- 10 8月, 2020 7 次提交
-
-
由 Beraldo Leal 提交于
This is waiting for #4062 to be solved. Signed-off-by: NBeraldo Leal <bleal@redhat.com>
-
由 Beraldo Leal 提交于
Adding here some some possible use case of the current API. So far I notice that we have two big cases: automatic job discovery vs manual/custom job. I understand that there is room for improvements but I would like to keep things consistent and intuitive here, because of that this work is separating the logic behind the constructors and the from_config() methods for both: Job() and TestSuite(). Signed-off-by: NBeraldo Leal <bleal@redhat.com>
-
由 Beraldo Leal 提交于
A lot of changes was made inside job and suite modules. This change will update the docstrings to map with the current code. Signed-off-by: NBeraldo Leal <bleal@redhat.com>
-
由 Beraldo Leal 提交于
Before this patch directories are not being created properly because uuid1 was not being called. Took the oportunity to replace to uuid4. Signed-off-by: NBeraldo Leal <bleal@redhat.com>
-
由 Beraldo Leal 提交于
With .get() we don't need this, since we have default values. Signed-off-by: NBeraldo Leal <bleal@redhat.com>
-
由 Beraldo Leal 提交于
This is a natural move since with this flexibility would be nice to have test_parameters per test suite. Signed-off-by: NBeraldo Leal <bleal@redhat.com>
-
由 Beraldo Leal 提交于
If job.config is present when creating a test suite then we will extend this configuration. This will make it possible to passa a very long job config and only small/specific configs for each test suite with the changes. Signed-off-by: NBeraldo Leal <bleal@redhat.com>
-
- 08 8月, 2020 6 次提交
-
-
由 Beraldo Leal 提交于
Since we have one variant per suite now, this plugin needs to get the list of variants instead of only one. Signed-off-by: NBeraldo Leal <bleal@redhat.com>
-
由 Beraldo Leal 提交于
Signed-off-by: NBeraldo Leal <bleal@redhat.com>
-
由 Beraldo Leal 提交于
Now variants is for each Test Suite, we should log from there. Also jobdata.record only needs the job object pointer. Signed-off-by: NBeraldo Leal <bleal@redhat.com>
-
由 Beraldo Leal 提交于
Since that we are going to a more decoupled way, users should create they test suites using TestSuite() class. Signed-off-by: NBeraldo Leal <bleal@redhat.com>
-
由 Beraldo Leal 提交于
This will return a sum of all len(test_suite.tests) of the current job. Signed-off-by: NBeraldo Leal <bleal@redhat.com>
-
由 Beraldo Leal 提交于
This property will hold a list of TestSuite objects. I'm adding a test_suite property here for backward compatibility. But it will be removed soon. Signed-off-by: NBeraldo Leal <bleal@redhat.com>
-
- 07 8月, 2020 1 次提交
-
-
由 Beraldo Leal 提交于
This method could be useful for other modules and since that is not related to the job, IMO, makes sense to move to a more generic module. Signed-off-by: NBeraldo Leal <bleal@redhat.com>
-
- 06 8月, 2020 2 次提交
-
-
由 Cleber Rosa 提交于
Signed-off-by: NCleber Rosa <crosa@redhat.com>
-
由 Cleber Rosa 提交于
To make sure that all interface checks, current and future, will be run with all the existing runner implementations. Reference: https://github.com/avocado-framework/avocado/issues/4058Signed-off-by: NCleber Rosa <crosa@redhat.com>
-
- 05 8月, 2020 8 次提交
-
-
由 Cleber Rosa 提交于
Signed-off-by: NCleber Rosa <crosa@redhat.com>
-
由 Cleber Rosa 提交于
Signed-off-by: NCleber Rosa <crosa@redhat.com>
-
由 Cleber Rosa 提交于
Signed-off-by: NCleber Rosa <crosa@redhat.com>
-
由 Jan Richter 提交于
The logging system inside avocado doesn't expect changes of configuration between jobs. This commit brings feature for enabling and disabling logging configurations of the system. It saves the old configurations to the list for later use. Reference: #3997 Signed-off-by: NJan Richter <jarichte@redhat.com>
-
由 Cleber Rosa 提交于
By using the correct keys related to the filtering by tags configuration. Fixes: https://github.com/avocado-framework/avocado/issues/4054Signed-off-by: NCleber Rosa <crosa@redhat.com>
-
由 Cleber Rosa 提交于
And reuse the core verbose option, like many other subcommands do. Signed-off-by: NCleber Rosa <crosa@redhat.com>
-
由 Cleber Rosa 提交于
Signed-off-by: NCleber Rosa <crosa@redhat.com>
-
由 Cleber Rosa 提交于
