641 lines
26 KiB
Rust
641 lines
26 KiB
Rust
//! Schema compilation.
|
|
//! The main idea is to compile the input JSON Schema to a validators tree that will contain
|
|
//! everything needed to perform such validation in runtime.
|
|
pub(crate) mod context;
|
|
pub(crate) mod options;
|
|
|
|
use crate::{
|
|
error::ErrorIterator,
|
|
keywords::{self, custom_keyword::compile_custom_keyword_validator},
|
|
output::Output,
|
|
paths::{InstancePath, JSONPointer},
|
|
primitive_type::{PrimitiveType, PrimitiveTypesBitMap},
|
|
schema_node::SchemaNode,
|
|
validator::Validate,
|
|
Draft, ValidationError,
|
|
};
|
|
use ahash::AHashMap;
|
|
use context::CompilationContext;
|
|
use once_cell::sync::Lazy;
|
|
use options::CompilationOptions;
|
|
use serde_json::Value;
|
|
use std::sync::Arc;
|
|
use url::Url;
|
|
|
|
pub(crate) const DEFAULT_ROOT_URL: &str = "json-schema:///";
|
|
|
|
/// The structure that holds a JSON Schema compiled into a validation tree
|
|
#[derive(Debug)]
|
|
pub struct JSONSchema {
|
|
pub(crate) node: SchemaNode,
|
|
config: Arc<CompilationOptions>,
|
|
}
|
|
|
|
pub(crate) static DEFAULT_SCOPE: Lazy<Url> =
|
|
Lazy::new(|| url::Url::parse(DEFAULT_ROOT_URL).expect("Is a valid URL"));
|
|
|
|
impl JSONSchema {
|
|
/// Return a default `CompilationOptions` that can configure
|
|
/// `JSONSchema` compilation flow.
|
|
///
|
|
/// Using options you will be able to configure the draft version
|
|
/// to use during `JSONSchema` compilation
|
|
///
|
|
/// Example of usage:
|
|
/// ```rust
|
|
/// # use crate::jsonschema::{Draft, JSONSchema};
|
|
/// # let schema = serde_json::json!({});
|
|
/// let maybe_jsonschema: Result<JSONSchema, _> = JSONSchema::options()
|
|
/// .with_draft(Draft::Draft7)
|
|
/// .compile(&schema);
|
|
/// ```
|
|
#[must_use]
|
|
pub fn options() -> CompilationOptions {
|
|
CompilationOptions::default()
|
|
}
|
|
|
|
/// Compile the input schema into a validation tree.
|
|
///
|
|
/// The method is equivalent to `JSONSchema::options().compile(schema)`
|
|
pub fn compile(schema: &Value) -> Result<JSONSchema, ValidationError> {
|
|
Self::options().compile(schema)
|
|
}
|
|
|
|
/// Run validation against `instance` and return an iterator over `ValidationError` in the error case.
|
|
#[inline]
|
|
pub fn validate<'instance>(
|
|
&'instance self,
|
|
instance: &'instance Value,
|
|
) -> Result<(), ErrorIterator<'instance>> {
|
|
let instance_path = InstancePath::new();
|
|
let mut errors = self.node.validate(instance, &instance_path).peekable();
|
|
if errors.peek().is_none() {
|
|
Ok(())
|
|
} else {
|
|
Err(Box::new(errors))
|
|
}
|
|
}
|
|
|
|
/// Run validation against `instance` but return a boolean result instead of an iterator.
|
|
/// It is useful for cases, where it is important to only know the fact if the data is valid or not.
|
|
/// This approach is much faster, than `validate`.
|
|
#[must_use]
|
|
#[inline]
|
|
pub fn is_valid(&self, instance: &Value) -> bool {
|
|
self.node.is_valid(instance)
|
|
}
|
|
|
|
/// Apply the schema and return an `Output`. No actual work is done at this point, the
|
|
/// evaluation of the schema is deferred until a method is called on the `Output`. This is
|
|
/// because different output formats will have different performance characteristics.
