Add `sncast get tx-trace` by franciszekjob · Pull Request #4546 · foundry-rs/starknet-foundry

franciszekjob · 2026-08-18T09:56:15Z

Towards #3557

Introduced changes

Add sncast get tx-trace command

Checklist

franciszekjob · 2026-08-20T10:28:07Z

@@ -163,6 +165,20 @@ impl OutputBuilder {
                        field_value,
                    )
                    .unwrap();
+
+                    for line in value_lines {
+                        if line.is_empty() {
+                            content.push('\n');
+                        } else {
+                            writeln!(
+                                content,
+                                "{}{}",
+                                " ".repeat(field_width + 1),
+                                style(line).yellow()
+                            )
+                            .unwrap();
+                        }
+                    }


note: This change is needed to align continuation lines with the field value, keeping multiline values (such as revert reason) readable.

franciszekjob · 2026-08-20T11:35:13Z

-starknet-rust = "0.19.0"
+# TODO: Restore version once starknet-rust releases new version
+starknet-rust = { git = "https://github.com/software-mansion/starknet-rust", rev = "56efca00feffac6beb39f464eca71afa42c1813b" }


note: Temporary, due to change in the content type in NoTraceAvailable error.

ddoktorski · 2026-08-27T09:51:04Z

+
+fn decode_invocation_json(
+    invocation: &FunctionInvocation,
+    json: &mut Value,


Why decode_execute_invocation_json takes Option and decode_invocation_json not?

Changed, 54735d2

ddoktorski · 2026-08-27T09:51:28Z

+                json.get_mut("validate_invocation"),
+                decoder,
+            );
+            if let Some(json) = json.get_mut("constructor_invocation") {


Shouldn't we rather throw an error if required field is missing? 🤔

Good catch, e82b05b

ddoktorski · 2026-08-27T09:53:55Z

+    where
+        S: Serializer,
+    {
+        let mut json = serde_json::to_value(&self.trace).map_err(S::Error::custom)?;


What is the rationale that it is first parsed to json instead of direct serialization of the struct like for other commands e.g. get tx-receipt?

Direct serialization would return raw felts. We convert to JSON first to preserve the original trace schema while replacing selectors, calldata etc. with decoded values. Mentioned tx-receipt (and others) do not require such transformations.

ddoktorski · 2026-08-27T09:56:32Z

+    let builder = append_calls(builder, &invocation.calls, decoder, indent);
+    let builder = builder
+        .with_indent(indent)
+        .felt_field("Class Hash", &invocation.class_hash)


Let's use padded_felt_field for class hashes and contract addresses

ddoktorski · 2026-08-27T10:00:57Z

+}
+
+#[must_use]
+pub fn contract_addresses_by_class_hash(


It seems that this and related code should be placed in starknet_commands/get/tx_trace.rs not in reponse dir

ddoktorski · 2026-08-27T10:18:47Z

+fn provider_error_message(error: &ProviderError) -> String {
+    match error {
+        ProviderError::StarknetError(error) => error.message().to_string(),
+        error => error.to_string(),
+    }
+}


Does it make sense to fix it on the starknet-rust side?

Yeah it does, changed in software-mansion/starknet-rust#160

ddoktorski · 2026-08-27T10:27:32Z

+    builder
+}
+
+fn append_full_invocation(


The current display seems to be in this order:

root Call Type, Calldata, Caller Address

root Calls

nested call details

root Class Hash, Contract Address, Entry Point Selector etc.

which is very hard to read, please fix it so all root data go first

ddoktorski · 2026-08-27T10:29:06Z

+
+## `--full`
+Optional.
+Conflicts with: [`--json`](../common.md#--json--j)


If --json is passed directly after sncast there is no conflict and --full is ignored

Yeah, right. In #4567 I made --json flag global, so the conflict could be defined at clap level. However, clap does not detect it when --json appears before the subcommand. I think the cleanest solution is to validate the --json / --full conflict at runtime remove the now-unnecessary change from mentioned PR.

