Skip to content

Latest commit

 

History

History
134 lines (87 loc) · 3.71 KB

CHANGELOG.md

File metadata and controls

134 lines (87 loc) · 3.71 KB
Raw

Unreleased

v4.3.0 2023-01-30

  • Development: purs v0.15.7, node-streams-aff v5.0.0, assorted improvements.

v4.2.0 2022-12-27

  • Add doc comments from the .proto messages and enums to the generated .purs code.
  • Improved dev environment for purescript-language-server and ide-purescript with PURS_IDE_SOURCES.
  • Use spago2nix_nativeBuildInputs.

v4.1.0 2022-12-10

  • Nix flake. With one-step spago2nix.
  • Upgrade dependencies.
  • Delete mkUint8Array. Upstreamed this to purescript-arraybuffer.
  • Delete manyArray, parseMaybe, parseEither, parseExceptT. Upstreamed these to purescript-parsing.
  • Refactor manyLength. Get rid of the FFI, and a slight speedup in the benchmarks.
  • Delete Runtime.positionZero.

v4.0.0 2022-06-02

  • Change the values of UnknownBits64 and UnknownBits32 to Bytes.
  • Upgrade int64

v3.0.0 2022-05-27

  • PureScript v0.15.0
  • Protobuf v3.21.0
  • Replace text-encoding dependency with web-encoding. Consequently we require Node.js ≥ v11.
  • Replace longs dependency with int64.
  • Remove the only NPM dependency: longs.js.
  • Rework the async I/O in the protoc plugin.

v2.1.2 2021-09-10

Now the codegen Nix derivation is working properly.

v2.1.0 2021-07-25

Reorganized the spago files so that programs which import generated code can use spago.dhall for the minimum necessary dependencies instead of spago-library.dhall.

Reorganized the PureScript modules so that all modules except Library are now explicity Internal modules.

v2.0.0 2021-07-22

Upgraded to PureScript 0.14.3.

For byte fields we’re now using the DataBuff type from arraybuffer-builder v2.1.0 which allows us to avoid array copies, and so may speed up encoding and decoding in some cases.

Created a Protobuf.Prelude module to eliminate many import warnings in the generated code.

Breaking changes

The Bytes field type was formerly a wrapper for ArrayBuffer, and now it is a wrapper for DataBuff, which may be either ArrayBuffer or DataView. This changes the API and will require some changes in dependent code.

Bugfixes

Fixed a bug which sometimes incorrectly errored with “index out of bounds” when decoding packed repeated fields of double, float, fixed32, or sfixed32.

Fixed a bug which would cause a compilation failure in generated code for enums with negative values.

v1.8.0

Large (×1000) speedup for decoding packed repeated fields of

  • double
  • float
  • fixed32
  • fixed64
  • sfixed32
  • sfixed64

v1.7.0

Upgraded from Protobuf v3.14.0 to v3.15.8.

Now supporting proto3 field presence.

v1.6.0

Upgraded from Protobuf v3.9.2 to v3.14.0.

To conform with v3.14.0, we’ve added a new type class Default for awareness of Protobuf default values.

v1.5.0

In the code generator, when protoc hands us nonsense descriptors, then report the error instead of generating uncompilable code.

Export Library.manyLength, it’s generally useful.

v1.4.0

Improve Protobuf.Runtime.manyLength for stack-safety in the case of packed repeated fields.

v1.3.0

Newtype Protobuf.Library.Bytes wrapper for ArrayBuffer, so that all messages can have Eq and Show instances.

Nix default.nix derivation for including the code generation step as part another derivation.

v1.2.0

Preserve unknown fields, so now all Google binary-wire-format proto3 conformance tests pass.

v1.1.0

Added parsing labels so that parsing failure messages include a path to the protobuf definition in which the failure occurred.

v1.0.0

First release