Bug #14115: MinGW (ruby-loco) build stopped svn 60832 2017-11-18

• Status: Closed
• Assignee: -
• Target version: -
• ruby -v: svn 60832 2017-11-18
• Backport: 2.3: UNKNOWN, 2.4: UNKNOWN

[MSP-Greg (Greg L)]

Previous build was successful at

ruby 2.5.0dev (2017-11-18 trunk 60828) [x64-mingw32]

Last lines of build log are:

compiling ../ruby/win32/win32.c
compiling ../ruby/win32/file.c
compiling prelude.c
compiling ../ruby/array.c
linking miniruby.exe
generating encdb.h
make: *** [uncommon.mk:704: .rbconfig.time] Segmentation fault
make: *** Waiting for unfinished jobs....
encdb.h updated
==> ERROR: A failure occurred in build().
    Aborting...
Command exited with code 1

Current Appveyor mswin build is frozen/locked 1.0.6109

[usa (Usaku NAKAMURA)]

• Status changed from Open to Closed

Applied in changeset trunk|r60837.

---

Cannot call rb_thread_call_with{out,}_gvl before running VM

• dir.c (opendir_without_gvl, with_gvl_gc_for_fd, opendir_at): check the VM is
  already initialized before calling rb_thread_call_with{out,}_gvl().
  [Bug #14115]

[normalperson (Eric Wong)]

[e-mail] wrote:

Bug #14115: MinGW (ruby-loco) build stopped svn 60832 2017-11-18
https://bugs.ruby-lang.org/issues/14115

Re: r60837

Why not change init order so rb_cThread gets defined earlier on?
That would be cleaner than checking every call. I also never
noticed the problem on pthreads, so I'm not sure why it only
manifests on w32