echangelog should parse out the name only, and not extra GECOS info. this could expose information that developers may not want to put in every changelog entry. "A lot of organisations use the gecos field of their passwd file to store the real name of their users, telephone numbers, addresses or similar information" i.e. from my example /etc/passwd: jnc:x:1555:155:Eric Shattow,room n/a,800-555-9555,800-555-4555,Joe No Castle:/home/jnc:/bin/bash and when i'm trying to do an echangelog: jnc@necro bio2jack $ echangelog --- bio2jack-0.3-r1.ebuild 2005-01-29 19:17:18.903802728 -0600 +++ bio2jack-0.3-r1.ebuild.new 2005-01-29 19:20:26.929218520 -0600 @@ -1 +1 @@ -# Copyright 1999-2004 Gentoo Foundation +# Copyright 1999-2005 Gentoo Foundation --- bio2jack-0.4.ebuild 2004-12-19 07:31:16.000000000 -0600 +++ bio2jack-0.4.ebuild.new 2005-01-29 19:20:26.935217608 -0600 @@ -1 +1 @@ -# Copyright 1999-2004 Gentoo Foundation +# Copyright 1999-2005 Gentoo Foundation --- ChangeLog 2005-01-08 18:28:17.000000000 -0600 +++ ChangeLog.new 2005-01-29 19:20:26.929218520 -0600 @@ -1,7 +1,11 @@ # ChangeLog for media-libs/bio2jack -# Copyright 1999-2004 Gentoo Foundation; Distributed under the GPL v2 +# Copyright 1999-2005 Gentoo Foundation; Distributed under the GPL v2 # $Header: /var/cvsroot/gentoo-x86/media-libs/bio2jack/ChangeLog,v 1.13 2004/12/29 00:39:43 ciaranm Exp $ + 29 Jan 2005; Eric Shattow,room n/a,800-555-9555,800-555-4555,Joe No Castle + <jnc@gentoo.org> bio2jack-0.3-r1.ebuild: + closes bug #74980 + 29 Dec 2004; Ciaran McCreesh <ciaranm@gentoo.org> : Change encoding to UTF-8 for GLEP 31 compliance
I vote no on this one. Parsing GECOS is a losing battle. I believe echangelog will use ECHANGELOG_USER if set so any developer who wants to "protect" their information can just set that.
All GECOS styles I have seen use the person's name as the first part, and comma seperated. Everything in the passwd entry for name field should be parsed up to the first comma, if there is one. That doesn't seem like a losing battle to me. This should be trivial to implement.
> This should be trivial to implement. Great, please attach the patch
> Great, please attach the patch I might set the incoming field seperator from awk to be a comma, and take the first field. Is there a better way?
Fixed in cvs. Your suggestion doesn't make any sense in a Perl program, btw, and certainly doesn't constitute a patch. Next time please make an effort.