You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@cassandra.apache.org by Michael Kjellman <kj...@apple.com> on 2017/11/28 18:49:25 UTC
[PROPOSAL] Migrate to pytest from nosetests for dtests
I'd like to propose we move from nosetest to pytest for the dtests. It looks like nosetests is basically abandoned, the python community doesn't like it, it hasn't been updated since 2015, and pytest even has nosetests support which would help us greatly during migration (https://docs.pytest.org/en/latest/nose.html).
Thoughts?
best,
kjellman
Re: [PROPOSAL] Migrate to pytest from nosetests for dtests
Posted by kurt greaves <ku...@instaclustr.com>.
+1nb, Never been that fond of nose from a usability perspective, and I
wouldn't be surprised if at least some of the problems running dtests were
related to issues in nose. I can't imagine it would be a lot of work to
port to py.test, if someone wants to do it they can go right ahead.
On 29 November 2017 at 17:55, Michael Kjellman <mkjellman@internalcircle.com
> wrote:
> s/handling/hanging
>
> > On Nov 29, 2017, at 9:54 AM, Michael Kjellman <
> mkjellman@internalcircle.com> wrote:
> >
> > i keep seeing nose randomly handing after a test successfully completes
> execution. i’m very far from a python guru but i spent a few hours with gdb
> trying to debug the thing and get python stacks and got symbolicated native
> stacks but it’s random but root causing while nose is sitting on a lock
> forever alludes me. some tests are more reproducible than others. some i
> see fail 1 in 10 runs.
> >
> > the net of it all though is this makes people not trust dtests because
> it randomly hangs and shows tests with “failures” that actually succeeded.
> >
> > i’m not a huge fan of just blindly upgrading to fix a problem but in
> this case I found that there is quite a lot of mistrust and dislike for
> nosetests in the python community with most projects already moving to
> pytest. and if it is some complicated set of interactions between threads
> we use in the tests and how nose works do we really want to even debug it
> when the project appears to be abandoned?
> >
> > i think regardless of the root cause for making things more stable it
> seems like there is little motivation to stick around on nose...
> >
> > lmk!
> >
> > best,
> > kjellman
> >
> >> On Nov 29, 2017, at 5:33 AM, Philip Thompson <
> philip.thompson@datastax.com> wrote:
> >>
> >> I don't have any objection to this, really. I know I rely on a handful
> of
> >> nose plugins, and possibly others do, but those should be easy enough to
> >> re-write. I am curious though, what's the impetus for this? Is there
> some
> >> pytest feature we want that nose lacks? Is there some nosetest bug or
> >> restriction getting in the way?
> >>
> >>> On Tue, Nov 28, 2017 at 8:34 PM, Jon Haddad <jo...@jonhaddad.com> wrote:
> >>>
> >>> +1
> >>>
> >>> I stopped using nose a long time ago in favor of py.test. It’s a
> >>> significant improvement.
> >>>
> >>>> On Nov 28, 2017, at 10:49 AM, Michael Kjellman <kj...@apple.com>
> >>> wrote:
> >>>>
> >>>> I'd like to propose we move from nosetest to pytest for the dtests. It
> >>> looks like nosetests is basically abandoned, the python community
> doesn't
> >>> like it, it hasn't been updated since 2015, and pytest even has
> nosetests
> >>> support which would help us greatly during migration (
> >>> https://docs.pytest.org/en/latest/nose.html).
> >>>>
> >>>> Thoughts?
> >>>>
> >>>> best,
> >>>> kjellman
> >>>
> >>>
> >>> ---------------------------------------------------------------------
> >>> To unsubscribe, e-mail: dev-unsubscribe@cassandra.apache.org
> >>> For additional commands, e-mail: dev-help@cassandra.apache.org
> >>>
> >>>
> >
> > ---------------------------------------------------------------------
> > To unsubscribe, e-mail: dev-unsubscribe@cassandra.apache.org
> > For additional commands, e-mail: dev-help@cassandra.apache.org
> >
>
Re: [PROPOSAL] Migrate to pytest from nosetests for dtests
Posted by Michael Kjellman <mk...@internalcircle.com>.
s/handling/hanging
> On Nov 29, 2017, at 9:54 AM, Michael Kjellman <mk...@internalcircle.com> wrote:
>
> i keep seeing nose randomly handing after a test successfully completes execution. i’m very far from a python guru but i spent a few hours with gdb trying to debug the thing and get python stacks and got symbolicated native stacks but it’s random but root causing while nose is sitting on a lock forever alludes me. some tests are more reproducible than others. some i see fail 1 in 10 runs.
>
> the net of it all though is this makes people not trust dtests because it randomly hangs and shows tests with “failures” that actually succeeded.
>
> i’m not a huge fan of just blindly upgrading to fix a problem but in this case I found that there is quite a lot of mistrust and dislike for nosetests in the python community with most projects already moving to pytest. and if it is some complicated set of interactions between threads we use in the tests and how nose works do we really want to even debug it when the project appears to be abandoned?
>
> i think regardless of the root cause for making things more stable it seems like there is little motivation to stick around on nose...
>
> lmk!
>
> best,
> kjellman
>
>> On Nov 29, 2017, at 5:33 AM, Philip Thompson <ph...@datastax.com> wrote:
>>
>> I don't have any objection to this, really. I know I rely on a handful of
>> nose plugins, and possibly others do, but those should be easy enough to
>> re-write. I am curious though, what's the impetus for this? Is there some
>> pytest feature we want that nose lacks? Is there some nosetest bug or
>> restriction getting in the way?
