-
Notifications
You must be signed in to change notification settings - Fork 399
ChangeLog.devel
fge edited this page Feb 15, 2013
·
40 revisions
- Fix incorrect pointer reporting in subschemas for keywords
disallow
,type
(draft v3),allOf
,anyOf
,oneOf
. -
host-name
format attribute in v3 ishostname
in v4: fix - When encountering a JSON Reference, the array/object subschema calculation should be done on the resolved reference, not the reference itself.
- Some minor code cleanups.
- Fix two bugs with
InlineSchemaTree
where it wouldn't find refs which it should have found. - Create, test all configuration classes.
- Create a set of unchecked exceptions for configuration errors.
- Create the first main user class:
JsonValidator
. - Some code cleanup.
- Package reorganization.
- Some Javadoc completed for already frozen parts.
- Allow easy processing exception customization; use it in existing processors.
- More tests and strenghtening of the core API.
- Create a new schema loading related configuration; use it in one core component.
- Complete draft v3/draft v4 validation, with
format
. - Implement a first, crude version of a complete validator.
- New class
ProcessorSelector
: select a processor according to aPredicate
, with an optional default. - Replace tree implementations with thread safe versions.
- Add methods to
ProcessingReport
interface to get/set log/exception thresholds. - Remove all traces of the old API.
One-liner: add HTTPS support by default to the implementation. Requested by issue #37.
- Write JSON Schemas for SMD, JSON-RPC 2.0, JSON Patch.
- Complete validation chain: ref resolving, syntax validation, digesting, keyword building, keyword validation. Test program.
- Implement all draft v4 keywords, except format.
- Fix error in syntax validation where subschema errors were incorrectly computed.
- More tests
- Complete draft v4 syntax validation, tested
- Some changes to the core processing architecture
- Dedicated dictionary classes for storage (of syntax validators for now)
- Brain fart release, or how to use queues properly...
- Add syntax checking for the "definitions" keyword.
- Bump JSON Pointer support to version 9.
- Start complete refactoring: stateful JSON trees and schema trees, processing context, processors, reports, exceptions; start with ref resolving.
- Fix fatal error reporting with
type
(draft v3),oneOf
(draft v4),not
(draft v4): they were swallowed silently.
- Fix bug with duplicate ids when using inline schema addressing (
IllegalArgumentException
). - Update JSON Pointer to support draft version 8.
- Namespace change.
- Remove obsolete code.
- Recursive schema syntax validation.
- No more Message.Builder argument to SyntaxChecker instances (use
AbstractSyntaxChecker#newMsg()
instead). - Make
JsonNodeEquivalence
robust. - Fix
JsonValidatorCache
bug where it would consider all value nodes, and empty arrays, as valid schemas. Add test. - Some code factorization.
- Dependencies updates.
- Drop custom NumberNode/ArrayNode/ObjectNode implementations, use Guava's Equivalence instead; this makes the code much more robust.
- API updates to submit new keywords/format attributes; mark old API as deprecated, convert all tests/examples to use the new API.
- Add
properties
syntax checking for draft v4 :/ - Add official test suite to the tests.
- Update JSON Pointer draft.
- Update
jackson-databind
dependency update.