InstantiationProgressHandler reports incorrect bytesTotal

[jr200]

What happens?

Using react to instantiate duckdb, I pass in a basic InstantiationProgressHandler that logs to console. The output is below.

{startedAt: Mon Aug 04 2025 07:51:54 GMT+0100 (British Summer Time), updatedAt: Mon Aug 04 2025 07:51:54 GMT+0100 (British Summer Time), bytesTotal: 6285550, bytesLoaded: 9314009}
{startedAt: Mon Aug 04 2025 07:51:54 GMT+0100 (British Summer Time), updatedAt: Mon Aug 04 2025 07:51:54 GMT+0100 (British Summer Time), bytesTotal: 6285550, bytesLoaded: 9325068}


....


{startedAt: Mon Aug 04 2025 07:51:54 GMT+0100 (British Summer Time), updatedAt: Mon Aug 04 2025 07:51:54 GMT+0100 (British Summer Time), bytesTotal: 6285550, bytesLoaded: 32599231}
{startedAt: Mon Aug 04 2025 07:51:54 GMT+0100 (British Summer Time), updatedAt: Mon Aug 04 2025 07:51:54 GMT+0100 (British Summer Time), bytesTotal: 6285550, bytesLoaded: 32628187}

Should bytesLoaded be greater than bytesTotal? It's a small bug, but it means I can't display a trustworthy progress bar.

To Reproduce

I think I'm using standard instantiation code:

  const bundles = getJsDelivrBundles()
  const bundle = await selectBundle(bundles)

  const progressHandler = (progress: InstantiationProgress) => {
    console.log(progress)
  }

  const workerUrl = URL.createObjectURL(
    new Blob([`importScripts("${bundle.mainWorker!}");`], {
      type: 'text/javascript',
    })
  )

  const worker = new Worker(workerUrl)
  const db = new AsyncDuckDB(new ConsoleLogger(LogLevel.DEBUG), worker)
  await db.instantiate(bundle.mainModule, bundle.pthreadWorker, progressHandler)
  URL.revokeObjectURL(workerUrl)

Browser/Environment:

Vite and NextJS on Chrome

Device:

macbook pro laptop

DuckDB-Wasm Version:

1.29.1-dev132.0

DuckDB-Wasm Deployment:

local machine

Full Name:

Jayshan Raghunandan

Affiliation:

None

[vincent-octo]

Hi, could it be that bytesTotal is based on the gzip-compressed size of the WASM file, and bytesLoaded on the uncompressed size? At least it matches the sizes in my test. I would still consider your use case valid, so it would be good if bytesTotal and bytesLoaded where counting the same thing.

I am a new user of DuckDB-WASM so unfortunately don't know that much to propose a solution.

I saw that there is also progressUpdate in the API, though not sure how to use it.