Update CI workflows to latest 0.0.17 tag by finagolfin · Pull Request #1989 · swiftlang/swift-testing

finagolfin · 2026-10-27T16:58:16Z

Remove local config changes that aren't needed and push Android emulator testing down to API 26.

finagolfin · 2026-10-27T17:48:48Z

OK, this works, with these Android and Windows builds already broken before and the static linux SDK build never enabled before. Just let me know if I should remove that one and we can get this in, with the Android issue to be fixed upstream in the workflow.

finagolfin · 2026-10-27T18:57:46Z

Ready to go: all three failures were already broken before this pull, with the two Android issues to be fixed in other repos and the Windows error likely to be fixed in the next trunk snapshot, as it is currently lagging behind at Sep. 21.

grynspan · 2026-10-29T00:32:00Z

Need to wait for swiftlang/swift-docker#609 in order to get Android passing (yes?)

finagolfin · 2026-10-29T02:12:06Z

Need to wait for swiftlang/swift-docker#609 in order to get Android passing (yes?)

Yes, but since that and the Ubuntu 26.04 update issue linked above already broke Android testing, and this pull can't fix it, I don't think that should hold back getting this pull in.

Remove local config changes that aren't needed and push Android emulator testing down to API 26.

grynspan · 2026-10-29T14:10:56Z

Need to wait for swiftlang/swift-docker#609 in order to get Android passing (yes?)

Yes, but since that and the Ubuntu 26.04 update issue linked above already broke Android testing, and this pull can't fix it, I don't think that should hold back getting this pull in.

I know I'm being pedantic, but how do we verify this PR is correct if the jobs it modifies are failing? Hence why we probably need to wait.

grynspan · 2026-10-29T14:12:58Z

And now Embedded Swift/Wasm is failing due to a linker failure on swift_dynamicCastTest(), which I assume is a compiler regression.

finagolfin · 2026-10-29T14:55:14Z

how do we verify this PR is correct if the jobs it modifies are failing?

Simple, all four failures- at least I assume wasm does also, as that trunk snapshot SDK was just tagged- are failing in the exact same way for the reasons given above for all pulls already, so we know this workflow update does not cause those problems.

Since this pull didn't break those, I say we get this in for the other platforms where it likely simply improves matters under the hood, and we can fix those few failures in subsequent pulls, like the Docker pull you linked.

grynspan · 2026-10-29T15:27:15Z

#1993 works around the Embedded Swift regression.

grynspan · 2026-10-29T15:28:23Z

Since this pull didn't break those, I say we get this in for the other platforms where it likely simply improves matters under the hood, and we can fix those few failures in subsequent pulls, like the Docker pull you linked.

I get it, and I'm not saying this PR caused the breakage, but how do we verify that this PR works if the existing failures interfere with the CI jobs that this PR modifies?

Is there any reason to move quickly here? Can you quantify the harm in waiting for swift-docker#609 to land?

grynspan · 2026-10-29T15:29:20Z

      freebsd_os_versions: '["14.3"]'
      enable_linux_checks: false
      enable_windows_checks: false
-      enable_macos_checks: false


As with the Linux change, can we leave these alone? We're explicitly stating that the FreeBSD build does not trigger these platforms.

Up to you but these are all disabled by default, unlike linux and Windows.

They are now, but if we ever change that, the FreeBSD build will start building those platforms.

Very unlikely, but I can put them back if you're concerned about that.

finagolfin · 2026-10-29T16:12:38Z

how do we verify that this PR works if the existing failures interfere with the CI jobs that this PR modifies?

Well, we know this update didn't break any of the other platforms, while we pretty much know the reasons these few failures happen, which don't happen on other repos for eg the trunk Android SDK.

Do you plan on stopping merging all Testing pulls till these CI issues are fixed?

Is there any reason to move quickly here?

None, other than this workflow update didn't break anything.

Can you quantify the harm in waiting for swift-docker#609 to land?

No harm, I just don't see these few CI failures that we understand well as a reason to wait on those Docker pulls and so on to fix them, which may take awhile.

grynspan · 2026-10-30T13:31:15Z

how do we verify that this PR works if the existing failures interfere with the CI jobs that this PR modifies?

Well, we know this update didn't break any of the other platforms, while we pretty much know the reasons these few failures happen, which don't happen on other repos for eg the trunk Android SDK.

Do you plan on stopping merging all Testing pulls till these CI issues are fixed?

No, but in general those PRs aren't specific to one platform while that one platform is failing in CI. We are holding up Windows-specific PRs from @compnerd right now because Windows CI is failing.

