Feature #4512: [PATCH] ext/fcntl/fcntl.c: add F_DUPFD_CLOEXEC constant

• Status: Closed
• Assignee: kosaki (Motohiro KOSAKI)
• Target version: 2.0.0

[normalperson (Eric Wong)]

=begin
It's in POSIX.1-2008 and will be useful when Ruby becomes more
multi-threaded.

related: http://redmine.ruby-lang.org/issues/1291

If/when 1291 goes through I'd also like IO#dup to work with the dup3()
system call, but this patch is much less intrusive.
=end

[nahi (Hiroshi Nakamura)]

• Target version changed from 1.9.3 to 2.0.0

We're looking for a fcntl maintainer...

[normalperson (Eric Wong)]

Hiroshi Nakamura [e-mail] wrote:

We're looking for a fcntl maintainer...

Since fcntl is related to I/O and files (and IO#fcntl is implemented in
io.c), why not just give this to io.c/file.c maintainers?

--
Eric Wong

[kosaki (Motohiro KOSAKI)]

• Status changed from Open to Assigned
• Assignee set to kosaki (Motohiro KOSAKI)

why not just give this to io.c/file.c maintainers?

We don't have io.c/file.c maintainers too. Anyway I assigned it to me. Please stop annoying ping pong.

[kosaki (Motohiro KOSAKI)]

• Status changed from Assigned to Feedback

I've reviewed linux kernel source code and concluded this patch doesn't have the same issue with O_CLOEXEC when run on older kernels. So, I'll commit the patch if you make test case.

Thanks.

[normalperson (Eric Wong)]

• File 0001-ext-fcntl-fcntl.c-add-F_DUPFD_CLOEXEC-constant.patch 0001-ext-fcntl-fcntl.c-add-F_DUPFD_CLOEXEC-constant.patch added

Also pushed to "fcntl-dupfd-cloexec" branch of git://bogomips.org/ruby.git
http://bogomips.org/ruby.git/log/?h=fcntl-dupfd-cloexec

[kosaki (Motohiro KOSAKI)]

• Status changed from Feedback to Assigned

[kosaki (Motohiro KOSAKI)]

I'm waiting the discussion of #5041.

[kosaki (Motohiro KOSAKI)]

• Status changed from Assigned to Closed

Akr committed #5041. so we can close this ticket too.