@kitbellew

Thank you for identifying and reporting this problem.

The "JDK compliance" and "runtime" log files seem to have errors which present differently
but which may have the same underlying cause. Insufficient stack space is a candidate.

Assumptions

1. The GitHub Windows_2025 platform is correct and up to date. This is probably a good
   assumption, but still need to be listed.

2. A straight forward change in the .yml file fromWindows_2022 toWindows_2025 is all
   that is needed. That is, no increase in memory parameters or special parameters.

   This is probably a good first candidate assumption but needs verification.

Primary Evidence

JDK compliance

The "JDK compliance" Windows_2025 logs have an apparent failure block which I have not seen in
Windows_2022 logs:

025-09-26T09:13:20.5662784Z �[0m[�[0m�[31merror�[0m] �[0m�[0mTest org.scalanative.testsuite.javalib.lang.�[33mStackTraceElementTest�[0m.�[31misNativeMethod�[0m failed: java.util.�[31mNoSuchElementException�[0m: head of empty array, took 0.000 sec�[0m2025-09-26T09:13:20.5664549Z �[0m[�[0m�[31merror�[0m] �[0m�[0m    at <none>.(:0)�[0m2025-09-26T09:13:20.5972614Z �[0m[�[0m�[31merror�[0m] �[0m�[0m    at <none>.(:0)�[0m2025-09-26T09:13:20.6401181Z �[0m[�[0m�[31merror�[0m] �[0m�[0m    at <none>.(:0)�[0m2025-09-26T09:13:20.6728751Z �[0m[�[0m�[31merror�[0m] �[0m�[0m    at <none>.(:0)�[0m2025-09-26T09:13:20.7175229Z �[0m[�[0m�[31merror�[0m] �[0m�[0m    ... 31 more�[0m

I think I also tracked down a "PendingStackOverflowException" (concept from memory, probably not exact wording)
in similar log files for PR#4484.

runtime

I did not check all three Windows runtime logs here. I have examined a few in PR#4484. Here the
Windows_2025 Scala 2.13 log reports:

2025-09-26T09:19:01.4820812Z �[0m[�[0m�[33mwarn�[0m] �[0m�[0mForce close java.lang.RuntimeException: Process D:\a\scala-native\scala-native\unit-tests\native\.2.13\target\scala-2.13\tests-test.exe finished with non-zero value -1073740940 (0xc0000374)�[0m

Google says this means:

Windows exit code0xc0000374 signifies a heap corruption, meaning an application or system process attempted to access or modify memory in an invalid way.

(Remember also reports on Linux of people having seen signal 11 under load).

Next Steps

Beyond this is outside my areas of expertise or likely growth in economic time.
I standby to help if and as I can.