Under the 5 seconds timeout that is specified in the job. The best way we can control the environment in which those tests run is to skip them on general `make check` runs, and tag them so that they are executed one at a time on `selftests/pre_release/jobs/timesensitive.py`. Fixes: https://github.com/avocado-framework/avocado/issues/3689Signed-off-by: NCleber Rosa <crosa@redhat.com>
-
- 04 8月, 2020 5 次提交
-
-
由 Cleber Rosa 提交于
Signed-off-by: NCleber Rosa <crosa@redhat.com>
-
由 Jan Richter 提交于
Signed-off-by: NJan Richter <jarichte@redhat.com>
-
由 Jan Richter 提交于
avocado run command line can accept options --loaders and --external-runner in one time. But when we use external runner there is a specific loader and we don't use values from option --loaders. When user uses both these options, we have to warn him about this behaviour. Reference: #3899 Signed-off-by: NJan Richter <jarichte@redhat.com>
-
由 Cleber Rosa 提交于
Signed-off-by: NCleber Rosa <crosa@redhat.com>
-
由 Cleber Rosa 提交于
In ecf253c4 we removed the default parameters, but forgot to remove all the documentation. Signed-off-by: NCleber Rosa <crosa@redhat.com>
-
- 03 8月, 2020 11 次提交
-
-
由 Cleber Rosa 提交于
Signed-off-by: NCleber Rosa <crosa@redhat.com>
-
由 Cleber Rosa 提交于
Signed-off-by: NCleber Rosa <crosa@redhat.com>
-
由 Cleber Rosa 提交于
Signed-off-by: NCleber Rosa <crosa@redhat.com>
-
由 Cleber Rosa 提交于
Signed-off-by: NCleber Rosa <crosa@redhat.com>
-
由 Cleber Rosa 提交于
Signed-off-by: NCleber Rosa <crosa@redhat.com>
-
由 Narasimhan V 提交于
Handling yes for vgcreate command, since vg creation can fail if a filesystem is already present. Using the force parameter for this. Reported-by: NWen Xiong <wenxiong@linux.vnet.ibm.com> Signed-off-by: NNarasimhan V <sim@linux.vnet.ibm.com>
-
由 Cleber Rosa 提交于
So that the features they use, end up being tested as a whole. Note: these may and should become irrelevant once we have complete Job API feature coverage, but it should catch and avoid breakages for now. Signed-off-by: NCleber Rosa <crosa@redhat.com>
-
由 Cleber Rosa 提交于
Like the other example jobs. Signed-off-by: NCleber Rosa <crosa@redhat.com>
-
由 Cleber Rosa 提交于
With the introduction of the TestSuite class, this plugin need to use the right attributes, like `tests` and `size`. Signed-off-by: NCleber Rosa <crosa@redhat.com>
-
由 Cleber Rosa 提交于
Currently we pass/show the raw CmdError as the failure information, resulting in output like this: JOB ID : 989cf0b0c8d90fb593eb5bc6ff9291aa3248955c JOB LOG : /home/cleber/avocado/job-results/job-2020-08-02T16.31-989cf0b/job.log (1/1) /bin/false: FAIL: Command '/bin/false' failed.\nstdout: b''\nstderr: b''\nadditional_info: None (0.06 s) RESULTS : PASS 0 | ERROR 0 | FAIL 1 | SKIP 0 | WARN 0 | INTERRUPT 0 | CANCEL 0 JOB HTML : /home/cleber/avocado/job-results/job-2020-08-02T16.31-989cf0b/results.html JOB TIME : 0.14 s Which contains a duplicate of the command path (already in the test name) an always empty and useless "additional_info", needless new lines, and very bad formatting. This changes turns it into: JOB ID : 9cde24d36988d6441f755c14f79b2e908ae533ef JOB LOG : /home/cleber/avocado/job-results/job-2020-08-02T16.58-9cde24d/job.log (1/1) /bin/false: FAIL: Exited with status: '1', stdout: '' stderr: '' (0.08 s) RESULTS : PASS 0 | ERROR 0 | FAIL 1 | SKIP 0 | WARN 0 | INTERRUPT 0 | CANCEL 0 JOB HTML : /home/cleber/avocado/job-results/job-2020-08-02T16.58-9cde24d/results.html JOB TIME : 0.16 s Removing the superfluous, and adding the execution exit status, and better formatting. Signed-off-by: NCleber Rosa <crosa@redhat.com>
-
由 Cleber Rosa 提交于
Currently, when running `avocado run`, the 'HTML RESULT' line is printed, even if the job had no tests in the test suite and did not really run. The reason is that on 856c3b57 the TestSuite() class was introduced, and with that, a different exception started to being raised when an empty test suite was created. When the new exception is handled, the job is set to FAILed, instead of ERRORed. We could change the exception or their handling, but I think this side effect is actually a good thing. A Job with an empty test suite failed to proceed, but did not really ERRORed. ERROR is currently reserved, and used in most cases, when things that are not expected happen, and say, Avocado crashes. Signed-off-by: NCleber Rosa <crosa@redhat.com>
-