Runtime validation -> 7758ae0

Could you remind me why actually we can't use both these flags together? 🙏

ddoktorski · 2026-09-03T18:20:55Z

+Type:                     INVOKE
+Validate Invocation
+  Entry Point Selector:   __validate__
+  Contract Address:       0x[..]


Why not to show the exact values here?

ddoktorski · 2026-09-04T07:15:28Z

+                fetch_contract_classes(&provider, class_hashes(&trace)).await;
+
+            if !failures.is_empty() {
+                ui.print_warning(WarningMessage::new(format_class_fetch_warning(&failures)));


Would it be possible to make it part of decoding_warnings so all decoding errors are gathered in one place? It should work better for json output then

ddoktorski · 2026-09-04T07:20:55Z

+
+    fn selector(&self, invocation: &FunctionInvocation) -> String {
+        if let Some(abi) = self.sierra_abis.get(&invocation.class_hash.into_())
+            && let Some(function) =


invocation.class_hash.into_() is evaluated a few times in this function, let's just add

let class_hash: ClassHash = invocation.class_hash.into_();

also in the functions below

ddoktorski · 2026-09-04T07:23:17Z

+    format!("Some trace data is shown as raw felts:\n{details}")
+}
+
+#[must_use]


Suggested change

#[must_use]

ddoktorski · 2026-09-04T07:45:33Z

+        let builder = if decoder.decoding_warnings().is_empty() {
+            OutputBuilder::new()
+        } else {
+            let warning_message = format_decoding_warning(&decoder.decoding_warnings());


nit: decoder.decoding_warnings() is called twice

ddoktorski · 2026-09-04T07:56:04Z

+    sierra_abis: HashMap<ClassHash, Vec<AbiEntry>>,
+    legacy_class_hashes: HashSet<ClassHash>,
+    legacy_selectors: HashMap<(ClassHash, Felt), String>,
+    decoding_warnings: RefCell<BTreeSet<TraceDecodingWarning>>,


Can we omit using interior mutability here?

ddoktorski · 2026-09-04T07:58:47Z

+
+## `--full`
+Optional.
+Conflicts with: [`--json`](../common.md#--json--j)


Could you remind me why actually we can't use both these flags together? 🙏

ddoktorski · 2026-09-04T07:59:34Z

+    let results = stream::iter(class_hashes)
+        .map(|class_hash| async move {
+            match provider
+                .get_class(BlockId::Tag(BlockTag::PreConfirmed), *class_hash)


Why pre confirmed and not latest?

ddoktorski · 2026-09-04T08:01:09Z

+        .field("Calldata", &decoder.calldata(invocation))
+        .padded_felt_field("Caller Address", &invocation.caller_address)
+        .padded_felt_field("Class Hash", &invocation.class_hash)
+        .padded_felt_field("Contract Address", &invocation.contract_address)


Suggested change

.padded_felt_field("Contract Address", &invocation.contract_address)

.contract_address(&invocation.contract_address)

ddoktorski · 2026-09-04T08:03:06Z


 ## Overview

 Starknet Foundry `sncast` supports the inspection of transaction statuses on a given network with the `sncast get tx-status` command.


It can be updated

franciszekjob changed the title ~~WIP~~ Add sncast get tx-trace Aug 18, 2026

franciszekjob force-pushed the franciszekjob/3557-get-tx-trace branch 2 times, most recently from f44eb8d to e41c61f Compare August 19, 2026 12:46

franciszekjob commented Aug 20, 2026

View reviewed changes

franciszekjob force-pushed the franciszekjob/3557-get-tx-trace branch from 3fdf536 to bf9b114 Compare August 20, 2026 15:24

