ci: use windows 2025 runners #7048

ShortDevelopment · 2025-12-03T16:58:04Z

windows-2019 actions are unsupported as of 30.06.2025.

This PR updates the ci-config to run on the supported versions of the windows runner:

windows-2022 → windows-2025
windows-2019 → windows-2022

💥 Breaking changes

The default CC configuration on windows used the WinGlob api.
This config is broken on recent versions of Windows 11 as CC does not initialize WinRT using RoInitialize to allow for some optimizations.
CC now uses the version of libICU that is shipped with Windows by default
This removes compat of the default config with all windows versions < Version 1903 (May 2019 Update)
Should be fine as Windows 10 is out-of-support anyway...

🧪 Tests

Tests did assume win32 => WinGlob; that is obviously wrong now.
Some baselines were WinGlob only => updated
libICU did return incorrect values so two tests were disabled => fixed
The "new" behavior is consistent with recent node.js and bun.js
- Intl: platform.normalizeLanguageTag does not correctly handle valueless unicode extensions #4490
- Intl: Investigate different behavior between engines on CanonicalizeLocaleList #2964

See actions/runner-images#12045
Fixes #7045
Fixes #4490

rhuanjl · 2025-12-16T09:18:39Z

So, the Windows globalisation implementation doesn't work on latest windows AND the ICU alternative fails too, weirdly the test runner seems to be crashing out after one of the fails rather than finishing and saying how many tests failed.... that is awkward.

A couple of test fails look like things that are may be fairly easily fixable e.g. ICU version mismatches to amend tests for or inconsistent conditions in the test suite that expect windows to not be ICU etc. Hopefully it's not overall too big a job but I have no setup I can work on this on.

ShortDevelopment · 2025-12-16T09:23:53Z

I'll look into this in the next view weeks

rhuanjl · 2026-01-08T21:50:36Z

The windows-2019 test runners are now disabled.... So all CI runs now fail.

`windows-2019` → `windows-2022`

https://learn.microsoft.com/en-us/windows/win32/intl/international-components-for-unicode--icu-#version-1903-may-2019-update

ShortDevelopment · 2026-01-08T22:09:55Z

Build/Chakra.Build.Default.props

    <BuildChakraICUData Condition="'$(BuildChakraICUData)'=='' AND ('$(ChakraICU)'=='static' OR '$(ChakraICU)'=='shared')">true</BuildChakraICUData>

-    <IcuLibraryDependencies Condition="'$(ChakraICU)'=='windows'">icuuc.lib;icuin.lib</IcuLibraryDependencies>
+    <IcuLibraryDependencies Condition="'$(ChakraICU)'=='windows'">icu.lib</IcuLibraryDependencies>


See History of changes to the ICU library in Windows.

Would it be possible to do version detection and build with the older libs if someone is building on older windows? I know we can't test it but may be helpful to someone.

If you decide to support this config I could add another build-flag or detect the config based on the current windows version.

ShortDevelopment · 2026-01-08T22:15:15Z

test/DebuggerCommon/ES6_intl_simple_attach.js.dbg.baseline

              "constructor": "function <large string>",
              "resolvedOptions": "function <large string>",
-              "compare": "function <large string>"
+              "compare": "Error <large string>"


I'm not sure why these results changed...

Are we running this on linux/macOS as well?

I.e. is the windows change bringing windows in line with the other platforms?

Yes, this test was WinGlob-only but now is running on all platforms.
This baseline is generated by the debugger so there might be an issue with that?

rhuanjl · 2026-01-09T21:09:03Z

@ppenzin What do you think? This fixes INTL on windows newer than the May 2019 Update but breaks it on older versions - though the key change is just build config we could perhaps document how to build the older version?

rhuanjl

Thank you for the work on this. Hoping we'll merge it within a week though I have a few questions - see comments.

rhuanjl · 2026-01-10T11:26:55Z

test/DebuggerCommon/ES6_intl_simple_attach.js.dbg.baseline

              "constructor": "function <large string>",
              "resolvedOptions": "function <large string>",
-              "compare": "function <large string>"
+              "compare": "Error <large string>"


Are we running this on linux/macOS as well?

