ricardo@pardini.net>
logging: manual merge (2) of all lib/*.sh changes between revisions 117633687e and 3083038855
Signed-off-by: Ricardo Pardini <ricardo@pardini.net>
logging: renaming function files a bit more consistently
Signed-off-by: Ricardo Pardini <ricardo@pardini.net>
logging: removing leftover empty file after all the moving around
Signed-off-by: Ricardo Pardini <ricardo@pardini.net>
logging: really insist on set -e during library loading
Signed-off-by: Ricardo Pardini <ricardo@pardini.net>
logging: shellfmt again after rebasing master
Signed-off-by: Ricardo Pardini <ricardo@pardini.net>
logging: manual merge of all lib/*.sh changes between revisions f6143eff67 and f3388b9aee
Signed-off-by: Ricardo Pardini <ricardo@pardini.net>
logging: generic do_capturing_defs wrapper; Python parser
- enabled by passing CONFIG_DEFS_ONLY=yes; in this case does not build anything
- [WiP] Python3 info reader / matrix expander
- multithreaded version
Signed-off-by: Ricardo Pardini <ricardo@pardini.net>
logging: move some interactive parts of config into its own functions
- mostly from config-prepare;
- there is still a lot of others in main-config
Signed-off-by: Ricardo Pardini <ricardo@pardini.net>
logging: use chroot_custom for grub and flash-kernel extension logging
Signed-off-by: Ricardo Pardini <ricardo@pardini.net>
logging: use line buffering, fix runner output color for GHA
Signed-off-by: Ricardo Pardini <ricardo@pardini.net>
logging: wrap dpkg-deb; set TMPDIR (not in chroot); refactor kernel make
- And a huge amount of @TODO's added
- Add "debug" and "deprecation" `display_alert()` levels
- insist that `install_common` is now `install_distribution_agnostic`
- unrelated: realtek 8822CS is EXTRAWIFI=yes only now, sorry.
- many debug statements for desktop
Signed-off-by: Ricardo Pardini <ricardo@pardini.net>
logging: don't bail out on patching error
Signed-off-by: Ricardo Pardini <ricardo@pardini.net>
logging: bunch of fixes; no-stdin; traps; better stacks
- mostly no-stdin dialog handling (desktop et al)
- let ERR trap run together with unmount trap (EXIT etc)
Signed-off-by: Ricardo Pardini <ricardo@pardini.net>
logging: trap ERR very early, pass-in caller info
Signed-off-by: Ricardo Pardini <ricardo@pardini.net>
armbian-next: huge refactor, shellfmt, codesplit, logging/error handling
- *this commit changes most/all the lines of bash code in armbian/build*
- *author is set to IgorPec for historical reasons, rpardini is to blame for the actual changes*
- logging: refactorings, pt.4: autogen lib, shellfmt tool, extract cli
- shellfmt: auto-downloader and runner of shellfmt
- darwin/linux
- amd64/arm64
- find ~correct files to format
- run formatting
- check formatting soon
- refactor compile's CLI stuff out into function/cli
- gen, and use genned library with tool
- logging: refactoring pt3: HUGE split of everything into everything else
- plus rebasing fixes
- logging: refactorings, pt. 2: error handling
- fix: no errors during umount_chroot()
- no progress for CI=true builds
- WiP disable kernel hashing. too crazy.
- a few builds now working with "set -e"
- wtf. or something
- kernel logging and long_running stuff - a mess - needs moving around in files
- rewrite uboot compile loop without using subshells. remove ALL traps. refactor host command
- better logging for u-boot
- more fixes, u-boot
- more fixes for logging et al
- git stuff
- many (many) fixes
- new color scheme
- a monster. make sure unmounted at the end. remove set -e's, to-be-readded.
