1
0
mirror of https://github.com/RIOT-OS/RIOT.git synced 2024-12-29 04:50:03 +01:00

tests/doc: reorganization and improvements to tests doc

This commit is contained in:
krzysztof-cabaj 2022-05-17 10:21:14 -04:00
parent 897abd09f2
commit f13794fdcd

View File

@ -1,14 +1,90 @@
Running and creating tests {#running-and-creating-tests}
==========================
# Running and creating tests {#running-and-creating-tests}
There are a number of tests included in RIOT. They are located in the
[tests folder](https://github.com/RIOT-OS/RIOT/tree/master/tests). These tests
allow basic functionality to be verified as well as provide an example of
usage.
# Running automated tests
Implementing automated tests
----------------------------
Some tests can be performed automatically. The test automation scripts are
defined in the `<test_application>/tests/` folder. They are written in python
and interact through the serial (typically UART) with the test application code running on a
board to do the validation. It is recommended to flash the board with the
test just before running it because some platforms cannot be reset while
testing.
## Running single test
From the test application directory run:
BOARD=<board_of_your_choice> make flash test
An automated way of knowing if a test is available is to execute the
'test/available' target from the test application directory.
It executes without error if tests run by 'make test' are present.
make test/available
## Running all test for particular board
If you would like execute all tests for given board, you could use dedicated
script `compile_and_test_for_board.py`
Go to main RIOT directory and execute command:
./dist/tools/compile_and_test_for_board/compile_and_test_for_board.py . <board_of_your_choice> --with-test-only --jobs=4
More details concerning other available parameters provided by this tool can be found in
[README.md file](https://github.com/RIOT-OS/RIOT/tree/master/dist/tools/compile_and_test_for_board)
and directly in [compile_and_test_for_board.py](https://github.com/RIOT-OS/RIOT/tree/master/dist/tools/compile_and_test_for_board/compile_and_test_for_board.py) script.
## Running tests that require a preliminary manual configuration
Some tests need active monitoring or manual setup steps but still have some
automated scripts. The test automation scripts are defined in the
`<test_application>/tests-with-config/` folder.
For running them, follow the setup or analysis documentation and use the
`test-with-config` target.
## Running tests that require root privileges
Some tests require root privileges to launch their automated script. In this
case, the test automation scripts are defined in the
`<test_application>/tests-as-root/` folder.
For running them, follow the setup or analysis documentation and use the
`test-as-root` target.
## Cleaning intermediate files
After test execution intermediate files are not automatically deleted.
Execution of multiple tests, especially all for particular board could generate
many files. For example, after execution of all test for stm32f469i-disco board
(more than 230 tests) around 7.5 GB of intermediate files are created.
There are few methods for cleaning intermediate files.
If you would like to clean intermediate file only for particular board you should
go to main RIOT directory and execute one from these commands:
./dist/tools/compile_and_test_for_board/compile_and_test_for_board.py . <board_of_your_choice> --compile-targets clean
or
make BOARD=<board_of_your_choice> clean
If you would like to clean intermediate files for all boards go to main RIOT
directory and use this command.
@warning This command cleans all local files, for example, pkg downloads and
locally generared docs.
make distclean
# Implementing automated tests
The goal is to be able to run all tests in a sequential way for as many targets
as possible.
@ -34,32 +110,7 @@ script and use `test_utils_interactive_sync();` function (e.g.:
you can disable `test_utils_interactive_sync_shell` module in the application
`Makefile`: `DISABLE_MODULE += test_utils_interactive_sync_shell`.
Running automated tests
-----------------------
Some tests can be performed automatically. The test automation scripts are
defined in the `<test_application>/tests/` folder. They are written in python
and interact through the uart with the test application code running on a
board to do the validation. It is recommended to flash the board with the
test just before running it because some platforms cannot be reset while
testing.
From the test application directory run:
BOARD=<board_of_your_choice> make flash test
An automated way of knowing if a test is available is to execute the
'test/available' target from the test application directory.
It executes without error if tests run by 'make test' are present.
make test/available
Automated Tests Guidelines
--------------------------
## Automated Tests Guidelines
When using `pexpect` `$` is useless for matching the end of a line, instead use
`\r\n`([pexpect end-of-line](https://pexpect.readthedocs.io/en/stable/overview.html#find-the-end-of-line-cr-lf-conventions)).
@ -93,8 +144,7 @@ pattern like `\r\n`, `\s`, `\)`, etc..
child.expect(r'some string: (\d+) ,')
~~~
Use expect() instead of assert()
--------------------------------
## Use expect() instead of assert()
In order to make a test application functional in all cases, use `expect()`
instead of `assert()`. The former works like the latter, but will still be
@ -104,8 +154,7 @@ be compiled with that flag. Otherwise, the application would force compiling
all tested code with assertions enabled.
`expect()` is defined in the header `test_utils/expect.h`.
Interaction through the uart
----------------------------
## Interaction through the uart
Tests implemented with `testrunner` use the `cleanterm` target that
provides an interaction without adding extra text output or input handling.
@ -117,22 +166,3 @@ The expected behavior is verified with the test in `tests/test_tools`.
Tests cannot rely on having on all boards and terminal programs:
* unbuffered input
* allowing sending special characters like `ctrl+c/ctrl+d`
Running tests that require a preliminary manual configuration
-------------------------------------------------------------
Some tests need active monitoring or manual setup steps but still have some
automated scripts. The test automation scripts are defined in the
`<test_application>/tests-with-config/` folder.
For running them, follow the setup or analysis documentation and use the
`test-with-config` target.
Running tests that require root privileges
------------------------------------------
Some tests require root privileges to launch their automated script. In this
case, the test automation scripts are defined in the
`<test_application>/tests-as-root/` folder.
For running them, follow the setup or analysis documentation and use the
`test-as-root` target.