I.e. is the windows change bringing windows in line with the other platforms?

rhuanjl · 2026-01-10T11:28:26Z

test/Intl/GetCanonicalLocales.js

-            equal("de-DE-u-co-kn", "de-DE-u-co-yes-kn-true", gcl("de-de-u-kn-co")[0]);
-            equal("de-DE-u-co-phonebk-kn", "de-DE-u-co-phonebk-kn-true", gcl("de-de-u-kn-co-phonebk")[0]);
-            equal("de-DE-u-co-phonebk-kn-yes", "de-DE-u-co-phonebk-kn-true", gcl("de-DE-u-kn-yes-co-phonebk")[0]);
+            if (WScript.Platform.ICU_VERSION < 62) {


Why are we only testing this on old versions of ICU? Is this behaviour wrong or broken on new ICU?

Same point below.

It was actually broken in old icu version and behaving as expected using WinGlob.
See #4490

Is this being used as an always false condition to avoid testing broken behaviour?

The correct behavior is tested in the else branch.

Before these tests used a custom equals function that tested the correct behavior for WinGlob and the wrong behavior for libicu. I guess this is why the exclude_icu62AndAboveTestFailures was added.
I still kept the old tests as CC technically still supports this version of libicu.

I deduce the condition from the flag name (exclude_icu62AndAboveTestFailures); so it maybe should be <= 62?

rhuanjl · 2026-01-10T11:29:30Z

test/Intl/IntlReturnedValueTests.js.dbg.baseline

    "functionCallsReturn": {
      "[get format returned]": "function <large string>",
-      "[Intl.DateTimeFormat.prototype.format returned]": "string ‎2‎/‎1‎/‎2000"
+      "[Intl.DateTimeFormat.prototype.format returned]": "string 2/1/2000"


What has changed here?

The output contained bi-di markers which normally aren't used in locales like en-US.

This also seemed to be a WinGlob quirk:

ChakraCore/test/Intl/rlexe.xml

Lines 84 to 88 in 792ee76

<default>

<compile-flags> -Intl -debugLaunch -dbgbaseline:IntlReturnedValueTests.js.dbg.baseline</compile-flags>

<files>IntlReturnedValueTests.js</files>

<tags>Intl,require_winglob</tags> 

</default>

rhuanjl · 2026-01-10T11:31:05Z

test/CMakeLists.txt

 if (NO_ICU)
    set(TEST_ICU --not-tag exclude_noicu)
-elseif (NOT EMBED_ICU)
-    set(TEST_ICU --not-tag exclude_icu62AndAboveTestFailures)


Have you removed all uses of this flag?

Yes. This flag was only used for NumberFormat.js and GetCanonicalLocales.js (two of the tests that were failing initially)

rhuanjl · 2026-01-10T11:32:47Z

Build/Chakra.Build.Default.props

    <BuildChakraICUData Condition="'$(BuildChakraICUData)'=='' AND ('$(ChakraICU)'=='static' OR '$(ChakraICU)'=='shared')">true</BuildChakraICUData>

-    <IcuLibraryDependencies Condition="'$(ChakraICU)'=='windows'">icuuc.lib;icuin.lib</IcuLibraryDependencies>
+    <IcuLibraryDependencies Condition="'$(ChakraICU)'=='windows'">icu.lib</IcuLibraryDependencies>


Would it be possible to do version detection and build with the older libs if someone is building on older windows? I know we can't test it but may be helpful to someone.

rhuanjl · 2026-01-10T11:33:31Z

lib/Runtime/Library/IntlEngineInterfaceExtensionObject.cpp

        }

-#if defined(ICU_VERSION) && ICU_VERSION >= 61
+#if !defined(ICU_VERSION) || ICU_VERSION >= 61


Is this condition ever going to be false now?

This code guards the usage of apis that were not available before icu < 61.
Windows icu does not expose a compile-time ICU_VERSION as the exact version is dependent on the windows version at runtime. Instead available apis are determined based on the windows version CC is compiled against.

ChakraCore/lib/Runtime/PlatformAgnostic/ChakraICU.h

Lines 9 to 18 in 792ee76

#ifdef HAS_ICU

#ifdef WINDOWS10_ICU

// if WINDOWS10_ICU is defined, pretend like we are building for recent Redstone,

// even if that isn't necessarily true

#pragma push_macro("NTDDI_VERSION")

#undef NTDDI_VERSION

#define NTDDI_VERSION NTDDI_WIN10_RS5

#include <icu.h>

#pragma pop_macro("NTDDI_VERSION")

#else // ifdef WINDOWS10_ICU

The idea is: Run the new code on icu >= 61 if we know the version at built-time else assume the newer apis are available (on windows).