|
|
///
|
|
/// # Examples
|
|
///
|
|
/// "basic" output format
|
|
///
|
|
/// ```rust
|
|
/// # use crate::jsonschema::{Draft, JSONSchema, output::{Output, BasicOutput}};
|
|
/// let schema_json = serde_json::json!({
|
|
/// "title": "string value",
|
|
/// "type": "string"
|
|
/// });
|
|
/// let instance = serde_json::json!{"some string"};
|
|
/// let schema = JSONSchema::options().compile(&schema_json).unwrap();
|
|
/// let output: BasicOutput = schema.apply(&instance).basic();
|
|
/// let output_json = serde_json::to_value(output).unwrap();
|
|
/// assert_eq!(output_json, serde_json::json!({
|
|
/// "valid": true,
|
|
/// "annotations": [
|
|
/// {
|
|
/// "keywordLocation": "",
|
|
/// "instanceLocation": "",
|
|
/// "annotations": {
|
|
/// "title": "string value"
|
|
/// }
|
|
/// }
|
|
/// ]
|
|
/// }));
|
|
/// ```
|
|
#[must_use]
|
|
pub const fn apply<'a, 'b>(&'a self, instance: &'b Value) -> Output<'a, 'b> {
|
|
Output::new(self, &self.node, instance)
|
|
}
|
|
|
|
/// The [`Draft`] which this schema was compiled against
|
|
#[must_use]
|
|
pub fn draft(&self) -> Draft {
|
|
self.config.draft()
|
|
}
|
|
|
|
/// The [`CompilationOptions`] that were used to compile this schema
|
|
#[must_use]
|
|
pub fn config(&self) -> Arc<CompilationOptions> {
|
|
Arc::clone(&self.config)
|
|
}
|
|
}
|
|
|
|
/// Compile JSON schema into a tree of validators.
|
|
#[inline]
|
|
pub(crate) fn compile_validators<'a>(
|
|
schema: &'a Value,
|
|
context: &CompilationContext,
|
|
) -> Result<SchemaNode, ValidationError<'a>> {
|
|
let context = context.push(schema)?;
|
|
let relative_path = context.clone().into_pointer();
|
|
match schema {
|
|
Value::Bool(value) => match value {
|
|
true => Ok(SchemaNode::new_from_boolean(&context, None)),
|
|
false => Ok(SchemaNode::new_from_boolean(
|
|
&context,
|
|
Some(
|
|
keywords::boolean::FalseValidator::compile(relative_path)
|
|
.expect("Should always compile"),
|
|
),
|
|
)),
|
|
},
|
|
Value::Object(object) => {
|
|
// In Draft 2019-09 and later, `$ref` can be evaluated alongside other attribute aka
|
|
// adjacent validation. We check here to see if adjacent validation is supported, and if
|
|
// so, we use the normal keyword validator collection logic.
|
|
//
|
|
// Otherwise, we isolate `$ref` and generate a schema reference validator directly.
