This patch updates the vagrant box to ubuntu bionic 64 and switches back
to the official ubuntu image cause the issues with the official image
are now solved.
Signed-off-by: Angelo Compagnucci <angelo@amarulasolutions.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
While integrating proxy support in builder.py, a log flush
was left in the code. This commit cleans/removes that code.
Signed-off-by: Matthew Weber <matthew.weber@rockwellcollins.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Allow builder.py to inherit the system proxy settings from
the env if they are present.
Signed-off-by: Matthew Weber <matthew.weber@rockwellcollins.com>
Reviewed-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
The download wrapper is a purely internal helper, and is not supposed to
be callable manually. No need to offer some help.
Besides, the help text was way out-dated.
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
With cmake packages, we are only using TARGET_LDFLAGS for executables
and not for shared libraries.
This patch adds CMAKE_SHARED_LINKER_FLAGS and
CMAKE_MODULE_LINKER_FLAGS to the cmake toolchain file so that
buildroot TARGET_LDFLAGS are used for shared and module libraries.
Signed-off-by: Damien Thébault <damien.thebault@vitec.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Add the i686 package list to install when using pre-built 32 bits
binaries with a redhat/fedora host distribution (glibc.i686 and
zlib.i686).
Signed-off-by: David De Grave (Essensium/Mind) <david.degrave@mind.be>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
All pre-built Buildroot toolchains have been rebuilt with Buildroot
2018.05, so this commit updates the corresponding configuration
fragments to make sure the autobuilders use the new toolchains.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Currently, we install flake8 and its dependencies via pip. We
tried to be reproducible by pinning the version of those python
packages, but we did forget quite a few of them, and thus some
dependencies for flake8 are installed as uncontrolled versions.
Furthermore, before we install flake8 and its dependencies, we
forcibly update pip, setuptools, and wheels packages to their
latest versions. This explicitly breaks reproducibility.
While we could enforce a specific version of all those packages
and still grab them from PyPI, we can simply grab them from the
distribution-provided packages instead.
Since we're using a pinned version of stretch, this already
guarantees we'll reproducibly get the same versions over and
over again. Besides, we just need to list flake8 as a package to
install to automatically get all its dependencies (again, in a
reproducible way).
This has the slight unfortunate drawback of downgrading flake8
to version 3.2.1, from version 3.5.0, as well as downgrading a
few of flake8's dependencies, as noticed by Ricardo:
http://lists.busybox.net/pipermail/buildroot/2018-May/222376.html
However, as Ricardo said, there isn't "any serious limitation of
this old version, the release notes for a version in the between
mentions 'Dramatically improve the performance' but we have a
limited number of scripts and running on Gitlab for all of them
still takes less than 5 minutes".
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Ricardo Martincoski <ricardo.martincoski@gmail.com>
Acked-by: Ricardo Martincoski <ricardo.martincoski@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
As suggested in the docker best practices [0], order the package list
alphabetically, and list only one package per line.
This will be much usefull later, we need to update the list of installed
packages, like adding new ones for example.
[0] https://docs.docker.com/develop/develop-images/dockerfile_best-practices/#sort-multi-line-arguments
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Ricardo Martincoski <ricardo.martincoski@gmail.com>
Reviewed-by: Ricardo Martincoski <ricardo.martincoski@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Fix three issues with code style in our test infra:
- 'print' is now a function,
- exceptions need to be caught-assigned with the 'as' keyword,
- old-style "%s"%() formatting is deprecated.
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Ricardo Martincoski <ricardo.martincoski@gmail.com>
[Thomas: drop indices in format strings.]
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
'+' is a valid character in a url. The current dl-wrapper gets the
URI scheme by dropping everything after the last '+' character, with
the intension of finding 'git' from e.g. 'git+https://uri'.
If a uri has a '+' anywhere in it, it ends up using too much of the
string as a scheme, and fails to match the handler properly.
An example of where this form of URI is used is when using deploy tokens
in gitlab. It uses a form like https://<username>:<password>@gitlab.com/<group>/<repo.git>
where username for deploy token is of the form 'gitlab+deploy-token-<number>'.
Use the %% operator to search backwards until the last '+' character when
dropping the rest of the string as we know that the first '+'
in the string should be the scheme.
Signed-off-by: Robert Beckett <bbeckett@netvu.org.uk>
Acked-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
In commit 7517aef4d (support/docker: limit the number of layers),
we reduced the number of layers by coalescing multiple RUN commands
into less commands.
In doing so, we especially coalesced "apt-get update" with "apt-get
install".
However, the distribution we used is a pinned version of stretch, so
we know that running apt-get update will always yield the same apt
database.
If we split the two apt-get commands, then we can re-use any local
intermediate image when we need to update the list of packages to
install; this helps quite a bit when testing the docker files over
and over again, with just slight variants in the packages list.
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Ricardo Martincoski <ricardo.martincoski@gmail.com>
Cc: Peter Korsgaard <peter@korsgaard.com>
Reviewed-by: Ricardo Martincoski <ricardo.martincoski@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Use Python 3 style print calls, in order to make pkg-stats Python 3
compliant.
Signed-off-by: Matthew Weber <matthew.weber@rockwellcollins.com>
Reviewed-by: Ricardo Martincoski <ricardo.martincoski@datacom.ind.br>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Since the rework of the download infrastructure, the "file" download
helper gets passed an URL that starts with file://, but forgets to
strip it before passing it to "cp", causing a failure as the "cp"
program isn't prepared for file paths starting with file://. This is
fixed by stripping the file:// at the beginning of the URL.
In addition, the path passed to cp lacked a slash between the
directory path and the filename part of the url. This is fixed by
adding a slash at the appropriate places.
Fixes the following build failure when the "file" download method is
used:
cp: cannot stat 'file:///home/angelo/DEV/TOOLCHAINSarmv7-eabihf--glibc--bleeding-edge-2017.11-1.tar.bz2': No such file or directory
Signed-off-by: Angelo Compagnucci <angelo@amarulasolutions.com>
Reviewed-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
A person on IRC reported a build failure with the util-linux package,
looking like this:
for I in uname26 linux32 linux64 ; do \
cd /home/aep/consulting/chargery/tracker/output/target/usr/bin && ln -sf setarch $I ; \
done
[...]
/bin/sh: line 1: ./ln: cannot execute binary file: Exec format error
/bin/sh: line 1: ./ln: cannot execute binary file: Exec format error
/bin/sh: line 1: ./ln: cannot execute binary file: Exec format error
The issue was an empty path in the PATH variable, which means "current
working directory", causing a "ln" binary built by util-linux for the
target to be used instead of the system-provided "ln".
We already check a number of things in the PATH and LD_LIBRARY_PATH
variables in support/dependencies/dependencies.sh, but we were not
checking that PATH did not contain an empty path.
This commit fixes that and takes this opportunity to simplify the test
code for PATH and LD_LIBRARY_PATH.
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
[Thomas: improve commit log.]
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
We currently attempt a shallow clone, as tentative to save bandwidth and
download time.
However, now that we keep the git tree as a cache, it may happen that we
need to checkout an earlier commit, and that would not be present with a
shallow clone.
Furthermore, the shallow fetch is already really broken, and just
happens to work by chance. Consider the following actions, which are
basically what happens today:
mkdir git
git init git
cd git
git remote add origin https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git
git fetch origin --depth 1 v4.17-rc1
if ! git fetch origin v4.17-rc1:v4.17-rc1 ; then
echo "warning"
fi
git checkout v4.17-rc1
The checkout succeeds just because of the git-fetch in the if-condition,
which is initially there to fetch the special refs from github PRs, or
gerrit reviews. That fails, but we just print a warning. If we were to
ever remove support for special refs, then the checkout would fail.
The whole purpose of the git cache is to actually save bandwidth and
download time, but in the long run. For one-offs, people would
preferably use a wget download (e.g. with the github macro) instead of
a git clone.
We switch to always doing a full clone. It is more correct, and pays off
in the long run...
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Maxime Hadjinlian <maxime.hadjinlian@gmail.com>
Cc: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Cc: Ricardo Martincoski <ricardo.martincoski@gmail.com>
Cc: Arnout Vandecappelle <arnout@mind.be>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
When a git tree has had sub-dir <-> sub-module conversions, or has had
submodules added or removed over the course of time, checking out a
changeset across those conversions/additions/removals may leave
untracked files, or may fail because of a conflict of type.
So, before we checkout the new changeset, we forcibly remove the
submodules. The new set of submodules, if any, will be restored later.
Ideally, we would use a native git command: git submodule deinit --all.
However, that was only introduced in git 1.8.3 which, while not being
recent by modern standards, is still too old for some enterprise-grade
distributions (RHEL6 only has git-1.7.1).
So, instead, we just use git submodule foreach, to rm -rf the submodules
directory.
Again, we would ideally use 'cd $toplevel && rm -rf $path', but
$toplevel was only introduced in git 1.7.2. $path has always been there.
So, instead, we just cd back one level, and remove the basename of the
directory.
Eventually, we need to get rid of now-empty and untracked directories,
that were parents of a removed submodule. For example. ./foo/bar/ was a
submodule, so ./foo/bar/ was removed, which left ./foo/ around.
Yet again, recent-ish git versions would have removed it during the
forced checkout, but old-ish versions (e.g. 1.7.1) do not remove it with
the forced checkout.
Instead we rely on the already used forced-forced clean of directories,
untracked, and ignored content, to really get rid of extra stuff we are
not interested in.
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Maxime Hadjinlian <maxime.hadjinlian@gmail.com>
Cc: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Cc: Ricardo Martincoski <ricardo.martincoski@gmail.com>
Cc: Arnout Vandecappelle <arnout@mind.be>
Reviewed-by: Ricardo Martincoski <ricardo.martincoski@gmail.com>
Tested-by: Ricardo Martincoski <ricardo.martincoski@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Force the checkout to ignore and throw away any local changes. This
allows recovering from a previous partial checkout (e.g. killed by
the user, or by a CI job...)
git checkout -f has been supported since the inception of git, so we
can use it without any second thought.
Also do a forced-forced clean, to really get rid of all untracked stuff.
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Maxime Hadjinlian <maxime.hadjinlian@gmail.com>
Cc: Ricardo Martincoski <ricardo.martincoski@gmail.com>
Cc: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Cc: Arnout Vandecappelle <arnout@mind.be>
Reviewed-by: Ricardo Martincoski <ricardo.martincoski@gmail.com>
Tested-by: Ricardo Martincoski <ricardo.martincoski@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
In some cases, the repository may be in a state we can't automatically
recover from, especially since we must still support oldish git versions
that do not provide the necessary commands or options thereof.
As a last-ditch recovery, delete the repository and recreate the cache
from scratch.
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Maxime Hadjinlian <maxime.hadjinlian@gmail.com>
Cc: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Cc: Ricardo Martincoski <ricardo.martincoski@gmail.com>
Cc: Arnout Vandecappelle <arnout@mind.be>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Check that the given cset is indeed something we can checkout. If not,
then exit early.
This will be useful when a later commit will trap any failing git
command to try to recover the repository by doing a clone from scratch:
when the cset is not a commit, it does not mean the repository is broken
or what, and re-cloning from scratch would not help, so no need to trash
a good cache.