So if we build on windows ICU_VERSION will be undefined? And therefore this condition will be true?

Please could you add a comment to explain that as it's a little counter intuitive. (Same for the equivalent condition below)

So if we build on windows ICU_VERSION will be undefined? And therefore this condition will be true?

Exactely

Please could you add a comment to explain that as it's a little counter intuitive. (Same for the equivalent condition below)

Sure: 1684f20

ShortDevelopment force-pushed the ci/use-windows-2025/7045 branch from acc22ee to 22393f5 Compare December 3, 2025 23:36

ShortDevelopment force-pushed the ci/use-windows-2025/7045 branch from 22393f5 to 659a32f Compare December 24, 2025 14:12

ShortDevelopment added 6 commits January 8, 2026 23:05

windows-2022 → windows-2025

39a6bbb

`windows-2019` → `windows-2022`

Use windows icu by default

4551683

Link against "new" icu.lib on windows.

22d8dbe

https://learn.microsoft.com/en-us/windows/win32/intl/international-components-for-unicode--icu-#version-1903-may-2019-update

Fix Intl.getCanonicalLocales tests

4a3f000

Check icu version dynamically for windows

106af5f

Assume windows version will always be >= 62

a487d1d

ShortDevelopment force-pushed the ci/use-windows-2025/7045 branch from efab8ed to 4135302 Compare January 8, 2026 22:05

ShortDevelopment commented Jan 8, 2026

View reviewed changes

ShortDevelopment added 2 commits January 8, 2026 23:18

Update ES6_intl_simple_attach.js baseline

8fc4708

Remove require_winglob

18eaa7e

ShortDevelopment force-pushed the ci/use-windows-2025/7045 branch from 4135302 to 18eaa7e Compare January 8, 2026 22:30

ShortDevelopment added 2 commits January 9, 2026 00:18

Fix copyright

14c1aa7

Remove bi-di chars from IntlReturnedValueTests.js baseline

a931ead

ShortDevelopment force-pushed the ci/use-windows-2025/7045 branch from a240230 to a931ead Compare January 8, 2026 23:21

ShortDevelopment marked this pull request as ready for review January 9, 2026 00:33

rhuanjl reviewed Jan 10, 2026

View reviewed changes

Add comment to explain ICU_VERSION conditional-compilation.

1684f20

	<default>
	<compile-flags> -Intl -debugLaunch -dbgbaseline:IntlReturnedValueTests.js.dbg.baseline</compile-flags>
	<files>IntlReturnedValueTests.js</files>
	<tags>Intl,require_winglob</tags> <!-- require_winglob because the baseline contains embedded bi-di markers -->
	</default>

	#ifdef HAS_ICU
	#ifdef WINDOWS10_ICU
	// if WINDOWS10_ICU is defined, pretend like we are building for recent Redstone,
	// even if that isn't necessarily true
	#pragma push_macro("NTDDI_VERSION")
	#undef NTDDI_VERSION
	#define NTDDI_VERSION NTDDI_WIN10_RS5
	#include <icu.h>
	#pragma pop_macro("NTDDI_VERSION")
	#else // ifdef WINDOWS10_ICU

ci: use windows 2025 runners #7048

Are you sure you want to change the base?

ci: use windows 2025 runners #7048

Uh oh!

Conversation

ShortDevelopment commented Dec 3, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

💥 Breaking changes

🧪 Tests

Uh oh!

rhuanjl commented Dec 16, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

ShortDevelopment commented Dec 16, 2025

Uh oh!

rhuanjl commented Jan 8, 2026

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

rhuanjl commented Jan 9, 2026

Uh oh!

rhuanjl left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

rhuanjl Jan 10, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

ShortDevelopment Jan 10, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

ShortDevelopment commented Dec 3, 2025 •

edited

Loading

rhuanjl commented Dec 16, 2025 •

edited

Loading

rhuanjl Jan 10, 2026 •

edited

Loading

ShortDevelopment Jan 10, 2026 •

edited

Loading