litrs

As announced, this is basically just v0.5. I'm really happy to finally have litrs in v1.x state, as the library was already fit for production usage for a long time. Now the version reflects that.

• Breaking: Mark Literal as #[non_exhaustive]
• Document semver guarantees

As announced, this is basically just v0.5. I'm really happy to finally have litrs in v1.x state, as the library was already fit for production usage for a long time. Now the version reflects that.

• Breaking: Mark Literal as #[non_exhaustive]
• Document semver guarantees

• Add Literal::raw_input

• Add Literal::raw_input

Note: this is supposed to be essentially a 1.0 release candidate. If there are no problems with this release, it will be released as 1.0 after some time.

• Breaking Bump MSRC from 1.54 to 1.56
• Breaking: Remove proc-macro2 from default features
• Breaking: Make StringLit<String>::into_value return String instead of Cow<'static, str>.
  Make ByteStringLit<String>::into_value return Vec<u8> instead of Cow<'static, [u8]>.
• Add support for C string literals
• Limit number of # as delimiter of raw string literals to 256 (to align with spec)
• Fix license in Cargo.toml to use proper SPDX format (in #19, thanks @atouchet)
• Fix error span for InvalidStartOfUnicodeEscape errors

Note: this is supposed to be essentially a 1.0 release candidate. If there are no problems with this release, it will be released as 1.0 after some time.

• Breaking Bump MSRC from 1.54 to 1.56
• Breaking: Remove proc-macro2 from default features
• Breaking: Make StringLit<String>::into_value return String instead of Cow<'static, str>. Make ByteStringLit<String>::into_value return Vec<u8> instead of Cow<'static, [u8]>.
• Add support for C string literals (also breaking due to new enum variant)
• Limit number of # as delimiter of raw string literals to 256 (to align with spec)
• Fix license in Cargo.toml to use proper SPDX format (in #19, thanks @atouchet)
• Fix error span for InvalidStartOfUnicodeEscape errors

• Fix incorrect byte string value with non-ASCII \xHH byte string escapes
  • ByteStringLit::parse(r#"b"\xff""#).unwrap().value() would return [0xc3, 0xbf] instead of [0xff].
• Remove CR LF normalization to align with spec
  • This is technically a breaking change, but released with a minor version bump as it just aligns with the specification. It is extremely unlikely to affect any real world use case.
• Fix incorrect error span for out-of-range Unicode escapes

• Fix incorrect byte string value with non-ASCII \xHH byte string escapes
  • ByteStringLit::parse(r#"b"\xff""#).unwrap().value() would return [0xc3, 0xbf] instead of [0xff].
• Remove CR LF normalization to align with spec
  • This is technically a breaking change, but released with a minor version bump as it just aligns with the specification. It is extremely unlikely to affect any real world use case.
• Fix incorrect error span for out-of-range Unicode escapes

• Fixed incorrectly labeling 27f32 a float literals in docs.
• Added hint to integer literal docs about parsing as u128.

• Fixed incorrectly labeling 27f32 a float literals in docs.
• Added hint to integer literal docs about parsing as u128.

Added

• Add ability to parse literals with arbitrary suffixes (e.g. "foo"bla or 23px)
• Add suffix() method to all literal types except BoolLit
• Add IntegerBase::value
• Add from_suffix and suffix methods to FloatType and IntegerType
• Add FromStr and Display impls to FloatType and IntegerType

Changed

• Breaking: Mark FloatType and IntegerType as #[non_exhaustive]
• Breaking: Fix integer parsing for cases like 27f32. Literal::parse
  and IntegerLit::parse will both identify this as an integer literal.
• Breaking: Fix float parsing by correctly rejecting inputs like 27f32. A
  float literal must have a period OR an exponent part, according to the spec.
  Previously decimal integers were accepted in FloatLit::parse.
• Improved some parts of the docs

Removed

• Breaking: Remove OwnedLiteral and SharedLiteral

Added

• Add ability to parse literals with arbitrary suffixes (e.g. "foo"bla or 23px)
• Add suffix() method to all literal types except BoolLit
• Add IntegerBase::value
• Add from_suffix and suffix methods to FloatType and IntegerType
• Add FromStr and Display impls to FloatType and IntegerType

Changed

• Breaking: Mark FloatType and IntegerType as #[non_exhaustive]
• Breaking: Fix integer parsing for cases like 27f32. Literal::parse and IntegerLit::parse will both identify this as an integer literal.
• Breaking: Fix float parsing by correctly rejecting inputs like 27f32. A float literal must have a period OR an exponent part, according to the spec. Previously decimal integers were accepted in FloatLit::parse.
• Improved some parts of the docs

Removed

• Breaking: Remove OwnedLiteral and SharedLiteral

Breaking

• Bump MSRV (minimal supported Rust version) to 1.54

Added

• Add raw_input and into_raw_input to non-bool *Lit types
• Add impl From<*Lit> for pm::Literal (for non-bool literals)
• Add impl From<BoolLit> for pm::Ident

Fixed

• Fix link to reference and clarify bool literals (#7)

Internals

• Move lots of parsing code into non-generic functions (this hopefully reduces compile times)
• To implement [into_]raw_input for integer and float literals, their
  internals were changed a bit so that they store the full input string now.

Breaking

• Bump MSRV (minimal supported Rust version) to 1.54

Added

• Add raw_input and into_raw_input to non-bool *Lit types
• Add impl From<*Lit> for pm::Literal (for non-bool literals)
• Add impl From<BoolLit> for pm::Ident

Fixed

• Fix link to reference and clarify bool literals (#7)

Internals

• Move lots of parsing code into non-generic functions (this hopefully reduces compile times)
• To implement [into_]raw_input for integer and float literals, their internals were changed a bit so that they store the full input string now.

See CHANGELOG.md

Changed

• Minor internal code change to bring MSRV from 1.52 to 1.42

See CHANGELOG.md

Changed

• Fixed (byte) string literal parsing by:
  • Correctly handling "string continue" sequences
  • Correctly converting \n\r into \n

See CHANGELOG.md

Changed

• Fixed the expected value of the error returned from TryFrom<TokenTree> impls in some cases

Changed

• Breaking: rename Error to ParseError. That describes its purpose more closely and is particular useful now that other error types exist in the library.

Removed

• Breaking: remove proc-macro feature and instead offer the corresponding impls unconditionally. Since the feature didn't enable/disable a dependency (proc-macro is a compiler provided crate) and since apparently it works fine in no_std environments, I dropped this feature. I don't currently see a reason why the corresponding impls should be conditional.

Added

• TryFrom<TokenTree> for litrs::Literal impls
• From<*Lit> for litrs::Literal impls
• TryFrom<proc_macro[2]::Literal> for *Lit
• TryFrom<TokenTree> for *Lit
• InvalidToken error type for all new TryFrom impls

Added

• From impls to create a Literal from references to proc-macro literal types:
  • From<&proc_macro::Literal>
  • From<&proc_macro2::Literal>
• Better examples in README and repository

Added

• Everything