The latest updates on your projects. Learn more about Vercel for GitHub.

Project	Deployment	Preview	Comments	Updated (UTC)
rollup	Ready	Preview	Comment	Sep 21, 2025 6:09pm

Thank you for your contribution! ❤️

You can try out this pull request locally by installing Rollup via

npm install rollup/rollup#improve-tla-deadlock-detection

Notice: Ensure you have installed the latest nightly Rust toolchain. If you haven't installed it yet, please see https://www.rust-lang.org/tools/install to learn how to download Rustup and install Rust.

or load it into the REPL:
https://rollup-f3tn119vl-rollup-js.vercel.app/repl/?pr=6121

Performance report

• BUILD: 7136ms, 815 MB
  • initialize: 0ms, 25 MB
  • generate module graph: 2660ms, 628 MB
    • generate ast: 1388ms (-46ms, -3.2%), 622 MB
  • sort and bind modules: 403ms, 679 MB
  • mark included statements: 4067ms, 815 MB
    • treeshaking pass 1: 2401ms, 814 MB
    • treeshaking pass 2: 471ms, 836 MB
    • treeshaking pass 3: 404ms, 817 MB
    • treeshaking pass 4: 390ms, 841 MB
    • treeshaking pass 5: 388ms, 815 MB
• GENERATE: 1028ms, 914 MB
  • initialize render: 0ms, 938 MB (+3%)
  • generate chunks: 256ms (+20ms, +8.5%), 890 MB (+7%)
    • optimize chunks: 0ms, 921 MB (+12%)
  • render chunks: 741ms, 900 MB
  • transform chunks: 20ms, 914 MB
  • generate bundle: 0ms, 914 MB

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 98.73%. Comparing base (2029f63) to head (5c0adf7).
⚠️ Report is 1 commits behind head on master.

@@            Coverage Diff             @@
##           master    #6121      +/-   ##
==========================================
- Coverage   98.74%   98.73%   -0.01%     
==========================================
  Files         271      271              
  Lines       10644    10630      -14     
  Branches     2850     2846       -4     
==========================================
- Hits        10510    10496      -14     
  Misses         89       89              
  Partials       45       45              

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

Looks good to me! That’s a really nice and simple approach. Although other non-TLA dynamic imports will also be considered as TLA, it completely eliminates this kind of issue.

I mean the alternative would be to either force comments upon people who want stuff to "just work", or to lift analysis on a completely different level. My hope is that top-level await is not that common anyway. And in the worst case, it will just be like just turning off the dynamic import optimization, i.e. it will be like most other bundlers.

This PR has been released as part of [email protected]. You can test it via npm install rollup.