* ccache: Fix for distcc cross-compiling / various improvements
* Make ccache work when cross-compiling with distcc (fix#716)
* Allow to configure the ccache size in "pmbootstrap init"
* Moved ccache stats code from pmb/build/other.py to
pmb/helpers/frontend.py
* Grouped job count, ccache size and timestamp based rebuilds
together to "build options" and allow to skip them
* Sorted config options that had to be modified anyway
alphabetically
* Improve comment in arch-bin-masquerade APKBUILD
This is a follow-up to #935.
* fix regression #941: pmbootstrap doesn't automatically pick the
right architecture for building when none is specified
* remove obsolete --noarch-arch parameter
* Rename pmb/build/package.py to pmb/build/_package.py, so we can
access the functions it contains in testcases, and still use
pmb.build.package()
* Refactor the entire file. Instead of one big function that does
too many things, we have many small ones now, that are tested
in the testsuite and easier to modify
* Whenever building a package, pmbootstrap does not only build and
install the "makedepends" (like we did before), now it does the
same for the "depends". That's required to be compatible with
abuild. The old behavior can still be used with 'pmbootstrap
build --ignore-depends'.
* Because of that change, noarch packages can no longer be built in
the native chroot if we need them for a foreign chroot. A device-
package depending on a kernel would pull in the same kernel for
the native architecture otherwise.
* Running 'pmbootstrap build device-...' without '--ignore-depends'
and without a matching '--arch' displays a note that explains
this change to the user and tells how to use it instead.
* Noarch packages no longer get symlinked. That was only
implemented for packages built in the native chroot, and now that
is not always the case anymore. Symlinking these packages creates
packages with broken dependencies anyway (e.g.
device-samsung-i9100 can't be installed in x86_64, because
linux-samsung-i9100 is armhf only).
* Rename "carch" to "arch" wherever used. Naming it "carch"
sometimes is confusing with no benefit.
* Add a testcase for the aarch64 qemu workaround (because it failed
first and I needed to know for sure if it is working again).
* Improved some verbose logging, which helped with development of
this feature.
* Removed the old "build" test case (which was disabled in
testcases_fast.sh) as the new "build_package" test case covers its
functionallity.
* Only build indexes if the packages folder exists for that arch (Travis
couldn't run a test case otherwise)
* add my own build key
* enable the repo in the config
* update the README file
* Adjust testcase, that validates the keys and enable it in testcases_fast.sh
* Only save/load keys to/from the config file, which we ask for during
'pmbootstrap init', so the binary repo gets used even if a config file
already exists (this also removes a workaround, that deletes the work
folder path from the config dictionary before writing it)
* Download missing APKINDEX.tar.gz files with Python code, before
attempting to build packages (so we know which ones aleady exist in
the binary packages repository)
* Consider APKINDEX files older than 4 hours as outdated and download
them again (also in Python code)
* Provide 'pmbootstrap update' to force-update the APKINDEX files
* Travis: more logging output on failure
* Only allow keys from config_keys to be used by "pmbootstrap config"
When not specifying an architecture for `pmbootstrap build`, and the `APKBUILD`
says that it can't be built for the native arch, it gets built for the right
foreign arch. `pmbootstrap` did not properly detect if packages were already
built in that case, and tried to build them again.
(I've noticed that while building packages for the binary repo #871.)
Use any `linux-` package, that is not available for `x86_64` on a `x86_64` PC
and build it twice. It should get properly detected now:
```shell
pmbootstrap build linux-amazon-thor
pmbootstrap build linux-amazon-thor # should not get built again
```
* Allow to specify a custom username in "pmbootstrap init"
* Build chroots have "pmos" instead of "user" as username now
* Installation user UID is 1000 now (as in all other Linux distributions)
* Adjust autologins
* postmarketos-base: enable wheel group for sudo, removed previous sudoers file
* Implement safe upgrade path:
We save the version of the work folder format now, in $WORK/version.
When this file does not exist, it defaults to 0.
In case it does not match the currently required version
(pmb.config.work_version), then ask the user if it should
automatically be upgraded.
* apkindex:
* Also parse the architecture field
* symlink_noarch_package:
* Renamed to symlink_noarch_packages
* Always work on all packages (so we don't need to guess which
subpackages have been generated after a certain build)
* Get invoked when running 'pmbootstrap index'
* Use 'apk index' to generate one index, where the architecture
does not get rewritten (abuild does that by default, due to
Alpine's repos not having a 'noarch' folder and diverging from
that doesn't make things easier for us). That goes super fast,
and then we know which packages are noarch packages and can
create the symlinks.
* Made output less verbose:
* Use -q for 'apk index' when calling it directly (when it gets
called by abuild we can't control that)
* Output that the APKINDEXes get reindexed only to the 'pmbootstrap
log'.
