Move documentation to doc/

6 files changed, 246 insertions, 4 deletions

diff --git a/Makefile b/Makefile
index c935c6c..f9d1707 100644
--- a/Makefile
+++ b/Makefile
@@ -46,10 +46,10 @@ SRC +=		$(LIB_SRC:%=lib/%)
 DIRT +=		$(LIB_IN:%=lib/%) lib/backup-files{,.o}
 
 DOC_IN :=	README
-DOC_SRC :=	$(DOC_IN:%=%.in)
-DOC :=		$(DOC_IN) docco.txt
-SRC +=		$(DOC_SRC) docco.txt
-DIRT +=		$(DOC_IN)
+DOC_SRC :=	$(DOC_IN:%=doc/%.in)
+DOC :=		$(DOC_IN) doc/docco.txt
+SRC +=		$(DOC_SRC) doc/docco.txt
+DIRT +=		$(DOC_IN:%=doc/%)
 
 MAN1 :=		bin/guards.1
 
diff --git a/README.in b/doc/README.in
index 83f252c..83f252c 100644
--- a/README.in
+++ b/doc/README.in
diff --git a/doc/consider-for-docs-1.mail b/doc/consider-for-docs-1.mail
new file mode 100644
index 0000000..ff45891
--- /dev/null
+++ b/doc/consider-for-docs-1.mail
@@ -0,0 +1,109 @@
+From: Andreas Gruenbacher <agruen@suse.de>
+Organization: SuSE Linux AG
+Subject: Patch Management Scripts
+Date: Sat, 18 Jan 2003 18:40:32 +0100
+
+Hello,
+
+this is to announce that the patch-scripts package is available at
+http://www.suse.de/~agruen/patch-scripts/. The current version is 0.11.
+
+These scripts were based on Andrew Morton's patch scripts
+(http://www.zip.com.au/~akpm/linux/patches/), but have been very
+heavily modified since then.
+
+I am sure that there still a couple of bugs left, but these should
+really be ironed out before long.
+
+
+The scripts support:
+
+*	Managing a stack of patches, applying and
+	removing patches from the stack, adding new patches, and
+	modifying them under a minimal kind of version control.
+
+*	They should be reasonably robust (hopefully!).
+
+*	Works for ordinary packages with only a few patches, and
+	also for the kernel with several hundered patches.
+
+*	The refresh script is careful not to drop any documentation
+	preceding the actual patch in the patch files.
+
+*	Simple to set up to get into "working mode" quickly.
+
+*	Any applied patch can be refreshed. This is trivial for the
+	topmost patch. If a file is modified by multiple patches,
+	all changes to the file affect only the topmost patch that
+	changed that file; patches further below will not include
+	any such changes when being rediffed.
+
+*	Several queries are supported, such as:
+	- topmost, preceding, following patch(es),
+	- files a patch creates or modifies
+	- patches that modify a given file
+	- changes a patch includes
+	- changes to a patch in the working directory
+
+
+For those of you who are interested, here is a how to get started, after
+grabbing and installing the patch-scripts package.
+
+
+Patching an ordinary package:
+
+	*	The usual RPM contains a tarball and a set of
+		patches. Unpack the tarball and copy the
+		patches into the patches/ sub-directory in the
+		package.
+
+	*	Create a file named series in the patches/ directory.
+		The series file contains the list of patches for
+		the package.
+
+		The patches are applied with -p1 by default; for -p0
+		or other style patches the series file must have
+		-p0 or similar after the patch name (but on the same
+		line).
+
+	*	Run `pushpatch -a' to apply all the patches. (Note
+		that the first run is slightly slower because lists
+		of files modified by the patches are created.)
+
+	*	Proceed with the README file in the patch-scripts
+		package.
+
+
+Patching the kernel:
+
+	*	The kernel lives in a CVS repository. Check out this
+		tree, say in a directory called $KCVS.
+
+	*	Unpack the kernel tarball, e.g., in $SRC. Create the
+		patches/ directory, and add soft links to the various
+		patches.* 	directories in the CVS into the patches
+		directory, e.g.,
+
+			$ cd $SRC/patches
+			$ ln -s $KCVS/patches.* .
+
+	*	Create a series file. This is done like that:
+
+			$ (cd $K_SRC/scripts ; \
+			   guards $(arch-symbols)) \
+			   < ../series.conf \
+			   > $SRC/patches/series
+
+	*	Change to $SRC, and run `pushpatch -qa' to apply
+		the patches (-q gives much less output).
+
+	*	As above, check the README in the patch-scripts
+		package, and try out what the individual scripts do.
+
+Cheers,
+Andreas.
+
+------------------------------------------------------------------
+ Andreas Gruenbacher                                SuSE Linux AG
+ mailto:agruen@suse.de                     Deutschherrnstr. 15-19
+ http://www.suse.de/                   D-90429 Nuernberg, Germany
diff --git a/doc/consider-for-docs-2.mail b/doc/consider-for-docs-2.mail
new file mode 100644
index 0000000..bad5c27
--- /dev/null
+++ b/doc/consider-for-docs-2.mail
