Add `CodamaPdaHelpers` derive by grod220 · Pull Request #86 · codama-idl/codama-rs

grod220 · 2026-03-25T18:14:37Z

Adds #[derive(CodamaPdaHelpers)] which generates runtime PDA helper methods from existing #[codama(seed(...))] attributes. This reduces repeated manual PDA boilerplate program devs are writing.

API

use codama::{CodamaPda, CodamaPdaHelpers};

#[derive(CodamaPda, CodamaPdaHelpers)]
#[codama(seed(type = string(utf8), value = "mint"))]
#[codama(seed(name = "unwrappedMint", type = public_key))]
#[codama(seed(name = "wrappedTokenProgram", type = public_key))]
pub struct WrappedMintPda;

// ==== In program ==== 

let (address, bump) = WrappedMintPda::derive_program_address(
    &unwrapped_mint,
    &wrapped_token_program,
    &program_id,
)?;

let signer_seeds: [Seed; 4] = WrappedMintPda::signer_seeds(
    &unwrapped_mint,
    &wrapped_token_program,
    &bump_bytes,
);

Implemented macro expansion that parses struct-level #[codama(seed(...))] directives and generates:
- seeds
- seeds_with_bump
- signer_seeds
- (wrapping Address methods below)
- derive_address
- create_program_address
- find_program_address
- try_find_program_address
- derive_program_address
Supports constant seeds (string(utf8), integer number(...)) and variable/linked seeds
Handles seed name deduplication, identifier collision detection, reserved bump, and Rust keyword escaping

Discussion points

Better name?
Should all of those Address methods be shipped? Aka, should we drop find_program_address?
The intention was for program usage, but might clients also need this?
An impl vs external crate helpers
How does this sit next to the CPI helpers PR (link?)

grod220 · 2026-03-25T18:15:30Z

+#[proc_macro_derive(CodamaPdaHelpers, attributes(codama))]
+pub fn codama_pda_helpers_derive(input: TokenStream) -> TokenStream {


@febo better naming proposals?

Whatever we land on with the derive name, just want to flag that IMO it should start with Codama since it purely relies on codama attributes.

grod220 · 2026-03-25T18:16:19Z

+/// (wrapping `Address` methods)
+/// - `derive_address`
+/// - `create_program_address`
+/// - `find_program_address`
+/// - `try_find_program_address`
+/// - `derive_program_address`


@febo thoughts on which Address methods this should include? This sort of wraps everything, but maybe it should be a blessed subset.

lorisleiva

I'll go through this thoroughly asap but I just wanted to share some early first impression feedback.

lorisleiva · 2026-03-26T17:16:56Z

@@ -0,0 +1,97 @@
+use heck::ToSnakeCase;


Since we already have an exported CamelCaseString struct in codama-nodes maybe it would be cleaner / more consistent to also add a SnakeCaseString struct there instead of introducing a new dependency.

lorisleiva · 2026-03-26T17:18:28Z

+#[proc_macro_derive(CodamaPdaHelpers, attributes(codama))]
+pub fn codama_pda_helpers_derive(input: TokenStream) -> TokenStream {
+    #[cfg(not(target_os = "solana"))]
+    {
+        pda_helpers::codama_pda_helpers_derive_impl(input.into())
+            .unwrap_or_else(codama_errors::CodamaError::into_compile_error)
+            .into()
+    }
+    #[cfg(target_os = "solana")]
+    {
+        input
+    }
+}
+


It would be nice if this macro was on its own crate. Maybe something like codama-pda-helpers or whatever the name ends up being here.

That way, if anyone wants to use a slightly different version of our generated code, it's easier for them to fork that crate on its own.

aditya-able · 2026-08-24T09:27:19Z

@lorisleiva @grod220 I opened #NNN with the SnakeCaseString struct from the comment above, so this can drop the heck dependency.

Gabe, if you are still on this I will leave the ident.rs change to you. If you have moved on, happy to pick it up, rebase, and switch it over.

grod220 · 2026-08-24T12:33:22Z

@aditya-able, I've moved on to some other things, but if you find it useful, I encourage you to push things forward! Feel free to take over this branch/PR or another 👍

Same shape as CamelCaseString: same derives, new<T: AsRef<str>>, the String and &str conversions, Deref and AsRef, and the tests in the same order. Word boundaries deliberately do not follow CamelCaseString. to_camel_case reads a run of uppercase letters as one word, turning tokenAMint into tokenAmint and SPLToken into spltoken. In snake_case that would give token_amint and spltoken, which make poor identifiers, and tokenAMint and tokenBMint show up often enough in real PDA seeds to matter. So an uppercase run splits before its last letter when a lowercase letter follows. Digits carry the case of whatever preceded them, so a digit ends a word only when an uppercase letter follows a lowercase run. seed2Bump gives seed2_bump while mint_2, this123 and 2Mint are left alone. That combination is exactly heck::ToSnakeCase, checked by differential fuzzing against heck over 1,000,021 inputs with zero mismatches, so codama-idl#86 can drop the dependency without a change in behaviour.