Reported-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Cc: Maxime Hadjinlian <maxime.hadjinlian@gmail.com>
Cc: Ricardo Martincoski <ricardo.martincoski@gmail.com>
Cc: Arnout Vandecappelle <arnout@mind.be>
Reviewed-by: Ricardo Martincoski <ricardo.martincoski@gmail.com>
Tested-by: Ricardo Martincoski <ricardo.martincoski@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
That way, we can pushd earlier, which will help with last-ditch recovery
in a followup commit.
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Maxime Hadjinlian <maxime.hadjinlian@gmail.com>
Cc: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Cc: Ricardo Martincoski <ricardo.martincoski@gmail.com>
Cc: Arnout Vandecappelle <arnout@mind.be>
Reviewed-by: Ricardo Martincoski <ricardo.martincoski@gmail.com>
Tested-by: Ricardo Martincoski <ricardo.martincoski@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
We really want the user not to use our git cache manually, or their
changes (committed or not) may eventually get lost.
So, add a warning file, not unlike the one we put in the target/
directory, to warn the user not to use the git tree.
Ideally, we would have carried this file in support/misc/, but the git
backend does not have access to it: the working directory is somewhere
unknown, and TOPDIR is not exported in the environment.
So, we have to carry it in-line in the backend instead.
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Maxime Hadjinlian <maxime.hadjinlian@gmail.com>
Cc: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Cc: Ricardo Martincoski <ricardo.martincoski@gmail.com>
Cc: Arnout Vandecappelle <arnout@mind.be>
Reviewed-by: Ricardo Martincoski <ricardo.martincoski@gmail.com>
Tested-by: Ricardo Martincoski <ricardo.martincoski@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
When the run-time tests to build rust and rust-bin packages are run via Docker,
the $USER environment variable is not set, which makes cargo fail when
initializing the test project.
So add it to make cargo happy.
Signed-off-by: Eric Le Bihan <eric.le.bihan.dev@free.fr>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
This script causes a large number of flake8 warnings, is rarely used
(but even never used), and is going to be replaced at some point by
the improved pkg-stats that will give details about the upstream
version available for all packages, not just X.org packages.
Therefore, let's drop the xorg-release script in order to silence all
those flake8 warnings:
support/scripts/xorg-release:36:1: E302 expected 2 blank lines, found 1
support/scripts/xorg-release:58:27: E201 whitespace after '{'
support/scripts/xorg-release:58:44: E203 whitespace before ':'
support/scripts/xorg-release:58:54: E202 whitespace before '}'
support/scripts/xorg-release:63:1: E305 expected 2 blank lines after class or function definition, found 1
support/scripts/xorg-release:64:15: E261 at least two spaces before inline comment
support/scripts/xorg-release:67:32: E261 at least two spaces before inline comment
support/scripts/xorg-release:86:1: E302 expected 2 blank lines, found 1
support/scripts/xorg-release:95:1: E302 expected 2 blank lines, found 1
support/scripts/xorg-release:107:1: E302 expected 2 blank lines, found 1
support/scripts/xorg-release:115:20: W601 .has_key() is deprecated, use 'in'
support/scripts/xorg-release:123:34: E201 whitespace after '{'
support/scripts/xorg-release:124:46: E203 whitespace before ':'
support/scripts/xorg-release:124:50: E202 whitespace before '}'
support/scripts/xorg-release:127:1: E302 expected 2 blank lines, found 1
support/scripts/xorg-release:141:15: W601 .has_key() is deprecated, use 'in'
support/scripts/xorg-release:146:21: W601 .has_key() is deprecated, use 'in'
support/scripts/xorg-release:176:1: E305 expected 2 blank lines after class or function definition, found 1
support/scripts/xorg-release:180:1: W391 blank line at end of file
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Acked-by: Bernd Kuhls <bernd.kuhls@t-online.de>
Tested-by: Ricardo Martincoski <ricardo.martincoski@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
The version of the ARM Trusted Firmware from Marvell was a Git branch,
not a Git commit, leading to unreproducible results. So let's use a
Git commit instead, which is the latest available from the branch that
was previously used.
More specifically, this branch has recently seen a fix that is needed
for ATF to build properly with recent gcc versions:
c96ec59f8b
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
This test case currently fails to build with:
./build/juno/release/bl1/context_mgmt.o: In function `cm_prepare_el3_exit':
context_mgmt.c:(.text.cm_prepare_el3_exit+0x54): undefined reference to `cm_set_next_context'
context_mgmt.c:(.text.cm_prepare_el3_exit+0x54): relocation truncated to fit: R_AARCH64_JUMP26 against undefined symbol `cm_set_next_context'
This issue has been fixed upstream in commit
10c252c14b7f446c0b49ef1aafbd5d37804577dd, available since v1.3. So
while we bump, let's bump to the latest version of ATF, v1.5.
Fixes:
https://gitlab.com/buildroot.org/buildroot/-/jobs/64360659
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Since commit 5563a1c6a4
("support/check-uniq-files: support weird locales and filenames"), the
'csv' Python module is no longer used by the check-uniq-files.
Due to this, flake8 complains with:
support/scripts/check-uniq-files:4:1: F401 'csv' imported but unused
Fix this by dropping the useless csv import.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
This commit updates all the pre-built Buildroot toolchains, which have
all been rebuilt with Buildroot as of commit 046c5e2. The initial
motivation for this update is that an upcoming bump of procps-ng uses
fopencookie(), which has only been introduced in musl 1.1.19, which
itself started being used in Buildroot after the 2018.02 release.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Older versions of tar (e.g. 1.27.1) incorrectly interpret the escaping
of the regexp separator, and generate broken tarballs.
For example, given the following transform expression:
--transform="s/^\.\//squashfs-e38956b92f738518c29734399629e7cdb33072d3\//"
the resulting paths in the generated tarball would be:
squashfs-e38956b92f738518c29734399629e7cdb33072d3\/
i.e. a directory which last character is indeed a '\'.
We fix that by using a separator which is very unlikely to occur in a
filename.
Fixes:
http://autobuild.buildroot.org/results/742/7427f34e5c9f6d043b0fe6ad2c66cc0f31d2b24f/
and probably a slew of others as well...
Take this opportunity to fix indentation on the following line
(leading spaces, not TABs).
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Ricardo Martincoski <ricardo.martincoski@gmail.com>
Cc: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Cc: Maxime Hadjinlian <maxime.hadjinlian@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Check:
- the daemon is started;
- a message is collected;
- the daemon does not issue a warning message on startup.
When the .conf file version does not match the package version a warning
message shows up on serial on every boot. This message is generated by
syslog-ng before it is running, so it is not logged to
/var/log/messages. So in order to test the message is generated, restart
the server. It makes the message appears on /var/log/messages (since the
server is already running) where its existence can be easily tested
using grep.
Signed-off-by: Ricardo Martincoski <ricardo.martincoski@datacom.ind.br>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
The different versions of git will behave in different ways when
fetching remote references, as summarised by the table below:
| ancient git | new git
--------------------------------------------------------------------
git fetch | fetch all refs but tags | fetches all refs but tags
git fetch -t | fetches only tags | fetch all refs and tags
(git-fetch may still fetch tags, but only if reachable from a branch)
So, to cover all the bases, we do a simple fetch, to be sure we have
branches, followed by the existing fetch -t, to get extra tags.
Fixes:
http://autobuild.buildroot.net/results/0a2/0a238a7f55ea56c33b639ad03ed5796143426889/build-end.log
Reported-by: Ricardo Martincoski <ricardo.martincoski@gmail.com>
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Ricardo Martincoski <ricardo.martincoski@gmail.com>
Cc: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Cc: Maxime Hadjinlian <maxime.hadjinlian@gmail.com>
Cc: Arnout Vandecappelle <arnout@mind.be>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
There are cases where a repository might be broken, e.g. when a previous
operation was killed or otherwise failed unexpectedly.
We fix that by always initialising the repository, as suggested by
Ricardo. git-init is safe on an otherwise-healthy repository:
Running git init in an existing repository is safe. It will not
overwrite things that are already there. [...]
Using git-init will just ensure that we have the strictly required files
to form a sane tree. Any blob that is still missing would get fetched
later on.
Reported-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Reported-by: Ricardo Martincoski <ricardo.martincoski@gmail.com>
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Ricardo Martincoski <ricardo.martincoski@gmail.com>
Cc: Maxime Hadjinlian <maxime.hadjinlian@gmail.com>
Cc: Arnout Vandecappelle <arnout@mind.be>
Cc: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Acked-by: Ricardo Martincoski <ricardo.martincoski@gmail.com>
Tested-by: Ricardo Martincoski <ricardo.martincoski@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
git always look directories up until it finds a repository. In case
the git cache is broken, it may no longer be identified as a repository,
and git will look higher in the directories until it finds one.
In the default conditions, this would be Buildroot's own git tree
(because DL_DIR is a subdir of Buildroot), but in some situations may
very well be any repository the user has Buildroot in, like a
br2-external tree...
So, we force git to use our git cache and never look elsewhere, as
Suggested by Ricardo.
Use GIT_DIR, as it has been there for ages now, while --git-dir was
only introduced later (even if most distros ship an later version),
as suggested by Arnout.
Also fix the one call to git that was not using the wrapper.
Reported-by: Ricardo Martincoski <ricardo.martincoski@gmail.com>
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Ricardo Martincoski <ricardo.martincoski@gmail.com>
Cc: Maxime Hadjinlian <maxime.hadjinlian@gmail.com>
Cc: Arnout Vandecappelle <arnout@mind.be>
Cc: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Acked-by: Ricardo Martincoski <ricardo.martincoski@gmail.com>
Tested-by: Ricardo Martincoski <ricardo.martincoski@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
In preparation for the removal of the Blackfin architecture, drop the
autobuilder toolchain configuration that was testing Blackfin.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Reviewed-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
${raw_name} is never defined in dl-wrapper, and therefore the value
passed to the -N option is always empty. This causes a problem for the
'cvs' backend, which uses the value of this option as the CVS module
to be downloaded.
If the name of the CVS module is omitted, all the CVS modules from
that CVS repository are downloaded, which creates a tarball with a lot
more contents, and the actual useful contents in a sub-directory,
obviously breaking patches that should be applied, and the entire
build process that follows.
Fixes:
http://autobuild.buildroot.net/results/fcee0e3d7eeeb373313b1794092c729b1b052348/
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Tested-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Acked-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
When calling the backend-specific helper scripts, the remaining
options are in ${@}. However, in order to let the helper script know
that those remaining options should not be parsed, but instead passed
as-is to the download tool, they must be separated from the main
options by "--".
Without this, packages that use <pkg>_DL_OPTS, such as the
amd-catalyst package, cannot download their tarball anymore.
Fixes:
http://autobuild.buildroot.net/results/de818f6e4c8e63d5e8a49c445d10c34eccc40410/
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Tested-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Acked-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
The /lib/firmware directory contains random firmware for various
devices. It happens that some of them might be or appear to be ELF
files, but they shouldn't be checked by fix-rpath. For example, one of
the Qualcomm VPU firmware file appears to be an ELF file, but patchelf
isn't happy about it:
$ ./output/host/bin/patchelf --print-rpath output/target/lib/firmware/qcom/venus-4.2/venus.b00
patchelf: patchelf.cc:387: void ElfFile<Elf_Ehdr, Elf_Phdr, Elf_Shdr, Elf_Addr, Elf_Off, Elf_Dyn, Elf_Sym>::parse() [with Elf_Ehdr = Elf32_Ehdr; Elf_Phdr = Elf32_Phdr; Elf_Shdr = Elf32_Shdr; Elf_Addr = unsigned int; Elf_Off = unsigned int; Elf_Dyn = Elf32_Dyn; Elf_Sym = Elf32_Sym]: Assertion `shstrtabIndex < shdrs.size()' failed.
