Age | Commit message (Collapse) | Author |
|
yocto-autobuilder2
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
When the parsing of the .conf files in buildset-config fails print the
error message from ConfigParser's exception so that the user has a chance
of understanding which file failed to parse and why.
Signed-off-by: Joshua Lock <joshua.g.lock@intel.com>
|
|
We need to store the passed cfile otherwise passing a cfile at
initialisation is broken,the code will not automatically attempt the
default cfile if this is set.
Signed-off-by: Joshua Lock <joshua.g.lock@intel.com>
|
|
The buildtools buildset isn't building a build-appliance, this may be a
copy/paste error.
Signed-off-by: Joshua Lock <joshua.g.lock@intel.com>
|
|
Use the new buildinfo option of CreateAutoConf to ensure that images used
by QA contain a build info file at /etc/build which enables them to
identify the image from within the running OS.
[YOCTO #12331]
Signed-off-by: Joshua Lock <joshua.g.lock@intel.com>
|
|
For QA purposes it's useful to be able to understand details about the
system from within a running image. Therefore add an option to enable the
image-buildinfo class with IMAGE_BASENAME and IMAGE_NAME included in the
written variables.
Signed-off-by: Joshua Lock <joshua.g.lock@intel.com>
|
|
The prior commit appends the release number to the SSTATE_MIRRORS variable
but does so without adding a separating / (due to a typo). Fix that issue.
Signed-off-by: Joshua Lock <joshua.g.lock@intel.com>
|
|
We recently switched to publishing sstate artefacts into a per-release
directory, therefore we need to append the same directory name to the value
we set in SSTATE_MIRRORS to ensure a generated eSDK is pointing to the
correct shared state mirror and able to install objects.
Signed-off-by: Joshua Lock <joshua.g.lock@intel.com>
|
|
Setting SRCREV without changing PV isn't correct and can lead to unexpected
issues. Ensure we're in-keeping with best practices and set the PV for
build-appliance-image whenever we are setting SRCREV.
Signed-off-by: Joshua Lock <joshua.g.lock@intel.com>
|
|
Signed-off-by: Joshua Lock <joshua.g.lock@intel.com>
|
|
Ensure we don't try and build refkit when the user requests a release build.
Signed-off-by: Joshua Lock <joshua.g.lock@intel.com>
|
|
Add Sumo, the next (2.5) YP release name.
Remove release names for releases which are no longer supported.
Signed-off-by: Joshua Lock <joshua.g.lock@intel.com>
|
|
Signed-off-by: Joshua Lock <joshua.g.lock@intel.com>
|
|
Rocko introduces yocto-check-layer-wrapper which creates a separate
build dir to run yocto-check-layer inside.
Signed-off-by: Stephano Cetola <stephano.cetola@linux.intel.com>
|
|
Signed-off-by: Joshua Lock <joshua.g.lock@intel.com>
|
|
Drop this buildset, the Mars version of Eclipse is no longer supported.
Signed-off-by: Joshua Lock <joshua.g.lock@intel.com>
|
|
We try to run distrodata.Distrodata.test_checkpkg periodically on the Yocto
Project autobuilders, yet this test was only introduced in the Rocko cycle.
Therefore filter it out if the layer version doesn't map to Rocko,
furthermore if we do filter it out and the list of tests thus becomes empty
we shouldn't run oe-selftest at all.
Signed-off-by: Joshua Lock <joshua.g.lock@intel.com>
|
|
Without this, builds of older releases fail due to the use of the new name
which doesn't exist there.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
Signed-off-by: Joshua Lock <joshua.g.lock@intel.com>
|
|
We build refkit images only with master and pre-integration branches, not
any release branches. Improve the way we detect Ross pre-integration
branches so that we better handle numebred pre-integration branches.
Signed-off-by: Joshua Lock <joshua.g.lock@intel.com>
|
|
This is the new name of the machine from Yocto BSP layer version 4 (sumo
release) on. We have code in place in CreateAutoConf to rename to the
older machine name on for morty..rocko branches.
Signed-off-by: Joshua Lock <joshua.g.lock@intel.com>
|
|
Signed-off-by: Joshua Lock <joshua.g.lock@intel.com>
|
|
We are no longer actively supporting Yocto Project releases older than the
morty branch. Simplify the logic of the buildstep by dropping any code
paths related to supporting versions older than morty.
Signed-off-by: Joshua Lock <joshua.g.lock@intel.com>
|
|
The unsupported recipe check is likely to often fail on the Yocto
Autobuilder when testing pre-integration branches that the refkit team
haven't worked with yet. Switch the unsupported recipes check to a warning
instead of an error so that we provide the recipe check information for the
refkit team without causing errors for our pre-integration checks.
Signed-off-by: Joshua Lock <joshua.g.lock@intel.com>
|
|
Signed-off-by: Joshua Lock <joshua.g.lock@intel.com>
|
|
Signed-off-by: Joshua Lock <joshua.g.lock@intel.com>
|
|
On the meta-intel autobuilder, we were only building musl for 64bit
architectures. This adds a buildset for core2-32 arch, and adds that
build to the nightly.
Signed-off-by: Stephano Cetola <stephano.cetola@linux.intel.com>
|
|
We were building with musl on 32 bit platforms but not on 64 bit.
This adds a 64 bit musl buildset that is similar to the 32 bit one
and also adds that buildset to the list of builds for nightly.
Signed-off-by: Stephano Cetola <stephano.cetola@linux.intel.com>
Signed-off-by: Joshua Lock <joshua.g.lock@intel.com>
|
|
Signed-off-by: Joshua Lock <joshua.g.lock@intel.com>
|
|
Add Rocko (2.4) to the "Name of the poky release" dropdown on the nightly
builder.
