Libguestfs - Feb 2017 - [PATCH 1/2] dib: preserve xattrs and SELinux attributes when exporting as tar

When extracting the content of the guest as tar, save also the extended
attributes and the SELinux attributes.  This makes sure guests exported
as tar, tgz, and as docker image will work fine afterwards.

This is what disk-image-create does as well.
---
 dib/output_format_docker.ml | 3 ++-
 dib/output_format_tar.ml    | 3 ++-
 dib/output_format_tgz.ml    | 4 ++--
 3 files changed, 6 insertions(+), 4 deletions(-)

diff --git a/dib/output_format_docker.ml b/dib/output_format_docker.ml
index 5303cf9..903ac65 100644
--- a/dib/output_format_docker.ml
+++ b/dib/output_format_docker.ml
@@ -38,7 +38,8 @@ let docker_run_fs (g : Guestfs.guestfs) _ temp_dir      | Some
t -> t in
   message (f_"Importing the image to docker as '%s'")
docker_target;
   let dockertmp = Filename.temp_file ~temp_dir "docker."
".tar" in
-  g#tar_out ~excludes:[| "./sys/*"; "./proc/*" |]
"/" dockertmp;
+  g#tar_out ~excludes:[| "./sys/*"; "./proc/*" |]
~xattrs:true ~selinux:true
+    "/" dockertmp;
   let cmd = [ "sudo"; "docker"; "import";
dockertmp; docker_target ] in
   if run_command cmd <> 0 then exit 1
 
diff --git a/dib/output_format_tar.ml b/dib/output_format_tar.ml
index d8d5bfa..132532d 100644
--- a/dib/output_format_tar.ml
+++ b/dib/output_format_tar.ml
@@ -23,7 +23,8 @@ open Output_format
 
 let tar_run_fs (g : Guestfs.guestfs) filename _    message (f_"Compressing
the image as tar");
-  g#tar_out ~excludes:[| "./sys/*"; "./proc/*" |]
"/" filename
+  g#tar_out ~excludes:[| "./sys/*"; "./proc/*" |]
~xattrs:true ~selinux:true
+    "/" filename
 
 let fmt = {
   defaults with
diff --git a/dib/output_format_tgz.ml b/dib/output_format_tgz.ml
index a74a4a6..155afb5 100644
--- a/dib/output_format_tgz.ml
+++ b/dib/output_format_tgz.ml
@@ -23,8 +23,8 @@ open Output_format
 
 let tgz_run_fs (g : Guestfs.guestfs) filename _    message (f_"Compressing
the image as tar.gz");
-  g#tar_out ~excludes:[| "./sys/*"; "./proc/*" |]
~compress:"gzip"
-    "/" filename
+  g#tar_out ~excludes:[| "./sys/*"; "./proc/*" |]
~xattrs:true ~selinux:true
+    ~compress:"gzip" "/" filename
 
 let fmt = {
   defaults with
-- 
2.9.3

Improve the grammar, and the wording, and make few details explicit.
No change in the actual content.
---
 dib/virt-dib.pod | 29 ++++++++++++++++-------------
 1 file changed, 16 insertions(+), 13 deletions(-)

diff --git a/dib/virt-dib.pod b/dib/virt-dib.pod
index 7083e78..be7550a 100644
--- a/dib/virt-dib.pod
+++ b/dib/virt-dib.pod
@@ -206,7 +206,7 @@ B<docker>.
 
 =item C<qcow2> (enabled by default)
 
-QEMU's qcow2.
+QEMU's qcow2.  This output format requires the C<qemu-img> tool.
 
 =item C<raw>
 
@@ -251,9 +251,9 @@ Set the path in the host where cache the resources used by
the
 elements of the C<extra-data.d> phase.  The default is
 F<~/.cache/image-create>.
 
-Please note that most of the resources fetched after C<extra-data>
-will be cached in the helper drive specified with I<--drive>;
-see also L</HELPER DRIVE>.
+Please note that most of the resources fetched in phases other than
+C<extra-data.d> will be cached in the helper drive specified with
+I<--drive>; see also L</HELPER DRIVE>.
 