Aborted (core dumped)
Even though patchelf definitely shouldn't crash, it anyway doesn't
make sense to check ELF files in /lib/firmware, so let's exclude this
directory from our check.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Acked-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
In case the git backend gets killed right in-between it finished
initialising the repository, but before it could add the remote,
we'd end up with a repository without the 'origin' remote, so we
would not be able to change its URL.
Another case that may happen (like in the build failure, below),
is that the repository was initialised with a previous version
of Buildroot, before the commit e17719264b (download/git: don't
require too-recent git) was applied, and that trepository was
still lying around...
Fixes:
http://autobuild.buildroot.org/results/25a/25aae054634368fadb265b97ebe4dda809deff6f/
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Maxime Hadjinlian <maxime.hadjinlian@gmail.com>
Cc: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Cc: Peter Korsgaard <peter@korsgaard.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
git has supported -C only since 1.8.5, and some distros have not yet
caught up after more than 4 years...
Fall back to entering the directory.
Fixes:
http://autobuild.buildroot.net/results/35f9f7a4adc6c2cad741079e4afdf1408c94703b
Reported-by: André Hentschel <nerv@dawncrow.de>
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Maxime Hadjinlian <maxime.hadjinlian@gmail.com>
Cc: André Hentschel <nerv@dawncrow.de>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
When a package contains a relative symlink which first component is '..'
(thus pointing one directory higher), for example package 'meh' contains
this symlink:
foo/bar -> ../buz
then it would be stored as 'meh-version./buz' because of the
transform-name pattern replacement.
Fix it to only match the leading './'.
Reported-by: Ricardo Martincoski <ricardo.martincoski@gmail.com>
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Ricardo Martincoski <ricardo.martincoski@gmail.com>
Cc: Peter Korsgaard <peter@korsgaard.com>
Cc: Maxime Hadjinlian <maxime.hadjinlian@gmail.com>
Reviewed-by: Ricardo Martincoski <ricardo.martincoski@gmail.com>
Tested-by: Ricardo Martincoski <ricardo.martincoski@gmail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Since c8ef0c03b0 (download: put most of the infra in dl-wrapper), the
backend for local files is now named after the scheme, which is 'file'
for a local file.
>From the same commit on, the directory part and the basename are now
passed separately, to let the backend reconstruct the full path when it
needs to do so, which is the case for the 'file' backend too.
Finaly, ff559846fd (support/download: Add support to pass options
directly to downloaders) introduced a nasty error, as it made use of
"${@}" when calling its internal function. Revert that mess now...
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Maxime Hadjinlian <maxime.hadjinlian@gmail.com>
Cc: Peter Korsgaard <peter@korsgaard.com>
Cc: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Cc: Arnout Vandecappelle <arnout@mind.be>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Commit "6d938bcb52 download: git: introduce cache feature" introduced a
typo that makes the tarball to contain files without the package
basename:
$ tar -tvf good-a238b1dfcd825d47d834af3c5223417c8411d90d.tar.gz
-rw-r--r-- 0/0 8 2017-10-14 02:10 ./file
Historically, all tarballs are generated with the basename:
$ tar -tvf good-a238b1dfcd825d47d834af3c5223417c8411d90d.tar.gz
-rw-r--r-- 0/0 8 2017-10-14 02:10 good-a238b1dfcd825d47d834af3c5223417c8411d90d/file
The hashes in the tree were calculated with the basename.
In the most common scenario, after the download ends the tarball is
generated, the hash mismatches and the download mechanism falls back to
use the tarball from http://sources.buildroot.net .
The problem can be reproduced by forcing the download of any git package
PKG that has a hash file to check against:
$ make defconfig
$ ./utils/config --set-str BR2_BACKUP_SITE ""
$ BR2_DL_DIR=$(mktemp -d) make PKG-dirclean PKG-source
Fix the typo so the basename is really added to the files, that was
clearly the intention of the code.
Signed-off-by: Ricardo Martincoski <ricardo.martincoski@gmail.com>
Cc: Arnout Vandecappelle <arnout@mind.be>
Cc: Maxime Hadjinlian <maxime.hadjinlian@gmail.com>
Cc: Peter Korsgaard <peter@korsgaard.com>
Cc: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Cc: Yann E. MORIN <yann.morin.1998@free.fr>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
This commit adds a new column in the HTML output containing the
current version of a package in Buildroot. As such, it isn't terribly
useful, but combined with the latest upstream version added in a
follow-up commit, it will become very useful.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Reviewed-by: Ricardo Martincoski <ricardo.martincoski@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
This commit adds the following options to the pkg-stats-new script:
-n, to specify a number of packages to parse instead of all packages
-p, to specify a list of packages (comma-separated) to parse instead
of all packages
These options are basically only useful when debugging/developing
this script, but they are very useful, because the script is rather
slow to run completely with all 2000+ packages, especially once
upstream versions will be fetched from release-monitoring.org.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Reviewed-by: Ricardo Martincoski <ricardo.martincoski@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
This commit adds a new version of the pkg-stats script, rewritten in
Python. It is for now implemented in a separate file called,
pkg-stats-new, in order to make the diff easily readable. A future
commit will rename it to pkg-stats.
Compared to the existing shell-based pkg-stats script, the
functionality and output is basically the same. The main difference is
that the output no longer goes to stdout, but to the file passed as
argument using the -o option. This allows stdout to be used for more
debugging related information.
The way the script works is that a first function get_pkglist()
returns a list of Package objects. Then, the function
package_init_make_info() uses 'make printvars' to gather information
about all packages, stored as class variables in the Package
class. Then, we iterate over all packages, and use various methods of
the Package class to retrieve all details about the package:
infrastructure, presence of hash file, presence of license
information, etc.
calculate_stats() then calculates global statistics (how packages have
license information, how packages have a hash file, etc.). Finally,
dump_html() produces the HTML output, using a number of sub-functions.
One improvement over the shell-based version is that we can use
regexps to exclude some .mk files. Thanks to this, we can exclude all
linux-ext-*.mk files, avoiding incorrect matches.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Reviewed-by: Ricardo Martincoski <ricardo.martincoski@gmail.com>
When the BR2_DL_DIR is a mountpoint (presumably shared between various
machine, or mounted from the local host when running in a VM), it is
possible that it does not support hardlinks (e.g. samba, or the VMWare
VMFS, etc...).
If the hardlink fails, fallback to copying the file. As a last resort,
if that also fails, eventually fallback to doing the download.
Note: this means that the dl-wrapper is no longer atomic-safe: the code
suffers of a TOCTTOU condition: the file may be created in-between the
check and the moment we try to ln/cp it. Fortunately, the dl-wrapper is
now run under an flock, so we're still safe. If we eventually go for a
more fine-grained implementation, we'll have to be careful then.
Reported-by: Arnout Vandecappelle <arnout@mind.be>
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Arnout Vandecappelle <arnout@mind.be>
Cc: Peter Korsgaard <peter@korsgaard.com>
Cc: Maxime Hadjinlian <maxime.hadjinlian@gmail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Now we keep the git clone that we download and generates our tarball
from there.
The main goal here is that if you change the version of a package (say
Linux), instead of cloning all over again, you will simply 'git fetch'
from the repo the missing objects, then generates the tarball again.
This should speed the 'source' part of the build significantly.
The drawback is that the DL_DIR will grow much larger; but time is more
important than disk space nowadays.
Signed-off-by: Maxime Hadjinlian <maxime.hadjinlian@gmail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
For existing setups, the global donload directory may have a lot of the
required archives, so look into there before attempting a download.
We simply hard-link them if found there and not in the new per-package
loaction. Then we resume the existing procedure (which means the new
hardlink will get removed if it happened to not match the hash).
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Maxime Hadjinlian <maxime.hadjinlian@gmail.com>
Cc: Peter Korsgaard <peter@korsgaard.com>
Cc: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
The infrastructure needs to give the 'dl_dir' to the dl-wrapper which in its
turn needs to give it to the helper. It will only be used by the 'git'
helper as of now.
Signed-off-by: Maxime Hadjinlian <maxime.hadjinlian@gmail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Fixes:
http://autobuild.buildroot.net/results/b18/b187e64a61918f17f69588e2355a03286bc5808e
tar 1.27 subtly changed the tar format when a GNU long link entry is added
(which is done for path elements > 100 characters). The code used to set
the permission mode of the link entry to 0:
header = start_private_header ("././@LongLink", size, time (NULL));
FILL (header->header.mtime, '0');
FILL (header->header.mode, '0');
FILL (header->header.uid, '0');
FILL (header->header.gid, '0');
FILL (header->header.devmajor, 0);
FILL (header->header.devminor, 0);
This got dropped in 1.27 by commit df7b55a8f6354e3 (Fix some problems with
negative and out-of-range integers), so the settings from
start_private_header() are used directly - Which are:
TIME_TO_CHARS (t < 0 ? 0 : min (t, MAX_OCTAL_VAL (header->header.mtime)),
header->header.mtime);
MODE_TO_CHARS (S_IFREG|S_IRUSR|S_IWUSR|S_IRGRP|S_IROTH, header->header.mode);
UID_TO_CHARS (0, header->header.uid);
GID_TO_CHARS (0, header->header.gid);
The end result is that tar >= 1.27 sets mode to 644.
The consequence of this is that we create different tar files when long path
names are encountered (which often happens when a package downloads a
specific sha1 from a git repo) depending on the host tar version used,
causing hash mismatches.
As a workaround, bump our minimum tar version to 1.27. It would be nicer to
only do this if we have packages from bzr/git/hg enabled, but that is an
exercise for later.
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Acked-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
The goal here is to simplify the infrastructure by putting most of the
code in the dl-wrapper as it is easier to implement and to read.
Most of the functions were common already, this patch finalizes it by
making the pkg-download.mk pass all the parameters needed to the
dl-wrapper which in turn will pass everything to every backend.
The backend will then cherry-pick what it needs from these arguments
and act accordingly.
It eases the transition to the addition of a sub directory per package
in the DL_DIR, and later on, a git cache.
[Peter: drop ';' in BR_NO_CHECK_HASH_FOR in DOWNLOAD macro and swap cd/rm
-rf as mentioned by Yann, fix typos]
Signed-off-by: Maxime Hadjinlian <maxime.hadjinlian@gmail.com>
Tested-by: Luca Ceresoli <luca@lucaceresoli.net>
Reviewed-by: Luca Ceresoli <luca@lucaceresoli.net>
Reviewed-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Currently all download helpers accepts the local output file, the remote
locations, the changesets and so on... as positional arguments.
This was well and nice when that's was all we needed.
But then we added an option to quiesce their verbosity, and that was
shoehorned with a trivial getopts, still keeping all the existing
positional arguments as... positional arguments.
