William Jamieson comments

Results 130 comments of


                                            William Jamieson

JP-2286: Refactor the Bayesian evidence step of residual fringe correction

> How does this PR relate to all the other ones currently open for residual fringe work, such as #6475 , #6761 , and #6739 ? Can this one be...

Behavior when attempting to override an official dialect is not well defined when recursing into subschemas containing the official dialect's $schema identifier

@Julian, sorry for not creating an asdf independent reproducer yesterday. The following is a reproducer of the issue, though I think it can be simplified: ```python import jsonschema.validators as mvalidators...

Behavior when attempting to override an official dialect is not well defined when recursing into subschemas containing the official dialect's $schema identifier

Note that changing https://github.com/python-jsonschema/jsonschema/blob/2f15a98f9243b77ff8f546e8b28cbbdabd7fef23/jsonschema/validators.py#L238 to ```python result = NewValidator(**changes) result.VALIDATORS.update(self.VALIDATORS) return result ``` Will fix the first part (`try`/`except` block) of the reproducer, but causes the second run of `check_schema...

Behavior when attempting to override an official dialect is not well defined when recursing into subschemas containing the official dialect's $schema identifier

A better fix than the one in https://github.com/python-jsonschema/jsonschema/issues/994#issuecomment-1234482128 is to replace https://github.com/python-jsonschema/jsonschema/blob/2f15a98f9243b77ff8f546e8b28cbbdabd7fef23/jsonschema/validators.py#L228 with ```python BaseValidator = validator_for(schema, default=cls) NewValidator = extend( validator=BaseValidator, validators=self.VALIDATORS ) ``` It seems to pass my...

Behavior when attempting to override an official dialect is not well defined when recursing into subschemas containing the official dialect's $schema identifier

> A better fix than the one in [#994 (comment)](https://github.com/python-jsonschema/jsonschema/issues/994#issuecomment-1234482128) is to replace > Better is relative, it appears to have created a recursion error (don't have time to explore...

Behavior when attempting to override an official dialect is not well defined when recursing into subschemas containing the official dialect's $schema identifier

The bit of asdf at issue is [here](https://github.com/asdf-format/asdf/blob/bf954a3921df6b3e5f36c211c18257ac97d4f423/asdf/schema.py#L705-L750). In asdf, we require all schemas to follow the our own "[yaml-schema](https://github.com/asdf-format/asdf-standard/blob/master/resources/schemas/stsci.edu/yaml-schema/draft-01.yaml)" meta-schema which is asdf's extension of the draft 4 meta-schema....

Behavior when attempting to override an official dialect is not well defined when recursing into subschemas containing the official dialect's $schema identifier

Tagging @eslavich for comments about asdf.

Behavior when attempting to override an official dialect is not well defined when recursing into subschemas containing the official dialect's $schema identifier

@eslavich I think you have changed what is being validated. What is breaking is our validation of `schema` against its respective `meta_schemas`. You created some data which you are validating...

Behavior when attempting to override an official dialect is not well defined when recursing into subschemas containing the official dialect's $schema identifier

ASDF managed to resolve this issue with asdf-format/asdf#1203. Those changes maybe useful to resolving the issue.

jsonschema 4.10.0 is breaking asdf

There seams to be at least two errors: 1. This Depreciation warning: https://github.com/asdf-format/asdf/runs/7860409541?check_suite_focus=true#step:8:29720 2. This `_id_of` get error: https://github.com/asdf-format/asdf/runs/7860422378?check_suite_focus=true#step:9:28059