Python3 support #11
Python3 support #11
Conversation
dnephin
force-pushed
the
python3
branch
2 times, most recently
from
83a4152 to
1373f65
Compare
|
I believe I've fixed the bug in jsonschema, I'm going to submit a PR |
|
Opened python-jsonschema/jsonschema#201, let's see if we get a response. For now we could just mark these failing tests as "broken on py3" so they get skipped |
| resolver = RefResolver( | ||
| 'file://{0}'.format(schema_path), | ||
| schema, | ||
| store={'http://swagger.io/v2/schema.json': schema}) |
There was a problem hiding this comment.
How will the store mapping help? Also, the schema could either be 1.2 or a 2.0.
There was a problem hiding this comment.
Oops, this should be store=schema['id']
I believe it should avoid a network call. The way refs are looked up, they use urljoin(current_scope, json_pointer) where json_pointer is some #/path/to/ ref. If anything uses the absolute path to the schema, it would cause a network call.
I guess this might not happen very often in practice. It might have been triggered by the bug that is being fixed in jsonschema under python 3.
There was a problem hiding this comment.
But the key id is not always present in the schema (it is present in 2.0 schema but not in api_declaration and resource_listing schema of 1.2).
Also, from signature of RefResolver, store takes a dict (id would be a str).
There was a problem hiding this comment.
sorry, I meant store={schema['id']: schema}
This should already be covered by tests. I'll double check that it works with 1.2.
There was a problem hiding this comment.
Thinking about this again api_declaration and resource_listing aren't valid jsonschema, which is why they don't have id. They're being used as the "json object" being validated against the swagger spec schema.
I can change this to schema.get('id') just to be safe.
There was a problem hiding this comment.
I'll just revert this part of it. It's not really important. We can revisit it if we ever hit this case (of things using absolute urls).
|
lgtm. Ship it may be after the python-jsonschema/jsonschema#203 merge? |
swagger_spec_validator. This commit only covers validation of swagger 2.0 spec and in no way handles serving of 2.0 compatible endpoint. It currently breaks py33 and py34 tests. This should be rectified with the merge of Yelp/swagger_spec_validator#11.
swagger_spec_validator. This commit only covers validation of swagger 2.0 spec and in no way handles serving of 2.0 compatible endpoint. It currently breaks py33 and py34 tests. This should be rectified with the merge of Yelp/swagger_spec_validator#11.
…python3 Conflicts: swagger_spec_validator/__about__.py
|
A new release of |
|
Ya, we can try that. Right now we're just installing any unpinned version. I don't think We'll have to pin the same version in the pyramid_swagger testsuite if we want to use it there as well. |
|
Ok, tests pass with pinned version of jsonschema. Look good? |
|
yeah lgtm. |
Related to #10
There are definition still some test flakes here, but they magically stopped happening locally so I decided to open this PR for now.
The error was:
which is strange because
simpleTypesdoesn't appear anywhere in this repo, but it does exist injsonschema/schemas/draft4.json.