We check if origin/HEAD is present. In case that reference is
missing, we show a meaningful error message now, with an explanation
on how to add it. Also moved find_out_of_sync_files_tracked() to
pmb.helpers.git
* Check kernel config
* Allow specifying multiple kernel packages, and also no packages
which defaults to scanning all kernel configs (it is super fast
anyway)
* Add the check to Travis CI
* Adjust existing kernel configs, so they pass the kconfig_check.
(We've had to put in a lot of defaults in the aarch64
linux-postmarketos configs, that's why the diff is a bit unclean.)
* Increase modified kernel pkgrels
Contrary to abuild, pmbootstrap only installs makedepends, and
keeps the installed packages around - both hacks save lots of time.
However, they may introduce missing makedepends in the APKBUILDs,
that the authors of the APKBUILDs do not notice because it works
for them.
This PR adss a strict mode, which will always clean the chroots
before building a package, and also remove all installed dependencies
after the package was built. You can use the following syntax to
only zap once, but build many packages at once:
`pmbootstrap build --strict hello-world 0xffff heimdall`
It also builds dependencies properly without leaving makedepends
behind.
* Don't specify pkgnames from "provides" as dependencies
Always use the regular pkgname. That way, we avoid listing all
kinds of so: files as dependencies (because Alpine automatically
adds them as depends= to the package database). This fixes building
weston, and reproducing the build with `pmbootstrap challenge`.
Additional changes.
* Clear the parsed APKINDEX cache for the current pmbootstrap
session after building a package
* Avoid rebuilding a package, in case it was already built due to
circular dependencies
I've replaced all instances in the code of `os.path.abspath`
with `os.path.realpath`, as this does the same as `abspath`
plus resolving symlinks.
See also: https://stackoverflow.com/a/40311142
Previously, distutils.version.LooseVersion was used, because it was
sort of close enough to how Alpine parses versions.
This new version uses the exact same algorithm, as `apk` does, and
it passes *all* of `apk`'s testcases for version checking (previously
we simply skipped the ones, that did not pass).
* Remove pmb/helpers/version.py left-over (it is in parse now)
* Make asserts consistent, do not use unnecessary parenthesis
Changes:
* Removed the apkindex_files cache. That particular cache caused
bug #189 and didn't bring any real-world performance improvements
(tested 3x with that cache and 3x without, no significant speed
difference). I decided to remove it instead of keeping it/adding
even more code to resolve the bug.
* Fix the check for already built packages: always use the architecture,
that the package should be built for instead of the architecture of
the build environment (e.g. use armhf, even when building a noarch
package in the x86_64 chroot). This partially resolves#341.
* Make pmb.chroot.apk.install_is_necessary() more robust: If the binary
package is missing, although it should be there, print a warning and
build it with force.
We can allow both ways:
* pmbootstrap menuconfig linux-motorola-titan
* pmbootstrap menuconfig motorola-titan
The former will show a tip about the second
This also prevents users from running menuconfig on aports like
'device-motorola-titan' or 'mkbootimg'
If you really want to build busybox, I recommend turning the
"timestamp based rebuilds" feature off, otherwise it will build
for all architectures all the time whenever you change something,
even if you do not increase the version number (that's the idea
of that feature). This is, because busybox is a dependency for
basiscally everything, so it must get updated whenever you install
something, in case it was out of date.
It is easier to simply rename the package.
I've done some refactoring while debugging #209.
* Unused file `pmb/build/crosscompiler.py` removed (that was a
left over from `_pmb_build_in_native_chroot` hack
* Do verbose logging in distccd, when `pmbootstrap --verbose` is
being invoked
* Restart distccd, when the commandline has changed (e.g. when the
currently running version was not verbose, and the new one is
verbose.) Prior to this change, it only got restarted, when the
architecture changed (so it did not allow changing the job count
on the fly for example).
* Insert missing whitespace in arguments help.
Thanks, Pablo Castellano!
* Fixed typo
* Guide user about what are the next steps after flashing kernel and password
* Increased postmarketos-mkinitfs' pkgver and rewording
Changes requested:
https://github.com/postmarketOS/pmbootstrap/pull/179
TLDR: Always rebuild/install packages when something changed when executing "pmbootstrap install/initfs/flash", more speed in dependency resolution.
---
pmbootstrap has already gotten some support for "timestamp based rebuilds", which modifies the logic for when packages should be rebuilt. It doesn't only consider packages outdated with old pkgver/pkgrel combinations, but also packages, where a source file has a newer timestamp, than the built package has.
I've found out, that this can lead to more rebuilds than expected. For example, when you check out the pmbootstrap git repository again into another folder, although you have already built packages. Then all files have the timestamp of the checkout, and the packages will appear to be outdated. While this is not largely a concern now, this will become a problem once we have a binary package repository, because then the packages from the binary repo will always seem to be outdated, if you just freshly checked out the repository.
