Initial Version #1
Conversation
* Added the basic Python package stuff (setup.py, etc). * Added a dummy test to setup Travis-CI.
reStructuredText is the prefered format of PyPI.
We don't need coverage of the `_version.py` file.
So we fail faster.
Until pypa/pipenv#857 is fixed, pipenv and hypothesis don't quite play will together.
From the hypothesis [docs](https://hypothesis.readthedocs.io/en/latest/supported.html#testing-frameworks): > However you should probably not use Coverage, Hypothesis and PyPy together. > Because Hypothesis does quite a lot of CPU heavy work compared to normal tests, > it really exacerbates the performance problems the two normally have working together.
Hypothesis tests just don't seem to work well in pypy.
Seem to be running into [this](pypa/setuptools#951) with Python 3.4 tests.
| @@ -0,0 +1,24 @@ | |||
| [[source]] | |||
|
|
|||
| url = "https://pypi.python.org/simple" | |||
There was a problem hiding this comment.
Rather trivial, but this URL redirects to add a tailing / . This makes me wonder how much energy and time amending this would save v.s. how much it takes to report+act on this.
There was a problem hiding this comment.
Depends on what pipenv's doing with it. I bet it's just urljoin()ing it with each package name.
pytest_localstack/container.py
Outdated
| for line in logs_generator: | ||
| if self.encoding is not None and isinstance(line, bytes): | ||
| line = line.decode(self.encoding) | ||
| self.logger.log(self.log_level, line.rstrip()) |
There was a problem hiding this comment.
pedantic: I am wondering if line.rstrip() is reasonable to relax to line or line[:-len(os.linesep)]. I am happy with it as it is anyway.
There was a problem hiding this comment.
thinking of it, it may be good in some imaginary scenario to give additional context to the logger, e.g. extra=container. Then again using loggers like myapp.tests.test_name.container_name would probably be a better way for people to get that information if they really wanted it.
There was a problem hiding this comment.
Which line separator you see would depending on the container. Probably have to do
line = re.sub(r'(?:\r?\n|\n\r)$', '', line)- Remove only single newline from each line in DockerTailer
The extra context would be good to do later. Raise an issue!
There was a problem hiding this comment.
Good point, I wasn't thinking with portals. Err, containers.
|
@trickv @nabadger I think we're going to need a Mintel user for https://pypi.python.org/pypi and https://testpypi.python.org/pypi to be one of the package owners and to let Travis-CI push package versions. |
|
Since this is basically functional, I'm thinking of merge to master and handling to remaining TODOs as separate issues (pending review). |
setup.cfg
Outdated
| # H401: docstring should not start with a space | ||
| # H403: multi line docstrings should end on a new line | ||
| # H404: multi line docstring should start without a leading new line | ||
| E133,D104,D105,D107,D203,D212,D213,D301,D400,D404,D401,D402,F812,H101,H202,H401,H403,H405 |
There was a problem hiding this comment.
Are all of these ignore rules really required? :)
There was a problem hiding this comment.
I'll check. It is partially cargo-culted. Recording why we're ignoring them would also be a good idea.
- Review flake8 ignored rules
Makefile
Outdated
| PIPENV := PIPENV_VENV_IN_PROJECT=1 pipenv | ||
| WITH_PIPENV := $(PIPENV) run | ||
|
|
||
| help: ## print this help |
There was a problem hiding this comment.
I don't know much about makefiles, but don't you also need to mark help as being .PHONY? It is in the link below.
The cache is huge; it's probably not speeding anything up at this point.
All these tests are expensive. We should only run them for - master - releases - pull requests [skip ci]
jtdoepke
force-pushed
the
initial
branch
2 times, most recently
from
77a130d
to
a616f71
Compare
No need to hold up all the tests waiting for lint.
This PR is the ongoing work to populate an initial version of this library.
TODO
Boilerplate
CONTRIBUTING.rstDocs
Testing
Review
Deployment