# EPF6 - Week 21

## Overview
- Attended the Lodestar Weekly Standup and [shared updates](https://github.com/ChainSafe/lodestar/wiki/Lodestar-Weekly-Standups-2025#kurtosis-migration-and-testing)
- Focused on **debugging inconsistencies** between the table renderer outputs from the **Kurtosis-based simulation** and the **original Lodestar testing framework**.
- Discovered that the **Kurtosis migration branch** on GitHub was slightly behind the Lodestar **reference branch** `unstable`, missing key updates recently introduced upstream.
- Updated the network configuration in multiFork.test.ts so that all forks now start from slot zero, activating all fork rules from genesis.
- After applying these updates, the simulation output no longer signals errors, and the table renderer output now closely matches Lodestar’s original testing framework.

## Main points

### 1. Table Renderer Output Debugging

This week’s primary goal was to identify why the **Kurtosis simulation’s table renderer output** differed from the results generated by the **regular Lodestar simulation**.

Upon inspection, several inconsistencies were traced back to outdated logic in the local Kurtosis branch, which was missing recent updates from the reference branch used by my mentor.

**Findings:**
- The local Kurtosis branch was **behind the reference branch**, lacking the latest network configuration updates and minor bug fixes.
- The missing commits affected how fork activation epochs were parsed and displayed in the final table report.

> [!TIP] Discovery
> <font color="#27F535">After syncing with the **updated reference branch**, the output discrepancies were reduced significantly.</font>

### 2. Network Configuration Update

The network configuration in `multiFork.test.ts` was updated from the previous fork layout to a version where **all forks start from slot zero**, meaning all protocol upgrades (**Altair**, **Bellatrix**, **Capella**, **Deneb**, and **Electra**) are active right from **genesis**:

- network_params:
-   altair_fork_epoch: 2
-   bellatrix_fork_epoch: 4
-   capella_fork_epoch: 6
-   deneb_fork_epoch: 8
-   electra_fork_epoch: 10

+ network_params:
+   altair_fork_epoch: 0
+   bellatrix_fork_epoch: 0
+   capella_fork_epoch: 0
+   deneb_fork_epoch: 0
+   electra_fork_epoch: 0

### Results and Current Observations

After these updates, the output from the **Kurtosis-based simulation** shows a noticeable improvement in alignment with the **Lodestar testing framework’s table renderer**, though results are still being evaluated.

:::warning
:warning: **Note**

The simulation is currently running with the three synchronization assertions commented out: <code>assertRangeSync(env)</code>, <code>assertCheckpointSync(env)</code>, and <code>assertUnknownBlockSync(env)</code>.
These assertions will be revisited and addressed as next steps after the fellowship.
:::

After these updates, the output from the Kurtosis-based simulation appears **stable** and **visually consistent** with the Lodestar testing framework’s table renderer.

**Lodestar testing framework output:**

**Lodestar with Kurtosis output:**

**Preliminary Observations:**
- The final simulation summary no longer reports **immediate** errors.
- The table renderer output appears **structurally similar** to Lodestar’s native simulation output.
- The simulation now runs through post-Electra behavior **without obvious inconsistencies**, but **additional tests** are required to confirm full parity.

> The hope is that this improvement could represent a valuable step for the team and potentially serve as an initial entry point for an eventual integration of Kurtosis into Lodestar’s testing framework!

## Key Learnings

| Area | Insight |
|------|----------|
| **Branch Alignment** |Keeping the Kurtosis branch synced with Lodestar’s main branch is essential to avoid outdated network parameters and misaligned results.|
| **Network Configuration** | Starting all forks from slot zero activates all upgrades at genesis, providing faster validation of post-Electra behavior.|
| **Output Consistency** | Once aligned, the Kurtosis simulation produces table renderer outputs similar to Lodestar’s framework|
|**Debugging Workflow**|Incremental debugging across frameworks helps isolate branch drift issues and improve reliability.|

## Final EPF week TODOs

As this marks the final week of the fellowship, the focus shifts from **implementation** to the **final communication**.

The remaining time is dedicated to preparing the final presentation and **consolidating findings**.

The goal is to highlight the technical insights gained throughout the process, from investigating Lodestar's codebase to the development of Kurtosis-based setup and identifying possible solid pathways for future integration and improvement.

This wrap-up phase is also about defining which outcomes hold **the greatest value for Lodestar’s team**.

It has been an amazing experience, full of learning, collaboration, and problem-solving. I’m deeply grateful for this opportunity and for the chance to work alongside such brilliant minds!

Next step, **Devconnect 2025 and EPF day in Buenos Aires**!