|
|
let maybe_reference = object
|
|
.get("$ref")
|
|
.filter(|_| !keywords::ref_::supports_adjacent_validation(context.config.draft()));
|
|
if let Some(reference) = maybe_reference {
|
|
let unmatched_keywords = object
|
|
.iter()
|
|
.filter_map(|(k, v)| {
|
|
if k.as_str() == "$ref" {
|
|
None
|
|
} else {
|
|
Some((k.clone(), v.clone()))
|
|
}
|
|
})
|
|
.collect();
|
|
|
|
let validator = keywords::ref_::compile(object, reference, &context)
|
|
.expect("should always return Some")?;
|
|
|
|
let validators = vec![("$ref".to_string(), validator)];
|
|
Ok(SchemaNode::new_from_keywords(
|
|
&context,
|
|
validators,
|
|
Some(unmatched_keywords),
|
|
))
|
|
} else {
|
|
let mut validators = Vec::with_capacity(object.len());
|
|
let mut unmatched_keywords = AHashMap::new();
|
|
let mut is_if = false;
|
|
let mut is_props = false;
|
|
for (keyword, subschema) in object {
|
|
if keyword == "if" {
|
|
is_if = true;
|
|
}
|
|
if keyword == "properties"
|
|
|| keyword == "additionalProperties"
|
|
|| keyword == "patternProperties"
|
|
{
|
|
is_props = true;
|
|
}
|
|
// first check if this keyword was added as a custom keyword
|
|
// it may override existing keyword behavior
|
|
if let Some(f) = context.config.get_custom_keyword_constructor(keyword) {
|
|
let validator = compile_custom_keyword_validator(
|
|
&context,
|
|
keyword.clone(),
|
|
f(),
|
|
subschema.clone(),
|
|
schema.clone(),
|
|
)?;
|
|
validators.push((keyword.clone(), validator));
|
|
} else if let Some(validator) = context
|
|
.config
|
|
.draft()
|
|
.get_validator(keyword)
|
|
.and_then(|f| f(object, subschema, &context))
|
|
{
|
|
validators.push((keyword.clone(), validator?));
|
|
} else {
|
|
unmatched_keywords.insert(keyword.to_string(), subschema.clone());
|
|
}
|
|
}
|
|
if is_if {
|
|
unmatched_keywords.remove("then");
|
|
unmatched_keywords.remove("else");
|
|
}
|
|
if is_props {
|
|
unmatched_keywords.remove("additionalProperties");
|
|
unmatched_keywords.remove("patternProperties");
|
|
unmatched_keywords.remove("properties");
|
|
}
|
|
let unmatched_keywords = if unmatched_keywords.is_empty() {
|
|
None
|
|
} else {
|
|
Some(unmatched_keywords)
|
|
};
|
|
Ok(SchemaNode::new_from_keywords(
|
|
&context,
|
|
validators,
|
|
unmatched_keywords,
|
|
))
|
|
}
|
|
}
|
|
_ => Err(ValidationError::multiple_type_error(
|
|
JSONPointer::default(),
|
|
relative_path,
|
|
schema,
|
|
PrimitiveTypesBitMap::new()
|
|
.add_type(PrimitiveType::Boolean)
|
|
.add_type(PrimitiveType::Object),
|
|
)),
|
|
}
|
|
}
|
|
|
|
#[cfg(test)]
|
|
mod tests {
|
|
use super::JSONSchema;
|
|
use crate::{
|
|
compilation::options::CustomKeywordValidator, error::ValidationError, paths::JSONPointer,
|
|
ErrorIterator,
|
|
};
|
|
use num_cmp::NumCmp;
|
|
use regex::Regex;
|
|
use serde_json::{from_str, json, Value};
|
|
use std::{
|
|
borrow::Cow,
|
|
fs::File,
|
|
io::Read,
|
|
path::Path,
|
|
sync::{Arc, Mutex},
|
|
};
|
|
|
|
fn load(path: &str, idx: usize) -> Value {
|
|
let path = Path::new(path);
|
|
let mut file = File::open(path).unwrap();
|
|
let mut content = String::new();
|
|
file.read_to_string(&mut content).ok().unwrap();
|
|
let data: Value = from_str(&content).unwrap();
|
|
let case = &data.as_array().unwrap()[idx];
|
|
