0.11.0

AST changes to reflect changes in rustc

• Rename ExprKind::Vec to ExprKind::Array
• Change the content of ExprKind::Closure from a Block to Expr
• Support expression values for break statements
• Support attributes on fields in a struct literal expression (FieldValue) and pattern (FieldPat)
• Add a new type of where predicate - EqPredicate
• Combine Ty::ObjectSum and Ty::PolyTraitRef into Ty::TraitObject
• Add Ty::Mac to support macros in type position

Other breaking changes

• Remove the custom-derive registry that was used for expanding derives on stable rust; real custom derives will be stable very soon
• Unify BlockCheckMode and Unsafety (#66, thanks @shepmaster)

Non-breaking changes

• Rename parse_macro_input to parse_derive_input and MacroInput to DeriveInput now that there are other types of procedural macros that are not derives; the old names will continue to work because they have appeared in so many Macros 1.1 guides (#63, thanks @SimonSapin)
• Support parsing macro invocations that use namespaced paths like a::b!(...)
• The AST walker syn::visit::Visitor now supports all AST types (#79, thanks @mystor)

0.10.6

• #[derive(Hash)] for all the syntax tree types (#73)

0.10.5

• Add syn::parse_ident (#71)

0.10.4

(yanked)

0.10.3

• Export FieldValue which was mistakenly not public (#67, thanks @shepmaster)

0.10.2

• Make the fields of Expr public (#65, thanks @shepmaster)

0.10.1

• Export ExprKind which was mistakenly not public (#64, thanks @KodrAus)

0.10.0

Significant new features

• Expand custom derives on stable Rust
• Optionally pretty-print expanded output

Lots of minor enhancements

• More efficient implementation of Generics::split_for_impl
• Support for out-of-line modules i.e. mod foo; in addition to mod foo { ... }
• From impls for constructing the various Lit variants (#50, thanks @killercup)
• Parsing of raw byte strings
• Parsing of floating-point literals
• Parsing of integer literals containing underscores for visual alignment
• Parsing of hex, octal, and binary integer literals
• More restrictive parsing of variable names
• Nicer output for string literals containing the zero byte \0
• Parsing of doc comments inside of macro invocations
• Parsing of const trait and impl methods
• Parsing of bare function types containing lifetimes
• Parsing of module inner attributes
• Parsing of hex escapes greater than \x7F in byte strings and byte literals
• Parsing of trailing commas in a match expression
• Parsing of slice patterns
• Parsing of unsafety and ABI associated with a bare function type
• Parsing of variadic extern functions
• Parsing of negative integer literal patterns
• Parsing of inner attributes within functions and methods
• Parsing of trailing commas in use declarations
• Parsing of attributes on statements
• Parsing of where clauses on type aliases
• Parsing of left-to-right mark and right-to-left mark codepoints
• Parsing of placement syntax
• Parsing of shebang at the top of a file
• Parsing of shorthand syntax within struct literal expressions
• Parsing of trailing commas inside generics
• Parsing of carriage returns in multiline string literals
• Parsing of pub(crate) and pub(restricted) visibility modifiers
• Parsing of byte order marks
• Parsing of trailing commas in slice patterns

0.10.0-rc1

Release 0.10.0-rc1

0.9.2

• Add syn::parse_token_trees to parse tokens from a string. Token trees are much simpler than an AST. This makes them easier to work with when they are sufficient for the task at hand, such as for example expanding a function-like macro. (#52, thanks @SimonSapin)