support Django 2.2

[kennir]

Modify sortedm2m field will raise a exception after upgrade to Django 2.2

TypeError: _add_items() got an unexpected keyword argument 'through_defaults'

How can I fix it

[pleasedontbelong]

I changed the code of _add_items https://github.com/gregmuellegger/django-sortedm2m/blob/master/sortedm2m/fields.py#L95 to:

def _add_items(self, source_field_name, target_field_name, *objs, through_defaults=None):
        # source_field_name: the PK fieldname in join table for the source object
        # target_field_name: the PK fieldname in join table for the target object
        # *objs - objects to add. Either object instances, or primary keys of object instances.
        through_defaults = through_defaults or {}

        # If there aren't any objects, there is nothing to do.
        from django.db.models import Max, Model
        if objs:
            # Django uses a set here, we need to use a list to keep the
            # correct ordering.
            new_ids = []
            for obj in objs:
                if isinstance(obj, self.model):
                    if not router.allow_relation(obj, self.instance):
                        raise ValueError(
                            'Cannot add "%r": instance is on database "%s", value is on database "%s"' %
                            (obj, self.instance._state.db, obj._state.db)
                        )
                    fk_val = self.through._meta.get_field(
                        target_field_name).get_foreign_related_value(obj)[0]
                    if fk_val is None:
                        raise ValueError(
                            'Cannot add "%r": the value for field "%s" is None' %
                            (obj, target_field_name)
                        )
                    new_ids.append(fk_val)
                elif isinstance(obj, Model):
                    raise TypeError(
                        "'%s' instance expected, got %r" %
                        (self.model._meta.object_name, obj)
                    )
                else:
                    new_ids.append(obj)

            db = router.db_for_write(self.through, instance=self.instance)
            manager = self.through._default_manager.using(db)
            vals = (manager
                    .values_list(target_field_name, flat=True)
                    .filter(**{
                        source_field_name: self._fk_val,
                        '%s__in' % target_field_name: new_ids,
                    }))
            for val in vals:
                if val in new_ids:
                    new_ids.remove(val)
            _new_ids = []
            for pk in new_ids:
                if pk not in _new_ids:
                    _new_ids.append(pk)
            new_ids = _new_ids
            new_ids_set = set(new_ids)

            if self.reverse or source_field_name == self.source_field_name:
                # Don't send the signal when we are inserting the
                # duplicate data row for symmetrical reverse entries.
                signals.m2m_changed.send(sender=rel.through, action='pre_add',
                    instance=self.instance, reverse=self.reverse,
                    model=self.model, pk_set=new_ids_set, using=db)

            # Add the ones that aren't there already
            with atomic(using=db):
                fk_val = self._fk_val
                source_queryset = manager.filter(**{'%s_id' % source_field_name: fk_val})
                sort_field_name = self.through._sort_field_name
                sort_value_max = source_queryset.aggregate(max=Max(sort_field_name))['max'] or 0

                manager.bulk_create([
                    self.through(**through_defaults, **{
                        '%s_id' % source_field_name: fk_val,
                        '%s_id' % target_field_name: pk,
                        sort_field_name: sort_value_max + i + 1,
                    })
                    for i, pk in enumerate(new_ids)
                ])

            if self.reverse or source_field_name == self.source_field_name:
                # Don't send the signal when we are inserting the
                # duplicate data row for symmetrical reverse entries.
                signals.m2m_changed.send(sender=rel.through, action='post_add',
                    instance=self.instance, reverse=self.reverse,
                    model=self.model, pk_set=new_ids_set, using=db)

it seems to work but it's not ready to be a PR

[MisterGlass]

I am also running into this issue. Are there any updates? I see a few PRs attempting to address it, they all seem to fail CI due to an issue with the CI build

[npadgen]

Looking at issue #112 I think @gregmuellegger may have abandoned this project. Could someone fork the project? My project depends heavily on this library!

[MisterGlass]

Maybe we should reach out on twitter or email & ask what they would prefer? There is contact info in the project README

[npadgen]

Maybe we should reach out on twitter or email & ask what they would prefer? There is contact info in the project README

I’ve just pinged him on twitter.

[MisterGlass]

@npadgen Did the maintainer respond?

[vprinz]

No, but my colleague created PR https://github.com/Seleznev-nvkz/django-sortedm2m/pull/1/files It is temporary solution but it works. We think that maintainer stopped develop projectfor installing - pip install git+git://github.com/Seleznev-nvkz/django-sortedm2m.git@refactoring/drop_old_versions04.06.2019, 18:23, "Yair Silbermintz" <notifications@github.com>:@npadgen Did the maintainer respond? —You are receiving this because you are subscribed to this thread.Reply to this email directly, view it on GitHub, or mute the thread. -- 

[ppawlak]

Hello,

Considering the original author of this package had been completely silent for a long time I decided to fork this repo and publish the package under a new name on PyPI.

I could not come up with a good name so I just named it django-sorted-m2m (note the additional -). I hope it does not create too much confusion. If someone has a better idea, please let me know.

The forked repo includes pull request #134 by @joehybird to support Django 2.2. Kudos to him!

My repo is at: https://github.com/Ponytech/django-sorted-m2m
PyPI link: https://pypi.org/project/django-sorted-m2m

Anyone interested into taking part in this fork is more than welcome!

[richardbarran]

FYI I have asked Jazzband about the possibility of moving this project over to their community, even if @gregmuellegger doesn't answer pings.

[vprinz]

Hello, thank you! So, we wait for him answer 20.06.2019, 22:20, "Richard Barran" <notifications@github.com>:FYI I have asked Jazzband about the possibility of moving this project over to their community, even if @gregmuellegger doesn't answer pings.—You are receiving this because you are subscribed to this thread.Reply to this email directly, view it on GitHub, or mute the thread.  -- С уважением,Павликов Валерий 

[richardbarran]

Greg has answered!
See this thread.

[vprinz]

It sounds great! Tanks a lot. We will follow for Jazzband26.06.2019, 15:11, "Richard Barran" <notifications@github.com>:Greg has answered! See this thread. —You are receiving this because you are subscribed to this thread.Reply to this email directly, view it on GitHub, or mute the thread. -- С уважением,Павликов Валерий

[chadsaun]

So now that Greg answered, is anyone able to move in pull request #134 by @joehybird?

[kegan]

Will there be a release soon?