Issue #19568 has been updated by luke-gru (Luke Gruber). It's my understanding that these tests should not be run, they should be skipped by the method `skip_unsupported_platforms!`, at least they are skipped when I run `make test-all`. Were they being run before? You can check this with: `$ make test-all TESTOPTS="test/rubygems/test_gem_ext_cargo_builder.rb"`. I get the output: ``` Finished tests in 1.098123s, 4.5532 tests/s, 9.1064 assertions/s. 5 tests, 10 assertions, 0 failures, 0 errors, 10 skips ``` I don't know how the tests are being run in mssys2, I'm assuming it's not just `make test-all` but something else? If so, the logic should be changed in the method `skip_unsupported_platforms!` ---------------------------------------- Bug #19568: Windows mingw rust, failures with test/rubygems/test_gem_ext_cargo_builder.rb https://bugs.ruby-lang.org/issues/19568#change-102621 * Author: MSP-Greg (Greg L) * Status: Open * Priority: Normal * ruby -v: master * Backport: 2.7: UNKNOWN, 3.0: UNKNOWN, 3.1: UNKNOWN, 3.2: UNKNOWN ---------------------------------------- The Windows ucrt & mingw builds recently started failing tests in: test/rubygems/test_gem_ext_cargo_builder.rb TestGemExtCargoBuilder Checking changed files from the last good build showed: array.c gems/bundled_gems iseq.c lib/irb.rb lib/irb/cmd/nop.rb lib/irb/completion.rb lib/irb/ext/tracer.rb lib/irb/extend-command.rb lib/ruby_vm/rjit/compiler.rb lib/ruby_vm/rjit/context.rb lib/ruby_vm/rjit/insn_compiler.rb lib/ruby_vm/rjit/jit_state.rb rjit_c.c rjit_c.h rjit_c.rb test/irb/test_cmd.rb tool/leaked-globals tool/rjit/bindgen.rb vm_callinfo.h yjit.c yjit/bindgen/src/main.rs yjit/src/codegen.rs yjit/src/cruby_bindings.inc.rs There have been no changes to ruby-loco during this time. ruby-loco uses rust installed by dtolnay/rust-toolchain@stable, which are the tools provided by the Rust organization. I also tried MSYS2 rust. I've created two files showing the errors logs: https://github.com/MSP-Greg/ruby-loco/releases/download/error_logs/windows_u... https://github.com/MSP-Greg/ruby-loco/releases/download/error_logs/windows_u... I'm not familiar with Rust, so I have no idea what the issue is. I'm guessing this might be an MSYS2 issue, but I'm not sure... -- https://bugs.ruby-lang.org/

Issue #19568 has been updated by MSP-Greg (Greg L). @hsbt Thanks. JFYI, the problem also exists in [oxidize-rb/rb-sys](https://github.com/oxidize-rb/rb-sys/actions/workflows/ci.yml)'s CI. It may be MSYS2, but probably a Rust issue. As mentioned above, I couldn't find a change to RubyGems between the last successful ruby-loco build and the first failing build. ---------------------------------------- Bug #19568: Windows mingw rust, failures with test/rubygems/test_gem_ext_cargo_builder.rb https://bugs.ruby-lang.org/issues/19568#change-102646 * Author: MSP-Greg (Greg L) * Status: Third Party's Issue * Priority: Normal * Assignee: hsbt (Hiroshi SHIBATA) * ruby -v: master * Backport: 2.7: UNKNOWN, 3.0: UNKNOWN, 3.1: UNKNOWN, 3.2: UNKNOWN ---------------------------------------- The Windows ucrt & mingw builds recently started failing tests in: test/rubygems/test_gem_ext_cargo_builder.rb TestGemExtCargoBuilder Checking changed files from the last good build showed: array.c gems/bundled_gems iseq.c lib/irb.rb lib/irb/cmd/nop.rb lib/irb/completion.rb lib/irb/ext/tracer.rb lib/irb/extend-command.rb lib/ruby_vm/rjit/compiler.rb lib/ruby_vm/rjit/context.rb lib/ruby_vm/rjit/insn_compiler.rb lib/ruby_vm/rjit/jit_state.rb rjit_c.c rjit_c.h rjit_c.rb test/irb/test_cmd.rb tool/leaked-globals tool/rjit/bindgen.rb vm_callinfo.h yjit.c yjit/bindgen/src/main.rs yjit/src/codegen.rs yjit/src/cruby_bindings.inc.rs There have been no changes to ruby-loco during this time. ruby-loco uses rust installed by dtolnay/rust-toolchain@stable, which are the tools provided by the Rust organization. I also tried MSYS2 rust. I've created two files showing the errors logs: https://github.com/MSP-Greg/ruby-loco/releases/download/error_logs/windows_u... https://github.com/MSP-Greg/ruby-loco/releases/download/error_logs/windows_u... I'm not familiar with Rust, so I have no idea what the issue is. I'm guessing this might be an MSYS2 issue, but I'm not sure... -- https://bugs.ruby-lang.org/