Same shape as CamelCaseString: same derives, new<T: AsRef<str>>, the String and &str conversions, Deref and AsRef, and the tests in the same order. Word boundaries deliberately do not follow CamelCaseString. to_camel_case reads a run of uppercase letters as one word, turning tokenAMint into tokenAmint and SPLToken into spltoken. In snake_case that would give token_amint and spltoken, which make poor identifiers, and tokenAMint and tokenBMint show up often enough in real PDA seeds to matter. So an uppercase run splits before its last letter when a lowercase letter follows. Digits carry the case of whatever preceded them, so a digit ends a word only when an uppercase letter follows a lowercase run. seed2Bump gives seed2_bump while mint_2, this123 and 2Mint are left alone. That combination matches heck::ToSnakeCase, checked by differential fuzzing over 1,000,021 ASCII inputs with zero mismatches. On a further 400,000 Unicode inputs the only divergence is a word-final Greek capital sigma, which heck lowercases to the final form while this lowercases per character, as to_camel_case does. So codama-idl#86 can drop the dependency.

senzenn · 2026-09-02T13:51:27Z

@lorisleiva happy to take this over if it's still wanted Gabe gave the go-ahead above. Before I rebase, two v2 questions, since the answers change what I'd build:

Timing: codama-rs is going to 1.0.0 to freeze the v1 API before the v2 work lands. Should this go into 1.x, or wait and target v2 directly?
Scope: is the seed attribute inventory settled yet specifically whether PDA seed pairing joins? If seed references become path expressions, the seed parsing here changes shape, and I want to know if that reaches seed(name = ...). [ref the issue where that's tracked]

On casing: agreed on #136, no new type in codama-nodes. If you want to keep heck out, I have to_snake_case as a plain function in codama-syn-helpers alongside the other syn helpers. It matches heck::ToSnakeCase on ASCII, fuzzed over a million inputs with zero mismatches. Equally fine leaving heck in it's one line and not the interesting part here.

If the answer is wait for v2, I'd rather spend the time on the identifier rename instead. Point me at whichever is more useful.

lorisleiva · 2026-09-03T13:12:47Z

@senzenn Since v2 is just around the corner I'd rather not make any radical changes to v1 in the meantime since we otherwise risk to miss some features when forward-porting them to v2.

grod220 commented Mar 25, 2026

View reviewed changes

Add CodamaPdaHelpers derive

ee25ff3

grod220 force-pushed the pda-helpers branch from faba10f to ee25ff3 Compare March 25, 2026 18:24

grod220 requested review from buffalojoec, febo and lorisleiva March 25, 2026 18:25

lorisleiva reviewed Mar 26, 2026

View reviewed changes

senzenn mentioned this pull request Aug 24, 2026

Add SnakeCaseString to codama-nodes #136

Closed

senzenn mentioned this pull request Sep 3, 2026

feat:snake case syn helper #139

Closed

Navigation Menu

Add `CodamaPdaHelpers` derive - #86

Add `CodamaPdaHelpers` derive#86
grod220 wants to merge 1 commit into
codama-idl:mainfrom
grod220:pda-helpers

grod220 commented Mar 25, 2026 •

edited

Loading

Uh oh!

grod220 Mar 25, 2026

Uh oh!

lorisleiva Mar 26, 2026

Uh oh!

grod220 Mar 25, 2026

Uh oh!

lorisleiva left a comment

Uh oh!

lorisleiva Mar 26, 2026

Uh oh!

lorisleiva Mar 26, 2026

Uh oh!

aditya-able commented Aug 24, 2026

Uh oh!

grod220 commented Aug 24, 2026

Uh oh!

senzenn commented Sep 2, 2026

Uh oh!

lorisleiva commented Sep 3, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

4 participants

		#[proc_macro_derive(CodamaPdaHelpers, attributes(codama))]
		pub fn codama_pda_helpers_derive(input: TokenStream) -> TokenStream {

Conversation

grod220 commented Mar 25, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

API

Discussion points

Uh oh!

grod220 Mar 25, 2026

Choose a reason for hiding this comment

Uh oh!

lorisleiva Mar 26, 2026

Choose a reason for hiding this comment

Uh oh!

grod220 Mar 25, 2026

Choose a reason for hiding this comment

Uh oh!

lorisleiva left a comment

Choose a reason for hiding this comment

Uh oh!

lorisleiva Mar 26, 2026

Choose a reason for hiding this comment

Uh oh!

lorisleiva Mar 26, 2026

Choose a reason for hiding this comment

Uh oh!

aditya-able commented Aug 24, 2026

Uh oh!

grod220 commented Aug 24, 2026

Uh oh!

senzenn commented Sep 2, 2026

Uh oh!

lorisleiva commented Sep 3, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

4 participants

grod220 commented Mar 25, 2026 •

edited

Loading