- remove set -e "for safety reasons"
- more alerts. we gotta know whats failing
- some more logging stuff and fixes for error checking
- more logging and error handling stuff
- fixes; some set -e's
- more logging stuff
- logging: refactoring codebase, pt.1: functions-only
- Refactor the hell out of everything into functions
- rename build-all-ng to build-multi; other fixes, extensions init
- slight nudge
- some were already good, like this one.
- syntax fixes
- some need a little nudge
- another clean one
- some just need a better name (and splitting later)
- syntax fixes
- some were already good, like this desktop one
- some were already good, like this other one
- some were already good, like this one.
- debootstrap is gone.
- extract functions from compile.sh
- add logging to main_default_build
- more stuff
- cleanups and refactors of main.sh
- logging: first steps
- logging: pt. 0: shellfmt everything
- add riscv64 do SRC_ARCH/ARCH/ARCHITECTURE mess; add warn
* Add some Wireless modules to RK3588 legacy kernel
* Add some Wireless modules to RK3588 legacy kernel
* Add some Wireless modules to RK3588 legacy kernel
This will be changed in NEXT in any case, but in order to generate (probably last) roofs cache before switch, we need to bump compression to max or files will not go into GH releases.
* Added Debian 12 Bookworm.
* Link identical
* Make minimal truly minimal again
Move netplan.io to CLI image only
* Add netplan.io to Bullseye standard CLI
Co-authored-by: Konstantin Litvinov <klitvinov@piesoft.us>
Co-authored-by: Igor <igor@armbian.com>
* Extend BUILD_ONLY value to single value "any"
- build-tasks.sh:
- build_task_is_enabled():
- handle value "any" adequate to logic for empty string
- build_validate_buildOnly():
- handle value "any" adequate to logic for empty string
- add specific error message in case "any" appears in a list of task names
- add function build_only_value_for_kernel_only_build():
- provide a constant list of task names for kernel only package build
supposed to be used unique at several places
- backward_compatibility_build_only():
- propagate KERNEL_ONLY="no" to BUILD_ONLY="any"
- propagate KERNEL_ONLY="yes" to BUILD_ONLY="$(build_only_value_for_kernel_only_build)"
- update log messages accordingly
- interactive.sh:
- re-factor interactive_config_ask_kernel_only()
-> interactive_config_ask_build_only()
- interactive_config_ask_build_only():
- provide three pre-defined choices:
- "u-boot" - "U-boot package only"
- "$(build_only_value_for_kernel_only_build)" - "U-boot and kernel packages"
- "any" - "Full OS image for flashing"
- un-comment call of interactive_config_ask_build_only
in interactive_config_ask_kernel()
* Extend BUILD_ONLY value to single value "default"
- build-tasks.sh:
- change BUILD_ONLY setting of "any" to "default"
to be closer to parallel WIP of #4526
- interactive.sh:
- interactive_config_ask_build_only():
- improve menu selection texts for unique quick selection first characters:
- "u-boot" - "U-boot package only"
- "$(build_only_value_for_kernel_only_build)" - "Kernel and U-boot packages only"
- "default" - "Full OS image for flashing"
* Update sequence of BUILD_ONLY selection
- start with default to "Kernel and U-Boot packages only"
Co-authored-by: Igor <igor@armbian.com>
* Update prepare-host.sh
* Updated so it´s equal to prepare-host.sh
Made the file as advised equal to prepare-host.sh. There where also some other Ubuntu distributions not added to this file.
* fixed "unrecognized symbol [["
the old way caused the compilation to abort with a syntax error. idk why, but it did, this is how I got it to work.
* fixed syntax error
improvement to previous fix
* Moving repository handling from main script
- cleaning
- working
TBD: handling of parameters
* Not needed anymore
* Paramter handling, edit config in mktemp
* Typo
* Improve string handling
* Cleanining
* Compacting
* Change to backslash compatible sed replacement
* Publish all, update selected to avoid having erros at update
* Add support for html index generation
* Bugfix
* More arch
* Don't run publish if there is nothing in the incoming
* Add repo sync control file creation
* Fix hardcoded value
* Disable kernel bootsplash. We don't use it anymore and its time to go rest
* Switch to more recent commits
* Fix remaining drivers and fix shellcheck errors
* Rename default-build.sh -> build-tasks.sh
This commit is just a file rename without any
further code change before this script will be
refactored to improve build task structure
for partial building.
