Repository navigation
fix(scan): check the numbers in ExpressionAttributeValues as the service does #562
There are no files selected for viewing
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1,170 @@ | ||
| // Copyright 2026 ExtendDB contributors | ||
| // SPDX-License-Identifier: Apache-2.0 | ||
|
|
||
| //! The `Scan` checks that run on the request body alone, before the body is | ||
| //! deserialized and before the table is looked up, in the service's order. | ||
| //! | ||
| //! Measured 2026-10-25 (us-east-1, each request twice): | ||
| //! | ||
| //! 1. the range constraints on `Limit` (at least 1), `Segment` (0 to 999999) | ||
| //! and `TotalSegments` (1 to 1000000), each checked when present, whether or | ||
| //! not the other segment field is. A `Limit` below 1 inside the int32 range | ||
| //! answers `1 validation error detected: Value at 'Limit' failed to satisfy | ||
| //! constraint: Member must have value greater than or equal to 1`; below | ||
| //! int32 the service saturates the value and answers the older form | ||
| //! `Value '-2147483648' at 'limit' ...`. The service aggregates several | ||
| //! range faults into one message; this module reports the first in the | ||
| //! order `Limit`, `Segment`, `TotalSegments`, with the `Segment` and | ||
| //! `TotalSegments` texts the previous release used; | ||
| //! 2. the expression and non-expression parameter mix; | ||
| //! 3. the `ReturnConsumedCapacity` enum; | ||
| //! 4. every number in `ExpressionAttributeValues`, `NS` members and values | ||
| //! nested in `L` and `M` included, whether a `FilterExpression` reads the | ||
| //! value, nothing reads it, or the request carries no expression at all; a | ||
| //! Unicode decimal digit counts as the digit it stands for; | ||
| //! 5. the `Select` enum. Both enums invalid answers the `ReturnConsumedCapacity` | ||
| //! error alone, so the two are not aggregated. | ||
| //! | ||
| //! Each of these is reported ahead of the next one, ahead of the table lookup, | ||
|
Collaborator
There was a problem hiding this comment. DynamoDB reports the undefined placeholder, unused value and filter syntax checks ahead of Select: |
||
| //! and ahead of the checks the handler makes after deserializing the body | ||
| //! (Segment without TotalSegments and the reverse, Segment not below | ||
| //! TotalSegments, the undefined placeholder, unused value, filter syntax, | ||
| //! IndexName, ConsistentRead, Select with ProjectionExpression and | ||
| //! ExclusiveStartKey checks). A field of the wrong JSON type anywhere in the | ||
| //! body is reported ahead of all of them, as a `SerializationException`. | ||
|
|
||
| use std::collections::HashMap; | ||
|
|
||
| use extenddb_core::error::DynamoDbError; | ||
| use extenddb_core::types::{AttributeValue, ScanInput}; | ||
| use serde_json::Value; | ||
|
|
||
| use crate::query_key_values::{ | ||
|
Collaborator
There was a problem hiding this comment. Merged with the current |
||
| non_empty, present, shape_fault_first, validate_expression_attribute_value_numbers, | ||
| }; | ||
|
|
||
| /// The `Scan` checks that run on the raw request body, in the service's order: | ||
| /// the `Limit`, `Segment` and `TotalSegments` ranges, the expression and | ||
| /// non-expression parameter mix, the `ReturnConsumedCapacity` enum, every | ||
| /// number in `ExpressionAttributeValues`, then the `Select` enum. | ||
| /// | ||
| /// The checks read the raw body, so a field that does not deserialize (an | ||
| /// invalid `Select`, for one) does not stop the ones before it. When a check | ||
| /// fails, the body is deserialized as a [`ScanInput`] and a | ||
| /// `SerializationException` from that (a field of the wrong JSON shape | ||
| /// anywhere in the body) is the answer instead, as the service reports a shape | ||
| /// fault ahead of every validation; see [`shape_fault_first`]. A body that | ||
| /// passes every check is not deserialized here. | ||
| /// | ||
| /// # Errors | ||
| /// | ||
| /// Returns `DynamoDbError::ValidationException` with the service's text, or | ||
| /// `DynamoDbError::SerializationException` for a body of the wrong shape. | ||
| pub fn validate_scan_request_body(body: &Value) -> Result<(), DynamoDbError> { | ||
| scan_request_checks(body).map_err(|check| shape_fault_first::<ScanInput>(body, check)) | ||
|
Collaborator
There was a problem hiding this comment. A JSON null where an AttributeValue belongs turns an enum error into SerializationException. Request: |
||
| } | ||
|
|
||
| /// The `Scan` checks in the service's order; see | ||
