Issue #20758 has been updated by eightbitraptor (Matthew Valentine-House). Status changed from Open to Closed Closing this as a duplicate. My apologies. ---------------------------------------- Bug #20758: Build failure with macOS Sequoia (15.0) and Xcode 16 https://bugs.ruby-lang.org/issues/20758#change-109891 * Author: eightbitraptor (Matthew Valentine-House) * Status: Closed * Backport: 3.1: UNKNOWN, 3.2: UNKNOWN, 3.3: UNKNOWN ---------------------------------------- ## Observed behaviour Test failures in `TestProcess#test_daemon_noclose`. Output as follows: ``` Run options: --seed=56670 "--ruby=./miniruby -I../src/lib -I. -I.ext/common ../src/tool/runruby.rb --extout=.ext -- --disable-gems" --excludes-dir=../src/test/.excludes --name=!/memory_leak/ # Running tests: [ 99/158] TestProcess#test_daemon_nocloseobjc[89134]: +[NSPlaceholderMutableString initialize] may have been in progress in another thread when fork() was called. objc[89134]: +[NSPlaceholderMutableString initialize] may have been in progress in another thread when fork() was called. We cannot safely call it or ignore it in the fork() child process. Crashing instead. Set a breakpoint on objc_initializeAfterForkError to debug. = 0.00 s 1) Failure: TestProcess#test_daemon_noclose [/Users/mattvh/git/ruby/src/test/ruby/test_process.rb:1882]: <"ok\n" + "/\n"> expected but was <"">. Finished tests in 8.625520s, 18.3177 tests/s, 112.5729 assertions/s. 158 tests, 971 assertions, 1 failures, 0 errors, 1 skips ruby -v: ruby 3.4.0dev (2024-09-23T12:23:58Z master 4c3727c6ac) +PRISM [arm64-darwin24] make: *** [yes-test-all] Error 1 ``` ## Expected behaviour I expect the tests to all pass with no errors. ## Notes This started happening for me immediately after upgrading my laptop to the latest macOS using the software update tool. Versions reported are: ``` Command Line Tools for Xcode 16.0 23/09/2024, 14:24:42 macOS Sequoia 15.0 15.0 23/09/2024, 14:41:14 ``` This appears to be the same issue as seen in https://bugs.ruby-lang.org/issues/18912 I can replicate it using the same minimal repro that @katei described in https://bugs.ruby-lang.org/issues/18912#note-15 ``` Process.daemon(false, true) Dir.pwd ``` and can confirm that exporting the variable `OBJC_DISABLE_INITIALIZE_FORK_SAFETY=YES` suppresses the error. This behaviour was tested on this commit https://github.com/ruby/ruby/commit/4c3727c6ac -- https://bugs.ruby-lang.org/