Adding yet more options while keeping positional arguments will not be
very easy, even if we do not envision any new option in the foreseeable
future (but 640K ought to be enough for everyone, remember? ;-) ).
Change all helpers to accept a set of generic options (-q for quiet and
-o for the output file) as well as helper-specific options (like -r for
the repository, -c for a changeset...).
Maxime:
Changed -R to -r for recurse (only for the git backend)
Changed -r to -u for URI (for all backend)
Change -R to -c for cset (for CVS and SVN backend)
Add the export of the BR_BACKEND_DL_GETOPTS so all the backend wrapper
can use the same option easily
Now all the backends use the same common options.
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Maxime Hadjinlian <maxime.hadjinlian@gmail.com>
Cc: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Reviewed-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Reviewed-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
graph-depends currently spits out a graph in .dot format. However, as
part of the upcoming introduction of <pkg>-show-recursive-depends and
<pkg>-show-recursive-rdepends, we need graph-depends to be able to
display a flat list.
Signed-off-by: George Redivo <george.redivo@datacom.ind.br>
[Thomas:
- Rebase on top of graph-depends changes
- Do not display the package name itself in the list, only its
dependencies (or reverse dependencies)
- Display the result on a single line, instead of one package per
line, in order to match what <pkg>-show-depends and
<pkg>-show-rdepends are doing today.]
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
This will be useful for the upcoming recursive show-depends and
show-rdepends features.
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Reviewed-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Instead of hardcoded sys.stderr.write() calls. No functional change, but
allows us to easily implement a quiet option.
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Reviewed-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
The graph-depends was not very consistent in colors vs. colours: some
parts were using colours, some parts were using colors.
Let's settle on the US spelling, colors.
This change the user-visble option --colours to --colors, but it is
unlikely that a lot of users customize the colors through
BR2_GRAPH_DEPS_OPTS, so this user interface change is considered
reasonable.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Reviewed-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
The graph-depends script had no main() function, and the main code was
actually spread between the function definitions, which was a real
mess.
This commit moves the global code into a main() function, which allows
to more easily follow the flow of the script. The argument parsing
code is moved into a parse_args() function.
Most of the global variables are removed, and are instead passed as
argument when appropriate. This has the side-effect that the
print_pkg_deps() function takes a lot of argument, but this is
considered better than tons of global variables.
The global variables that are removed are: max_depth, transitive,
mode, root_colour, target_colour, host_colour, outfile, dict_deps,
dict_version, stop_list, exclude_list, arrow_dir.
The root_colour/target_colour/host_colour variables are entirely
removed, and instead a single colours array is passed, and it's the
function using the colors that actually uses the different entries in
the array.
The way the print_attrs() function determines if we're display the
root node is not is changed. Instead of relying on the package name
and the mode (which requires passing the root package name, and the
mode), it relies on the depth: when the depth is 0, we're at the root
node.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Some test cases don't use a full build as setup, so split the build()
method into configure() and build().
It allows a test case to perform configuration at the setup stage and
the build inside the test itself.
Call this new method just before build in the BRTest base class, to keep
the current behavior for existing test cases.
This change will be needed when adding a common class to test the git
download infra.
Signed-off-by: Ricardo Martincoski <ricardo.martincoski@datacom.ind.br>
Cc: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Use an empty environment when calling make, but import PATH so the
buildroot tree under test can find binaries from the host machine.
Since environment variables are now ignored, move the handling of
BR2_DL_DIR to the defconfig to keep the current precedence of -d:
BR2_DL_DIR | -d DIR | test downloads | BR downloads
------------+----------+------------------+--------------
unset | unset | [error] | [error]
unset | set | in $(DIR) | in $(DIR)
set | unset | in $(BR2_DL_DIR) | in $(BR2_DL_DIR)
set | set | in $(DIR) | in $(DIR)
Signed-off-by: Ricardo Martincoski <ricardo.martincoski@gmail.com>
Cc: Arnout Vandecappelle <arnout@mind.be>
Cc: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Cc: Yann E. MORIN <yann.morin.1998@free.fr>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
This commit reorganizes the toolchain-configs.csv so that the first
toolchains are a subset of "useful" toolchains to be tested by
contributors to validate a package. This subset is the one that will
be used by default by test-pkg.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Reviewed-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
In the situation where the hash is missing from the hash file, the
dl-wrapper downloads the file again and again until the developer
specifies the hash to complete the download step.
To avoid this situation, the freshly-downloaded file is not removed
anymore after a successful download.
After this change, the behaviour is as follows:
- Hash file doesn't exist, or file is in BR_NO_CHECK_HASH_FOR
=> always succeeds.
- Hash file exists, but file is not present
=> file is NOT removed, build is terminated immediately (i.e.
secondary site is not tried).
- Hash file exists, file is present, but hash mismatch
=> file is removed, secondary site is tried.
=> If all primary/secondary site downloads or hash checks fail, the
build is terminated.
Signed-off-by: Gaël PORTAY <gael.portay@savoirfairelinux.com>
[Arnout: extend commit log]
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Currently, some packages may register hooks to be run just before and
just after the generic tarball image is generated, because they need to
prepare the filesystem for read-only or read-write operation.
However, this means that, if any of the hooks or the image generation
fails, the target directory is left in a dangling, inconsistent state.
We fix that by doing a copy of target/, run the hooks on that copy,
generate the generic tarball image out of that, and get rid of the copy.
This way, we can guarantee consistency of the target directory, and we
can even ditch support for post-fs hooks (those that restore target/).
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Cc: Arnout Vandecappelle <arnout@mind.be>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Currently, when a filename contains characters not representable in the
user's locale, we fail hard, especially when the host python is python3.
This is because python2 and python3 handle encoding/decoding strings
differently, with python3 presumable doing the right thing, but it
breaks on some systems, while python2 presumable does the wrong thing,
but it works everywhere. (Just joking, obviously...)
Part of the issue being that the csv reader in python2 is broken with
UTF8.
We fix the issue by ditching the csv reader, and simply read the file in
binary mode, manually partitioning the lines on the first comma.
Then, we use the binary-encoded (really, un-encoded) package names and
filenames as values and keys, respectively.
Finally, for each filename or package we need to print, we try to decode
them with the defaults for the user settings, but catch any decoding
exception and fall back to dumping the raw, binary values. Which codec
is used by default differs between Python version, but in all cases
something sane is printed at least.
Thanks a lot to Arnout for the live help doing this patch. :-)
Reported-by: Jaap Crezee <jaap@jcz.nl>
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Arnout Vandecappelle <arnout@mind.be>
Cc: Jaap Crezee <jaap@jcz.nl>
[Arnout: commit log improvement]
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Commit c868095681 ("toolchain: fix
detection of SSP support") fixed the SSP check so that it does the
correct thing for nios2 toolchains. While this commit fixed the
description of the Sourcery NIOSII toolchain, it didn't fix the
description for the autobuilders of the br-nios2-glibc toolchain,
causing some build failures. This commit adjusts br-nios2-glibc.config
to indicate that the toolchain doesn't have SSP support.
Fixes:
http://autobuild.buildroot.net/results/6c44e328b7bffd8474d29d5bdf1ea109ec15f4ad/
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Now that DEPENDENCIES_HOST_PREREQ is no longer used anywhere, we can
kill it.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Acked-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
This moves the host-ccache dependency handling from
DEPENDENCIES_HOST_PREREQ to a proper package dependency. When
BR2_CCACHE=y, we add host-ccache as a regular dependency of all
packages except:
- The extractor packages host-tar, host-xz and host-lzip
- host-ccache itself
- host-skeleton, because all packages depend on it
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Reviewed-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Tested-by: Matt Weber <matthew.weber@rockwellcollins.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
This moves the host-lzip dependency handling from
DEPENDENCY_HOST_PREREQ to an extract dependency.
To achieve that, check-host-lzip.mk fills in the
BR2_LZIP_HOST_DEPENDENCY with host-lzip if building a host-lzip is
needed. The name BR2_LZIP_HOST_DEPENDENCY has been chosen because it
matches the name BR2_CMAKE_HOST_DEPENDENCY already used in
check-host-cmake.mk.
The BR2_LZIP_HOST_DEPENDENCY is added to all packages, except:
- host-lzip, because we would otherwise depend on ourself.
- host-tar, because lzip itself is delivered as a tarball, so we need
to have host-lzip depend on host-tar, and not host-tar depend on
host-lzip
- host-skeleton, because we need to have host-lzip depend on
host-skeleton, and not the opposite.
We also mutually exclude host-lzip and host-xz from dependending on
each other, to avoid a circular dependency.
In addition, we modify lzip.mk to explicitly build host-lzip without
ccache. We generally took the approach of building host-ccache *after*
all the extractors have been built.
[Peter: fix s/host-tar/host-lzip/ typo, fix s/xz/lzip/ typo]
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Reviewed-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
This moves the host-xz dependency handling from
DEPENDENCY_HOST_PREREQ to an extract dependency.
To achieve that, check-host-xz.mk fills in the
BR2_XZ_HOST_DEPENDENCY with host-xz if building a host-xz is
needed. The name BR2_XZ_HOST_DEPENDENCY has been chosen because it
matches the name BR2_CMAKE_HOST_DEPENDENCY already used in
check-host-cmake.mk.
The BR2_XZ_HOST_DEPENDENCY is added to all packages, except:
- host-xz, because we would otherwise depend on ourself.
- host-tar, because xz itself is delivered as a tarball, so we need
to have host-xz depend on host-tar, and not host-tar depend on
host-xz
- host-skeleton, because we need to have host-xz depend on
host-skeleton, and not the opposite.
In addition, we modify xz.mk to explicitly build host-xz without
ccache. We generally took the approach of building host-ccache *after*
all the extractors have been built.
[Peter: fix s/host-tar/host-xz/ typo]
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Reviewed-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
This moves the host-tar dependency handling from
DEPENDENCY_HOST_PREREQ to an extract dependency.
To achieve that, check-host-tar.mk fills in the
BR2_TAR_HOST_DEPENDENCY variable with host-tar if building a host-tar
is needed. The name BR2_TAR_HOST_DEPENDENCY has been chosen because it
matches the name BR2_CMAKE_HOST_DEPENDENCY already used in
check-host-cmake.mk.
The BR2_TAR_HOST_DEPENDENCY is added to all packages, except host-tar
itself (obviously) and host-skeleton, because we depend on
host-skeleton to install host-tar properly in HOST_DIR.
In addition, we modify tar.mk to explicitly build host-tar without
ccache: since ccache source code is available as a tarball, ccache
will obviously depend on host-tar if the system tar is insufficient.
Finally, to make things really clean, we also add
$(BR2_TAR_HOST_DEPENDENCY) to the dependencies of the tar filesystem
format, since it requires tar, so we'd better make sure we have a
suitable tar.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Reviewed-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Tested-by: Matt Weber <matthew.weber@rockwellcollins.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Use the latest version of the tool because it is actively maintained.
But use a fixed version of the tool and its dependencies to get stable
results. It can be manually bumped from time to time.
Before installing any Python packages, ensure pip, setuptools, and wheel
are up to date as recommended in the docs [1].