>>
>>> On Tue, Nov 28, 2017 at 8:34 PM, Jon Haddad <jo...@jonhaddad.com> wrote:
>>>
>>> +1
>>>
>>> I stopped using nose a long time ago in favor of py.test. It’s a
>>> significant improvement.
>>>
>>>> On Nov 28, 2017, at 10:49 AM, Michael Kjellman <kj...@apple.com>
>>> wrote:
>>>>
>>>> I'd like to propose we move from nosetest to pytest for the dtests. It
>>> looks like nosetests is basically abandoned, the python community doesn't
>>> like it, it hasn't been updated since 2015, and pytest even has nosetests
>>> support which would help us greatly during migration (
>>> https://docs.pytest.org/en/latest/nose.html).
>>>>
>>>> Thoughts?
>>>>
>>>> best,
>>>> kjellman
>>>
>>>
>>> ---------------------------------------------------------------------
>>> To unsubscribe, e-mail: dev-unsubscribe@cassandra.apache.org
>>> For additional commands, e-mail: dev-help@cassandra.apache.org
>>>
>>>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: dev-unsubscribe@cassandra.apache.org
> For additional commands, e-mail: dev-help@cassandra.apache.org
>
Re: [PROPOSAL] Migrate to pytest from nosetests for dtests
Posted by Michael Kjellman <mk...@internalcircle.com>.
i keep seeing nose randomly handing after a test successfully completes execution. i’m very far from a python guru but i spent a few hours with gdb trying to debug the thing and get python stacks and got symbolicated native stacks but it’s random but root causing while nose is sitting on a lock forever alludes me. some tests are more reproducible than others. some i see fail 1 in 10 runs.
the net of it all though is this makes people not trust dtests because it randomly hangs and shows tests with “failures” that actually succeeded.
i’m not a huge fan of just blindly upgrading to fix a problem but in this case I found that there is quite a lot of mistrust and dislike for nosetests in the python community with most projects already moving to pytest. and if it is some complicated set of interactions between threads we use in the tests and how nose works do we really want to even debug it when the project appears to be abandoned?
i think regardless of the root cause for making things more stable it seems like there is little motivation to stick around on nose...
lmk!
best,
kjellman
> On Nov 29, 2017, at 5:33 AM, Philip Thompson <ph...@datastax.com> wrote:
>
> I don't have any objection to this, really. I know I rely on a handful of
> nose plugins, and possibly others do, but those should be easy enough to
> re-write. I am curious though, what's the impetus for this? Is there some
> pytest feature we want that nose lacks? Is there some nosetest bug or
> restriction getting in the way?
>
>> On Tue, Nov 28, 2017 at 8:34 PM, Jon Haddad <jo...@jonhaddad.com> wrote:
>>
>> +1
>>
>> I stopped using nose a long time ago in favor of py.test. It’s a
>> significant improvement.
>>
>>> On Nov 28, 2017, at 10:49 AM, Michael Kjellman <kj...@apple.com>
>> wrote:
>>>
>>> I'd like to propose we move from nosetest to pytest for the dtests. It
>> looks like nosetests is basically abandoned, the python community doesn't
>> like it, it hasn't been updated since 2015, and pytest even has nosetests
>> support which would help us greatly during migration (
>> https://docs.pytest.org/en/latest/nose.html).
>>>
>>> Thoughts?
>>>
>>> best,
>>> kjellman
>>
>>
>> ---------------------------------------------------------------------
>> To unsubscribe, e-mail: dev-unsubscribe@cassandra.apache.org
>> For additional commands, e-mail: dev-help@cassandra.apache.org
>>
>>
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@cassandra.apache.org
For additional commands, e-mail: dev-help@cassandra.apache.org
Re: [PROPOSAL] Migrate to pytest from nosetests for dtests
Posted by Philip Thompson <ph...@datastax.com>.
I don't have any objection to this, really. I know I rely on a handful of
nose plugins, and possibly others do, but those should be easy enough to
re-write. I am curious though, what's the impetus for this? Is there some
pytest feature we want that nose lacks? Is there some nosetest bug or
restriction getting in the way?
On Tue, Nov 28, 2017 at 8:34 PM, Jon Haddad <jo...@jonhaddad.com> wrote:
> +1
>
> I stopped using nose a long time ago in favor of py.test. It’s a
> significant improvement.
>
> > On Nov 28, 2017, at 10:49 AM, Michael Kjellman <kj...@apple.com>
> wrote:
> >
> > I'd like to propose we move from nosetest to pytest for the dtests. It
> looks like nosetests is basically abandoned, the python community doesn't
> like it, it hasn't been updated since 2015, and pytest even has nosetests
> support which would help us greatly during migration (
> https://docs.pytest.org/en/latest/nose.html).
> >
> > Thoughts?
> >
> > best,
> > kjellman
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: dev-unsubscribe@cassandra.apache.org
> For additional commands, e-mail: dev-help@cassandra.apache.org
>
>
Re: [PROPOSAL] Migrate to pytest from nosetests for dtests
Posted by Jon Haddad <jo...@jonhaddad.com>.
+1
I stopped using nose a long time ago in favor of py.test. It’s a significant improvement.
> On Nov 28, 2017, at 10:49 AM, Michael Kjellman <kj...@apple.com> wrote:
>
> I'd like to propose we move from nosetest to pytest for the dtests. It looks like nosetests is basically abandoned, the python community doesn't like it, it hasn't been updated since 2015, and pytest even has nosetests support which would help us greatly during migration (https://docs.pytest.org/en/latest/nose.html).
>
> Thoughts?
>
> best,
> kjellman
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@cassandra.apache.org
For additional commands, e-mail: dev-help@cassandra.apache.org