Mailman 3 April 2016 - pytest-dev

Moving pytest-rerunfailures to pytest-dev
by Florian Pilz Oct. 27, 2016

Oct. 27, 2016

Hi, I’m asking whether `pytest-rerunfailures` should be moved to the `pytest-dev` organisation (https://github.com/gocept/pytest-rerunfailures <https://github.com/gocept/pytest-rerunfailures>). Since we (gocept) only rarely take time to work on issues / pull requests of `pytest-rerunfailures` we would prefer to move the extremely useful plugin to `pytest-dev`. There we still can work on it, but are not the only ones taking care of the project. So if you think `pytest-rerunfailures` is a great addition to `pytest` and should be moved to the pytest-dev organization, let us know. :) Best wishes Florian ––––––––– Florian Pilz · fp(a)gocept.com · Software Developer gocept gmbh & co. kg · Forsterstraße 29 · 06112 Halle (Saale) · Germany http://gocept.com <http://gocept.com/> · Tel +49 345 122 9889 7 Python, Pyramid, Plone, Zope · consulting, development, hosting, operations

3 8

Testing of console scripts
by Vasily Kuznetsov June 13, 2016

June 13, 2016

Hi! I was recently writing some tests for a bunch of console scripts. I ended up having a parametrised function that can run a script for real (using subprocess.run) or can import its main function, mock the environment and execute it in the same process (this is much faster). I've done similar things before and in general this seems to be a reasonably common use case. I couldn't find a pytest plugin for testing console scripts, so I'm wondering if it would be useful to create one. What do you think? Cheers, Vasily

5 12

Timing tests and fixtures
by Florian Schulze April 28, 2016

April 28, 2016

Hi! Is there a plugin to measure the execution time of tests and fixtures and the usage count of fixtures? There seems to be a profile plugin, but that is way to detailed for my need. I want to find slow tests and fixtures to see if I can optimize them. Regards, Florian Schulze

3 3

Participation in Pytest sprint
by Ana Ribeiro April 23, 2016

April 23, 2016

Hello! My name is Ana and I am an almost 20 years old girl from Brazil (my birthday is on may) and this year I am going to work with Pytest this summer in an internship program named "Outreachy" ( https://github.com/davehunt/pytest-html#outreachy). I am going to London to participate in "All Hands" Mozilla meeting and my mentor (Dave Hunt) said that after this week he is going to Germany to participate in this sprint ( http://pytest.org/latest/announce/sprint2016.html), and we think that is a great idea if I go as well since I will have the chance to meet so many expert people and get help of you all on the development of my project. The funding I need is to cover my travel expenses from London to Freiburg (I looked for some flights and trains and I think it is about 200€) and a place to stay (I would love to stay with everyone in the same house/flat) and maybe money for some food or some food. Regards, Ana

3 7

pytest sprint
by Sivan Greenberg April 22, 2016

April 22, 2016

Hi All, I'm really interested in the sprint! specifically in contributing to a standardized server testing fixture as I miss the ability to configure a full www-root structure with pytest-localserver and pytest-server-fixtures seems broken through pip (but I did not investigate too deep to confirm this is not just in my setup ;)). And indeed, the fixture system in general needs more helpb to be understandable. Also, as I'm using pytest to test PyQt based mechanized browsers, I'd love to see what work needs to be done to have something around that standardized as well - namely, QApplication setup and teardown (modulescope breaks its) and anything I can help off the list of items to work on from the list. I'll be coming from Tel Aviv, a flight to Frankfurt / Berlin would be around 450$ or a bit less, Skyscanner does not show me how can one get from Tel Aviv to Freiburg via air travel. Perhaps a train is needed? :) (Is Freiburg in Switzerland , or ?.. A bit confused after my googling). Many thanks, (and thanks for pytest!) -Sivan

3 5

Optional static typing plugin?
by Alessandro Amici April 21, 2016

April 21, 2016

Hey, I'm starting using python 3 optional static typing and the mypy checker in one of my projects and it's working quite nicely, but I could not find a mypy, or any other static typing checker, plugin for pytest, so I'm now left to run two independent tools for every quality control check. Is there a way to check optional static typing from within pytest, that I overlooked? Is anybody working on a mypy plugin of anything comparable? In anybody else interested? Cheers, Alessandro

2 2

issue tracker labeling for sprint preps
by Ronny Pfannschmidt April 20, 2016

April 20, 2016

Hi all, for obsessive planing i created https://github.com/pytest-dev/pytest/labels/sprint-candidate i hope it can be a useful tool, lets take a look at how things turn out. -- Ronny

1 0

Re: [pytest-dev] pytest overall execution timeout option
by Shankar Hiremath April 20, 2016

