GH-49995: [R][Wasm] Fix threading issue under Wasm/Emscripten by amoeba · Pull Request #49996 · apache/arrow

amoeba · 2026-05-19T21:17:17Z

Rationale for this change

#49982 added a basic Wasm build but after doing some testing I ran into an issue with threading.

What changes are included in this PR?

The main fix: Modified CanRunWithCapturedR so it's always false under Emscripten. This change is the only change that was strictly needed to fix my issue. I'm not a pro with this part of the codebase so scrutiny welcome.
I also set options(arrow.use_threads = FALSE) when running under Wasm just in case since it makes sense to not use threads in a Wasm context

Are these changes tested?

Yes. Crossbow jobs show the build works and I have a demo at https://amoeba.github.io/arrow-r-wasm/.

Are there any user-facing changes?

No.

GitHub Issue: [R][Wasm] Fix Error: thread constructor failed: Not supported under Wasm #49995

amoeba · 2026-05-19T21:17:33Z

@github-actions crossbow submit test-r-wasm

github-actions · 2026-05-19T21:19:52Z

Revision: dadf0f5

Submitted crossbow builds: ursacomputing/crossbow @ actions-9b0263bfa9

Task	Status
test-r-wasm

amoeba · 2026-05-19T21:45:44Z

@github-actions crossbow submit test-r-wasm

amoeba · 2026-05-19T21:46:30Z

With dadf0f5, Parquet reading works with my local test. I'm testing the changes separately so I know what's actually fixing things here.

github-actions · 2026-05-19T21:47:55Z

Revision: abaacb8

Submitted crossbow builds: ursacomputing/crossbow @ actions-510511a107

Task	Status
test-r-wasm

amoeba · 2026-05-20T03:57:29Z

I marked this as ready for review. The PR description hopefully explains all that's needed and I put together a demo at https://amoeba.github.io/arrow-r-wasm/

jonkeane · 2026-05-20T12:31:01Z

+bool CanRunWithCapturedR() {
+#ifdef __EMSCRIPTEN__
+  // Threading is not supported under Emscripten/WASM. Always take the
+  // synchronous path to avoid attempting pthread_create which will fail
+  // with "thread constructor failed: Not supported".
+  return false;
+#else
+  return MainRThread::GetInstance().Executor() == nullptr;
+#endif


Is it possible to add a test (or maybe it's that we actually need to run the tests under wasm(???) to catch this?

It looks like maybe we can at least run some R code in the container if not run the test suite. I'll report back.

amoeba · 2026-05-20T19:40:10Z

@github-actions crossbow submit test-r-wasm

amoeba · 2026-05-20T19:41:21Z

I'm testing trying to install, load the package, and run the test suite in CI here. I tried testing locally but I'm running into what appears to be some qemu errors since I'm running Docker on Mac so we'll see how this runs in CI.

github-actions · 2026-05-20T19:42:41Z

Revision: 19efaf3

Submitted crossbow builds: ursacomputing/crossbow @ actions-6a9811d919

Task	Status
test-r-wasm

amoeba · 2026-05-20T23:28:29Z

@github-actions crossbow submit test-r-wasm

github-actions · 2026-05-20T23:30:44Z

Revision: 47738fa

Submitted crossbow builds: ursacomputing/crossbow @ actions-8e088c6ac8

Task	Status
test-r-wasm

amoeba · 2026-05-20T23:58:48Z

The job runs but we don't actually run any tests,

Running testthat suite under webR...
No installed testthat tests found for arrow.
Results: 0 passed, 0 skipped, 0 failed, 0 errors
✓ testthat suite passed
✓ All tests passed!

More soon!

amoeba · 2026-05-21T03:33:26Z

@github-actions crossbow submit test-r-wasm

github-actions · 2026-05-21T03:36:06Z

Revision: 9f03a2a

Submitted crossbow builds: ursacomputing/crossbow @ actions-d71cf926ef

Task	Status
test-r-wasm

amoeba · 2026-05-21T03:42:04Z

I couldn't find any examples of people running their test suites under Wasm but I'm going to see if I can't get it working in 9f03a2a. If that doesn't work easily, I think it'd be good enough to just write a simple test script to run instead.

amoeba · 2026-05-21T03:55:02Z

That went shockingly well,

Results: 6376 passed, 92 skipped, 1 failed, 3 errors

Here are the errors/failures:

── 1. Error ('test-dataset-dplyr.R:76:3'): filter() on timestamp columns ───────
Error in `compute.arrow_dplyr_query(x)`: Invalid: Cannot locate or parse timezone 'UTC': discover_tz_dir failed to find zoneinfo

── 2. Error ('test-dataset-dplyr.R:124:3'): filter() on date32 columns ─────────
Error in `compute.arrow_dplyr_query(x)`: Invalid: Cannot locate or parse timezone 'UTC': discover_tz_dir failed to find zoneinfo

── 3. Failure ('test-dplyr-filter.R:425:3'): filter() with across() ────────────
Expected `via_table <- rlang::eval_tidy(expr, rlang::new_data_mask(rlang::env(.input = arrow_table(tbl))))` not to throw any warnings.

── 4. Error ('test-dplyr-query.R:650:3'): Scalars in expressions match the type 
Error: Error: NotImplemented: Function 'greater' has no kernel matching input types (timestamp[us, tz=UTC], string)

amoeba · 2026-05-21T04:01:16Z

@github-actions crossbow submit test-r-wasm

github-actions · 2026-05-21T04:04:06Z

Revision: 80a2fb1

Submitted crossbow builds: ursacomputing/crossbow @ actions-a3bfe97816

Task	Status
test-r-wasm

amoeba · 2026-05-21T15:00:54Z

@github-actions crossbow submit test-r-wasm

github-actions · 2026-05-21T15:07:05Z

Revision: b8810d9

Submitted crossbow builds: ursacomputing/crossbow @ actions-1810f2dac4

Task	Status
test-r-wasm

Try to fix issue

dadf0f5

github-actions Bot added awaiting review Awaiting review Component: R labels May 19, 2026

test not setting usethreads

abaacb8

set back

bafaf05

amoeba marked this pull request as ready for review May 20, 2026 03:56

amoeba requested review from jonkeane and thisisnic as code owners May 20, 2026 03:56

amoeba changed the title ~~GH-49995: [R][Wasm] Fix threading issue under Wasm/Emscritpen~~ GH-49995: [R][Wasm] Fix threading issue under Wasm/Emscripten May 20, 2026

jonkeane reviewed May 20, 2026

View reviewed changes

github-actions Bot added awaiting changes Awaiting changes and removed awaiting review Awaiting review labels May 20, 2026

wip, run tests under wasm

19efaf3

amoeba requested review from assignUser, kou and raulcd as code owners May 20, 2026 19:39

github-actions Bot added awaiting change review Awaiting change review and removed awaiting changes Awaiting changes labels May 20, 2026

set NODE_PATH

47738fa

try this for tests

9f03a2a

tidy up a bit

dd555bc

configure tzdb

80a2fb1

amoeba added 2 commits May 20, 2026 21:05

add skip

dee814c

add debug info for ci

b8810d9

Conversation

amoeba commented May 19, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Rationale for this change

What changes are included in this PR?

Are these changes tested?

Are there any user-facing changes?

Uh oh!

amoeba commented May 19, 2026

Uh oh!

github-actions Bot commented May 19, 2026

Uh oh!

amoeba commented May 19, 2026

Uh oh!

amoeba commented May 19, 2026

Uh oh!

github-actions Bot commented May 19, 2026

Uh oh!

amoeba commented May 20, 2026

Uh oh!

jonkeane May 20, 2026

Choose a reason for hiding this comment

Uh oh!

amoeba May 20, 2026

Choose a reason for hiding this comment

Uh oh!

amoeba commented May 20, 2026

Uh oh!

amoeba commented May 20, 2026

Uh oh!

github-actions Bot commented May 20, 2026

Uh oh!

amoeba commented May 20, 2026

Uh oh!

github-actions Bot commented May 20, 2026

Uh oh!

amoeba commented May 20, 2026

Uh oh!

amoeba commented May 21, 2026

Uh oh!

github-actions Bot commented May 21, 2026

Uh oh!

amoeba commented May 21, 2026

Uh oh!

amoeba commented May 21, 2026

Uh oh!

amoeba commented May 21, 2026

Uh oh!

github-actions Bot commented May 21, 2026

Uh oh!

amoeba commented May 21, 2026

Uh oh!

github-actions Bot commented May 21, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

amoeba commented May 19, 2026 •

edited

Loading