case.get("schema").unwrap().clone()
|
|
}
|
|
|
|
#[test]
|
|
fn only_keyword() {
|
|
// When only one keyword is specified
|
|
let schema = json!({"type": "string"});
|
|
let compiled = JSONSchema::compile(&schema).unwrap();
|
|
let value1 = json!("AB");
|
|
let value2 = json!(1);
|
|
// And only this validator
|
|
assert_eq!(compiled.node.validators().len(), 1);
|
|
assert!(compiled.validate(&value1).is_ok());
|
|
assert!(compiled.validate(&value2).is_err());
|
|
}
|
|
|
|
#[test]
|
|
fn validate_ref() {
|
|
let schema = load("tests/suite/tests/draft7/ref.json", 1);
|
|
let value = json!({"bar": 3});
|
|
let compiled = JSONSchema::compile(&schema).unwrap();
|
|
assert!(compiled.validate(&value).is_ok());
|
|
let value = json!({"bar": true});
|
|
assert!(compiled.validate(&value).is_err());
|
|
}
|
|
|
|
#[test]
|
|
fn wrong_schema_type() {
|
|
let schema = json!([1]);
|
|
let compiled = JSONSchema::compile(&schema);
|
|
assert!(compiled.is_err());
|
|
}
|
|
|
|
#[test]
|
|
fn multiple_errors() {
|
|
let schema = json!({"minProperties": 2, "propertyNames": {"minLength": 3}});
|
|
let value = json!({"a": 3});
|
|
let compiled = JSONSchema::compile(&schema).unwrap();
|
|
let result = compiled.validate(&value);
|
|
let errors: Vec<ValidationError> = result.unwrap_err().collect();
|
|
assert_eq!(errors.len(), 2);
|
|
assert_eq!(
|
|
errors[0].to_string(),
|
|
r#"{"a":3} has less than 2 properties"#
|
|
);
|
|
assert_eq!(errors[1].to_string(), r#""a" is shorter than 3 characters"#);
|
|
}
|
|
|
|
#[test]
|
|
fn custom_keyword_definition() {
|
|
// Define a custom validator that verifies the object's keys consist of
|
|
// only ASCII representable characters.
|
|
struct CustomObjectValidator;
|
|
impl CustomKeywordValidator for CustomObjectValidator {
|
|
fn validate<'instance>(
|
|
&self,
|
|
instance: &'instance Value,
|
|
instance_path: JSONPointer,
|
|
subschema: Arc<Value>,
|
|
subschema_path: JSONPointer,
|
|
_schema: Arc<Value>,
|
|
) -> ErrorIterator<'instance> {
|
|
if subschema.as_str().map_or(true, |str| str != "ascii-keys") {
|
|
let error = ValidationError {
|
|
instance: Cow::Borrowed(instance),
|
|
kind: crate::error::ValidationErrorKind::Schema,
|
|
instance_path,
|
|
schema_path: subschema_path,
|
|
};
|
|
return Box::new(Some(error).into_iter()); // Invalid schema
|
|
}
|
|
let mut errors = vec![];
|
|
for (key, _value) in instance.as_object().unwrap() {
|
|
if !key.is_ascii() {
|
|
let error = ValidationError {
|
|
instance: Cow::Borrowed(instance),
|
|
kind: crate::error::ValidationErrorKind::Format { format: "ASCII" },
|
|
instance_path: instance_path.clone(),
|
|
schema_path: subschema_path.clone(),
|
|
};
|
|
errors.push(error);
|
|
}
|
|
}
|
|
Box::new(errors.into_iter())
|
|
}
|
|
|
|
fn is_valid(&self, instance: &Value, subschema: &Value, _schema: &Value) -> bool {
|
|
if subschema.as_str().map_or(true, |str| str != "ascii-keys") {
|
|
return false; // Invalid schema
|
|
}
|
|
for (key, _value) in instance.as_object().unwrap() {
|
|
if !key.is_ascii() {
|
|
return false;
|
|
}
|
|
}
|
|
true
|
|
}
|
|
}
|
|
|
|
// Define a JSON schema that enforces the top level object has ASCII keys and has at least 1 property
|
|
let schema =
|
|