[1] https://packaging.python.org/tutorials/installing-packages/
Signed-off-by: Ricardo Martincoski <ricardo.martincoski@gmail.com>
Cc: Arnout Vandecappelle <arnout@mind.be>
Cc: Peter Korsgaard <peter@korsgaard.com>
Cc: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Cc: Yann E. MORIN <yann.morin.1998@free.fr>
Acked-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
When using a merged /usr, the kernel module path is really
/usr/lib/modules, as /lib is a symlink to usr/lib .
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Cc: Peter Korsgaard <peter@korsgaard.com>
Reviewed-by: Peter Korsgaard <peter@korsgaard.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Some packages (mostly, out-of-tree) may want to install binary blobs for
another architecture, outside the locations we currently exclude, like
in /opt or whatever...
Add support in check-bin-arch to accept any arbitrary location, that
individual package can each request to excude from the check, when they
are installed.
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Peter Korsgaard <peter@korsgaard.com>
Cc: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Recent systemd bump has broken DBus dameon and DBus applications can no
longer find the daemon. So we want to catch those kind of failures
early.
We also want to check that the system as a whole is stable: no unit
should be failed.
Finally, ensure that we can read the jounrnal, even when we are doing our
tricks on read-only systems.
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Its use has been globbed into the more generic
BR2_NEEDS_HOST_UTF8_LOCALE option now.
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Cc: Peter Korsgaard <peter@korsgaard.com>
Cc: Julius Kriukas <julius@kriukas.lt>
Cc: Christian Stewart <kidovate@gmail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Although the UTF-8 locales in mainstream distributions all are suffixed
with just 'utf8', the nomenclature is a bit ambiguous with the way they
are to be specified with the various LC_* variables, suffixed there with
'UTF-8'.
Also, POSIX, ISO, and IEC do not enforce any specific suffix in LC_*
variables:
http://pubs.opengroup.org/onlinepubs/9699919799/basedefs/V1_chap08.html#tag_08_02
"""
If the locale value has the form:
language[_territory][.codeset]
it refers to an implementation-provided locale, where settings of
language, territory, and codeset are implementation-defined.
"""
To avoid any confusion, use a regexp that is a bit more lax when
matching locales.
Also, quote the regexp, so that the '?' and '$' are not interpreted by
the shell.
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Peter Korsgaard <peter@korsgaard.com>
Cc: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
The script support/scripts/check-uniq-files uses the argparse Python
module. In most recent Python versions (starting with 2.7), the
argparse module is part of the standard library, and we already check
for the availability of Python in
support/dependencies/dependencies.sh.
However, when running on an ancient distribution with Python 2.6, the
argparse module is not part of the Python standard library, but
available as an external module. Without this module, the build fails,
because check-uniq-files, which is used in target-finalize, fails to
run.
To avoid this failure, this commit adds a check in
support/dependencies/dependencies.sh to verify that the argparse
module is available.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Acked-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Using {} in format strings is only supported in sufficiently recent
Python versions. Python 2.6 doesn't support this, and only format
strings with numbered arguments: {0}, {1}, etc.
Python 2.7:
$ python -c 'print("foo {}".format(12))'
foo 12
$ python -c 'print("foo {0}".format(12))'
foo 12
Python 2.6:
$ python -c 'print("foo {}".format(12))'
Traceback (most recent call last):
File "<string>", line 1, in <module>
ValueError: zero length field name in format
$ python -c 'print("foo {0}".format(12))'
foo 12
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Acked-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
When BR2_REPRODUCIBLE is enabled or host uses non UTF-8 capable locale
building systemd fails with an error:
[1/1080] Generating systemd.bg.catalog with a meson_exe.py custom command.
FAILED: catalog/systemd.bg.catalog
/buildroot/output/host/bin/python3 /buildroot/output/host/bin/meson --internal exe /buildroot/output/build/systemd-236/build/meson-private/meson_exe_sed_232a0623cc7ce2cd67ec72ed784b76307102ed76.dat
Warning: You are using 'ANSI_X3.4-1968' which is not a Unicode-compatible locale.
You might see errors if you use UTF-8 strings as filenames, as strings, or as file contents.
Please switch to a UTF-8 locale for your platform.
...
UnicodeDecodeError: 'ascii' codec can't decode byte 0xd0 in position 1079: ordinal not in range(128)
package/pkg-generic.mk:247: recipe for target '/buildroot/output/build/systemd-236/.stamp_built' failed
make: *** [/buildroot/output/build/systemd-236/.stamp_built] Error 1
This patch changes default host system locale from C to C.UTF-8 when
building systemd package to fix this issue. It also introduces
BR2_NEEDS_HOST_C_UTF8_LOCALE flag that checks if this locale is available on
the host system. If locale is not available error message is show and build
process is stopped.
Signed-off-by: Julius Kriukas <julius@kriukas.lt>
[Thomas: use C.UTF-8 instead of en_US.UTF-8.]
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Tar 1.30 changed the --numeric-owner output for filenames > 100 characters,
leading to hash mismatches for the tar archives we create ourselves from
git. This is really a fix for a bug in earlier tar versions regarding
deterministic output, so it is unlikely to be reverted in later versions.
For more details, see:
http://lists.busybox.net/pipermail/buildroot/2018-January/211222.html
To work around this issue, blacklist tar 1.30+ similar to how we do it for
pre-1.17 versions so Buildroot falls back to building host-tar.
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
host-{cmake,lzip,xz} needs host-tar to extract their source code tarball, so
we need to ensure that host-tar gets added to DEPENDENCIES_HOST_PREREQ
before these in case they are both needed, otherwise the tools will fail to
extract.
With the upcoming change to blacklist modern tar versions this situation is
likely to trigger more often.
The real solution to this issue is the <foo>_EXTRACT_DEPENDENCIES rework,
but that series is a bit too intrusive to add this close to 2018.02, so
therefore this hack.
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Reviewed-by: Matt Weber <matthew.weber@rockwellcollins.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
The script check-bin-arch fails as follows on a config for PowerPC e6500
(64-bit CPU) with BR2_ARCH="powerpc" (32-bit userland desired):
ERROR: architecture for "/lib/modules/..../lib/libcrc32c.ko"
is "PowerPC64", should be "PowerPC"
This situation is perfectly acceptable: the kernel is 64-bit and so are its
modules, even though userland is 32-bit.
To keep check-bin-arch and its caller simple, just skip /lib/modules/
entirely, like is done for /lib/firmware and some others.
Signed-off-by: Thomas De Schampheleire <thomas.de_schampheleire@nokia.com>
Reviewed-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Currently 'run-tests -l' is broken. It breaks 'make .gitlab-ci.yml' that
in turn breaks the job in GitLab.
TestRustBase is not a test case by itself, so it can't have a method
with the name starting with "test_" otherwise nose2 assumes it is a test
case.
Move the test_run method from the base class to the derived classes.
While at it, update .gitlab-ci.yml with the new test cases.
Fixes:
https://gitlab.com/buildroot.org/buildroot/-/jobs/52000035
Reported-by: Yegor Yefremov <yegorslists@googlemail.com>
Signed-off-by: Ricardo Martincoski <ricardo.martincoski@gmail.com>
Cc: Eric Le Bihan <eric.le.bihan.dev@free.fr>
Cc: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Cc: Yann E. MORIN <yann.morin.1998@free.fr>
Cc: Yegor Yefremov <yegorslists@googlemail.com>
Reviewed-by: Yegor Yefremov <yegorslists@googlemail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Some packages build C++ 32bits host-tools and need the g++-multilib to
be installed on the build machine. As example, qt5webengine builds a C++
host-tool when target is 32bits.
Add the check for g++-multilib to the dependencies script; and update
the Dockerfile to install g++-multilib package.
Signed-off-by: Gaël PORTAY <gael.portay@savoirfairelinux.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
While we have several defconfigs building internal toolchains with
uClibc, we don't have any building internal toolchain with glibc and
musl. However, having such defconfigs is nice when we bump the C
library version, in order to immediately get feedback on build
failures.
Note that while the ARC internal defconfig uses glibc, it uses the
special ARC glibc version, so it doesn't test version bumps of the
upstream glibc C library.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
To test the support for the Rust language, the following tests are added:
- building Rust compiler and Cargo from source.
- installing a pre-built Rust compiler and building Cargo from source.
For each test, a Rust test program is built and installed in the root file
system of a ARM vexpress QEMU system. The test is declared OK if the program can
be run properly from the test system.
Signed-off-by: Eric Le Bihan <eric.le.bihan.dev@free.fr>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Currently, we refer to the latest version of the image, which means we
can't guarantee any reproducibility. Also, it measn we can't have a
separate images for the maintenance branches (especially the LTS) and
master.
Update the comment in the Dockerfile to create and push tagged images.
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Arnout Vandecappelle <arnout@mind.be>
Cc: Peter Korsgaard <peter@korsgaard.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Since we're now using a specific base image tag, we need to also use a
specific, stable repository to get additional packages from for this
image.
As such, use the Debian snapshot that matches the base image.
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Arnout Vandecappelle <arnout@mind.be>
Cc: Peter Korsgaard <peter@korsgaard.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Currently, we are using debian:stable, which is subject to change with
time, as new stable versions of Debian are released/updated.
Use the latest tagged stable release, stretch-20171210 as of today, as
the base distribution to use.
This will ease reproducible builds in the future.
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Arnout Vandecappelle <arnout@mind.be>
Cc: Peter Korsgaard <peter@korsgaard.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
This image is not built very often, and when it is, it is important to
see what's going on, so don't be silent when installing packages from
the distro, and since that can take a bit of time it thus serves as
progress report...
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Arnout Vandecappelle <arnout@mind.be>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
The official documentation [0] suggests limiting the number of layers
generated from a dockerfile. A layer is created for each RUN (and COPY
and ADD) command. But we are only ever interested in the final image,
so the intermediate layers are useless to us.
Limit the number of RUN commands to limit the number of generated
layers.
[0] https://docs.docker.com/develop/develop-images/dockerfile_best-practices/#minimize-the-number-of-layers
Reported-by: Peter Korsgaard <peter@korsgaard.com>
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Peter Korsgaard <peter@korsgaard.com>
Cc: Arnout Vandecappelle <arnout@mind.be>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Fix these warnings:
E122 continuation line missing indentation or outdented
E302 expected 2 blank lines, found 1
F401 'os' imported but unused
Signed-off-by: Ricardo Martincoski <ricardo.martincoski@gmail.com>
Cc: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Fix these warnings:
E302 expected 2 blank lines, found 1
E305 expected 2 blank lines after class or function definition, found 1
E713 test for membership should be 'not in'
Signed-off-by: Ricardo Martincoski <ricardo.martincoski@gmail.com>
Cc: Thomas De Schampheleire <thomas.de_schampheleire@nokia.com>
Cc: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Fix these warnings:
E302 expected 2 blank lines, found 1
E305 expected 2 blank lines after class or function definition, found 1
Signed-off-by: Ricardo Martincoski <ricardo.martincoski@gmail.com>
Cc: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Cc: Yegor Yefremov <yegorslists@googlemail.com>
Reviewed-by: Yegor Yefremov <yegorslists@googlemail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Fix these warnings:
E201 whitespace after '['
E202 whitespace before ']'
E302 expected 2 blank lines, found 1
E305 expected 2 blank lines after class or function definition, found 1
Ignore these warnings:
E402 module level import not at top of file
Signed-off-by: Ricardo Martincoski <ricardo.martincoski@gmail.com>
Cc: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Cc: Yann E. MORIN <yann.morin.1998@free.fr>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Fix these warnings:
E128 continuation line under-indented for visual indent
E302 expected 2 blank lines, found 1
E305 expected 2 blank lines after class or function definition, found 1
Signed-off-by: Ricardo Martincoski <ricardo.martincoski@gmail.com>
Cc: Yann E. MORIN <yann.morin.1998@free.fr>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Fix these warnings:
E122 continuation line missing indentation or outdented
E127 continuation line over-indented for visual indent
E128 continuation line under-indented for visual indent
E202 whitespace before ']'
E221 multiple spaces before operator
E225 missing whitespace around operator
E231 missing whitespace after ','
E302 expected 2 blank lines, found 1
E305 expected 2 blank lines after class or function definition, found 1
E502 the backslash is redundant between brackets
E713 test for membership should be 'not in'
Signed-off-by: Ricardo Martincoski <ricardo.martincoski@gmail.com>
Cc: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Pass BR2_EXTERNAL value via -e option.
