This might have saved me a couple of minutes of debugging, at least.
While working with the aportgen code I had two distinct issues:
1. APKINDEX was outdated, so I got an error about not being able to
download the supposedly latest version of BusyBox as it no longer
existed on Alpine's mirror.
2. APKINDEX had not been created by any other command, so aportgen
would tell me that it couldn't find the "busybox" package.
The changes in this commit resolves both of these issues from my
testing.
These were almost identical across the Busybox, GRUB EFI, and musl
generators. Unify them by creating functions they all (and any future
aportgen generators) can call instead of having to maintain three
separate copies of this code.
This can otherwise fail if the package was only built for some
architectures so far, but not for the architecture that is being generated.
E.g. right now musl-1.2.5-r6 was built already for x86_64 but not aarch64,
which causes errors like the following when running "pmbootstrap aportgen
musl-x86_64" on aarch64:
>>> musl-x86_64: Fetching musl-1.2.5-r5-x86_64-edge.apk::http://dl-4.alpinelinux.org/alpine//edge/main/x86_64/musl-1.2.5-r5.apk
Connecting to dl-4.alpinelinux.org (147.75.40.42:80)
Connecting to dl-4.alpinelinux.org (147.75.40.42:443)
wget: server returned error: HTTP/1.1 404 Not Found
>>> ERROR: musl-x86_64: checksum failed
Cease merging pmbootstrap.cfg into args, implement a Context type to let
us pull globals out of thin air (as an intermediate workaround) and rip
args out of a lot of the codebase.
This is just a first pass, after this we can split all the state that
leaked over into Context into types with narrower scopes (like a
BuildContext(), etc).
Signed-off-by: Caleb Connolly <caleb@postmarketos.org>
Introduce a new module: pmb.core to contain explicitly typed pmbootstrap
API. The first component being Suffix and SuffixType. This explicitly
defines what suffixes are possible, future changes should aim to further
constrain this API (e.g. by validating against available device
codenames or architectures for buildroot suffixes).
Additionally, migrate the entire codebase over to using pathlib.Path.
This is a relatively new part of the Python standard library that uses a
more object oriented model for path handling. It also uses strong type
hinting and has other features that make it much cleaner and easier to
work with than pure f-strings. The Chroot class overloads the "/"
operator the same way the Path object does, allowing one to write paths
relative to a given chroot as:
builddir = chroot / "home/pmos/build"
The Chroot class also has a string representation ("native", or
"rootfs_valve-jupiter"), and a .path property for directly accessing the
absolute path (as a Path object).
The general idea here is to encapsulate common patterns into type hinted
code, and gradually reduce the amount of assumptions made around the
codebase so that future changes are easier to implement.
As the chroot suffixes are now part of the Chroot class, we also
implement validation for them, this encodes the rules on suffix naming
and will cause a runtime exception if a suffix doesn't follow the rules.
Instead of building cross packages for the native architecture, add a
new get_cross_packages_arches() function that returns the desired
architectures no matter what the current native architecture is.
So gcc-x86_64 will always have arch="aarch64", and gcc-aarch64 will
always have arch="x86_64".
We can tweak the function in the future if e.g. it becomes viable to
cross compile from riscv64. Or even better, remove this and the cross
compile packages upstreamed into alpine (see pmb#551).
Link: https://lists.sr.ht/~postmarketos/pmbootstrap-devel/%3C20231126104823.3723-1-ollieparanoid@postmarketos.org%3E
Replace "args.arch_native" with the direct function call in order to
avoid passing "args" to all functions. This is a step to get rid of this
args-passed-to-all-functions pattern in pmbootstrap.
Packages like binutils-*, busybox-static-*, gcc-*, grub-efi-*, musl-*
are only needed for the native architecture during cross compilation.
Don't bother with trying to build them for other arches to save time and
to avoid getting stuck frequently at "armv7/binutils-aarch64" etc.
A few people like to use pmbootstrap on aarch64 hosts (e.g. PineBook
Pro), so let's make it available for aarch64 again when we can build
aarch64 packages natively in CI and bpo. (They do get stuck there right
now, because of qemu user emulation.)
Related: https://gitlab.com/postmarketOS/build.postmarketos.org/-/issues/75
Use a deterministic mirror URL instead of CDN for aportgen. Otherwise we
may generate a pmaport that wraps an apk from Alpine (e.g. musl-armv7)
locally with one up-to-date mirror given by the CDN. But then the build
will fail if CDN picks an outdated mirror for CI or BPO.
While at it, also remove unnecessary "#!/usr/bin/env python3" in files
that only get imported, and adjust other empty/comment lines in the
beginnings of the files for consistency.
This makes files easier to read, and makes the pmbootstrap codebase more
consistent with the build.postmarketos.org codebase.
After initializing the build environment, the cache_distfiles folder
currently is writable by everyone (which is not ideal, fix following
soon). The aportgen code for `busybox-static-*` and `musl-*` copies
the foreign arch `.apk` file to the distfiles, but it executes this
action as regular user and not as root. This only works as long as
build initialization ran before (which may not be the case on Travis
and expecting this to run before is a bug in general).
With this commit, the copy action gets executed as root, so it works
in any case. I'm commiting this directly (without a PR), because it
is a super simple fix and it unblocks our continuous integration.
Local testing:
$ pmbootstrap -y zap
$ sudo rm -r ~/.local/var/pmbootstrap/cache_distfiles
$ pmbootstrap aportgen musl-armhf
* The APKINDEX parser used to return a dictionary with one package for
a given package name. This works for the installed packages database,
because there can only be one provider for a package. But when
parsing packages from binary repositories, we need to support
multiple providers for one package. It is now possible to get a
dictionary with either multiple providers, or just a single provider
for each package.
* Dependency parsing logic has been adjusted, to support multiple
providers. For multiple providers, the one with the same package
name as the package we are looking up is prefered. If there is none
(eg. "so:libEGL.so.1" is provided by "mesa-egl"), it prefers packages
that will be installed anyway, and after that packages that are
already installed. When all else fails, it just picks the first one
and prints a note in the "pmbootstrap log".
* Added testcases for all functions in pmb.parse.apkindex and
pmb.parse.depends
* pmbootstrap chroot has a new "--add" parameter to specify packages
that pmbootstrap should build if neccessary, and install in the
chroot. This can be used to quickly test the depencency resolution
of pmbootstrap without doing a full "pmbootstrap install".
Fixes#1122.
