Charles Plessy [Tue, 12 Oct 2010 15:14:42 +0000 (00:14 +0900)]
Clarification of the format of control files, Closes: #501930, #593909.
- Specifies field names similarly to RFC 822/5832;
- Distinguishes simple, folded and mulitiline fields;
- Clarifies paragraph separators (#501930);
- The order of paragraphs is significant;
- Fields can have different types or purposes in different control files;
- Moved the description of comments from §5.2 to §5.1;
- Documented that relationship fields can only be folded in debian/control.
Rather than using imperative tense to describe changes as if they
were commit messages, switch to descriptive statements about the
contents of Policy or imperatives for packagers where appropriate.
Only the most recent entries have been reviewed.
Add a cross-reference for Pre-Depends requirements
* Add a cross-reference to the Pre-Depends requirement in 3.5 to section
7.2 where Pre-Depends is defined. Thanks, Mattias Ellert and Jonathan
Nieder. (Closes: #599944)
Jonathan Nieder [Tue, 1 Mar 2011 23:58:48 +0000 (15:58 -0800)]
Add a reminder of Pre-Depends policy to §7.2 "Binary Dependencies"
Section 3.5 says:
You should not specify a Pre-Depends entry for a package before this
has been discussed on the debian-devel mailing list and a consensus
about doing that has been reached.
Add a reminder of this to section 7.2, where the semantics of
Pre-Depends are discussed in more detail.
Requested-by: Mattias Ellert <mattias.ellert@fysast.uu.se> Fixes: http://bugs.debian.org/599944 Signed-off-by: Jonathan Nieder <jrnieder@gmail.com>
Steve Langasek [Mon, 21 Mar 2011 09:17:14 +0000 (02:17 -0700)]
Fix multiarch FHS exception for i386 in light of recent discussions
The current value of DEB_HOST_GNU_TYPE on i386 is unsuitable for
cross-distro standardization, because it varies according to the default CPU
target of the toolchain. Discussion with the toolchain and dpkg maintainers
yielded an alternative solution, a new dpkg-architecture variable
DEB_HOST_MULTIARCH which is committed to dpkg upstream in commit af3153d09aa3ed5597d6d415e5ab7cc3ba972e7c and will be included in the upload
of dpkg 1.16.0. Update Policy to document this new requirement for
multiarch.
* Add introductory paragraphs for each archive area explaining briefly
the purpose of that archive area. Based on a proposal by CJ
Fearnley. (Closes: #594542)
Update the ldconfig footnote listing /etc/ld.so.conf directories
* Update the ldconfig footnote listing the /etc/ld.so.conf directories
to remove the libc5 compatibility directories and mention the
multiarch triplet directories. Based on a patch by Charles Plessy.
(Closes: #597074)
* Add emacs23 to the build dependencies and remove the files generated
from org source from the revision control repository. Build and clean
files from org source unconditionally. Add Process.{txt,html} to the
list of files generated from org source. (Closes: #594274)
* Move Build-Depends-Indep to Build-Depends (there's no reason to use
-Indep in a package that builds only architecture-independent binary
packages), wrap it, and remove version restrictions that are older
than the version in oldstable.
Russ Allbery [Fri, 20 Aug 2010 22:36:19 +0000 (15:36 -0700)]
Document current behavior of conflicting conffiles
Document that the new package takes over the conffile from the removed
package when the packages conflict, and that the new package may see
left-over configuration files from the old package. Reorder the section
to put the preferred solutions first.
Russ Allbery [Wed, 18 Aug 2010 20:55:34 +0000 (13:55 -0700)]
Clarify required rules in debian/rules
* Use the word "implemented" to describe required targets in
debian/rules, which is clearer about allowing wildcard rules. List
the required rules in their own paragraph rather than with the
paragraph discussing non-interactivity, and explicitly mark all rules
as either required or optional. (Closes: #536790)
Russ Allbery [Thu, 12 Aug 2010 18:23:22 +0000 (11:23 -0700)]
Better define build architecture and host architecture
* Be more verbose in defining the build architecture and the host
architecture and consistently refer to architecture rather than
machine. (Closes: #591857)
Mention that orphaned packages are an exception to the dictate that every
package have a maintainer, and avoid saying that packages "should" be
orphaned in particular situations. Be purely descriptive about the state
and describe only the package requirements for indicating an orphaned
package.
Incorporate maintainer script updates from Jonathan Nieder's review
Clarify and in some cases fix the summaries of what maintainer scripts
can rely upon based on a comprehensive review from Jonathan Nieder,
and include a wording fix from Ben Finney.
Add detailed breakdown of what maintainer scripts can assume
For each possible action in a maintainer script, state what the
maintainer script can assume in that case. Remove references to
using Pre-Depends for controlling what's available during package
removal. Be clear that no postrm actions can rely on dependencies
being available. Remove references saying that prerm can rely on
dependencies being configured.