[pytest-dev] Fwd: Continuing on failures

[Brian Okken]

Victor,

This would be brilliant. There is a fairly large need for a soft assert
feature.
pytest-check would not be needed if it existed.

I personally would not want it to be an overriding of assert, but perhaps a
different function that is used almost identically to assert, including
failing the test, but just not stopping execution.
This is also the reason why many people resort to subtest.

-Brian Okken

On Fri, Mar 27, 2020 at 9:19 AM Victor Maryama <victor.maryama at gmail.com>
wrote:

> Oh sure, I see now what you mean.
>
> Back then when I was thinking about this subject it seemed to me that soft
> assertions could be something implemented in pytest core itself using the
> assertion rewriting module (which in turn would put the assertion calls
> around the context manager), along the lines of what Bruno mentioned.
> Except that I probably would not go for a generic hook, but just the soft
> assertion feature itself.
>
>
> On Fri, Mar 27, 2020, 5:09 PM Maximilian Roos <m at maximilianroos.com>
> wrote:
>
>> Hi Victor,
>>
>> > What about pytest-assume as a context manager around the assert line?
>>
>> Yes, this could work with either a context manager or a function like
>> pytest-check uses.
>> But it couples the plugin to the tests; it's no longer possible to run
>> the plugin on normal tests with a plain `assert`, which is some of the
>> brilliant philosophy of pytest.
>>
>> Without such a hook, I think the function or context manager is the best
>> path forward, and is a reasonable way to build a better prototype before
>> adding hooks deep in pytest internals.
>>
>> Thank you!
>> Max
>>
>> On Fri, 27 Mar 2020 at 09:26, Victor Maryama <victor.maryama at gmail.com>
>> wrote:
>>
>>> What about pytest-assume as a context manager around the assert line?
>>> That way you get the soft assertions feature and pytest still rewrites
>>> the assertions.
>>>
>>> On Fri, Mar 27, 2020, 12:47 PM Bruno Oliveira <nicoddemus at gmail.com>
>>> wrote:
>>>
>>>> Hi Maximilian,
>>>>
>>>> On Thu, Mar 26, 2020 at 7:29 PM Maximilian Roos <m at maximilianroos.com>
>>>> wrote:
>>>>
>>>>>
>>>>> To prefix: as a long time user and evangelizer of pytest, a big thank
>>>>> you for the superb tool.
>>>>>
>>>>
>>>> Thanks for the kind words, we appreciate it.
>>>>
>>>>
>>>>> As per that README, it only works on the first failure, because the
>>>>> test doesn't continue execution beyond that. *Is there any way of
>>>>> allowing execution to continue, without ripping up the internals of pytest?*
>>>>> My prior is "no", since the assertion rewrite still retains an assertion,
>>>>> and there's no hook for changing that.
>>>>>
>>>>
>>>> Not at the moment I'm afraid.
>>>>
>>>> I think it should be possible to implement a hook that transforms the
>>>> AST, which our rewriter would then call. I think, with this hook in place,
>>>> you would be able to rewrite assert statements into a function call of your
>>>> plugin, which would then be free to do whatever it wanted. If you are
>>>> interested, I suggest writing up a proposal in the issue tracker so it can
>>>> be discussed.
>>>>
>>>> Cheers,
>>>> Bruno
>>>>
>>>>
>>>>> Ref: a similar question I asked @okken
>>>>> https://github.com/okken/pytest-check/issues/32
>>>>>
>>>>> Thank you,
>>>>> Max
>>>>> _______________________________________________
>>>>> pytest-dev mailing list
>>>>> pytest-dev at python.org
>>>>> https://mail.python.org/mailman/listinfo/pytest-dev
>>>>>
>>>> _______________________________________________
>>>> pytest-dev mailing list
>>>> pytest-dev at python.org
>>>> https://mail.python.org/mailman/listinfo/pytest-dev
>>>>
>>>> _______________________________________________
> pytest-dev mailing list
> pytest-dev at python.org
> https://mail.python.org/mailman/listinfo/pytest-dev
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/pytest-dev/attachments/20200327/e99085b0/attachment.html>