0.5

[oscartbeaumont]

Minor:

• Make specta_serde errors good. Include path to procedure where it broke and which type it broke on.
• Remove BigInt warnings

Core:

• rspc_procedure
  • Split out of rspc_core
  • Allow downcasting (DynOutput::value) a non from_*_stream type?
  • Handle invariants of DynInput::new_value. Option::None will reach havoc.
  • Make serde/related crates an optional depedency (cause Binario doesn't need it)
  • ResolverError::new forces usage of Serde for value. Can we allow dyn Any?
  • Error should have a variant for a custom error by middleware (Eg. Zer). I'm not sure we can make it generate a type but it should be able to serialize any data so a JS package could throw types back onto it later.
  • Should DynOutput have a variant for ProcedureError? We can allow manually grabbing it but it would make the serialization code in userspace much cleaner.
  • Simplify ProcedureStream's constructor code
  • ProcedureStream::from_error is broken with flushing logic. The todo!() in ProcedureStream::poll_inner
  • Make flush work
  • trait impls of ProcedureStream::map related types
  • type_name on DynOutput make it useful or remove it
  • Rename DynInput and DynOutput?
  • Drop all Any bounds cause they are implied
  • Rename - require_manual_stream -> block_streaming, stream -> unblock_streaming, etc.
  • Procedures::new
  • impl Fn() -> TCtx or TCtx: Clone. rspc_invalidation makes this annoying.
  • wait_util function similar to the Cloudflare Workers/Vercel feature. How would a CDN implement it properly?
  • Should we call it State or Data?
  • Finish Debug impls + clippy run
  • Documentation
  • Audit for erased_serde in public API

Ecosystem:

• rspc_devtools
  • fix it
  • Should we provide a way to set CORS header. If we leave it on the user the URL being the same is a problem?
• rspc_tracing:
  • Drop `rspc_core::Procedure::with_logger``
  • Add Logger in rspc_core
  • Allow Router::setup to inject a Logger instance for rspc-tracing
  • Logger hook should create span with procedure metadata (ensure that span is tracked for the future somehow)
  • Middleware should only add input/result to existing span.

Features

• Zer::from_request errors handled as an rspc_core::ProcedureError??
• rspc - cleanup module structure cause it's a mess
• SFM's
  • What if a SFM tries to call .invalidate?? We can't queue it again without conflicting keys
  • What if a SFM calls flush?? Cause we are well past that point.
  • stream.resolved() || stream.flushable() is bad
  • Delay executing SFM queries until the batch resolves
  • Deduplicating SFM keys for rspc_invalidation
  • Prevent calling .invalidate after streaming has started. What should happen?
  • rspc_invalidation can we skip serde_json::Value?
  • rspc_invalidate Invalidate::Any
  • rspc_invalidate Invalidate::Many
• Wire protocol
  • Allow extensions to add extra data
  • Custom streaming response format. Differentiate errors vs values and reference to which query/mutation the result is referring to.
  • Poll the runtime while waiting for the next FormData item in the input.
  • Support WebSocket upgrades too
  • Are we going to keep rspc-axum/rspc-actix-web/rspc-http. Heavily dependant on how out of control the code in userspace ends up as.
  • File will bridge AsyncWrite/AsyncBufWrite to Stream<Item = Vec<u8>> so it can reuse the existing streaming machinery. This will allow interleaving multiple file downloads in one response.
    • In pratice this is how Hyper works so as much as I hate the overhead of buffering it's gonna happen anyway.
    • Yield AsBuffer(Vec<u8>) and then we can know to set the Content-Type in the transport layer correctly.
• New syntax
  • If TError is fixed maybe we put it on the Router cause it's a bit wierd with BaseProcedure?
  • ResolverError takes two args which should both be self and anyhow::Error is not Clone. I think we can do smart stuff with constructors to solve this.
  • rspc::Error2 is super unsafe! It lets you choose to use #[error(...)] for the repr which is a string not TError.
  • Drop rspc::Error2 in favor of Into?
  • Manual flushing option build on the ProcedureStream flushing stuff
  • ProcedureBuilder::subscription implemented
  • Type exporting (right now they are hardcoded to DataType::Unknown)
  • Support exporting statics #285
  • Improve error handling DX #305 (might require Specta major)
  • ProcedureMeta::state called within a Middleware::setup should panic when accessed as it can't be valid.
  • Correctly determine key in Procedure2::builder.
  • specta_serde integration prototype
  • Jump to definition on routers
  • IntoResponse: ! future compat warning
  • Move Procedure2 into modern::procedure or keep in root?
• Middleware system
  • Middleware::setup and Extension::setup are overriding which is major cringe alert.
  • Extension map function working. Might need TError generic?
  • Break out into rspc_middleware - ResolverInput, ResolverOutput, Middleware, Extension and related stuff
  • Stop any of the middleware depending on rspc and instead use rspc_middleware
  • Prevent applying rspc-invalidation to mutations and subscriptions. Should we make it typesafe?
• Handling serializer errors
  • ProcedureStreamMap error handling
  • ProcedureError::Serializer or not? If not clear out commented out stuff.
  • I suppose we hardcode an error response to send if we can't serialize the first error.
  • rspc_tauri should handle first-level serializer errors through rspc instead of throwing into Tauri.
• Tauri
  • What's the new integration going to look like?
  • Fix @rspc/tauri with legacy system
  • Maybe copy File from rspc_http? Idk if their IPC can handle it but I suppose supporting it anyway would be worthwhile.
• WithDate
  • trait impls
  • Generalise implementation to any Serialize?
  • Use thread_local to enable the special rspc encoding so the type doesn't mess with users other stuff.
  • Require Accept: text/x-rspc for it to use the special encoding.
• This is blocked on TypeMap + TypeCollection specta#294 & Language in practice specta#297 & releasing specta-rust.

Breaking Changes:

• Determine future for rspc_axum/rspc_tauri. Userspace maybe?
• Drop rspc_legacy + old bindings format

Unsorted:

• Improve Procedure's Arcing.
• Should Procedure2::error being Option<DataType> or not?
• Ability to downcast instead of Serialize on error. Similar to the regular Ok type.
• Defining middleware that don't depend on rspc.