Rename validation and serialization schema types to be unique#1629
Rename validation and serialization schema types to be unique#1629
Conversation
`validator-` or `serializer-` is prepended, and the existing ones are deprecated.
Viicos
force-pushed
the
unique-schema-types
branch
from
e1c4af5 to
ec4e4a1
Compare
davidhewitt
left a comment
davidhewitt
left a comment
There was a problem hiding this comment.
Looking at this, I wonder if we should not change validators, and just change the serialization ones to ser-function-plain etc?
That would still be good enough to make the functional split, without potentially causing a lot of churn in validation error messages.
| } | ||
| } | ||
|
|
||
| fn get_new_type(py: Python, type_: &str) -> PyResult<String> { |
There was a problem hiding this comment.
To avoid allocating:
| fn get_new_type(py: Python, type_: &str) -> PyResult<String> { | |
| fn get_new_type<'a>(py: Python, type_: &'a str) -> PyResult<&'a str> { |
(and then drop the .to_owned() calls below)
| }; | ||
| } | ||
|
|
||
| fn get_new_type(py: Python, type_: &str) -> PyResult<String> { |
There was a problem hiding this comment.
Same here:
| fn get_new_type(py: Python, type_: &str) -> PyResult<String> { | |
| fn get_new_type<'a>(py: Python, type_: &'a str) -> PyResult<&'a str> { |
|
|
||
| assert str(exc_info.value) == ( | ||
| '1 validation error for function-plain[my_function()]\n' | ||
| '1 validation error for validator-function-plain[my_function()]\n' |
There was a problem hiding this comment.
I guess we should just double-check we're happy with this new format, as it's potentially breaking to downstream test suites and if we change this we should aim to do so once and not again?
(e.g. would it be better if this just read 1 validation error for my_function()?)
| if schema['type'] not in { | ||
| 'validator-function-before', | ||
| 'validator-function-after', | ||
| 'validator-function-wrap', | ||
| 'validator-function-plain', | ||
| }: |
There was a problem hiding this comment.
This exclusion is because these types are no longer supported in serialization, because of the split?
Do we need a similar test for serialization schemas?
2 participants
validator-orserializer-is prepended, and the existing ones are deprecated.Change Summary
Partly fixes pydantic/pydantic#11087.
Related issue number
Checklist
pydantic-core(except for expected changes)