jsonptr

Added

• Adds Pointer::split_at_offset, the byte-offset-based replacement for the now-deprecated Pointer::split_at.

Deprecated

• Deprecated ResolveError name.
• Deprecated Pointer::split_at in favor of Pointer::split_at_offset. The split_at name is being reserved so it can be reintroduced as a position (index) based split by 1.0, consistent with the rest of the API (e.g. Pointer::get).

Fixed

• Delete::delete no longer panics when a pointer's final token resolves to an array index equal to the array's length (e.g. - on any array, or a numeric index equal to the length). It now returns None and leaves the document unmodified, as documented. Fixes #120.
• EncodingError and ParseIndexError now implement Diagnostic, which unifies the API for errors originating from parse-like operations.
• Fixes returning an incorrect error type when parsing a Token that terminates with ~. This now correctly classifies the error as a Tilde error.
• Disabled default features for the serde dependency, which enables using this library in no-std environments even with the serde feature enabled.

Removed

• Some methods were removed from InvalidCharacterError, as that type no longer holds a copy of the input string internally. This is a breaking change. To access equivalent functionality, use the Diagnostic API integration.

Changed

• Token::from_encoded now accepts owned or borrowed strings (any type that implements Into<Cow<'_, str>>).
• Sealed the Diagnose trait.
• Implementation of the Default trait for Pointer now doesn't constrain the lifetime.

Changed

• Removes accidentally enabled default features "miette" and "toml"

Added

• Adds method into_buf for Box<Pointer> and impl From<PathBuf> for Box<Pointer>.
• Adds unsafe associated methods Pointer::new_unchecked and PointerBuf::new_unchecked for external zero-cost construction.
• Adds Pointer::starts_with and Pointer::ends_with for prefix and suffix matching.
• Adds new ParseIndexError variant to express the presence non-digit characters in the token.
• Adds Token::is_next for checking if a token represents the - character.
• Adds InvalidEncoding to represent the two possible encoding errors when decoding a token.
• Adds diagnotic::Diagnostic trait to facilitate error reporting and miette integration. All errors intended for usage with assign::Assign or resolve::Resolve must implement this trait.
• Adds diagnostic::Report<T> to capture the input for PointerBuf::parse and to facilitate miette integration for all errors.
• Adds "miette" feature flag to enable miette integration for error reporting.

Changed

• Pointer::get now accepts ranges and can produce Pointer segments as output (similar to slice::get).
• Bumps minimum Rust version to 1.79.
• PointerBuf::parse now returns RichParseError, an alias to Report<ParseError> which contains the allocated string as well as the error. Use Report::original for matches or `Report::
• Renames ParseError::NoLeadingBackslash to ParseError::NoLeadingSlash (sorry for the churn, I spaced hard - @chanced).
• Adds field position to variants of resolve::Error and assign::Error to indicate the token index of where the error occurred.
• Renames ParseError::is_no_leading_backslash to ParseError::is_no_leading_slash.
• Renames assign::AssignError to assign::Error
• Renames resolve::ResolveError to resolve::Error
• Renames InvalidEncodingError to EncodingError

Fixed

• Make validation of array indices conform to RFC 6901 in the presence of non-digit characters.

Deprecated

• ParseError::is_no_leading_backslash renamed to ParseError::is_no_leading_slash.
• assign::AssignError renamed to assign::Error
• resolve::ResolveError renamed to resolve::Error
• InvalidEncodingError renamed to EncodingError

Fixed

• Token::to_index now fails if the token contains leading zeros, as mandated by the RFC.

Changed

• ParseIndexError is now an enum to reflect the new failure mode when parsing indices.

Changed

• README tweak.

This is a breaking release including:

• #30 and #37 by @asmello
• #41 by @chanced & @asmello

Added

• New slice type Pointer that enables zero-copy usage patterns
• New const constructor const fn Pointer::from_static for compile-time allocated Pointers
• Zero-allocation Pointer::root singleton pointer
• Quickcheck-based testing
• New methods: Pointer::split_front, Pointer::split_back, Pointer::parent, Pointer::strip_suffix
• Implemented Display and Debug for ParseError
• Adds Pointer::split_at which utilizes character offsets to split a pointer at a separator
• Adds specific error types ParseError, ResolveError, AssignError

Changed

• JSON Pointers with leading "#" are no longer accepted. Previously, the erroneous leading hashtag was allowed during parsing but discarded.
• Assign, Resolve, ResolveMut, Delete all now use associated types Value and Error, allowing for more impls other than JSON
• Debug implementation now preserves type information (e.g. prints PathBuf("/foo/bar") instead of "/foo/bar") - Display remains the same
• Original Pointer type renamed to PointerBuf
• Error types now use character offset indexing instead of owned copies of Pointer and Token.
• Pointer::root is now PointerBuf::new
• Pointer::new is now PointerBuf::from_tokens (and takes an IntoIterator argument - arrays still work)
• Pointer::union is now PointerBuf::intersection
• Token type has been simplified and is now by default a borrowed type (use Token::to_owned or Token::into_owned to make it owned)
• Assign::assign now returns Result<Option<Assign::Value>, AssignError>, where Option<Assign::Value> is the replaced value

Fixed

• Fixes #28: Pointer::union is misleadingly named

Removed

• Removes Assignment
• Removes MaybePointer
• Removes Error
• Removes impl Deref<Target=&str> from Pointer
• Removes optional dependencies of url, fluent-uri and uniresid as well as the TryFrom implementations for their respective types
• Removed Token::as_key and Token::as_str - use Token::decoded().as_ref() to achieve the same effect
• Several redundant or error-prone trait implementations were removed from Token