franciszekjob commented Aug 20, 2026

View reviewed changes

Comment thread docs/src/starknet/multicall.md

franciszekjob commented Aug 20, 2026

View reviewed changes

Comment thread crates/sncast/tests/e2e/get/tx_receipt.rs

franciszekjob commented Aug 20, 2026

View reviewed changes

Comment thread crates/sncast/src/starknet_commands/get/tx_trace.rs

franciszekjob marked this pull request as ready for review August 20, 2026 15:58

franciszekjob requested a review from a team as a code owner August 20, 2026 15:58

franciszekjob requested review from MKowalski8, ddoktorski and integraledelebesgue and removed request for MKowalski8 August 20, 2026 15:58

integraledelebesgue reviewed Aug 24, 2026

View reviewed changes

Comment thread crates/sncast/src/response/get/tx_trace.rs Outdated

franciszekjob changed the base branch from master to franciszekjob/global-json-flag August 25, 2026 12:44

franciszekjob force-pushed the franciszekjob/3557-get-tx-trace branch 2 times, most recently from bb2b7d8 to a3154a4 Compare August 25, 2026 12:45

franciszekjob mentioned this pull request Aug 26, 2026

Cache fetched contract classes #4571

Open

ddoktorski requested changes Aug 27, 2026

View reviewed changes

franciszekjob force-pushed the franciszekjob/3557-get-tx-trace branch from 00e1a51 to ff96770 Compare August 27, 2026 11:05

Base automatically changed from franciszekjob/global-json-flag to master August 27, 2026 13:25

franciszekjob force-pushed the franciszekjob/3557-get-tx-trace branch 3 times, most recently from c6d690a to 7bbf106 Compare September 1, 2026 09:06

franciszekjob added 2 commits September 1, 2026 12:25

WIP

628fc2e

Make self review

a706356

franciszekjob added 25 commits September 1, 2026 12:26

Update asserted output in test_get_tx_receipt_reverted

fa3145f

Update multicall template output formatting

6d25694

Change MAX_CONCURRENT_CLASS_REQUESTS to 4

c194c09

Little refactor in TraceDecoder::new()

bb0eee3

Rename r -> trace param

bd86add

Update changelog

8e50f3b

Make --full conflict with --json

c54f64d

Simplify TransactionTraceResponse::text()

24bd9c5

Use padded felt field for class hashes and addresses

06bae77

Rename ContractClassesFetchResponse -> FetchedContractClasses

d325363

Simplify contract class fetching for transaction traces

b5a7753

Simplify required invocation JSON decoding

b333d37

Error on missing required trace fields

8299704

Move trace class fetching logic to command module

e5cc8c6

Display parent invocation data before nested calls

e04da86

Validate --full conflict with global --json at runtime

d6919af

Add detailed trace decoding warnings to text and JSON output

8b16679

Fix docs snippets test

2138784

Rename mock tx hash

7fb3484

Add test_full_trace_with_state_diff

f63df3d

Remove provider_error_message

c455979

Fix tests

63fef69

Rename issues -> warnings

a401305

Update starknet-rust sha

e67460f

Update Cargo.lock

295bedc

franciszekjob force-pushed the franciszekjob/3557-get-tx-trace branch from 62ead40 to 295bedc Compare September 1, 2026 10:37

franciszekjob requested review from ddoktorski and integraledelebesgue September 1, 2026 12:44

integraledelebesgue approved these changes Sep 2, 2026

View reviewed changes

ddoktorski reviewed Sep 4, 2026

View reviewed changes

	.padded_felt_field("Contract Address", &invocation.contract_address)
	.contract_address(&invocation.contract_address)


		## Overview

		Starknet Foundry `sncast` supports the inspection of transaction statuses on a given network with the `sncast get tx-status` command.

Uh oh!

Conversation

franciszekjob commented Aug 18, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Introduced changes

Checklist

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

franciszekjob Aug 31, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

franciszekjob commented Aug 18, 2026 •

edited

Loading

franciszekjob Aug 31, 2026 •

edited

Loading