| /// [`validate_scan_request_body`]. | ||
| fn scan_request_checks(body: &Value) -> Result<(), DynamoDbError> { | ||
| validate_ranges(body)?; | ||
|
|
||
| extenddb_core::validation::validate_no_expression_param_mixing( | ||
|
Collaborator
There was a problem hiding this comment. The service reports an invalid ReturnConsumedCapacity ahead of the parameter mix, as main does. Request: |
||
| &[ | ||
| ("AttributesToGet", non_empty(body, "AttributesToGet")), | ||
| ("ScanFilter", non_empty(body, "ScanFilter")), | ||
| ("ConditionalOperator", present(body, "ConditionalOperator")), | ||
| ], | ||
| &[ | ||
| ( | ||
| "ProjectionExpression", | ||
| present(body, "ProjectionExpression"), | ||
| ), | ||
| ("FilterExpression", present(body, "FilterExpression")), | ||
| ], | ||
| )?; | ||
|
|
||
| crate::validate_enum_fields( | ||
| body, | ||
| &[crate::EnumField { | ||
| json_name: "ReturnConsumedCapacity", | ||
| valid: &["INDEXES", "TOTAL", "NONE"], | ||
| clause: crate::EnumClause::Named("returnConsumedCapacity"), | ||
| }], | ||
| )?; | ||
|
|
||
| if let Some(raw) = body.get("ExpressionAttributeValues") | ||
|
Collaborator
There was a problem hiding this comment. The ExpressionAttributeNames checks run only in typed deserialization, and |
||
| && let Ok(values) = serde_json::from_value::<HashMap<String, AttributeValue>>(raw.clone()) | ||
| { | ||
| validate_expression_attribute_value_numbers(&values)?; | ||
| } | ||
|
|
||
| crate::validate_enum_fields( | ||
| body, | ||
| &[crate::EnumField { | ||
| json_name: "Select", | ||
| valid: &[ | ||
| "SPECIFIC_ATTRIBUTES", | ||
| "COUNT", | ||
| "ALL_ATTRIBUTES", | ||
| "ALL_PROJECTED_ATTRIBUTES", | ||
| ], | ||
| clause: crate::EnumClause::Named("select"), | ||
| }], | ||
| ) | ||
| } | ||
|
|
||
| /// The range constraints on `Limit`, `Segment` and `TotalSegments`, the first | ||
| /// fault in that order. | ||
| /// | ||
| /// `Limit` answers the service's text: inside the int32 range the field is | ||
| /// capitalised and the value is not echoed; below int32 the service saturates | ||
| /// the value and uses the older `Value '<n>' at 'limit'` form (measured | ||
| /// 2026-10-25, `Limit: -2147483649`). `Segment` and `TotalSegments` keep the | ||
| /// texts of the previous release. | ||
| fn validate_ranges(body: &Value) -> Result<(), DynamoDbError> { | ||
| let get = |field: &str| body.get(field).and_then(Value::as_i64); | ||
| if let Some(limit) = get("Limit") | ||
| && limit < 1 | ||
| { | ||
| let text = if limit >= i64::from(i32::MIN) { | ||
|
Collaborator
There was a problem hiding this comment. DynamoDB uses the older form when the map holds a value it cannot read. |
||
| "Value at 'Limit' failed to satisfy constraint: Member must have value greater than or equal to 1".to_owned() | ||
| } else { | ||
| format!( | ||
| "Value '{}' at 'limit' failed to satisfy constraint: Member must have value greater than or equal to 1", | ||
| i32::MIN | ||
| ) | ||
| }; | ||
| return Err(DynamoDbError::ValidationException(format!( | ||
| "1 validation error detected: {text}" | ||
| ))); | ||
| } | ||
| if let Some(segment) = get("Segment") { | ||
|
Collaborator
There was a problem hiding this comment. These checks now answer requests that main answered with the pairing text, so they need the service's texts. DynamoDB, twice each:
This branch answers |
||
| if segment < 0 { | ||
| return Err(DynamoDbError::ValidationException(format!( | ||
| "1 validation error detected: Value '{segment}' at 'segment' failed to satisfy constraint: Member must have value greater than or equal to 0" | ||
| ))); | ||
| } | ||
| if segment > 999_999 { | ||
| return Err(DynamoDbError::ValidationException(format!( | ||
| "1 validation error detected: Value '{segment}' at 'segment' failed to satisfy constraint: Member must have value less than or equal to 999999" | ||
| ))); | ||
| } | ||
| } | ||
| if let Some(total) = get("TotalSegments") { | ||
| if total < 1 { | ||
| return Err(DynamoDbError::ValidationException( | ||
| "The parameter TotalSegments should be greater than or equal to 1".to_owned(), | ||
| )); | ||
| } | ||
| if total > 1_000_000 { | ||
| return Err(DynamoDbError::ValidationException(format!( | ||
| "1 validation error detected: Value '{total}' at 'totalSegments' failed to satisfy constraint: Member must have value less than or equal to 1000000" | ||
| ))); | ||
| } | ||
| } | ||
| Ok(()) | ||
| } | ||
|
|
||
| #[cfg(test)] | ||
| mod tests; | ||
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
No test fails with this check disabled.
TotalSegments: 1accepted has no row either.