Signed-off-by: Joshua Lock <joshua.g.lock@intel.com>
|
|
Add some extra edge cases for:
* meta-qt3: it won't have a rocko (or newer) branch
* meta-gplv2: it only has a pyro and master branch
Fix bitbake version map to use the master branch of bitbake when selecting
the master entry in the dropdown.
Signed-off-by: Joshua Lock <joshua.g.lock@intel.com>
|
|
This script name was changed in the following commit:
b46e05677b342df44829ffe8bcfbfc954e906030
This patch updates the script name to match and bumps the supported
layer to version 11, the Rocko series.
[YOCTO #12110]
Signed-off-by: Stephano Cetola <stephano.cetola@linux.intel.com>
Signed-off-by: Joshua Lock <joshua.g.lock@intel.com>
|
|
Add a convenience dropdown to the "Force build" form for the nightly
builder which lets the AB user select an actively maintained Yocto
Project branch and have the branch fields in the cloned repositories
populated correctly.
Unfortunately I couldn't determine a good way to do this without directly
modifying the buildbot sources. Sadly this is unclean and will impact
efforts to update buildbot in future... Once we have moved to a newer
buildbot there's a clean set of hooks to implement custom web status UI's
[YOCTO #11804]
Signed-off-by: Joshua Lock <joshua.g.lock@intel.com>
|
|
We were carrying two versions of sqlalchemy_migrate, this one was unused.
Signed-off-by: Joshua Lock <joshua.g.lock@intel.com>
|
|
We were carrying two version of SQLAlchemy, this one is unused.
Signed-off-by: Joshua Lock <joshua.g.lock@intel.com>
|
|
We no longer have a scripts directory, so don't add it to PATH
Signed-off-by: Joshua Lock <joshua.g.lock@intel.com>
|
|
73dec5e984b99ea2fe08e940be0fad811bdd8129 was an attempt to add
RPM_GPG_SIGN_CHUNK = "1" to the generated auto.conf, but syntax was
incorrect. Add this variable to atextappend so that it's included in
auto.conf
Signed-off-by: Joshua Lock <joshua.g.lock@intel.com>
|
|
* switch to SQLAlchemy-0.7.0, 0.8.0b2 isn't compatible with the
sqlalchemy_migrate 0.7.2 we ship
* add setuptools
Signed-off-by: Joshua Lock <joshua.g.lock@intel.com>
|
|
If the user-site directory doesn't exist the logic to create autobuilder.pth
will fail, ensure we create the directory before we try and create a file
there.
Signed-off-by: Joshua Lock <joshua.g.lock@intel.com>
|
|
Exporting PYTHONPATH into the environment which starts bitbake is causing
problems since the port to Python 3.
The PYTHONPATH environment variable is used by both Python 2 and Python 3
and having a PYTHONPATH which points to a site-packages with 2.7 code in
causes Python 3 errors where the code in site-packages has syntax which is
incompatible with Python 3 (for further details see
https://bugzilla.yoctoproject.org/show_bug.cgi?id=12076).
Instead of exporting PYTHONPATH and polluting the environment we build in
we now create an autobuilder.pth file in the user-site location.
The down side of this is that we modify how the host OS behaves, at least
for the user who invokes yocto-autobuilder-setup, which is contrary to the
original design goals of yocto-autobuilder.
[YOCTO #12076]
Signed-off-by: Joshua Lock <joshua.g.lock@intel.com>
|
|
The intel-iot-refkit project tracks master and doesn't maintain stable
releases, therefore we should skip building refkit images when building
branches which aren't recognised as master or a pre-integration test
branch for master.
[YOCTO #12063]
Signed-off-by: Joshua Lock <joshua.g.lock@intel.com>
|
|
This step can't have worked before now as it was mistakenly using the
poky/oe-core checkout directory (build), instead of its build child, for
BUILDDIR.
Fix the path usage and improve the code to handle multiple entries in the
buildstep's files parameter.
Signed-off-by: Joshua Lock <joshua.g.lock@intel.com>
|
|
Make use of the new CheckYoctoCompat layer early in the cycle to determine
whether cloned layers are Yocto Compatible v2.
Signed-off-by: Joshua Lock <joshua.g.lock@intel.com>
|
|
With the introduction of the Yocto Compatible v2 it will be useful for users
of yocto-autobuilder to be able to run the yocto-compat-layer-wrapper script
against their layers as part of their CI/build/release process.
This simple buildstep runs yocto-compat-layer-wrapper against all non-core
(i.e. openembedded-core, poky and poky-eclipse) layers cloned during the
CheckOutLayers buildstep.
[YOCTO #12047]
Signed-off-by: Joshua Lock <joshua.g.lock@intel.com>
|
|
Store the names of checked out layers in a buildbot property so that we
can access the list of checkout out layers from other buildsteps.
Signed-off-by: Joshua Lock <joshua.g.lock@intel.com>
|
|
This helps trace properties and their lineage when viewing them in the
web UI.
Signed-off-by: Joshua Lock <joshua.g.lock@intel.com>
|
|
This helps trace properties and their lineage when viewing them in the
web UI.
Signed-off-by: Joshua Lock <joshua.g.lock@intel.com>
|
|
Without an explicit timeout value requests will not timeout and can hang
indefinitely. We have observed this on a heavily loaded wiki (when the
database was in the process of being backed up).
To start we'll use a 60s timeout for both GET and POST.
[YOCTO #11974]
Signed-off-by: Joshua Lock <joshua.g.lock@intel.com>
|
|
It's not used in more than the one site where it was defined but feels a
little cleaner this way.
Signed-off-by: Joshua Lock <joshua.g.lock@intel.com>
|
|
Fix a couple of minor typos in comments.
Signed-off-by: Joshua Lock <joshua.g.lock@intel.com>
|