[email protected]

0.15.2 - 2025-11-10

Fixes

• Return arbitrary::Error::IncorrectFormat for unsupported floats- tninesling, pull/1005
  When generating floats for GraphQL documents, we were naively unwrapping the
  conversion from f64 to serde_json::Number. This would panic when
  arbitrary returned f64::INFINITY of f64::NAN because the Number
  conversion only works when its input is finite. In this case, the underlying
  bytes used to generate the value are considered to be in an invalid format.
  So, we return arbitrary::Error::IncorrectFormat to tell fuzzers to use a
  different seed in the future.

Maintenance

• Apply new clippy rules from Rust 1.90 - goto-bus-stop, pull/1001

• bump minimum arbitrary version - goto-bus-stop, pull/1007

[email protected]

Features

• Allow coercing Int variables to Float - tninesling, pull/1011

  The GraphQL spec allows coercing Int values to Float in input positions (see
  input coercion). There are a couple things to note about this.

  • Strings are not allowed to be coerced in this position, even if they are
    numeric.
  • Ints can only be converted to Float when they are "representable by finite
    IEEE 754" floating point numbers.

  Since an IEEE 754 floating point double (f64) has 53 bits of precision, it can
  safely represent up to the value 2^53 - 1 as a finite value. Beyond that, the
  round trip from integer to float and back will lose information. This is
  represented with a new MAX_SAFE_INT constant which is often included in
  other languages like JavaScript's Number.MAX_SAFE_INT. When, we encounter an
  Int variable in a Float position, we ensure that its value is finitely
  representable.

  There is some nuance in that the spec does not say all floats have to be
  within this range. So, this implementation allows explicitly passed floats
  which are greater than that bound, only applying the integer conversion limit
  when coercing a value.

Fixes

• Validate missing fragments when parsing standalone executable documents - Abdel-Monaam-Aouini, pull/1003

  When validating standalone executable documents, the use of undefined fragment
  definitions will return a validation error. Previously, executable documents
  like the following would pass validation without errors, despite
  CompanyFragment being undefined.

  query {
    company {
      user {
        ...UserFragment
      }
      ...CompanyFragment
    }
  }
  fragment UserFragment on User {
    id
    name
  }

Maintenance

• Add benchmark for parsing and validation when a type has many extensions tninesling, pull/1000

  Introduces a new benchmark for query parsing and validation when a type has
  many extensions. We made an update in [email protected] to expose
  .iter_origins() for AST nodes, and we reimplemented .extensions() in
  terms of .iter_origins(). We were concerned that this may have caused a
  performance regression in parsing, but running this new benchmark against
  main with 1.28.0 as the base indicates no change in performance.

[email protected]

Features

• Add ignore_builtin_redefinitions method to SchemaBuilder- dariuszkuc, pull/990 and pull/994

  This allows input SDL to contain built-in types.

  The following SDL will result in a validation error, for example:

    let schema = r#"
      type __Directive {
        name: String!
        description: String!
        isRepeatable: String!
        args: __InputValue
        locations: String!
      }
      type Query {
        foo: String
      }
      "#;
  
    let valid = Schema::parse_and_validate(schema, "schema.graphql")?

  Error:

    Error: the type `__Directive` is defined multiple times in the schema
      ╭─[ built_in.graphql:87:6 ]
      │
   87 │ type __Directive {
      │      ─────┬─────  
      │           ╰─────── previous definition of `__Directive` here
      │
      ├─[ schema.graphql:2:6 ]
      │
    2 │ type __Directive {
      │      ─────┬─────  
      │           ╰─────── `__Directive` redefined here
      │ 
      │ Help: remove or rename one of the definitions, or use `extend`
  ────╯

  However, when using the ignore_builtin_redefinitions method, this successfully passes validation given the same schema:

    let builder = SchemaBuilder::new().ignore_builtin_redefinitions();
    let _ = builder
        .parse(schema, "schema.graphql")
        .build()
        .expect("schema parsed successfully");

Fixes

• Fix handling of orphan root type extensions - dariuszkuc, pull/993

  SchemaBuilder's adopt_orphan_extensions method allows users to define type
  extensions without an existing type definition. But before this fix, orphan
  RootTypeOperation extensions would result in an invalid schema despite
  adopt_orphan_extensions being enabled. Using this method now generates a
  valid schema for all lone extensions.

• Fix directive definition validation with nested types arguments - dariuszkuc, pull/987

  Directive definition with nested argument types resulted in a stack overflow, for example

    directive @custom(input: NestedInput) on OBJECT | INTERFACE
  
    input NestedInput {
      name: String
      nested: NestedInput
    }
    
    type Query @custom(input: {name: "hello", nested: {name: "hello"}}) {
      foo: String
    }
    
    query myQuery {
      foo
    }

  This fix ensures the above example is possible and does not result in a validation error.

• Fix iter_origins() to be a pub method - duckki, pull/989

  Previously added ::iter_origins() methods on Schema and Type Definitions was not made pub.

