Fix add_field_serializer_for_reverse_relations clearing validators #1302
+80
−5
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
When a model has a reverse relation, Ormar synthesizes a field for it. In order to get Pydantic to see the new field, it calls
DecoratorInfos.build
(inadd_field_serializer_for_reverse_relations
) to pick up the new serializer, and thenmodel_rebuild
. Unfortunately,DecoratorInfos.build
is not idempotent. When it processes a model class, it collects the information in__pydantic_decorators__
, and it strips out the decorators from the class members, to save time in case it has to process the class again (as the base class of a different class). If it is called again, it starts collecting the decorators from scratch, ignoring the existing__pydantic_decorators__
. Since the class members have already had their decorators stripped, they are not collected again. The net result is that any pydantic decorators that had already been processed are lost.This has the result of ignoring
@field_validator
etc. if a model has a reverse relation that is processed after the model.This PR fixes the issue by storing the preexisting
__pydantic_decorators__
and merging the new ones into it. This fix should keep working even if Pydantic changesDecoratorInfos.build
to be idempotent.