This will prevent merge_config.sh from silently eating any symbols defined in
external trees on a clean buildroot tree invocation.
Signed-off-by: Ilya Kuzmich <ilya.kuzmich@gmail.com>
Acked-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Currently, only post-build and post-image scripts were tested, each with
their own test-script.
The two test-scripts only differ in the name of the log file they
create, and it is based on the name of the script, so it is easy to
share the script.
This allows us to easily re-use it for testing post-fakeroot scripts.
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Cc: Ricardo Martincoski <ricardo.martincoski@gmail.com>
Reviewed-by: Ricardo Martincoski <ricardo.martincoski@gmail.com>
Tested-by: Ricardo Martincoski <ricardo.martincoski@gmail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Buildroot's "make nconfig" command stopped working a while ago on
Gentoo systems. Running the command would result in a crash.
The issue is caused by lxdialog's cflags which are also used to build
nconfig; It would detect *ncursesw* and turn on WIDECHAR support --
but the Makefile would still link to plain *ncurses* while building
nconfig (which was built without WIDECHAR support).
This would cause a crash after using *wattrset* on a WINDOW instance.
WIDECHAR *wattrset* would try to set the _color member in the WINDOW
struct which does not exist in the NON-WIDECHAR ncurses instance. It
would end up clobbering data outside the struct (usually _line entries).
An upstream patch fixes the issue, so we're applying it to Buildroot's
kconfig.
Signed-off-by: Guillermo A. Amaral <g@maral.me>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Currently, our jobs on the gitlab-ci infra are running as root, which is
problematic for two reasons:
- this is not the usual way Buildroot is built;
- it may miss issues where running as non-root is problematic.
So, complement our Dockerfile with directives to add a new user and run
everything as that user, as demonstrated by this build job:
https://gitlab.com/ymorin/buildroot-ci/-/jobs/46929562
Additional, enforce an UTF-8 locale while running.
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Arnout Vandecappelle <arnout@mind.be>
Cc: Peter Korsgaard <peter@korsgaard.com>
Cc: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
BR2_TOOLCHAIN_BUILDROOT_INET_IPV6 has been removed with the commit [1].
Since this option is still in br-powerpc-internal-full config-fragment,
the powerpc configuration is droped by autobuild-run script:
WARN: toolchain can't be used
Missing: BR2_TOOLCHAIN_BUILDROOT_INET_IPV6=y
[Sat, 06 Jan 2018 03:03:43] WARN: failed to generate configuration
[1] 4bcacfd2c0
Signed-off-by: Romain Naour <romain.naour@gmail.com>
Cc: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Commit 4932c8a7cc introduced the
core-dependencies target to make sure that the dependencies.sh script
runs before we attempt to compile any host tool, so that the absence of
a compiler is properly detected. However, this relied on the
left-to-right evaluation of dependencies. This will no longer be true
when we enable top-level parallel build.
Fix this by letting DEPENDENCIES_HOST_PREREQ depend on
core-dependencies.
Note that it is not possible to remove the
dependencies <- core-dependencies. Indeed, it is possible that
DEPENDENCIES_HOST_PREREQ is completely empty, and in that case we still
need to check core-dependencies.
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
dependencies.sh uses HOSTCC_NOCCACHE directly, and this variable is
exported from the top-level Makefile, so there is no need to pass
HOSTCC to it. HOSTCC is not used at all in dependencies.sh.
Thus, we also no longer need to apply the HOSTCC override for
core-dependencies. The core-depencies rule doesn't use HOSTCC or
HOSTCXX.
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Make sure that the pie charts produced by 'graph-build' and 'graph-size'
targets are sorted on the size of each piece of the pie. Otherwise, making
visual analysis is difficult, as one needs to look at the legends of each
piece and do the sorting manually in their head.
Signed-off-by: Thomas De Schampheleire <thomas.de_schampheleire@nokia.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
It needs mkzftree from zisofs-tools, so we add a dependency to it, and
we call that one explicitly (to avoid using the one from the host in
PATH).
It also needs the the uncompressed kernel image, but because it is
already in target/ so it gets compressed by mkzftree. We have two
options:
- compress everything but the kernel image,
- compress everything, kernel included, and recopy it later.
We choose the latter, because it is the simplest solution. So, we always
define the kernel-copy hook, but only register it when needed.
Finally, it needs a kernel with support for transparent
(de)compression, so we update the existing test config.
Reported-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
[Thomas: really set the transparent compression option, take into
account the renaming of the option.]
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
The Marvell U-Boot needs DTC to build, so the test case fails to build
on systems where dtc isn't available. We add
BR2_TARGET_UBOOT_NEEDS_DTC=y to make sure that we build our own DTC.
Fixes:
https://gitlab.com/buildroot.org/buildroot/-/jobs/44126707
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
These new tests only do build tests, but allow to quickly verify that
the ATF/U-Boot combinations for vexpress, Allwinner and Marvell
platforms all continue to build properly.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
This merges the next branch accumulated during the 2017.11 release
cycle back into the master branch.
A few conflicts had to be resolved:
- In the DEVELOPERS file, because Fabrice Fontaine was added as a
developer for libupnp in master, and for libupnp18 in
next. Resolution is simple: add him for both.
- linux/Config.in, because we updated the 4.13.x release used by
default in master, while we moved to 4.14 in next. Resolution: use
4.14.
- package/libupnp/libupnp.hash: a hash for the license file was added
in master, while the package was bumped into next. Resolution: keep
the hash for the license file, and keep the hash for the newest
version of libupnp.
- package/linux-headers/Config.in.host: default version of the kernel
headers for 4.13 was bumped to the latest 4.13.x in master, but was
changed to 4.14 in next. Resolution: use 4.14.
- package/samba4/: samba was bumped to 4.6.11 in master for security
reasons, but was bumped to 4.7.3 in next. Resolution: keep 4.7.3.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Some packages (ex: skeleton-init-systemd) have a zero size so we cannot
divide by the package size. In that case make their percent zero
explicitly and avoid a ZeroDivisionError exception.
Signed-off-by: Andrey Yurovsky <yurovsky@gmail.com>
Acked-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Currently, we do nothing about packages that touch the same file: given
a specific configuration, the result is reproducible (even though it
might not be what the user expected) because the build order is
guaranteed.
However, when we later introduce top-level parallel build, we will no
longer be able to guarantee a build order, by the mere way of it being
parallel. Reconciliating all those modified files will be impossible to
do automatically. The only way will be to refuse such situations.
As a preliminary step, introduce a helper script that detects files that
are being moified by two or more packages, and reports them and the
impacted packages, at the end of the build.
The list being reported at the end of the build will make it prominently
visible in autobuilder results, so we can assess the problem, if any.
Later on, calling that helper script can be done right after the package
installation step, to bail out early.
Thanks Arnout for the pythonist way to write default dictionaries! ;-)
Note: doing it in python rather than a shell script is impressively
faster: where the shell script takes ~1.2s on a minimalist build, the
python script only takes ~0.015s, that is about 80 times faster.
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Arnout Vandecappelle <arnout@mind.be>
Cc: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Cc: Peter Korsgaard <peter@korsgaard.com>
Cc: Baruch Siach <baruch@tkos.co.il>
Cc: Peter Seiderer <ps.report@gmx.net>
[Thomas: rename script without .py extension.]
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Instead of overriding the _svn command and injecting --non-interactive,
change the default value of BR2_SVN to include this flag so the end user
can choose not to use the flag.
This change helps users behind corporate system rules which may not
allow them to locally cache credentials and require interactive mode.
Signed-off-by: Sam Voss <sam.voss@rockwellcollins.com>
[Originally implemented by]
CC: "Yann E. MORIN" <yann.morin.1998@free.fr>
Acked-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
This is needed to get the autobuilders to use a toolchain that
includes the fix merged in 9d544feb8a
("fwup: fix for ARC toolchain").
In addition, this new toolchain version also fixes for real the RPATH
issue that should have been fixed by
f90f28a6df, but wasn't done properly.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
All Buildroot pre-built toolchains have been rebuilt with Buildroot
2017.11-rc1, so that they have the latest version of
glibc/musl/uClibc, and also the latest gcc/binutils updates.
Specifically, this will fix the build failures on Blackfin that were
due to the missing accept4() support:
http://autobuild.buildroot.net/results/8b5a72dd7cde685f6f68f46aeee8b1b60c96d559/
(openobex)
http://autobuild.buildroot.net/results/b19dd9ed29944d7f79c6f824669e3baaa0bb045a/
(libiio)
In terms of changes to the toolchains:
- AArch64 glibc toolchain changed to use 4.4 kernel headers instead
of 4.1, in order to increase the variety of kernel header versions
being tested.
- Most configurations now use 4.13 kernel headers instead of 4.12
(except the configurations that were explicitly using an older
kernel headers version)
- The mips64 n64 configuration is changed from using gcc 4.9 to gcc
5, since another ARM configuration already tests gcc 4.9.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Fixes:
http://autobuild.buildroot.org/results/2af/2af7412846c576089f8596857ab8c81ac31c1bed/
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Cc: André Hentschel <nerv@dawncrow.de>
Reviewed-by: André Hentschel <nerv@dawncrow.de>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Currently, HOSTCC and HOSTCXX are set to their _NOCACHE variants in the
'dependencies' target. This is needed because at that time, ccache is
not built yet - host-ccache is one of the dependencies. However, because
this override is only specified for the 'dependencies' target (and
thereby gets inherited by its dependencies), the override is only
applied when the package is reached through the 'dependencies' target.
This is not the case when one of DEPENDENCIES_HOST_PREREQ is built
directly from the command line, e.g. when doing 'make host-ccache'. So
in that case, ccache will be built with ccache... which fails of
course.
To fix this, directly apply the override to the DEPENCIES_HOST_PREREQ
targets.
Note that this only fixes the issue for 'make host-ccache', NOT for
e.g. 'make host-ccache-configure'.
Signed-off-by: Alfredo Alvarez Fernandez <alfredo.alvarez_fernandez@nokia.com>
[Arnout: improve commit message]
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
glibc is now supported for ARC so let's kick-start autobuilders
with glibc toolchain for ARC HS.
