v2.0.0

Release to fix limitations I faced developing rescript-envsafe. Shouldn't be breaking for most users.

• Add S.Int.port built-in refinement
• Rename action type to migration. Still don't like the name though 😅
• Add advanced preprocessing functionality, might be useful when you build sometimes on top of rescript-struct
• Remove [ReScript Struct] prefix from error message created by S.Error.toString
• Remove custom error RescriptStructError for panic and throw normal Errors with [rescript-struct] prefix in message
• Fix matching of Literal in the tagged type. I mentioned the problem in the form post
• Fix Defaulted and Json structs with parseAsyncInStepsWith. They will synchronously return error that happened before asynchronous operation
• The Json struct factory is more performant and creates structs with Json name instead of String
• Creating structs became 20% faster. Now to create a struct and parse is just a little slower than only parsing for zod
• Add rescript-envsafe to the ecosystem of rescript-struct
• Update LICENSE copyright

v1.0.0

Final API changes:

• Rename Record struct to Object
• Rename S.default to S.defaulted
• Redesign undocumented Metadata module
• Rename tagged_t to tagged
• Move Defaulted, Deprecated and unknownKeys from S.tagged to their own modules
• Add highlights to the documentation

v0.27.0

• Remove type: module from package.json

v0.26.0

• Add exception-based parseOrRaiseWith and serializeOrRaiseWith
• Remove undocumented isAsyncSerialize
• Update advancedTransform API
• Improve library tree-shaking
• Use ESM
• Remove beta banner from the README

v0.25.0

• Start using [email protected]
• Update API of the Deprecated struct factory to make it similar to built-in refinements
• Change Date struct tagged_t from Instance(unknown) to Date
• Add helper S.name to get a struct name. Useful for custom error messages

v0.24.0

• Add name property to the custom struct factory for better error messages.
• Remove GADT from the tagged_t type.

v0.23.0

Breaking changes:

• Remove different parsing modes. The reason is that the Safe mode became comparably fast to the Migration mode.
• Use S.Record.strip as default instead of S.Record.strict.
• Remove S.superTransform and S.Error.make.
• Change S.transform and S.custom API from result based to exception based. Use the new S.Error.raise to exit with failure.

New features:

• Finalize and document S.refine function.
• Add async refinements with S.asyncRefine.
• Add async transforms with S.advancedTransform.
• Add S.parseAsyncWith and S.parseAsyncInStepsWith to parse structs with asynchronous refinements or transforms.
• Add the UnexpectedAsync error when trying to synchronously parse structs with asynchronous refinements or transforms.
• Increase parsing performance by 20%.
• Update roadmap

v0.22.0

• Rename mode type to parsingMode
• Rename Unsafe parsing mode to Migration, fix its description
• Make serializing performance for common use-cases 30% faster

v0.21.0

• Remove different modes for serializing
• [dev] Update rescript-ava package

v0.20.1

• Don't test for an excess key when parsing failed with another error
• Add sub-error location to the InvalidUnion error string