[email protected]

Features

• Add iter_origin() methods to schema elements - duckki, pull/978
• Add public API for resolver-based execution - SimonSapin, pull/983

Fixes

• Fix serialization in non-standard orphan extensions mode - duckki, pull/984

[email protected]

Features

• Reject @skip/@include on subscription root fields - dariuszkuc SimonSapin and goto-bus-stop, pull/963

  This implements a GraphQL spec RFC, rejecting
  subscriptions in validation that can be invalid during execution.

• New shorthand methods for mutable directive argument access - tninesling, pull/967

  Introduces new methods:

  • DirectiveList::get_all_mut
  • DirectiveList::argument_by_name_mut
  • DirectiveList::specified_argument_by_name_mut

Fixes

• Update ariadne trait implementations - lrlna, pull/960

  [email protected] release changed their type signature for ariadne::Cache trait, which required an
  update to apollo-compiler's implementation of ariadne::Cache<FileId>.

  This release also had a slight change to path formatting, so if you had any snapshots in your
  tests, you can expect a change from this:

  Error: `typeFragment1` contains too much nesting
      ╭─[overflow.graphql:11:11]
  

  to this (notice the extra white space around the file path):

  Error: `typeFragment1` contains too much nesting
      ╭─[ overflow.graphql:11:11 ]
  

• Harden stack overflow protection - goto-bus-stop, pull/966

  Closes a theoretical gap in stack overflow protection when processing long fragment chains, and
  significantly improves validation performance on documents with thousands of fragment definitions.

[email protected]

0.8.1 - 2024-08-20

Features

• make SyntaxToken a publicly accessible API - g-plane, pull/894

This makes it easier for pretty-printers, linters and the like to work with the
parser API.

[email protected]

• Update apollo-parser dependency to 0.8.0
• Update apollo-compiler dependency to =1.0.0-beta.20

[email protected]

BREAKING

This release removes the Error::new constructor. We recommend not creating instances of
apollo_parser::Error yourself at all.

Fixes

• add missing location information for lexer errors - PhoebeSzmucer, pull/886, issue/731
  Unexpected characters now raise an error pointing to the character itself, instead of the start of the input document.

[email protected]

Fixes

• Fix variable validation in operation directives - goto-bus-stop in pull/888.
  • query ($var: Int) @directive(arg: $var) is now accepted - it used to raise an unused variable error for $var.

Maintenance

• Make unused variable validation more efficient - goto-bus-stop in pull/887.

[email protected]

1.0.0-beta.19 - 2024-07-19

BREAKING

• Remove deprecated Name reexports - SimonSapin in pull/877.
  apollo_compiler::Name should now be imported instead of:

  • apollo_compiler::ast::Name
  • apollo_compiler::schema::Name
  • apollo_compiler::executable::Name
    These other paths emitted deprecation warnings in 1.0.0-beta.18 and are now removed.

• Move operations of an ExecutableDocument into a new struct - SimonSapin in pull/879.

  • doc.anonymous_operation → doc.operations.anonymous
  • doc.named_operations → doc.operations.named
  • doc.get_operation() → doc.operations.get()
  • doc.get_operation_mut() → doc.operations.get_mut()
  • doc.insert_operation() → doc.operations.insert()
  • doc.all_operations() → doc.operations.iter()
    This change makes get_mut() borrow only doc.operations instead of the entire document, making it possible to also mutate doc.fragments during that mutable borrow.

• Move or rename some import paths - SimonSapin in pull/885. Moved from the crate root to the new(ly public) apollo_compiler::parser module:

  • Parser
  • SourceFile
  • SourceMap
  • FileId

  Moved to the parser module and renamed:

  • NodeLocation → SourceSpan
  • GraphQLLocation → LineColumn

• Return ranges of line/column locations - dylan-apollo in pull/861. SourceSpan contains a file ID and a range of UTF-8 offsets within that file. Various APIs can convert offsets to line and column numbers, but often only considered the start offset. They are now changed to consider the range of start and end positions.

  Added, returning Option<Range<LineColumn>>:

  • SourceSpan::line_column_range
  • Diagnostic::line_column_range
  • Name::line_column_range

  Removed:

  • LineColumn::from_node
  • Diagnostic::get_line_column
  • SourceFile::get_line_column

• Use a fast hasher - o0Ignition0o in pull/881. Configured all hash-based collections used in apollo-compiler to use ahash, which is faster than the default standard library hasher. apollo_compiler::collections provides type aliases for collections configured with the same hasher as collections in various parts of the public API.

Features

• Add a few helper methods - SimonSapin in pull/885:
  • executable::OperationMap::from_one
  • schema::SchemaDefinition::iter_root_operations()
  • schema::ExtendedType::is_leaf

Fixes

• Fix potential hash collision bug in validation - goto-bus-stop in pull/878
• Fix validation for undefined variables nested in values - goto-bus-stop in pull/885