json!({ "custom-object-type": "ascii-keys", "type": "object", "minProperties": 1 });
|
|
let json_schema = JSONSchema::options()
|
|
.with_custom_keyword("custom-object-type", || Box::new(CustomObjectValidator))
|
|
.compile(&schema)
|
|
.unwrap();
|
|
|
|
// Verify schema validation detects object with too few properties
|
|
let instance_err_not_object = json!({});
|
|
assert!(json_schema.validate(&instance_err_not_object).is_err());
|
|
assert!(!json_schema.is_valid(&instance_err_not_object));
|
|
|
|
// Verify validator succeeds on a valid custom-object-type
|
|
let instance_ok = json!({ "a" : 1 });
|
|
assert!(json_schema.validate(&instance_ok).is_ok());
|
|
assert!(json_schema.is_valid(&instance_ok));
|
|
|
|
// Verify validator detects invalid custom-object-type
|
|
let instance_err_non_ascii_keys = json!({ "å" : 1 });
|
|
assert!(json_schema.validate(&instance_err_non_ascii_keys).is_err());
|
|
assert!(!json_schema.is_valid(&instance_err_non_ascii_keys));
|
|
}
|
|
|
|
#[test]
|
|
fn custom_format_and_override_keyword() {
|
|
// prepare a custom format checker
|
|
// in this case, the format is "currency"
|
|
// checks that a string has some number of digits followed by a dot followed by
|
|
// exactly 2 digits.
|
|
const CURRENCY_RE_STR: &str = "^(0|([1-9]+[0-9]*))(\\.[0-9]{2})$";
|
|
fn currency_format_checker(s: &str) -> bool {
|
|
Regex::new(CURRENCY_RE_STR).unwrap().is_match(s)
|
|
}
|
|
// Define a custom keyword validator that overrides "minimum"
|
|
// so that "minimum" may apply to "currency"-formatted strings as well
|
|
struct CustomMinimumValidator;
|
|
impl CustomKeywordValidator for CustomMinimumValidator {
|
|
fn validate<'instance>(
|
|
&self,
|
|
instance: &'instance Value,
|
|
instance_path: JSONPointer,
|
|
subschema: Arc<Value>,
|
|
subschema_path: JSONPointer,
|
|
schema: Arc<Value>,
|
|
) -> ErrorIterator<'instance> {
|
|
let subschema: &Value = &subschema;
|
|
let limit = match subschema {
|
|
Value::Number(limit) => limit,
|
|
_ => {
|
|
let error = ValidationError {
|
|
instance: Cow::Borrowed(instance),
|
|
kind: crate::error::ValidationErrorKind::Schema,
|
|
instance_path,
|
|
schema_path: subschema_path,
|
|
};
|
|
return Box::new(Some(error).into_iter()); // Invalid schema
|
|
}
|
|
};
|
|
let mut errors = vec![];
|
|
let valid = match instance {
|
|
// numeric comparison should happen just like original behavior
|
|
Value::Number(instance) => {
|
|
if let Some(item) = instance.as_u64() {
|
|
!NumCmp::num_lt(item, limit.as_f64().unwrap())
|
|
} else if let Some(item) = limit.as_i64() {
|
|
!NumCmp::num_lt(item, limit.as_f64().unwrap())
|
|
} else {
|
|
let item = instance.as_f64().expect("Always valid");
|
|
!NumCmp::num_lt(item, limit.as_f64().unwrap())
|
|
}
|
|
}
|
|
// string comparison should cast currency-formatted
|
|
Value::String(instance) => {
|
|
let mut valid = true;
|
|
if let Some(schema) = schema.as_object() {
|
|
if let Some(format) = schema.get("format") {
|
|
if format == "currency" && currency_format_checker(instance) {
|
|
// all preconditions for minimum applying are met
|
|
let as_f64 = instance
|
|
.parse::<f64>()
|
|
.expect("format validated by regex checker");
|
|
println!("1 {:#?} {:#?}", as_f64, limit.as_f64().unwrap());
|
|
