[pytest-dev] xdist: removing boxed with degraded backward compat

[Bruno Oliveira]

But will pytest-xdist still keep the option around? IIUC the conflict will
only happen if the user is using the new pytest-boxed together with an old
pytest-xdist version, correct?

Another idea: how about naming it something else? pytest-forked (heh)? That
would be a completely separate plugin with a separate option (--forked).

Cheers,


On Fri, Jul 28, 2017 at 1:00 PM RonnyPfannschmidt <
opensource at ronnypfannschmidt.de> wrote:

> Hi Bruno,
>
> its the other way around boxed jsut takes the fork using mess out of
> xdist so i can drop that code
>
> -- Ronny
>
>
> Am 28.07.2017 um 17:59 schrieb Bruno Oliveira:
> > Hey Ronny,
> >
> > Does this new "pytest-boxed"(?) plugin depends on pytest-xdist? If it
> > does I think it is just a matter of pinning that dependency as requiring
> > a minimum of pytest-xdist >= 2 so users will get a version conflict
> > error if they try to install incompatible versions (I'm assuming the
> > pytest-xdist release without --boxed will be 2.0).
> >
> > Cheers,
> >
> > On Fri, Jul 28, 2017 at 12:50 PM RonnyPfannschmidt
> > <opensource at ronnypfannschmidt.de
> > <mailto:opensource at ronnypfannschmidt.de>> wrote:
> >
> >     hi everyone,
> >
> >     its finally time to take and remove xdists boxed mode,
> >
> >     in order to prepare the removal i extracted it, but that creates one
> >     clear issue,
> >
> >     py.test messes up and has no way to protect against old xdist and
> boxed
> >     both registering --boxed
> >
> >     as such users in such a situation will get a completely horrible and
> >     entirely useless error message
> >
> >     i tried for a while to get boxed not to register its own option if
> xdist
> >     is there and wants it as well,
> >     but it seems impossible, in my own tests the plugin registered hook
> >     triggers after addoption
> >
> >     so i opted to simply opt out of the own option if the xdist.boxed
> module
> >     is availiable
> >
> >     -- Ronny
> >
> >     _______________________________________________
> >     pytest-dev mailing list
> >     pytest-dev at python.org <mailto: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/20170728/61bb7271/attachment.html>