Ruby Issue Tracking System

10/25/2026

05:16 AM Ruby Revision 349d9d65 (git): Initialize the fast fallback waiter fd before anything can raise

An invalid open_timeout raised before the waiter was assigned, so the cleanup closed fd 0 left by memset and took the process's stdin with it.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> hsbt (Hiroshi SHIBATA)

03:39 AM Ruby Revision 83800032 (git): [ruby/zlib] Fix Zlib.gunzip with a header truncated in the file name or comment

Reading the rest of the field went to the IO, which Zlib.gunzip does not have, instead of raising Zlib::GzipFile::Error as the other truncated header fields do.
https://github.com/ruby/zlib/commit/a16601dc10
Co-Authored-By: Claude Opus... hsbt (Hiroshi SHIBATA)

03:39 AM Ruby Revision b6ab1924 (git): [ruby/zlib] Test Zlib.gunzip with optional gzip header fields

https://github.com/ruby/zlib/commit/2fc5eb721f
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> hsbt (Hiroshi SHIBATA)

02:57 AM Ruby Revision 764d1d3a (git): win32: Fix a handle leak in rb_w32_read_internal

One of its early returns after `setup_overlapped` did not close the event.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> hsbt (Hiroshi SHIBATA)

02:57 AM Ruby Revision 47c12f20 (git): win32: Do not set the pipe mode when it is unchanged

`rb_fd_set_nonblock` returns early on other platforms when the descriptor is already non-blocking, but `rb_w32_set_nonblock2` called `SetNamedPipeHandleState` every time.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> hsbt (Hiroshi SHIBATA)

02:57 AM Ruby Revision c3b61b12 (git): win32: Wait for the cancelled pipe I/O before returning

An interrupted pipe read or write returned right after `CancelIo`, which only requests the cancellation. Wait for the I/O to end and return what it transferred.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> hsbt (Hiroshi SHIBATA)

10/23/2026

03:26 AM Ruby Revision b9ceb36f (git): Keep the program size error from OP_END with subexp calls

With a subexp call, the result of unset_addr_list_fix overwrote the error, so a pattern over the limit compiled without raising RegexpError.
Follow-up to cf35145fc46.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> hsbt (Hiroshi SHIBATA)

03:17 AM Ruby Revision af48f08f (git): Bump up to 3.3.13

hsbt (Hiroshi SHIBATA)

03:07 AM Ruby Revision 366f5fdc (git): Keep the program size error from OP_END with subexp calls

With a subexp call, the result of unset_addr_list_fix overwrote the error, so a pattern over the limit compiled without raising RegexpError.
Follow-up to cf35145fc46.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> hsbt (Hiroshi SHIBATA)

03:00 AM Ruby Revision f6496bca (git): Keep the program size error from OP_END with subexp calls

With a subexp call, the result of unset_addr_list_fix overwrote the error, so a pattern over the limit compiled without raising RegexpError.
Follow-up to cf35145fc46.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
(cherry pick... hsbt (Hiroshi SHIBATA)