Creating a new workload to execute the following tests on GFXBench.
* Car Chase
* Car Chase Offscreen
* Manhattan 3.1
* 1080p Manhattan 3.1 Offscreen
* 1440p Manhattan 3.1 Offscreen
* Tessellation
* Tessellation Offscreen
Now allows for specifying a `group` value for each section which will
cross product the sections within that group with the sections in each
other group. Additionally classifiers will automatically be added to
each job spec with the relevant group information.
In the case of a failure in the initialization of one output_processor the
remaining `initialize` methods may not get called causing variables to
not be initialized correctly.
The 'jsonB' datatype was only added in v9.4 so ensure that the Postgres
server to is running this or later and inform the user if this is not
the case.
The Output processor which is used to upload the results found in the
wa_output folder to a Postgres database, whose schema is defined by the
WA Create Database command.
Add a command to create a PostgreSQL database with supplied parameters
which draws its structure from the supplied schema (Version 1.1). This
database is of a format intended to be used with the forthcoming WA
Postgres output processor.
Leftover code from WA2 meant that logcat was cleared on the device by
the workload directly instead of using devlib, this caused issues if logcat was
still being cleared from other areas of the code.
Expose the context to the initialize and finalize functions for Output
Processors. This was found to be necessary for the upcoming PostgreSQL
Output Processor.
Previously when using the create command for adding
EnergyInstruemntBackends they were treated like any other plugin and
generated incorrect configuration. Now automatically add the
`energy_measurement` instrument and populate it's configuration with the
relevant defaults for the specified Backend.
In commit d2ece we are now tracking augmentations which are used during a
run in the run_config via the context when installing augmentations.
Update the Process command and its ProcessContext with a dummy method
to relect this change.
Ensure that a plug-ins config entry is converted to an identifier before being
stored in the PluginCache so that the relevant configuration can be
retrieved appropriately. For example this allows for both 'trace-cmd' and
'trace_cmd' to be used as config entries to provide configuration for the
'trace-cmd' plugin.
In commit 8da911 the initialization of the target was split into a
separate method of the TargetManger. Ensure that we call the relevant
method after creating the manager.
Split the tests into their own jobs so it is easier to see what test is
failing, also only run pep8 and pylint tests with python 3.6 as it is
uncessary to run with both python versions.
Add Context.write_config() to write the combined config into run output
__meta. Use it after instruments and result processors get installed to
make sure their configuration gets serialized in the output.