Is there any reason to move quickly here?

None, other than this workflow update didn't break anything.

I'm not saying it did, but we don't have any passing Android builds so how can we confirm?

Can you quantify the harm in waiting for swift-docker#609 to land?

No harm, I just don't see these few CI failures that we understand well as a reason to wait on those Docker pulls and so on to fix them, which may take awhile.

We can escalate the docker PR because it's blocking Swift Testing's CI.

grynspan · 2026-10-30T16:01:22Z

~~The Docker change has been merged. Is some other step required to get that fix in place? New toolchain?~~

Duh new toolchain.

finagolfin · 2026-10-30T17:02:10Z

this workflow update didn't break anything.

we don't have any passing Android builds so how can we confirm?

I simply meant that Android CI were already broken before this pull, so this pull didn't break them. 😉 If you mean that we can't be sure this pull won't break Android in some other additional way, that's unlikely because this workflow tag I'm pulling in is working well for other repos.

finagolfin · 2026-10-31T12:48:26Z

Downloaded the first passing trunk Android SDK bundle on the CI since the zlib pull was merged and checked its AArch64 libraries to make sure those zlib symbols are no longer there: they're gone. So as those Android bundles start getting tagged, these tests will start passing on the Android emulator again, now that all of them have finally been enabled by @niktishakov's workflow fix. 👍

finagolfin requested review from briancroom, grynspan, harlanhaskins, jerryjrchen and stmontgomery as code owners October 27, 2026 16:58

finagolfin mentioned this pull request Oct 27, 2026

Bump the swiftlang-actions group with 3 updates #1986

Closed

finagolfin commented Oct 27, 2026

View reviewed changes

Comment thread .github/workflows/pull_request.yml Resolved

finagolfin commented Oct 27, 2026

View reviewed changes

Comment thread .github/workflows/pull_request.yml Resolved

grynspan requested changes Oct 27, 2026

View reviewed changes

Comment thread .github/workflows/main_using_main.yml Resolved

finagolfin force-pushed the ci branch from 064c5c1 to c2601e9 Compare October 27, 2026 18:20

finagolfin requested a review from grynspan October 27, 2026 19:17

grynspan added linux 🐧 Linux support (all distros) android 🤖 Android support github_actions Pull requests that update GitHub Actions code labels Oct 29, 2026

grynspan added this to the Swift 6.5.0 (main) milestone Oct 29, 2026

grynspan assigned finagolfin Oct 29, 2026

finagolfin force-pushed the ci branch from c2601e9 to ebe62c0 Compare October 29, 2026 02:03

finagolfin commented Oct 29, 2026

View reviewed changes

Comment thread .github/workflows/main_using_release.yml Resolved

Update CI workflows to latest 0.0.17 tag

6b1eadb

Remove local config changes that aren't needed and push Android emulator testing down to API 26.

finagolfin force-pushed the ci branch from ebe62c0 to 6b1eadb Compare October 29, 2026 14:02

grynspan reviewed Oct 29, 2026

View reviewed changes

Merge branch 'main' into ci

4e16224

Merge branch 'main' into ci

5d668d9

Merge branch 'main' into ci

a648981

Conversation

finagolfin commented Oct 27, 2026

Uh oh!

Uh oh!

Uh oh!

finagolfin commented Oct 27, 2026

Uh oh!

Uh oh!

finagolfin commented Oct 27, 2026

Uh oh!

grynspan commented Oct 29, 2026

Uh oh!

Uh oh!

finagolfin commented Oct 29, 2026

Uh oh!

grynspan commented Oct 29, 2026

Uh oh!

grynspan commented Oct 29, 2026

Uh oh!

finagolfin commented Oct 29, 2026

Uh oh!

grynspan commented Oct 29, 2026

Uh oh!

grynspan commented Oct 29, 2026

Uh oh!

grynspan Oct 29, 2026

Choose a reason for hiding this comment

Uh oh!

finagolfin Oct 29, 2026

Choose a reason for hiding this comment

Uh oh!

grynspan Oct 30, 2026

Choose a reason for hiding this comment

Uh oh!

finagolfin Oct 30, 2026

Choose a reason for hiding this comment

Uh oh!

finagolfin commented Oct 29, 2026

Uh oh!

grynspan commented Oct 30, 2026

Uh oh!

grynspan commented Oct 30, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

finagolfin commented Oct 30, 2026

Uh oh!

finagolfin commented Oct 31, 2026

Uh oh!

Reviewers

Uh oh!

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

grynspan commented Oct 30, 2026 •

edited

Loading