Refers to #4421
* Prepare build-tasks.sh for moving each build step to a dedicated function
This commit is an intermediate update to reflect
NO functional code changes of the particular
build steps before moving the unmodified code
into the appropriate function by copy/paste.
Only build_get_boot_sources() has slightly changed
to an appropriate if / fi structure.
Note: This .sh file is bash syntactically not correct.
The purpose of this commit is to leverage code review.
Refers to #4421
* Update build-tasks.sh function do_default() for filtered function calls
Function do_default():
- Add $1 function parameter for filtering build tasks
and assign this to local variable $_buildOnly
- move "pseudo inline"-functions to the beginning of the script
- replace the "inline functions" by filtered function calls
- add BUILD_ONLY variable to the build epilog print
Refers to #4421
* Complete refactoring of build-tasks.sh for BUILD_ONLY support
- update indentation of functions build_get_boot_sources()
and build_get_kernel_sources() to meet code style
- rename original function do_default() -> build_main()
- create new do_default() for backward compatibility and
comment this one as deprecated
- call build_main "" from this deprecated do_default()
- cli-entrypoint.sh:
- update call of do_default by build_main "${BUILD_ONLY}"
closes#4421
* Print final runtime as min:seconds
As with filtered build runtimes
may get below a minute, we would like
to get the runtime printed with seconds.
* Delete duplicate messages
* Fix the confusion of the choice of building the kernel
* Fix a typo in comment on do_default()
* Backward compatibility of KERNEL_ONLY and new BUILD_ONLY keys.
* interactive_config: replace KERNEL_ONLY on BUILD_ONLY
* Fix: Correct verification when an image is created
* Abort with error if BUILD_ONLY contains invalid build task names
- add function build_validate_buildOnly()
- call this function from build_main
- fix position of KERNEL_ONLY / BUILD_ONLY contradiction message
(did log empty _buildOnly content all time)
- improve local variable names
* Improve final error message to list valid BUILD_ONLY task names
* Improve error message logging
- collect multiple invalid BUILD_ONLY task names in a single error message
- log final message for valid BUILD_ONLY task names on ext level
- simple exit with exit code 1 in case of error
* The RELEASE variable cannot be empty.
Additional conditions for checking RELEASE are unacceptable.
* Fix build_main() to use BUILD_ONLY inside instead of local var set from parameter $1
- build-tasks.sh:
- build_main():
- don't take parameter $1
- replace usage of $_buildOnly by $BUILD_ONLY
- added info messages on KERNEL_ONLY cases in case BUILD_ONLY was propagated
- build_validate_buildOnly():
- remove parameter $2
- assign _buildOnly from global BUILD_ONLY accordingly
- cli-entrypoint.sh:
- call build_main without parameter
* Choosing interactive_config if the goal is bootstrap
* fix: order of selection backward_compatibility_build_only
* Update function comments to match current state
- removed obsolete parameter of build_main call do_default()
- replaced :space: by :comma: in _kernel_buildOnly
to leverage copy / paste for final repeatable command line
printed after build
* Improve logic coding to filter build tasks
- build-tasks.sh:
- added functions:
- build_task_is_enabled()
- build_task_one_of_is_enabled()
- build_task_each_of_is_enabled()
- updated existing build task filter logic
to use function build_task_is_enabled
- config-prepare.sh, prepare-host.sh:
- replaced existing build task filter logic
to use function build_task_is_enabled
* Remove unused functions
- remove function build_task_one_of_is_enabled()
and build_task_each_of_is_enabled()
according to code review
Co-authored-by: The-going <48602507+The-going@users.noreply.github.com>
