From owner-svn-ports-all@FreeBSD.ORG Wed Apr 23 17:24:04 2014 Return-Path: Delivered-To: svn-ports-all@freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2001:1900:2254:206a::19:1]) (using TLSv1 with cipher ADH-AES256-SHA (256/256 bits)) (No client certificate requested) by hub.freebsd.org (Postfix) with ESMTPS id CCDB4A9C for ; Wed, 23 Apr 2014 17:24:04 +0000 (UTC) Received: from b2bfep15.mx.upcmail.net (b2bfep15.mx.upcmail.net [62.179.121.60]) by mx1.freebsd.org (Postfix) with ESMTP id 1815C113C for ; Wed, 23 Apr 2014 17:24:03 +0000 (UTC) Received: from edge11.upcmail.net ([192.168.13.81]) by b2bfep14-int.chello.at (InterMail vM.8.01.05.11 201-2260-151-128-20120928) with ESMTP id <20140423172257.KAFM28875.b2bfep14-int.chello.at@edge11.upcmail.net> for ; Wed, 23 Apr 2014 19:22:57 +0200 Received: from iznogoud.viz ([91.119.56.164]) by edge11.upcmail.net with edge id tVNw1n00V3YbPKv0BVNwB0; Wed, 23 Apr 2014 19:22:57 +0200 X-SourceIP: 91.119.56.164 Received: from wolfgang by iznogoud.viz with local (Exim 4.82 (FreeBSD)) (envelope-from ) id 1Wd0sq-0000Nx-0T; Wed, 23 Apr 2014 19:22:56 +0200 From: Wolfgang Jenkner To: Emanuel Haupt Subject: Re: svn commit: r351864 - head/shells/bash Date: Wed, 23 Apr 2014 19:06:35 +0200 References: <5356B13D.3010600@gmx.de> <3C27DBBC-628F-48F7-9EC4-82C7BA559D85@adamw.org> <20140423102806.daa72bd1fec855d882d47163@critical.ch> Message-ID: <85ha5kgdu8.fsf@iznogoud.viz> User-Agent: Gnus/5.13001 (Ma Gnus v0.10) Emacs/24.4.50 (berkeley-unix) MIME-Version: 1.0 Content-Type: text/plain Cc: olli hauer , svn-ports-all@FreeBSD.org, ports-committers@freebsd.org, Emanuel Haupt , Adam Weinberger X-BeenThere: svn-ports-all@freebsd.org X-Mailman-Version: 2.1.17 Precedence: list List-Id: SVN commit messages for the ports tree List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Wed, 23 Apr 2014 17:24:04 -0000 On Wed, Apr 23 2014, Emanuel Haupt wrote: > "Adam Weinberger" wrote: >> Also I'm confused by the statement "bash requires fdescfs(5) mounted >> on /dev/fd", as it runs just fine without it. Perhaps it should be >> changed to: >> bash will support a '/dev/fd/N' syntax if fdescfs(5) is >> mounted on /dev/fd > > Adam, this seems more accurate. Please feel free to commit. This is not quite true either, but kib@ perfectly explained the awkward situation in http://permalink.gmane.org/gmane.os.freebsd.devel.amd64/15636 Also (if you look at what started the thread) the main problem is not the /dev/fd/* syntax per se, but the process substitution feature, as mat@ already observed. So I'd suggest something like You should mount fdescfs(5) on /dev/fd if you are installing a precompiled package from an official pkg repository or, more generally, if bash was compiled with fdescfs(5) being mounted. The reason is that bash implements process substitution by means of the file-descriptor file system if that was mounted at build time. For details, please see `Process Substitution' in the bash(1) manual.