@@ -0,0 +1,77 @@
+From: Andreas Gruenbacher <agruen@suse.de>
+Organization: SuSE Linux AG
+To: quilt-dev@nongnu.org
+Subject: My current quilt 0.21 :)
+Date: Tue, 28 Jan 2003 22:05:15 +0100
+
+Hi,
+
+I have copied my current version to http://www.suse.de/~agruen/quilt/.
+Should we use this version to reload the CVS? I think we are nearing
+the stable development phase, so now this seems to make some sense.
+
+There are surely many new bugs. Please help shake them out :)
+
+I also didn't merge several improvements. Would it be possible that
+people contribute patches for the things they want included? I really
+don't have so much time to work on features I don't need so much.
+
+Here's a list of changes (some probably forgotten):
+
+	- There now is a quilt wrapper script. All the individual
+	  commands are run from this script. This is a pain for frequently
+	  used commands like push and pop. (I am using shell aliases
+	  for some of them, which works well.)
+
+	- The scripts have been split into many separate commands that
+	  are intended to do only a few things each.
+
+		$ quilt
+		Usage: quilt command [-h] ...
+		Commands are:
+		        add      files   patches   refresh  setup
+		        applied  import  pop       remove   top
+		        delete   new     previous  rest
+		        diff     next    push      series
+
+	- .gz and .bz2 compressed patches are now suported.
+
+	- Quilt setup takes a series file, unpacks archive(s), and
+	  copies patches in.
+
+	- For those working with RPMs, quilt setup can process an RPM
+	  spec file and produces the series file from that (by running the
+	  %prep section).
+
+	- Quilt diff can diff a specified patch, diff the changes
+	  done relative to after a patch, and create combined diffs.
+
+	- Quilt series shows which patches are applied,
+	  which is the topmost, and which are unapplied:
+
+		$ quilt series -v
+		+ samba-2.2.4-smbadduser
+		+ samba-2.2.3-smbsh
+		+ samba-2.2.6-doc
+		+ samba-2.2.6-pdbedit
+		+ samba-2.2.7a-removeldap
+		= samba-2.2.5-linkvfs
+		  samba-2.2.7-nettime
+		  samba-2.2.7a-acl-mapping-fixes
+		  samba-2.2.7a-glibc
+
+	- Quilt files shows which files are new/removed by a patch (none
+	  are removed here):
+
+		$ quilt files -v
+		  fs/Config.in
+		  fs/Makefile
+		+ fs/posix_acl.c
+		+ include/linux/posix_acl.h
+
+	- Quilt import still sucks. Some ideas are found in the TODO.
+
+
+So long,
+Andreas.
+
diff --git a/doc/consider-for-docs-3.mail b/doc/consider-for-docs-3.mail
new file mode 100644
index 0000000..4672ae2
--- /dev/null
+++ b/doc/consider-for-docs-3.mail
@@ -0,0 +1,56 @@
+From: Andreas Gruenbacher <agruen@suse.de>
+Organization: SuSE Linux AG
+To: Andrew Morton <akpm@digeo.com>
+Subject: Patch scripts 0.9 improvements
+Date: Tue, 7 Jan 2003 14:39:16 +0100
+
+Hi Andrew,
+
+I still find your patch management scripts very useful. There are a fewg
+small limitations that still keeps me from using them for the SuSEg
+kernel at the moment. Particularly I need a way to have sub-directoriesg
+in patches/ etc.
+
+Please find some improvements and ideas in the attached patch.
+
+Also I think it would be a good idea to keep all (non-temporary)g
+information pertaining to a patch in a single file, rather thang
+splitting into .patch, .txt, .pc files. I have a format like theg
+following in mind. A simple parser for that proposed format is alsog
+attached.
+
+	Summary: Test patch
+	Author: Andreas Gruenbacher <agruen@suse.de>
+	URL: http://www.suse.de/
+
+	%description
+	DESCRIPTION
+
+	%files
+	FILE1
+	FILE2
+	FILE3
+
+	%patch
+	PATCH
+
+I think that it doesn't make much sense to keep the .pc files (%filesg
+section) as part of the patch set; they can easily be regenerated fromg
+the .patch files. They are useful while working on a set of patchesg
+though. What do you think of always generating them on the fly, insteadg
+of using import_patch? Then the patches and series files could beg
+checked out from CVS, and patching could start immediately; no cruftg
+would assemble in the .pc files.
+
+Another idea: Maybe all the *~* files could live in their own directoryg
+tree, e.g., under pc/? Then the file names would become more simple,g
+e.g., from file.x~patch to ps/patch/file.x and file.x~dir_patch tog
+dir/patch/file.x, with a smaller chance of clashes. What do you think
+of that?
+
+Cheers,
+Andreas.
+
+
+[ Attachments removed ]
+
diff --git a/docco.txt b/doc/docco.txt
index 1860f13..1860f13 100644
--- a/docco.txt
+++ b/doc/docco.txt