[#7955] Failing tests in ruby since 1.8.2 — "Caleb Tennis" <caleb@...>
The following tests have been failing in Ruby for a long time, including
[#7978] Patch for Unix socket peer credentials — "James F. Hranicky" <jfh@...>
This patch adds support for getting the uid and gid of the peer
In article <200606091528.30171.jfh@cise.ufl.edu>,
On Friday 16 June 2006 11:51, Tanaka Akira wrote:
In article <200606161327.35948.jfh@cise.ufl.edu>,
On Saturday 17 June 2006 06:27, Tanaka Akira wrote:
In article <200607101352.16804.jfh@cise.ufl.edu>,
On Tuesday 11 July 2006 00:10, Tanaka Akira wrote:
Hi,
On Thursday 13 July 2006 22:48, nobu@ruby-lang.org wrote:
On Jul 18, 2006, at 12:27 PM, James F. Hranicky wrote:
On Tuesday 18 July 2006 15:52, Eric Hodel wrote:
[#7994] Ruby Kaigi date confusion — "Charles O Nutter" <headius@...>
I'm quite confused by the dates I have seen reported on various Ruby Kaigi
[#8013] Download page on ruby-lang has numeric URL — Hugh Sasse <hgs@...>
This is off-topic to ruby-core, but possibly core to ruby's uptake:
On Jun 19, 2006, at 3:32 AM, Hugh Sasse wrote:
[#8038] bug in $. ? — Wybo Dekker <wybo@...>
wybo>cat t
Wybo Dekker schrieb:
Pit Capitain wrote:
[#8050] Thank-you to the Rails Core Team — Dave Teare <devlists-ruby-core@...>
While we were listening to Dave Thomas' Keynote address today at
[#8061] Win32 Extension Issues Wanted! — "Austin Ziegler" <halostatue@...>
Everyone. I had a conversation with folks from Microsoft today about
[#8065] Core documentation patches — Alex Young <alex@...>
Hi there,
Hi,
Yukihiro Matsumoto wrote:
[#8073] 1.8.5p1 build failure on Solaris 10 — "Daniel Berger" <Daniel.Berger@...>
Solaris 10
Hi,
Yukihiro Matsumoto wrote:
>>>>> "D" == Daniel Berger <Daniel.Berger@qwest.com> writes:
ts <decoux@moulon.inra.fr> wrote on 28.06.2006 17:37:00:
Hi,
Yukihiro Matsumoto <matz@ruby-lang.org> wrote on 29.06.2006 20:02:11:
Hi,
Yukihiro Matsumoto <matz@ruby-lang.org> wrote on 29.06.2006 20:53:20:
ville.mattila@stonesoft.com wrote:
[#8087] optparse.rb to RDoc documentation patch — <noreply@...>
Patches item #4879, was opened at 2006-06-28 20:50
On Jun 28, 2006, at 11:50 AM, <noreply@rubyforge.org>
[#8102] Reorganizing configure.in by platform? — "Daniel Berger" <Daniel.Berger@...>
Hi,
Re: Patch for Unix socket peer credentials
On Sat, Jun 10, 2006 at 04:28:36AM +0900, James F. Hranicky wrote: > This patch adds support for getting the uid and gid of the peer > socket connected to a Unix domain socket. > > The patch is really just a suggestion, as you can't necessarily get the > same kinds of information across platforms. Unless I'm mistaken this is > what you can get for the following platforms: > > FreeBSD, OSX (getpeereid) : euid, egid > Linux (getsockopt) : pid, uid, gid > Solaris 10 (getpeerucred) : uid, euid, gid, egid, pid and more > > The patch adds checks to extconf.rb for the getpeereid and getpeerucred > functions and a check for the SO_PEERCRED macro. If none of these are > found calls to the peer_uid and peer_gid method raise an error. > > Questions and comments welcome, and I'm open to suggestions for > changes/improvements. I tried to do this on OS X recently using DGRAM AF_LOCAL sockets based on the examples from Stevens, and failed, possibly due to an error on my part, or maybe because the only way to do it is with getpeereid() on a connected stream-oriented socket? So, this is serendipitous, and very useful. Will try the patch this weekend. A bits of feedback: - Maybe instead of #peer_uid, and #peer_gid, have just #peer_id => [uid,gid], seems more convenient to me to get them both in one shot, matches (some of) the underlying system calls more closely, and also rids you of the duplicate code adn documentation (btw, thanks for including docs!). - Maybe instead of raising an error if not supported (which means you have to call it to find that out), just don't define the method at all? Kindof like socket options, you can do Socket.constants.include? to see if a sockopt is supported before even attempting to use it, it might be nice to do (Socket/UNIXSocket).methods.include? :peer_id to see if #peer_id is supported. - I think the methods should be lower down in the hierarchy, or at least duplicated in Socket, if I create an AF_LOCAL connection-oriented socket with Socket.new, I should be able to get the peer ID, too, I think. - I am using DGRAM unix domain sockets... it sure would be nice to have a way to determine the peer's credentials for received messages, any thoughts on that? Even non-ruby thoughts... do you know if it works on the Linux and OS X, and how to make it work? Please and Thank You, Sam