valid = !NumCmp::num_lt(as_f64, limit.as_f64().unwrap());
|
|
println!("valid {:#?}", valid);
|
|
}
|
|
}
|
|
}
|
|
valid
|
|
}
|
|
// in all other cases, the "minimum" keyword should not apply
|
|
_ => true,
|
|
};
|
|
if !valid {
|
|
let error = ValidationError {
|
|
instance: Cow::Borrowed(instance),
|
|
kind: crate::error::ValidationErrorKind::Minimum {
|
|
limit: subschema.clone(),
|
|
},
|
|
instance_path: instance_path.clone(),
|
|
schema_path: subschema_path.clone(),
|
|
};
|
|
errors.push(error);
|
|
}
|
|
Box::new(errors.into_iter())
|
|
}
|
|
|
|
fn is_valid(&self, instance: &Value, subschema: &Value, schema: &Value) -> bool {
|
|
let limit = match subschema {
|
|
Value::Number(limit) => limit,
|
|
_ => return false,
|
|
};
|
|
let valid = match instance {
|
|
// numeric comparison should happen just like original behavior
|
|
Value::Number(instance) => {
|
|
if let Some(item) = instance.as_u64() {
|
|
!NumCmp::num_lt(item, limit.as_f64().unwrap())
|
|
} else if let Some(item) = limit.as_i64() {
|
|
!NumCmp::num_lt(item, limit.as_f64().unwrap())
|
|
} else {
|
|
let item = instance.as_f64().expect("Always valid");
|
|
!NumCmp::num_lt(item, limit.as_f64().unwrap())
|
|
}
|
|
}
|
|
// string comparison should cast currency-formatted
|
|
Value::String(instance) => {
|
|
let mut valid = true;
|
|
if let Some(schema) = schema.as_object() {
|
|
if let Some(format) = schema.get("format") {
|
|
if format == "currency" && currency_format_checker(instance) {
|
|
// all preconditions for minimum applying are met
|
|
let as_f64 = instance
|
|
.parse::<f64>()
|
|
.expect("format validated by regex checker");
|
|
println!("1 {:#?} {:#?}", as_f64, limit.as_f64().unwrap());
|
|
valid = !NumCmp::num_lt(as_f64, limit.as_f64().unwrap());
|
|
println!("valid {:#?}", valid);
|
|
}
|
|
}
|
|
}
|
|
valid
|
|
}
|
|
// in all other cases, the "minimum" keyword should not apply
|
|
_ => true,
|
|
};
|
|
valid
|
|
}
|
|
}
|
|
|
|
// define compilation options that include the custom format and the overridden keyword
|
|
let mut options = JSONSchema::options();
|
|
let options = options
|
|
.with_format("currency", currency_format_checker)
|
|
.with_custom_keyword("minimum", || Box::new(CustomMinimumValidator));
|
|
|
|
// Define a schema that includes both the custom format and the overridden keyword
|
|
let schema = json!({ "minimum": 2, "type": "string", "format": "currency" });
|
|
let compiled = options.compile(&schema).unwrap();
|
|
|
|
// Control: verify schema validation rejects non-string types
|
|
let instance_err_not_string = json!(15);
|
|
assert!(compiled.validate(&instance_err_not_string).is_err());
|
|
assert!(!compiled.is_valid(&instance_err_not_string));
|
|
|
|
// Control: verify validator rejects ill-formatted strings
|
|
let instance_ok = json!("not a currency");
|
|
assert!(compiled.validate(&instance_ok).is_err());
|
|
assert!(!compiled.is_valid(&instance_ok));
|
|
|
|
// Verify validator allows properly formatted strings that conform to custom keyword
|
|
let instance_err_non_ascii_keys = json!("3.00");
|
|
assert!(compiled.validate(&instance_err_non_ascii_keys).is_ok());
|
|
assert!(compiled.is_valid(&instance_err_non_ascii_keys));
|
|
|
|
// Verify validator rejects properly formatted strings that do not conform to custom keyword
