pezkuwichain/pezkuwi-subxt
1
0
Fork 0
mirror of https://github.com/pezkuwichain/pezkuwi-subxt.git synced 2026-06-12 03:01:07 +00:00
Code Issues Packages Projects Releases Wiki Activity

subxt-core crate (#1466)

Browse Source 
* start migrating, broken

* first iteration of updating

* fmt and clippy

* add Composite<u32> decoding via scale value patch

* bump scale type gen versions

* fix decoding with new scale decode

* compiling with changed deps

* core utils, condig, client, metadata

* core crate compiling

* signer crate no once lock

* add core to no-std-tests, change imports

* broken commit, start pulling everything together in subxt

* port more things to subxt

* events in core crate, extrinsics sadly much more difficult

* almost all examples pass again

* dynamic values fix in examples

* fix no std issue and fmt

* remove unused dependencies

* fix lightclient impl

* runtime version refactor

* formatting and addressing nits

* more comments addressed

* update wasm example and no-std-signer tests

* other nits and error impl on signer errors

* fix feature flag

* fix runtime version refactor

* fix doc links

* fix integration tests

* fix feature flag gated client state

* fix native feature in CI

* fix lightclient utils

* make imports more lean in subxt-core

* integrate changes from subxt-core imports into subxt

* other changes in subxt simplify imports more

* fix  and docs

* doc false for cli

* fix clippy

* remove events block hash in tests

* codegen no-std support in generated code

* export alloc crate for no-std codegen

* fix doc test

* implement James comments

* remove std traits, use core traits instead

* address nits

* remove unusued dep in no-std tests

* fix Box import in no_std

* sp-crypto-hashing instead of sp-core-hashing

* bump scale-typegen, add no std codegen tests

* fix some things

* replace unmaintained derivative with derive_where to remove non-canonical warnings

* fmt

* remove unused dep

* fix deps

* update artifacts to fix type ID mismatches

* bump to latest scale-typegen

---------

Co-authored-by: James Wilson <james@jsdw.me>
This commit is contained in:
Tadeo Hepperle
2024-03-27 09:55:08 +01:00
committed by GitHub
parent 92c1ba7f66
commit a0cb14aa4f
106 changed files with 24329 additions and 26882 deletions
Download Patch File Download Diff File Expand all files Collapse all files
metadata/src/lib.rs
-39
View File
@@ -171,45 +171,6 @@ impl Metadata {
&OuterEnumHashes::empty(),
))
}
/// Ensure that every unique type we'll be generating or referring to also has a
/// unique path, so that types with matching paths don't end up overwriting each other
/// in the codegen. We ignore any types with generics; Subxt actually endeavours to
/// de-duplicate those into single types with a generic.
pub fn ensure_unique_type_paths(&mut self) {
let mut visited_path_counts = HashMap::<Vec<String>, usize>::new();
for ty in self.types.types.iter_mut() {
// Ignore types without a path (ie prelude types).
if ty.ty.path.namespace().is_empty() {
continue;
}
let has_valid_type_params = ty.ty.type_params.iter().any(|tp| tp.ty.is_some());
// Ignore types which have generic params that the type generation will use.
// Ordinarily we'd expect that any two types with identical paths must be parameterized
// in order to share the path. However scale-info doesn't understand all forms of generics
// properly I think (eg generics that have associated types that can differ), and so in
// those cases we need to fix the paths for Subxt to generate correct code.
if has_valid_type_params {
continue;
}
// Count how many times we've seen the same path already.
let visited_count = visited_path_counts
.entry(ty.ty.path.segments.clone())
.or_default();
*visited_count += 1;
// alter the type so that if it's been seen more than once, we append a number to
// its name to ensure that every unique type has a unique path, too.
if *visited_count > 1 {
if let Some(name) = ty.ty.path.segments.last_mut() {
*name = alloc::format!("{name}{visited_count}");
}
}
}
}
}
/// Metadata for a specific pallet.