Signed-off-by: Alexey Brodkin <abrodkin@synopsys.com>
Cc: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Cc: Arnout Vandecappelle <arnout@mind.be>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
The Linux 4.0 kernel doesn't build with gcc 6.x, which is used since
the toolchain update in commit
193dfffa83 ("support/testing: use more
recent toolchains"). So let's update to Linux 4.11 instead (like the
existing Qemu x86 defconfig does), and update the kernel configuration
file accordingly.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
With the hard disk crash of autobuild.b.o, we lost old toolchains, so
use the latest toolchains, which have been restored.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Currently, the download directory, when specified with the -d option, is
only used to store the files downloaded by the testing infra, not those
downloaded by Buildroot.
So, we end up with this situation:
BR2_DL_DIR | -d DIR | test downloads | BR downloads
------------+----------+------------------+--------------
unset | unset | [error] | [error]
unset | set | in $(DIR) | in $(TOP_DIR)/dl
set | unset | in $(BR2_DL_DIR) | in $(BR2_DL_DIR)
set | set | in $(DIR) | in $(BR2_DL_DIR)
This is not very consistent.
We change the behaviour so that the value of -d always takes precedence,
and is used by Buildroot as well, giving this new behaviour:
BR2_DL_DIR | -d DIR | test downloads | BR downloads
------------+----------+------------------+--------------
unset | unset | [error] | [error]
unset | set | in $(DIR) | in $(DIR)
set | unset | in $(BR2_DL_DIR) | in $(BR2_DL_DIR)
set | set | in $(DIR) | in $(DIR)
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Ricardo Martincoski <ricardo.martincoski@gmail.com>
Cc: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Fix typo 'selectes' -> 'selects'.
Additionally, change 'will exclude' to 'excludes' to align with 'selects'.
Signed-off-by: Thomas De Schampheleire <thomas.de_schampheleire@nokia.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Fix the remaining code style warnings from flake8:
- properly indent continuation lines;
- use proper code to test a parameter is not None.
Cc: Yann E. MORIN <yann.morin.1998@free.fr>
Signed-off-by: Ricardo Martincoski <ricardo.martincoski@gmail.com>
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Fix the trivial warnings from flake8:
- remove modules imported but unused;
- use 2 lines before class or module level method;
- remove blank line at end of file.
Signed-off-by: Ricardo Martincoski <ricardo.martincoski@gmail.com>
Acked-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Change all defconfig fragments to take advantage of
"cf3cd4388a support/tests: allow properly indented config fragment".
Make each defconfig fragment:
- start after a backslash;
- be declared as a multi-line string literal;
- be indented one level more than the variable that contains it.
Signed-off-by: Ricardo Martincoski <ricardo.martincoski@gmail.com>
Acked-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Move the strip out of leading spaces in defconfig fragments from the
BRTest class to the Builder class. It actually postpones the strip out,
consequentially allowing test cases to post-process the defconfig in
their own __init__ before calling the __init__ method from BRTest.
Moving this code to the Builder class also allows any new test class
that inherits from BRTest to reuse the same code even if the new class
overrides the setUp method.
At same time, prepend a newline to the jlevel handling otherwise it
would stop working for defconfig fragments that don't end in a newline.
Signed-off-by: Ricardo Martincoski <ricardo.martincoski@gmail.com>
Cc: Yann E. MORIN <yann.morin.1998@free.fr>
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
We sanity-check the host executables that they have a correct RPATH
pointing to the host libraries.
This is currently done by looking for all files in $(HOST_DIR) that
match the 'ELF executable' pattern (a bit more complex, but that's
idea).
However, when an executable is built with -fPIE of -fpie, it no longer
appears to be an 'ELF executable', but it rather looks like an 'ELF
sheard object' (like if it were an library.
So, we miss those files.
It turns out that the problem is a real one, because quite a few
mainline distros, expecially those based on Debian for example, have
already switched to generating PIE code by default, and thus we miss on
a whole class of systems..
We fix that by simply looking if we can find an ELF interpreter in each
file. If we there is one, this is an ELF executable; if not, it may be
anything else: we don't care (not even about ELF libraries).
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Arnout Vandecappelle <arnout@mind.be>
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
grub is no longer maintained: it is stuck at version 0.97 with huge
patches that have no opportunity to be applied upstream, as upstream
has even renamed it grub-legacy.
Besides, it no longer builds correctly with recent binutils versions,
and even the huge patches we could grab from Debian do not help the
slightest.
Since upstream really considers it dead, and there are at least two
alternatives (grub2 and syslinux), just remove grub.
Add a legacy entry.
Remove the test cases as well.
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Cc: Ricardo Martincoski <ricardo.martincoski@gmail.com>
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Patches with renames apply properly with patch >= 2.7, but not with
older patch versions. Since "git format-patch" by default generates
patches with renames, Buildroot developers often don't realize that
their patches will not apply properly on build machines that have
patch < 2.7. In order to prevent such a situation from happening
again, this commit adds some logic in apply-patches.sh to refuse
applying patches that contain renames.
Note that just searching for '^rename' is not sufficient, since the
patch commit message may contain the words "rename from" or "rename to"
as well. Therefore, the grep expression is made as accurate as possible,
checking both.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Reviewed-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Reviewed-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
[Arnout: spaces instead of tabs (suggested by Yann);
extend commit message.]
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Force gzip compression level 6 when calculating hash of a downloaded GIT repo.
To make sure the tar->gzip->checksum chain always provides consistent result.`
The script was relying on the default compression level, which must not be
necessarily consistent among different gzip versions. The level 6 is gzip's
current default compression level.
Signed-off-by: Petr Kulhavy <brain@jikos.cz>
Acked-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
When we list the available tests, we use test_dir, which is set from
the path of the script. However, when we run the tests, we use the
hard-coded path.
Ditto to find the config file.
For consistency, always use test_dir.
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Cc: Ricardo Martincoski <ricardo.martincoski@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Commit c96b8675ea
("support/scripts/check-bin-arch: ignore symbolic links") was bogus,
because it tested ${f}, which is the relative path of the file inside
${TARGET_DIR}, so we end up testing if ${f} on the system is a
symbolic link.
This commit fixes that by testing ${TARGET_DIR}/${f}.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Since commit da32b49f00
("instrumentation: extend packages-file-list.txt with symlinks and
directories"), the packages-file-list.txt also contains symbolic
links. Therefore, check-bin-arch is now also checking symbolic links.
However, symbolic links in $(TARGET_DIR) can have absolute path as
targets, such as:
$ ls -l output/target/sbin/ifdown
lrwxrwxrwx 1 thomas thomas 10 Sep 3 15:55 output/target/sbin/ifdown -> /sbin/ifup
Therefore, we are now potentially checking a host binary, which
obviously makes check-bin-arch fail.
This commit changes check-bin-arch to ignore symbolic links. Indeed,
we have two cases:
- The symbolic link really points to something that will in the
rootfs (such as /sbin/ifup above). In this case, /sbin/ifup will be
checked separately by check-bin-arch.
- The symbolic link doesn't point to something that will be in the
rootfs, and that is not a problem from the perspective of
check-bin-arch, which checks the architecture of target binaries.
Fixes:
http://autobuild.buildroot.net/results/16d384a0183d477646ac7692feb65f00dde7d068/
(vim)
http://autobuild.buildroot.net/results/50429c0f63a8befff9e20899327b9a8d754d99be/
(ifupdown)
http://autobuild.buildroot.net/results/1db65973e782bfa61abcbccd3501bfd235f77288/
(gawk)
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Reviewed-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Add a parameter to run-tests to act as a multiplier for all timeouts of
emulator.
It can be used to avoid sporadic failures on slow host machines as well
in elastic runners on the cloud.
Cc: Arnout Vandecappelle <arnout@mind.be>
Cc: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Ricardo Martincoski <ricardo.martincoski@gmail.com>
[Arnout: rename multiplier to timeout_multiplier everywhere]
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
We add the 3 following combinations:
- basic systemd, read-only, network w/ ifupdown
- basic systemd, read-only, network w/ networkd
- full systemd, read-only, network w/ networkd
The tests just verify what the /sbin/init binary is, and that we were
able to grab an IP address. More tests can be added later, for example
to check each systemd features (journal, tmpfiles...)
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
[Arnout: regenerate .gitlab-ci.yml]
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
The "builtin" kernel does not boot a systemd-based system, so
we resort to building the same one as currently used by our
qemu_arm_vexpress_defconfig.
We test the 8 following combinations:
- busybox, read-only, without network
- busybox, read-only, with network
- busybox, read-write, without network
- busybox, read-write, with network
- basic systemd, read-write, network w/ ifupdown
- basic systemd, read-write, network w/ networkd
- full systemd, read-write, network w/ networkd
- no init system, read-only, without network
The tests just verify what the /sbin/init binary is, and that we were
able to grab an IP address. More tests can be added later, for example
to check each systemd features (journal, tmpfiles...)
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Cc: Arnout Vandecappelle <arnout@mind.be>
[Arnout: update .gitlab-ci.yml]
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
The default audio backend for qemu is configured at compile time. It
generates annoying warning messages to qemu's stderr when running our
tests, like these:
pulseaudio: set_sink_input_volume() failed
pulseaudio: Reason: Invalid argument
pulseaudio: set_sink_input_mute() failed
pulseaudio: Reason: Invalid argument
Explicitly set the audio backend to "none" at runtime to remove those
messages from our logs. There is no command line argument for this, so
use an environment variable when starting qemu.
Signed-off-by: Ricardo Martincoski <ricardo.martincoski@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
The toolchain configuration used by the autobuilders to test the
internal toolchain backend on ARM contains an option that no longer
exists: BR2_UCLIBC_VERSION_NG. This option has been removed since
uClibc-ng has been made the one and only uClibc version supported.
Due to this option no longer existing, this toolchain configuration
was in fact never used, because the randgenconfig script validates
that a toolchain configuration is valid by checking that all lines are
still present in the final .config.
Therefore, this commit removes the bogus option, which will make sure
this toolchain config gets used again by our autobuilders.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
This column is not used by either genrandconfig or test-pkg, so remove
it.
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
We currently have a list of toolchain configurations that are used by
the autobuilders at [1]. However, this makes it a little more difficult
for people to use these configurations, and also to have a different
list of configurations for different branches. For example if a new
architecture is introduced, the 2017.02.x branch doesn't have support
for this architecture yet so it shouldn't try to run those configs.
Therefore, include the autobuild config fragments directly in Buildroot,
so they can be branched together with the rest. We create a new
directory under support/ to store them.
Generated with
wget -nd --no-parent --recursive http://autobuild.buildroot.net/toolchains/configs/
The index.html file is removed.
The toolchain-configs.csv file is adapted so the URLs become relative
paths pointing to the config fragments.
[1] http://autobuild.buildroot.net/toolchains/configs/toolchain-configs.csv
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
As suggested by Arnout in [1].
While at it, simplify the logic by always appending the BR2_JLEVEL and
defaulting to 0 (the value copied from Config.in is used for 5 years now
and is very unlikely to change).
[1] http://patchwork.ozlabs.org/patch/790525/
Suggested-by: Arnout Vandecappelle <arnout@mind.be>
Signed-off-by: Ricardo Martincoski <ricardo.martincoski@gmail.com>
Cc: Arnout Vandecappelle <arnout@mind.be>
Tested-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Reviewed-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Since commit cf3cd4388a the -j option is
silently ignored.
