| Age | Commit message (Collapse) | Author | Files | Lines |
|---|
|
change default logging mode to "slim".
the old default mode, which was "default", is
not any more. We change the default logging mode
to "slim", which means delete temp files per case.
also not to make users missunderstanding the meaning,
old option name "default" is now named as "origin".
Bug-AGL: SPEC-4753
Signed-off-by: Yan <yanxk.fnst@fujitsu.com>
Change-Id: I98ab15935970ab0298e8fe972ba9a8ecd3eca722
|
|
add new options: "default", "slim", "clear".
users can use export LOG_MODE="" to control how the
framework would delete all its temp files or folders.
"default": no influence, nothing would be removed.
"slim": test self-owned temp files and folders removed.
"clear": clean up each and every temp files and folders.
Bug-AGL: SPEC-4733
Signed-off-by: Yan <yanxk.fnst@fujitsu.com>
Change-Id: Ia3d8c2a971bbf97749448f732e848d9d15193459
|
|
Fixed problems:
1.When all the executed test suites are skipped, no new summary-
report is generated.
2.When there is test suite successfully executed, the generated
summary-report will contain the test suites which have logs
in the "/tmp-log" directory, rather than just the test suites
which has just executed.
Correction method:
1.Move the function in "conftest.py" to "agl_test_report.py"
and call it in the "agl-test" script to ensure that the
function will be executed.
2.Set the timestamp to record the information of each executed
test suite or skipped test suite to the file of
"test_list_timestamp". And generate a summar-report according
to the above file.
Bug-AGL: SPEC-4345
Signed-off-by: duerpei <duep.fnst@fujitsu.com>
Change-Id: I47bfd09706e37ce6bdc13f3f9f266bc62f74f777
|
|
'agl-test' modified, now, usr would not run dangerous
tests like this 'crashme' unless they do self selection
by '-m' pytest option. Set default to 'not dangerous'.
a new marker 'dangerous' has been add to the pytest.ini.
add new method into the agl_test_base.py, to help get path
of the 'spec.json', this file help stress tests like 'crashme'
to run under different stress levels.
for run_tests.py, modifications focus on how to run 'crashme'
under that different levels.
new marker 'order' would always ensure this 'crashme' test to
be placed as the so called 'last' test of all.
Bug-AGL: SPEC-4345
Signed-off-by: yanxk <yanxk.fnst@fujitsu.com>
Change-Id: Icb36d7c5c52430d89787f4373a2852b1b03ec8a6
|
|
Submit the demo code of agl-test-framework
The "agl-test framework" encapsulates pytest,which aims to provide a
unified test set execution entrance. It supports to run various test sets,
even these test sets come from different test frameworks, processing these
test logs uniformly, and generating complete test report.
In this way, it is convenient to test as many targets as possible,
in a wide range, so that the test covers a wider range of objects,
and is more comprehensive.
At present, we plan to support the following test sets in "agl-test":
1. Transplant test sets under Fuego and AGL-JTA
2. Retain the test set under pyagl and agl-ptest
(so will depend on "agl-ptest")
3. Migrate new test sets (with upstream)
4. Append new test sets (without upstream)
The output of test run is summarized by levels.
The first level is the summary of all test sets, and the second level is
the summary of a single test set. Now, they are displayed in HTML format,
and other formats also can be considered later.
Bug-AGL: SPEC-4345
Signed-off-by: duerpei <duep.fnst@fujitsu.com>
Change-Id: I25dfedcf8cdd373544c4fae677330defb5d21840
|
Yan
duerpei