vite-node sourcemap changes not reflected after HMR

[theoephraim]

Describe the bug

I am using vite-node directly and while sourcemaps appear to be generated correctly, and error line numbers are correct for an initial execution, after HMR and clearing the module from the cache, line numbers are no longer correct. While the file is being updated and the new code is executing, it seems like the original sourcemap is somehow sticking around?

Is there another method that must be called to make sure the new sourcemap is respected?

Reproduction

https://stackblitz.com/~/github.com/dmno-dev/vite-node-sourcemap-issue

System Info

System:
    OS: macOS 13.6.7
    CPU: (12) arm64 Apple M2 Max
    Memory: 3.00 GB / 96.00 GB
    Shell: 5.9 - /bin/zsh
  Binaries:
    Node: 20.11.1 - ~/Library/Caches/fnm_multishells/32938_1728091914366/bin/node
    Yarn: 1.22.22 - ~/Library/Caches/fnm_multishells/32938_1728091914366/bin/yarn
    npm: 10.2.4 - ~/Library/Caches/fnm_multishells/32938_1728091914366/bin/npm
    pnpm: 9.6.0 - ~/Library/Caches/fnm_multishells/32938_1728091914366/bin/pnpm
    bun: 1.1.21 - /opt/homebrew/bin/bun
  Browsers:
    Chrome: 129.0.6668.90
    Safari: 16.6
  npmPackages:
    vite: ^5.4.8 => 5.4.8

Used Package Manager

pnpm

Validations

• Follow our Code of Conduct
• Read the Contributing Guidelines.
• Read the docs.
• Check that there isn't already an issue that reports the same bug to avoid creating a duplicate.
• Check that this is a concrete bug. For Q&A open a GitHub Discussion or join our Discord Chat Server.
• The provided reproduction is a minimal reproducible example of the bug.

[hi-ogawa]

Can you reproduce the same with vite-node cli watch mode?

[hi-ogawa]

I wasn't sure about your custom handleHotUpdate, but I can reproduce with vite-node --watch https://github.com/hi-ogawa/reproductions/tree/main/vitest-6642-vite-node-watch-error-stack and I assume it's the same issue.

[theoephraim]

Nice! At least it's consistent. I also noticed that adding a log on the getSourceMap function, it only gets called once.

this seems to be the related PR

Perhaps @sheremet-va has an idea of what's going on?