 =item B<--install-type> TYPE
 
@@ -311,7 +311,7 @@ C<DIB_OFFLINE> is set to C<1>, signaling the
elements that they
 should use only cached resources when available.  Note also that,
 unlike with C<diskimage-builder> where elements may still be able
 to access to the network even with C<DIB_OFFLINE=>, under virt-dib
-network will be fully unaccessible.
+network will not be accessible at all.
 
 =item B<--name> NAME
 
@@ -324,7 +324,7 @@ current directory:
 
 =item F<$NAME.ext>
 
-For each output format, a disk image named after the outout image
+For each output format, a file named after the output image
 with the extension depending on the format; for example:
 F<$NAME.qcow2>, F<$NAME.raw>, etc.
 
@@ -351,7 +351,7 @@ Not applicable in ramdisk mode, see L</RAMDISK
BUILDING>.
 Don't delete the output files on failure to build.  You can use this
 to debug failures to run scripts.
 
-The default is to delete the output file if virt-dib fails (or,
+The default is to delete the output files if virt-dib fails (or,
 for example, some script that it runs fails).
 
 =item B<-q>
@@ -445,7 +445,7 @@ Enable tracing of libguestfs API calls.
 
 Unlike with C<diskimage-builder>, the environment of the host is
 B<not> inherited in the appliance when running most of the elements
-(i.e. all the ones different than C<extra-data.d>).
+(i.e. all except the ones in the C<extra-data.d> phase).
 
 To set environment for the elements being run, it is necessary to tell
 virt-dib to use them, with the option I<--envvar>.  Such option
@@ -481,7 +481,7 @@ on it, or the first partition on it will be used.  A disk
with
 the latter configuration can be easily created with L<guestfish(1)>
 like the following:
 
- guestfish -N filename.img=fs:ext4:10G
+ guestfish -N filename.img=fs:ext4:10G exit
 
 The above will create a disk image called F<filename.img>, 10G big,
 with a single partition of type ext4;
@@ -517,7 +517,7 @@ cached in the home directory of the user running the various
tools.
 =item F</virt-dib-*.log>
 
 These are the logs of the elements being run within the libguestfs
-appliance, which means all the hooks except C<extra-data.d>.
+appliance, which means all the phases except C<extra-data.d>.
 
 =back
 
@@ -569,7 +569,8 @@ Because of virt-dib runs most of the elements in its own
appliance,
 all the tools and libraries used by elements running outside the
 guest (typically C<root.d>, C<block-device.d>, and
C<cleanup.d>)
 need to be present in the appliance as well.  In case they
-are not, scripts will fail mostly with a C<command not found> error.
+are not, scripts will fail typically with a C<command not found>
+error.
 
 For tools and libraries packaged by the distribution, the easy
 solution is to tell libguestfs to include additional packages in the
@@ -580,7 +581,7 @@ additional packages:
 
 The actual path to the F<supermin.d> directory depends on the
 distribution; additional files can list more packages, each in
-its own line.
+its own line.  For more details, see L<supermin(1)>.
 
 =head1 COMPARISON WITH DISKIMAGE-BUILDER
 
@@ -604,7 +605,9 @@ available in C<diskimage-builder>, while some have
different names:
  -o IMAGENAME                  --name IMAGENAME
  -p PACKAGE(S)                 --extra-packages PACKAGE(S)
  -t FORMAT(S)                  --formats FORMAT(S)
- -x                            --debug N
+ -x                            --debug 1
+ -x -x                         --debug 2
+ -x -x [-x ...]                --debug 3/4/etc
 
 =item
 
-- 
2.9.3

On Fri, Feb 17, 2017 at 02:44:45PM +0100, Pino Toscano
wrote:
> Improve the grammar, and the wording, and make few details explicit.
> No change in the actual content.
ACK series.

Rich.

-- 
Richard Jones, Virtualization Group, Red Hat http://people.redhat.com/~rjones
Read my programming and virtualization blog: http://rwmj.wordpress.com
Fedora Windows cross-compiler. Compile Windows programs, test, and
build Windows installers. Over 100 libraries supported.
http://fedoraproject.org/wiki/MinGW