Hi all,

I believe I've got most of the kinks worked out with the D-BUS/HAL
installation. I'm sure there will be more issues (?), however, most
things seem to work well, and are properly documented at this point.

Here is a summary. Note there is a question at the end of this message
that requires input from the group. *Please* respond on this issue.

D-BUS Instructions:

1) Created a "HAL Requirements" section below the dependencies which
identifies what you must do to support HAL.

  a) Install GLib-2 before building D-BUS so that the libdbus-glib
  library is built (required by the HAL daemon)

  b) Install Pyrex (requires Python) before building D-BUS so that
  the D-BUS Python bindings are built. It is identified that this
  is optional and is only needed if you want the hal-device-manager
  program to work.

  c) Added text that says to remove the --disable-python switch on
  the configure command if you have Pyrex installed.

  d) Unfortunately, the Pyrex web site
  (http://nz.cosc.canterbury.ac.nz/~greg/python/Pyrex/) seems to be
  unavailable and the link in the book is broken. Googling show that
  this should be a good link, and as best I can tell should still be
  Pyrex web site. I'll wait a few days and see if it comes up before
  doing anything about it. I'll BZ it for now.

HAL Instructions:

1) Added a note after the D-BUS dependency that says "(see the HAL
Requirements)", and the "HAL Requirements" is a link to the D-BUS
page where the HAL stuff is outlined. This way there is information
on both the D-BUS and HAL pages that identify the special needs
building D-BUS to support HAL.

2) Moved Python to the Required Dependencies.

3) Moved GLib2 to the Required Dependencies.

4) Removed Expat from Optional Dependencies as it was redundant. It
must already be installed when you install XML::Parser.

5) Created a new section after "Command Explanations" and before
"Configuration" titled "Run-Time Dependencies". This section is where
the GNOME-Python requirement so that the hal-device-manager program
will work is listed. It mentions that it is a significant
installation and lists the build order of the many packages required
before you can even install the actual GNOME-Python package.

6) Changed the sed command so that "comment=managed" is inserted in
the /etc/fstab file when HAL updates the file.

7) Added information in the Configuration section about creating and
populating the /etc/hal/fdi/policy/30user/30user-policy.fdi file to
support the iocharset and codepage stuff.

8) I haven't done anything about trying to make it where the codepage
stuff is added *only* if a VFAT partition is discovered. Alex will
have to get back on this. I don't know if it is required or not.

#################################################
This is the question the group needs to help with
#################################################

9) Currently the book has instructions to create a hal user and group.
This really isn't necessary, and here is why. The haldaemon boot
script runs the hald daemon with the --retain-privileges switch so
the root user never hands the daemon off to the unprivileged user.

I tried everything I could think of to circumvent this, however no
matter what I did the unprivileged user could not perform any
callouts (such as update /etc/fstab). I went to extremes to see if
I could make it work and could not. This confirms information I
read about on the mailing list.

So, the daemon must run as root meaning we don't *really* need the
hal user and group. Here are the issues with that.

1) The documentation says to create the user and group.
2) Text at the end of the configure command say to create it before
   issuing 'make install'.
3) The configuration file has the unprivileged user's name in it,
   so when you start the D-BUS (yes, the D-BUS and not the HAL)
   daemon, there is a message. This could be fixed by editing the
   config file.
4) The HAL folks are working to see if they can't implement some
   solution so that an unprivileged user can run the daemon.
   Many the unpriv user and group *may* be needed in the future.

Now, I don't like the fact that we say to create the user and group
when we know it isn't used for anything. So, finally the question:

Should we continue to have instructions in the book to create this
user and group?

Or should we not create the user and group and make the very minor
change to the config file so the message is not displayed (see
issue #3 above). A very simple sed takes care of it.

Sorry for the long message, but this summary is mostly for DBUS/HAL
information archive purposes.

-- 
Randy

rmlscsi: [GNU ld version 2.15.94.0.2 20041220] [gcc (GCC) 3.4.3]
[GNU C Library stable release version 2.3.4] [Linux 2.6.10 i686]
10:40:00 up 98 days, 20:04, 3 users, load average: 0.03, 0.02, 0.22
-- 
http://linuxfromscratch.org/mailman/listinfo/blfs-dev
FAQ: http://www.linuxfromscratch.org/blfs/faq.html
Unsubscribe: See the above information page

Reply via email to