|
|
let instance_err_non_ascii_keys = json!("1.99");
|
|
assert!(compiled.validate(&instance_err_non_ascii_keys).is_err());
|
|
assert!(!compiled.is_valid(&instance_err_non_ascii_keys));
|
|
|
|
// Define another schema that applies "minimum" to an integer to ensure original behavior
|
|
let schema = json!({ "minimum": 2, "type": "integer" });
|
|
let compiled = options.compile(&schema).unwrap();
|
|
|
|
// Verify schema allows integers greater than 2
|
|
let instance_err_not_string = json!(3);
|
|
assert!(compiled.validate(&instance_err_not_string).is_ok());
|
|
assert!(compiled.is_valid(&instance_err_not_string));
|
|
|
|
// Verify schema rejects integers less than 2
|
|
let instance_ok = json!(1);
|
|
assert!(compiled.validate(&instance_ok).is_err());
|
|
assert!(!compiled.is_valid(&instance_ok));
|
|
}
|
|
|
|
#[test]
|
|
fn custom_keyword_with_inner_state() {
|
|
// Define a custom keyword validator that wraps "minimum"
|
|
// but maintains a counter of how many times the validator was applied.
|
|
struct CountingValidator {
|
|
count: Mutex<i64>,
|
|
}
|
|
|
|
impl CountingValidator {
|
|
fn increment(&self, amount: i64) {
|
|
let mut count = self.count.lock().expect("Lock is poisoned");
|
|
*count += amount;
|
|
}
|
|
}
|
|
|
|
impl CustomKeywordValidator for CountingValidator {
|
|
fn validate<'instance>(
|
|
&self,
|
|
_: &'instance Value,
|
|
_: JSONPointer,
|
|
subschema: Arc<Value>,
|
|
_: JSONPointer,
|
|
_: Arc<Value>,
|
|
) -> ErrorIterator<'instance> {
|
|
let amount = match &*subschema {
|
|
Value::Number(x) => x.as_i64().expect("countme value must be integer"),
|
|
_ => panic!("Validator requires numeric values"),
|
|
};
|
|
self.increment(amount);
|
|
Box::new(None.into_iter())
|
|
}
|
|
|
|
fn is_valid(&self, _: &Value, subschema: &Value, _: &Value) -> bool {
|
|
let amount = match subschema {
|
|
Value::Number(x) => x.as_i64().expect("countme value must be integer"),
|
|
_ => return false,
|
|
};
|
|
self.increment(amount);
|
|
true
|
|
}
|
|
}
|
|
|
|
// define compilation options that include the custom format and the overridden keyword
|
|
let count = Mutex::new(0);
|
|
let mut options = JSONSchema::options();
|
|
let options = options.with_custom_keyword("countme", || {
|
|
Box::new(CountingValidator {
|
|
count: Mutex::new(0),
|
|
})
|
|
});
|
|
|
|
// Define a schema that includes the custom keyword and therefore should increase the count
|
|
let schema = json!({ "countme": 3, "type": "string" });
|
|
let compiled = options.compile(&schema).unwrap();
|
|
|
|
// TODO: Communicate the increment changes via `validate` output, e.g. fail after N
|
|
// increments, etc.
|
|
// Because the schema has "countme" in it, whenever we run validation we should expect the validator's count to increase
|
|
let instance_ok = json!("i am a string");
|
|
assert_eq!(*count.lock().expect("Lock is poinsoned"), 0);
|
|
assert!(compiled.validate(&instance_ok).is_err());
|
|
assert_eq!(*count.lock().expect("Lock is poinsoned"), 3);
|
|
assert!(!compiled.is_valid(&instance_ok));
|
|
assert_eq!(*count.lock().expect("Lock is poinsoned"), 6);
|
|
|
|
// TODO compile a schema that doesn't have "countme" and ensure count does not increase
|
|
}
|
|
}
|