The project is available on GitHub at https://github.com/AnyBlok/anyblok_wms_base.
As far as workflow is concerned, we’re using standard GitHub pull requests and issues, and that’s it!
If you’re submitting a bugfix, we’d appreciate if you could also fill in an issue, because it’s more visible to other users.
If you want to know if we’d be interested in some improvements, just fill an issue.
Please add the standard MPL 2.0 header on any new files you’d contribute.
Don’t forget to add yourself to the Credits page! (unless you don’t want to).
We’re using flake8 with its default settings. This is enforced by the Travis CI system that runs on pull requests.
To check by yourself:
pip install flake8
flake8 anyblok_wms_base
We also recommend using the flake8 Git pre-commit hook.
The global tests coverage has been at 100% for almost all the history of the project, it’s checked by the continuous integration, and it must stay this way.
Don’t worry, we’ll help you getting your pull request to that level.
Of course, It doesn’t prove that all important situations are really tested, but we see it the other way around: a line uncovered is certainly not tested! In any case, full tests coverage allows us to very quickly pinpoint problems during refactorings.
The self-coverage of some Python packages of this project have also been at 100% for a fair amount of time, and it’s been really great for refactorings.
This means that running the tests of some Blok or subpackage gives 100% coverage of that package.
Ideally, we’d like it to stay that way, but it’s not readily enforced by the current continuous integration systems.
This is currently true for
wms-core and wms-quantity Bloks:
Note
There’s a special case for the forthcoming wms-inventory
Blok. Since it has functionality that depends on whether
wms-reservation is installed or not, full coverage can only be
achieved by runnig the tests twice, one without wms-reservation
and one with it. Of course run_all_tests.py does precisely this.
We expect more similar cases to appear in the future.
make sure you have a local PostgreSQL server running, with the standard
command-line interface (psql), accepting requests on UNIX
Socket Domain (this is the default in most installations)
make sure that your current system user has a corresponding PostgreSQL role under peer authentication. This is typically ensured by something like:
sudo -u postgres createuser -dSR `whoami`
you’ll need a virtual environment:
python3 -m venv /path/to/virtualenv
source /path/to/virtualenv/bin/activate
it’s necessary to install AnyBlok / Wms Base in the virtualenv, so that its Bloks are registered. From your Git clone, just do:
pip install -e .
To run all the tests (it takes about one or two minutes depending on your development rig) from your Git clone:
./run_all_tests.py
The tests run on a dedicated database that is created afresh each time.
Just do, e.g:
./test_partial.py wms-core
or for just a subpackage:
./test_partial.py wms-core --subpkg physobj
for more options:
./test_partial.py --help
The script ends with a coverage report and instructions on how to rerun more precise tests:
$ ./test_partial.py wms-core --subpkg physobj
(...)
Name Stmts Miss Cover
---------------------------------------------------------------
anyblok_wms_base/core/physobj/__init__.py 7 0 100%
anyblok_wms_base/core/physobj/main.py 241 0 100%
anyblok_wms_base/core/physobj/type.py 99 0 100%
---------------------------------------------------------------
TOTAL 347 0 100%
----------------------------------------------------------------------
Ran 38 tests in 3.200s
OK
For HTML coverage report, open file:///tmp/COVER-anyblok_wms_base.core.physobj/index.html
To run again with different options, maybe incompatible with the Nose plugin, or a precise test, do something based on:
anyblok_nose --db-driver-name postgresql --db-name test_awb_anyblok_wms_base.core.physobj -- anyblok_wms_base.core.physobj