Fix readonly related fields generating incorrect schema

[diesieben07]

This fixes the issue outlined in #274. I've added a test for this case as well.

[codecov]

Codecov Report

Merging #275 (23ca3a8) into master (f23de89) will increase coverage by 0.01%.
The diff coverage is 100.00%.

@@            Coverage Diff             @@
##           master     #275      +/-   ##
==========================================
+ Coverage   98.37%   98.39%   +0.01%     
==========================================
  Files          53       53              
  Lines        4438     4476      +38     
==========================================
+ Hits         4366     4404      +38     
  Misses         72       72              

Impacted Files	Coverage Δ
drf_spectacular/openapi.py	96.29% <100.00%> (-0.01%)	⬇️
tests/test_regressions.py	100.00% <100.00%> (ø)
drf_spectacular/settings.py	100.00% <0.00%> (ø)
tests/test_extend_schema.py	100.00% <0.00%> (ø)
drf_spectacular/utils.py	98.76% <0.00%> (+<0.01%)	⬆️
drf_spectacular/plumbing.py	96.50% <0.00%> (+0.01%)	⬆️

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update f23de89...23ca3a8. Read the comment docs.

[tfranzel]

@diesieben07 awesome catch! funny how such a bug can go unnoticed for so long.

however the bugfix does break in some cases. this test shows it with indirect_referenced_model_ro. field.field_name does not always apply.

for regressions we prefer consise minimalistic tests and put it in test_regressions.py.
also, i prefer to not add full yamls dumps for single bugs. i prepared this for your convenience. please use it instead.

def test_incorrect_foreignkey_type_on_readonly_field(no_warnings):
    class ReferencingModel(models.Model):
        id = models.UUIDField(primary_key=True)
        referenced_model = models.ForeignKey(SimpleModel, on_delete=models.CASCADE)
        referenced_model_ro = models.ForeignKey(SimpleModel, on_delete=models.CASCADE)
        referenced_model_m2m = models.ManyToManyField(SimpleModel)
        referenced_model_m2m_ro = models.ManyToManyField(SimpleModel)

    class ReferencingModelSerializer(serializers.ModelSerializer):
        indirect_referenced_model_ro = serializers.PrimaryKeyRelatedField(
            source='referenced_model',
            read_only=True,
        )

        class Meta:
            fields = '__all__'
            read_only_fields = ['id', 'referenced_model_ro', 'referenced_model_m2m_ro']
            model = ReferencingModel

    class ReferencingModelViewset(viewsets.ModelViewSet):
        serializer_class = ReferencingModelSerializer
        queryset = ReferencingModel.objects.all()

    schema = generate_schema('/x/', ReferencingModelViewset)
    properties = schema['components']['schemas']['ReferencingModel']['properties']

    assert properties['referenced_model']['type'] == 'integer'
    assert properties['referenced_model_ro']['type'] == 'integer'
    assert properties['referenced_model_m2m']['items']['type'] == 'integer'
    assert properties['referenced_model_m2m_ro']['items']['type'] == 'integer'
    assert properties['indirect_referenced_model_ro']['type'] == 'integer'

[diesieben07]

Thanks for the note about field_name, I had totally forgot about source. I've switched the code to your smaller regression test as well.

[tfranzel]

excellent work! thank you so much. still puzzled how this went unnoticed for so long.