April 20, 2016

Hi Bruno, I tried with the below approach provided by you, when i increase the sleep time inside the test_foo to 10 sec, then the test execution took 20.12 sec and then after further test case execution stooped. but what i required is to stop the test itself instead of waiting for the test to complete (in this case 10 sec sleep), i agree for this case pytest-timeout is the best option. My use case is little bit different: whenever there is a new commit, i am going to run smoke tests which are annotated with “pytest.mark.smoke” annotation in pytest (around 400 test cases) when the new commit code quality is of good then all of the smoke tests will get completed well within 2 hours duration, when there is a product issue due to new commit randomly few of the tests might get hanged for some time due to that, the overall test execution will take > 2hours. 1) if all tests passed well within 2 hour duration i will say “+1” to the commit, 2) if any test failed & all are well within 2 hour duration all tests completed i will say “-1” to the commit, 3) if tests are taking more time >2 hours or any test hanged, i want to stop the current running test & stop further test execution, i will say “-1” to the commit, Inorder to achieve the case-3, i need to have some mechanism of global timeout of 2 hours (i am ok with +/- 5 minutes difference accuracy) from pytest execution. Any suggestion or approach to achieve this will very helpful. Thanks in advance. def pytest_sessionstart(session): session.start_time = time.time() @pytest.fixture(autouse=True) def check_session_time(request): elapsed = time.time() - request.session.start_time if elapsed > 1.0: request.session.shouldstop = 'time limit reached: %0.2f seconds' % elapsed @pytest.mark.parametrize('i', range(10)) def test_foo(i): time.sleep(10) The output of pytest execution is as below: ================================================= test session starts ================================================= platform linux2 -- Python 2.7.6 -- py-1.4.31 -- pytest-2.6.2 plugins: timeout collected 10 items a.py .. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! Interrupted: time limit reached: 10.02 seconds !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! ============================================== 2 passed in 20.12 seconds ============================================= Thanks -Shankar On Apr 19, 2016, at 9:22 PM, Shankar Hiremath <shankarhiremath2006(a)gmail.com> wrote: > Hi Bruno, > > Thanks a lot, this meets my requirement. I will go ahead with the approach provided by you. > > Regards > -Shankar > > On Mon, Apr 18, 2016 at 6:56 PM, Bruno Oliveira <nicoddemus(a)gmail.com> wrote: > Hi Shankar, > > I don’t know of any plugin that does this out of the box, but the simplest way I found is to implement this is using an auto-use fixture: > > # contents of conftest.py > import pytest > import time > > def pytest_sessionstart(session): > session.start_time = time.time() > > @pytest.fixture(autouse=True) > def check_session_time(request): > elapsed = time.time() - request.session.start_time > if elapsed > 1.0: > request.session.shouldstop = 'time limit reached: %0.2f seconds' % elapsed > > # contents of test_foo.py > import pytest > import time > > @pytest.mark.parametrize('i', range(10)) > def test_foo(i): > time.sleep(0.5) > Running this produces: > > collected 10 items > > test_foo.py ... > > !!!!!!!!!!!!!!!! Interrupted: time limit reached: 1.11 seconds !!!!!!!!!!!!!!!! > ========================== 3 passed in 1.69 seconds =========================== > Hope that helps. > > Cheers, > Bruno. > > > On Mon, Apr 18, 2016 at 8:56 AM Shankar Hiremath <shankarhiremath2006(a)gmail.com> wrote: > Hi All, > > Is there any any plugin similar to "pytest-timeout” but at higher level (ex: complete suite level execution timeout) > > My requirement is to run pytest with test suites and if all tests finishes well within the 2 hours time duration then its good, else I want to stop the test execution. > ( I mean stop the current running test & further tests if any, but the pytest call backs should get executed, for example: pytest_sessionfinish need to be executed, before exiting pytest) > > Please suggest me if any existing plugins provides this feature, or any easy way to achieve this feature in pytest. > > Thanks > -Shankar > _______________________________________________ > pytest-dev mailing list > pytest-dev(a)python.org > https://mail.python.org/mailman/listinfo/pytest-dev >

4 10

Test discovery by subclassing
by Dave April 19, 2016

April 19, 2016

Hi, Instead of discovering tests by name matching is there any way to specify that all subclasses of X be collected? I realise this can be done by subclassing unittest.TestCase but that is a) pretty heavyweight and b) seems to break parameterised fixtures for me. Thanks, Dave

3 4

Re: [pytest-dev] pytest overall execution timeout option
by Bruno Oliveira April 18, 2016

April 18, 2016

On Mon, Apr 18, 2016 at 10:47 AM Ronny Pfannschmidt < ich(a)ronnypfannschmidt.de> wrote: > the given simple implementation is called at the setup of a test item and > has effect after teardown of that test item > > (note how py.test took 1.69 seconds, even tho the timeout is 1 second and > each test takes 0.5 seconds) > > the reason for that difference is that the exit condition is set at the > setup of the test that will take 0.5 seconds, and then is checked after the > teardown of that test in the py.test testloop > > to get more exact the check has to be done at test teardown, > > to get even more exact, you need to use tools similar to pytest-timeout > and calculate exact maximum remaining run times in a fixture > > it would be helpful if you outlined your use-case a bit more exactly > instead of the envisioned implementation as "global timeout" > Ronny is correct if one wants more control exactly when to stop. >From the brief description provided by Shankar I assumed a "soft timeout" was what he wanted, which is why I went for the simpler implementation which achieves that. Cheers, Bruno.

1 0