To combat this, git gets asked if the files from the aport we're looking at are in sync with upstream, or not. Only when the files are not in sync with upstream and the timestamps of the sources are newer, a rebuild gets triggered from now on.
In case this logic should fail, I've added an option during "pmbootstrap init" where you can enable or disable the "timestamp based rebuilds" option.
In addition to that, this commit also works on fixing #120: packages do not get updated in "pmbootstrap install" after they have been rebuilt. For this to work, we specify all packages explicitly for abuild, instead of letting abuild do the resolving. This feature will also work with the "timestamp based rebuilds".
This commit also fixes the working_dir argument in pmb.helpers.run.user, which was simply ignored before.
Finally, the performance of the dependency resolution is faster again (when compared to the current version in master), because the parsed apkbuilds and finding the aport by pkgname gets cached during one pmbootstrap call (in args.cache, which also makes it easy to put fake data there in testcases).
The new dependency resolution code can output lots of verbose messages for debugging by specifying the `-v` parameter. The meaning of that changed, it used to output the file names where log messages come from, but no one seemed to use that anyway.
...instead of running apk every time to get the list of installed
packages and their versions. The internal package database from
apk has the same format, as the extracted APKINDEX file (except
that it has more key-value pairs, which we ignore/do not need
right now). So the APKINDEX code has been extended to parse both
tar-packed APKINDEX files and regular text files in the APKINDEX
format.
This is required for #108, for a better detection of outdated
packages (because the internal package database saves the
package's timestamp, too). A nice benefit is, that this is faster
than calling apk every time and it doesn't fill up the log as much.
I've also used this improved function for determining the apk
version (for the outdated version check), and I've deleted
pmb.parse.other.package_split(), as it is not needed anymore.
* New commandline parameter --mirror-pmOS, where the binary repository
URL for postmarketOS can be specified (empty by default as of now,
this will be filled with the real URL once the repo works)
* Do not build packages, when they are in the binary repository and
the version of the package in the binary repository is up-to-date.
* Add a testcase for pmb.build.is_necessary().
...even if the pkgver and pkgrel have *not* changed. This should
make development much more intuitive. The detection works by looking
at the last modified timestamps, just like `make` does it.
...and some smaller fixes:
* make the diff output easier to read
* verify, that only .apk, .buildinfo.json and the APKINDEX have
changed inside the local repository folder. Because the file
names of these changes will be used to release files from
staging to release.
The following message will be printed out now, in case you attempt
to build a package for an architecture, that is not specified in
the "arch=..." line inside the APKBUILD:
ERROR: Architecture 'aarch64' is not supported for this package.
Please add it to the 'arch=' line inside the APKBUILD and try again:
/path/to/APKBUILD
* The supported architectures are inside the config now
* Symlinks get created for that list of supported architectures now.
* During initialization, the architecture from the selected device
gets checked against the list of supported architectures. When
it is not included, a meaningful exception gets raised.
* the aportgen and (cross-compiler) build tests make use of the
new variable now (they had armhf and aarch64 hardcoded previously).
* Fix hardcoded `armhf` in pmb/aportgen/binutils.py
* Generate aports: `binutils-aarch64`, `musl-aarch64`, `gcc-aarch64`
* Distccd: Remember the cross-compiler architecture (currently armhf
or aarch64), that the current distccd is running as, and restart
distccd with the correct architecture, in case a different arch
is needed than what it is currently running as. (Depending on the
cross-compiler arch, the PATH variable gets adjusted before
starting distccd)
* Testcases: add aport generation for aarch64, add cross-compiling
to aarch64
* pmb/parse/arch.py: Add aarch64 to the mapping
* Two new functions for getting a list of files and their timestamps
in the repo, and diffing that information to get a list of changed
files: pmb.helpers.repo.files() and pmb.helpers.repo.diff().
(I've put it in the helpers folder, because it is not specific to
one chroot, but to all chroots at once.)
* pmbootstrap challenge (new command introduced a few commits back to
verify, that the contents of an APK file are deterministic) uses
these functions to a) support subpackages and b) optionally
output a list of changed files (this gets used in the pmbuilder
script, which lives outside of this repository).
This commit is progress for #64 again.
We have "lazy reproducible builds" now. What I mean by that is, that
the resulting "apk" archive is not fully reproducible, but all binaries
inside it are. This is necessary to kick-off the binary repo, which is
in turn required to get the testsuite going on Travis. Read #64 for more
information.
Usage:
```
pmbootstrap build hello-world --buildinfo
pmbootstrap challenge /tmp/path/to/hello-world-1-r2.apk
```
The "--buildinfo" parameter generates a "buildinfo.json", which contains
the versions of all dependencies. It is not very optimizied, so this
is a performance bottleneck and takes 10 seconds (which is quite much
considering that the hello-world package builds in less than a second).
This can be improved in the future, and then the buildinfo parameter
may become the default.
Oliver Smith
lawl
drebrez
Pablo Castellano
Attila Szöllősi