handling of "evolving code" in get_msgs_until

[MonkeyOliver]

🐛 Bug Description

get_msgs_until in rdagent/log/ui/app.py does not change state.msgs[state.round]["evolving code"], resulting in the evolving code list always being empty in the UI.

To Reproduce

Steps to reproduce the behavior:

1. rdagent ui --port 19899 --log_dir "log path"
2. Check All Loops in a result

Expected Behavior

I was able to get it to display correctly by adding the following logic to get_msgs_until:

state.msgs[state.lround][msg.tag].append(msg)
# Handle evolving code (line 248)
if "evolving code" in tags:
    state.msgs[state.lround]["evolving code"].append(msg)

However, this addition affects other parts of the code. I commented out some potentially impacted code to make it work.

Screenshot

After some coding:

Environment

Note: Users can run rdagent collect_info to get system information and paste it directly here.

• Name of current operating system: Darwin
• Processor architecture: arm64
• System, version, and hardware information: macOS-15.5-arm64-arm-64bit
• Version number of the system: Darwin Kernel Version 24.5.0: Tue Apr 22 19:54:33 PDT 2025; root:xnu-11417.121.6~2/RELEASE_ARM64_T8122
• Python version: 3.10.18 (main, Jun 5 2025, 08:37:47) [Clang 14.0.6 ]
• Container ID: 9b88717505bfe8ba3a7e89d82c0d82f49b129c68aca2a8ed1478ba2c72f1cf2f
• Container Name: elastic_khorana
• Container Status: created
• Image ID used by the container: sha256:c4a64e22e380a82f2138ccd8c2207c9e44bf0e9b4ba7d787fa0d70f14239a2ac
• Image tag used by the container: ['local_qlib:latest']
• Container port mapping: {}
• Container Label: {'com.nvidia.volumes.needed': 'nvidia_driver', 'org.opencontainers.image.ref.name': 'ubuntu', 'org.opencontainers.image.version': '22.04'}
• Startup Commands: nvidia-smi
• RD-Agent version: 0.6.1

[SunsetWolf]

Hi, @MonkeyOliver , Thanks for the detailed report!

Could you try upgrading to the latest RD-Agent version (or pull the newest code from the repo) and see if the problem still occurs?

If it does, feel free to open a new issue with the updated logs — we’d be happy to help investigate further!