Otherwise mypy gets confused by the type later in the code, and I also
think it might be confusing behaviour in general that <options> gets
overwritten in this loop.
Instead of printing an URL with each warning, print one once at the
end. This gives less cluttered output, and we don't have information
about each and every option on that wiki page anyway (and given that we
now have so many options, I don't think it's good use of time to add it
there.)
Move all kconfig rules from pmb.config to a separate toml file
pmb/data/kconfigcheck.toml. This is a fallback, pmbootstrap now prefers
loading kconfigcheck.toml from the currently checked out pmaports branch
if it exists.
This finally allows having separate kconfig check rules per pmaports
branch and makes the workflow of adjusting these rules much more
pleasant as the rules and kernel configs can just be adjusted at the
same time in pmaports!
This patch also moves the definition of what rules should be checked for
community and main devices, those that have pmb:kconfigcheck-community
in their linux APKBUILD, to the new kconfigcheck.toml. This should make
it much more intuitive, previously one needed to find the place in the
pmbootstrap source and edit it there.
Furthermore the "enforce_check" logic is removed. Previously pmbootstrap
would print warnings for failed config checks in some cases, but not
exit with error which was very confusing. Now exit 0 means all checks
passed and exit 1 means, that there is at least one error.
Use toml for the file, as discussed in pmbootstrap issue 2165. Python
3.11 has a native toml reader, use tomli for previous Python versions
for compatibility.
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>
With the new chroot type, we can now write fancy paths in the pythonic
way. Convert most of the codebase over, as well as adding various other
type hints.
Signed-off-by: Caleb Connolly <caleb@postmarketos.org>
We use a custom verbose log level in pmbootstrap, unfortunately it isn't
possible to correctly type this due to some limitations in the logging
library [1], [2].
Given that our usecase is fairly simple, we can just wrap the module
with our own so we only have to tell mypy to ignore the error once
instead of at every callsite.
[1]: https://github.com/cryptax/droidlysis/issues/15
[2]: https://github.com/python/typing/discussions/980
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.
Make the handling of the custom NonBugError and BuildFailedError
exceptions more consistent with the handling of other exceptions, by
printing "ERROR: " infront of the actual error text. Then we don't need
to duplicate that where we raise the errors. pmbootstrap prints "ERROR"
in red.
The intent of 9a74848f19 was to not print
a stacktrace when this happened. However, I still opted to raise an
exception, so you still get a stacktrace, just with a more helpful
message. There's no need to give the user a big stacktrace here, so
raise a NonBugError instead of a RuntimeError.
Make sure that we disable legacy gadgets like USB_ETH. Our initramfs
uses configfs gadgets instead, currently RNDIS for USB networking.
In the future this can be expanded to more options like mass storage,
MIDI or whatever we're going to integrate into the OS that can be
configured by the user.
Reviewed-by: Oliver Smith <ollieparanoid@postmarketos.org>
Link: https://lists.sr.ht/~postmarketos/pmbootstrap-devel/%3C20230423163438.746054-1-luca@z3ntu.xyz%3E
Python passes all arguments by reference. For mutable objects such as
a list used here, changing the object will change the original one. If
components_list is not set, this means the default value gets modified.
This lead to kernels getting checked with the wrong required components.
For example, when checking a kernel from community first with the
default component_list, it would get extended with all options needed
for a kernel in community. When checking another kernel from the
testing category, also with the default component_list, it would now
check for the community options in the testing kernel.
Related: https://stackoverflow.com/a/986145
Reviewed-by: Luca Weiss <luca@z3ntu.xyz>
Link: https://lists.sr.ht/~postmarketos/pmbootstrap-devel/%3C20230402124338.64886-1-ollieparanoid@postmarketos.org%3E
Don't set the config_path_pretty variable and pass it through various
functions until it ends up being used in check_option. This was just the
directory name of the kernel config and the kernel config filename (or
in case of pmbootstrap kconfig check --file, just the same as the
config path).
Instead we can just print the filename of the kernel config, for example
"config-postmarketos-qcom-sdm845.aarch64". It is shorter and already
obvious to which package it belongs.
Reviewed-by: Clayton Craft <clayton@craftyguy.net>
Link: https://lists.sr.ht/~postmarketos/pmbootstrap-devel/%3C20230312151325.1968-3-ollieparanoid@postmarketos.org%3E
Currently when any device does not conform to the options they declare,
we fail the whole kconfig check.
Now that we start requiring more options, especially with
pmb:kconfigcheck-community it makes sense to relax these restrictions so
we're more free to edit kconfig options and don't have to adjust all
testing devices that may or may not be properly maintained.
As a side effect this patch makes it practically impossible to make
kconfig check actually fail for any testing device which might not be
optimal. If these use cases appear in the future we will want to adjust
pmbootstrap to allow for that.
Reviewed-by: Oliver Smith <ollieparanoid@postmarketos.org>
Link: https://lists.sr.ht/~postmarketos/pmbootstrap-devel/%3C20221105074432.13804-2-luca@z3ntu.xyz%3E
Add the new option that will be mandatory for all devices in
community/main category. This is just a combination of anbox + iwd +
nftables + containers + zram + netboot.
While the existing options could be removed we're keeping it for now
given that also some devices with downstream kernel might find some
options useful.
When using a kernel from Alpine the kernel configuration cannot be
found in pmaports. We cannot check the kernel config for missing
options in that case, but that's no reason to break the flasher
entirely.
Support full linux-* package names in argument completion for
"pmbootstrap kconfig ..." command-lines and get rid of related PROTIP
messages:
PROTIP: You can simply do 'pmbootstrap kconfig check postmarketos-allwinner'
This improves consistency, as in other places we expect the user to
supply full package names as well (e.g. pmbootstrap build).
iwd seems like a promising alternative to wpa_supplicant. It uses crypto
implementations from the kernel, so let's make kconfig check aware of
the options it needs.
Extend the kconfig check code to not only support booleans and arrays,
but also strings. This will be used for CONFIG_LSM with apparmor where
it's important that "apparmor" has a certain spot in the list.
