Fix CI

[h-vetinari]

Tests are currently failing on all platforms; on unix we get

[ RUN      ] TestAzureFileSystemOnAllScenarios/1.GetFileInfoObjectWithNestedStructure
unknown file: Failure
C++ exception with description "Fail to get a new connection for: http://127.0.0.1:10000. Could not connect to server" thrown in the test body.

whereas on windows the errors are predominantly due to missing a tzdb. Windows has apparently been failing for longer already, but this had been hidden behind an expansion bug. We need something like apache/arrow#48874 here.

[conda-forge-admin]

Hi! This is the friendly automated conda-forge-linting service.

I just wanted to let you know that I linted all conda-recipes in your PR (recipe/meta.yaml) and found it was in an excellent condition.

I do have some suggestions for making it better though...

For recipe/meta.yaml:

• ℹ️ The recipe is not parsable by parser conda-souschef (grayskull). This parser is not currently used by conda-forge, but may be in the future. We are collecting information to see which recipes are compatible with grayskull.
• ℹ️ The recipe is not parsable by parser conda-recipe-manager. The recipe can only be automatically migrated to the new v1 format if it is parseable by conda-recipe-manager.

_{This message was generated by GitHub Actions workflow run https://github.com/conda-forge/conda-forge-webservices/actions/runs/21290798182. Examine the logs at this URL for more detail.}

[raulcd]

The linux/macOS failures are related to Azure, could this be a problem with Azurite as in:

• [Python] Conda jobs segfault on Azure tests due to old nodejs version being installed apache/arrow#48894

I fixed this by pinning nodejs>=16
apache/arrow#48895

The TZ failures are a whole different beast :)

[h-vetinari]

Unfortunately, backporting apache/arrow#48895 has not fixed the failures here (nor in previous versions; see #1906 #1907 #1908)

[raulcd]

For the Linux and macOS failures, this is installing a really old nodejs version:

2026-01-20T21:50:29.1691560Z     nodejs:                           12.4.0-he1b5a44_0              conda-forge

This fails to start azurite:

2026-01-20T23:35:32.7060608Z $PREFIX/lib/node_modules/azurite/node_modules/fs-extra/lib/util/async.js:14
2026-01-20T23:35:32.7060979Z         (err) => err ?? new Error('unknown error')
2026-01-20T23:35:32.7061274Z                       ^
2026-01-20T23:35:32.7061394Z 
2026-01-20T23:35:32.7076632Z SyntaxError: Unexpected token ?
2026-01-20T23:35:32.7077725Z     at Module._compile (internal/modules/cjs/loader.js:718:23)
2026-01-20T23:35:32.7079514Z     at Object.Module._extensions..js (internal/modules/cjs/loader.js:785:10)
2026-01-20T23:35:32.7080000Z     at Module.load (internal/modules/cjs/loader.js:641:32)
2026-01-20T23:35:32.7080963Z     at Function.Module._load (internal/modules/cjs/loader.js:556:12)
2026-01-20T23:35:32.7081404Z     at Module.require (internal/modules/cjs/loader.js:681:19)
2026-01-20T23:35:32.7081799Z     at require (internal/modules/cjs/helpers.js:16:16)
2026-01-20T23:35:32.7082236Z     at Object.<anonymous> ($PREFIX/lib/node_modules/azurite/node_modules/fs-extra/lib/copy/copy.js:9:44)
2026-01-20T23:35:32.7083263Z     at Module._compile (internal/modules/cjs/loader.js:774:30)
2026-01-20T23:35:32.7088849Z     at Object.Module._extensions..js (internal/modules/cjs/loader.js:785:10)
2026-01-20T23:35:32.7089519Z     at Module.load (internal/modules/cjs/loader.js:641:32)

which requires minimum nodejs 16.

I am unsure why the resolver is trying to install a version that old, can we update the version here?

arrow-cpp-feedstock/recipe/meta.yaml

Lines 109 to 110 in d0a673f

	# for npm, to install azurite
	- nodejs *

[h-vetinari]

For the Linux and macOS failures, this is installing a really old nodejs version:

OK, I hadn't actually looked at that, because I had been confident that the floating pin would pull in a new (or at least new-ish) version. Forcing this to a newer nodejs (+ the backport of apache/arrow#48895, presumably) does fix the issues on linux/osx 🥳

So that leaves only apache/arrow#48874 to deal with

[h-vetinari]

I had been confident that the floating pin would pull in a new (or at least new-ish) version.

Related (at least) to conda-forge/nodejs-feedstock#436 not being merged.

[conda-forge-admin]

Hi! This is the friendly conda-forge automerge bot!

I considered the following status checks when analyzing this PR:

• linter: passed
• azure: failed

Thus the PR was not passing and not merged.