The configuration lines are processed using '\n'.join().
This function adds intervening occurrences of the separator, but the
resulting string does not end at a separator.
>>> "n".join(["a","b"])
'anb'
It results in a defconfig that does not end in a newline.
When BR2_JLEVEL is added by -j logic to the defconfig it ends up
concatenated to the last line of the defconfig.
BR2_TOOLCHAIN_EXTERNAL_CODESOURCERY_ARM=yBR2_JLEVEL=7
The resulting .config has the default BR2_JLEVEL=0.
Instead of just workaround this problem by adding a newline before
BR2_JLEVEL when -j is used, make the defconfig to end in a newline since
it is a more future-proof solution.
Signed-off-by: Ricardo Martincoski <ricardo.martincoski@gmail.com>
Cc: Arnout Vandecappelle <arnout@mind.be>
Cc: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Cc: Yann E. MORIN <yann.morin.1998@free.fr>
Tested-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Acked-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
The defconfig is composed on-the-fly by test infra + tests.
Dump it to the logfile before running 'make olddefconfig' so it can
easily analysed when debugging.
Signed-off-by: Ricardo Martincoski <ricardo.martincoski@gmail.com>
Cc: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Depending on Python implementation used for testing, time it takes to
perform a given test can vary pretty significantly. To accout for that
allow individual test functions to specify different timeout value.
Signed-off-by: Andrey Smirnov <andrew.smirnov@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
In order to be able to leverage the same test code for testing
different python interpreters (or wrappers around CPython) allow child
classes of TestPythonBase to override the name of the executable used
to run tests.
Signed-off-by: Andrey Smirnov <andrew.smirnov@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Add Python3 version of TestPython2 to make sure both versions of
Python get unit-tested.
Modify the code of libc_time_test() to support that change (convert
the code to use Python3-style "print").
Signed-off-by: Andrey Smirnov <andrew.smirnov@gmail.com>
[Thomas: update .gitlab-ci.yml.]
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Convert TestPythonBase to a true base class that only provides code
implementing various tests without defining tests themselves in a
"discoverable" form.
To retain correct testing functionality, add TestPython2 derived class
that uses code from TestPythonBase to define actual runnable test.
Signed-off-by: Andrey Smirnov <andrew.smirnov@gmail.com>
Reviewed-by: Ricardo Martincoski <ricardo.martincoski@gmail.com>
Tested-by: Ricardo Martincoski <ricardo.martincoski@gmail.com>
[Thomas: fix typo in commit log, update .gitlab-ci.yml, both pointed
by Ricardo.]
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Some commands take more than 5 seconds to complete under QEMU, so add
provisions to allow individual unit-test to specify different duration
to avoid false negative test failures.
Signed-off-by: Andrey Smirnov <andrew.smirnov@gmail.com>
Reviewed-by: Ricardo Martincoski <ricardo.martincoski@gmail.com>
Tested-by: Ricardo Martincoski <ricardo.martincoski@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
The user shouldn't need to pass this manually when creating a test-pkg
config file. It's an absolutely harmless option to enable always.
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
We really want test-pkg to do the test with a paranoid unsafe path.
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
This minimal configuration is also very useful outside test-pkg. In
addition, it will simplify the config merge in a later patch.
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Cc: Yann E. MORIN <yann.morin.1998@free.fr>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
This commit introduces the script "fix-rpath" able to scan a tree,
detect ELF files, check their RPATH and fix it in a proper way.
The RPATH fixup is done by the patchelf utility using the option
"--make-rpath-relative <root-directory>".
Signed-off-by: Samuel Martin <s.martin49@gmail.com>
Signed-off-by: Wolfgang Grandegger <wg@grandegger.com>
Acked-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
The normal shell does not support the bashism "< <(...)". Therefore
we use a normal pipe to find files containing a specific string.
Signed-off-by: Wolfgang Grandegger <wg@grandegger.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
This is because $(HOST_DIR)/usr is gone.
Signed-off-by: Wolfgang Grandegger <wg@grandegger.com>
Acked-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Currently, defining a config fragment in the runtime test infra requires
that the fragment not to be indented. This is beark, and causes grievance
when looking at the code (e.g. to fix it).
Just strip out all leading spaces/tabs when writing the configuration
lines into the config file, allowing in-line indented config fragments,
like so:
class TestFoo(bla):
config = bla.config + \
"""
FOO=y
# BAR is not set
"""
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Cc: Arnout Vandecappelle <arnout@mind.be>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
The current test fails because of a legacy option, renamed during the
recent ext overhaul.
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Sébastien Szymanski <sebastien.szymanski@armadeus.com>
Cc: Samuel Martin <s.martin49@gmail.com>
Cc: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Cc: Arnout Vandecappelle <arnout@mind.be>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
In commit b78b50465c, the initialisation
of BRTest.builddir was moved to the __init__ function. However, it is
set based on BRTest.outputdir and that is only set when the -o argument
is given to run-tests. When called as "run-tests -l", there is no -o
argument so BRTest.outputdir remains unset.
To fix, keep BRTest.builddir at None when BRTest.outputdir is None.
While we're at it, drop the direct access to the class member. If a
subclass wishes to set outputdir to something else before calling
BRTest.__init__, they are free to do so.
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Reported-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Tested-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
BRTest's setUp() method contains a few assignments that initialize its
member variables. Since we will want to use these in test case
overrides, move them to the __init__ function.
Also allow the config member to be overridden, rather than always
taking the class member.
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
This file is not a package per-se, it includes other .mk files that
are packages.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
The waf package infrastructure was not known by the pkg-stats script,
so let's add it.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
With 2000+ packages it's not trivial to identify i.e.:
- all packages that don't have a hash file;
- all packages that have patches;
- all packages that have code style warnings;
User experience can be improved by dynamically sorting the resulting
table.
There is an open-source solution that does that in the client-side and
requires minimal changes to our script: sorttable.js. The script is
MIT licensed as stated in its website.
Also add a hint to the user that the table can be sorted.
Signed-off-by: Ricardo Martincoski <ricardo.martincoski@gmail.com>
Cc: Arnout Vandecappelle <arnout@mind.be>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Now all packages have been updated to install things in $(HOST_DIR)/lib
instead of $(HOST_DIR)/usr/lib, there should no longer be any reason
to have $(HOST_DIR)/usr/lib in the RPATH, so we don't allow it any more.
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Cc: Wolfgang Grandegger <wg@grandegger.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
The tools are now installed in host/bin instead of host/usr/bin.
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
This silences the annoying warning that there is no hash file for our
own COPYING file.
Also change the message so that it is more obvious what we're doing.
Reported-by: Peter Korsgaard <peter@korsgaard.com>
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Luca Ceresoli <luca@lucaceresoli.net>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Since things are no longer installed in $(HOST_DIR)/usr, the callers
should also not refer to it.
This is a mechanical change with
git grep -l '$(HOST_DIR)/usr/bin' | xargs sed -i 's%$(HOST_DIR)/usr/bin%$(HOST_DIR)/bin%g'
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
These tests simply build a system with musl and uclibc toolchains, and
boot them under qemu. It allows to minimally validate that our support
for musl/uclibc external toolchains is working. We already had some
tests covering glibc toolchains, so we can now easily test that all
three C libraries are supported.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
---
This commit is part of the series, as I've written/used those tests to
validate that things are still working correctly with all of glibc,
uclibc and musl toolchains.
Remove the redundant usr/ component of the HOST_DIR paths. Since a
previous commit added a symlink from $(HOST_DIR)/usr to $(HOST_DIR),
everything keeps on working.
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Reviewed-by: Romain Naour <romain.naour@smile.fr>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Move toolchainfile.cmake and Buildroot.cmake from
$(HOST_DIR)/usr/share/buildroot to $(HOST_DIR)/share/buildroot.
Build-tested with a bunch of cmake packages.
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Reviewed-by: Romain Naour <romain.naour@smile.fr>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Since $(HOST_DIR)/usr/{bin,sbin} are now symlinks to
$(HOST_DIR)/{bin,sbin}, it makes no sense to check them - they are
already covered.
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Reviewed-by: Romain Naour <romain.naour@smile.fr>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
This is a step towards eliminating $(HOST_DIR)/usr. It allows us to
convert all packages installing things into $(HOST_DIR)/usr/lib without
affecting the rest.
To allow compatibility with packages that still use $(HOST_DIR)/usr as
the prefix, create a symlink from usr/lib to ../lib.
Note that the symlink creation will break when $(HOST_DIR)/usr/lib
already exists as a directory, i.e. when rebuilding in an existing
output directory. This is necessary: if we don't break it now, the
following commits (which remove the usr part from various variables)
_will_ break it.
At the same time as creating this symlink, we also have to update the
check-host-rpath script to accept both $(HOST_DIR)/usr/lib and
$(HOST_DIR)/lib, because depending on how the package derives the
path, it may be different.
Since there are some dependency chains that involve $(STAGING_DIR),
$(STAGING_DIR) may in fact be created before $(HOST_DIR). Since
$(STAGING_DIR) is a subdirectory of $(HOST_DIR), it is possible that the
newly added rule for $(HOST_DIR) never triggers. To make sure that the
rule does trigger, add an order-only dependency from $(STAGING_DIR) to
$(HOST_DIR).
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Reviewed-by: Romain Naour <romain.naour@smile.fr>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
It will install the script "relocate-sdk.sh" in the HOST_DIR
allowing to adjust the path to the SDK directory in all text
files after it has been moved to a new location.
Signed-off-by: Wolfgang Grandegger <wg@grandegger.com>
[Thomas:
- Fix shebang to be /bin/sh instead of /bin/bash, suggested by Arnout
- Use | instead of \ as a separator for sed expressions, suggested by
Arnout, discussed with Wolfgang and others
- Remove ./ at the beginning of LOCFILE, suggested by Arnout
- Fix comment about the path check being made before doing the
replacement, suggested by Arnout
- Fix indentation, suggested by Arnout.]
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
By default, cut prints the entire line if the specified delimiter is not
present at all:
$ printf "foo bar" | cut -d' ' -f2
bar
$ printf "foobar" | cut -d' ' -f2
foobar
In setlocalversion, cut is presented with the output of 'hg id' which has
the format:
"<revision> <tags-if-any>"
If the current revision is not tagged, the output of 'hg id' does not
contain the delimiter (space), cut prints the entire string, and
setlocalversion thinks the version is the tag.
As setlocalversion does not print anything for tagged versions, there is no
output overall, and no correct indication of the mercurial revision.
Fix by passing the extra cut option '--only-delimited', which suppresses
output if no delimiter is found.
This problem likely went unnoticed for so long, because the tag 'tip' (i.e.
most recent revision of the branch) is treated specially: in this case the
mercurial revision _is_ printed, i.e. the situation is treated as
'untagged'.
The problem is only seen when you are _not_ at the most recent revision in
your branch.
Signed-off-by: Thomas De Schampheleire <thomas.de_schampheleire@nokia.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
For Gitlab-CI, we want to avoid re-generating the minimal install to
be able to run tests all the time. So let's create a docker image that
we can post on Docker Hub and then pull.
For the time being, this is just what we need for running our CI. Later
we can produce something that is also useful for users.
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
