dragoran wrote:

Ralf Ertzinger wrote:

...

Hi.

For some time now plugging in a new device does not create an entry in /etc/fstab any more. The device shows up in nautilus just fine and can be mounted and unmounted. Working from the command line is quite a hassle, though (just like in pre-udev-days).

Is this intentional?

gnome-mount is used for mounts in the desktop now, so no fstab entrys are needed. why are you require them? you only have to know where the device is mounted and which device it is if you are working from the command line.

Which some of us do quite a bit of for specialized work. I can't use a Linux distro that doesn't have reasonable functionality when worked from the command line, e.g. from a terminal window, from a console, or via an ssh connection.

Hi.

On Wed, 22 Feb 2006 14:10:42 -0500, Jeff Spaleta wrote:

console is unreasonable in your estimation? Using the new gnome-mount command on the cmdline in these situations to mimic what the gnome desktop's automounter does is unreasonable?

gnome-mount --help

Hey, I need the complete Gnome library zoo just to mount a USB stick! This is progress for sure.

Hi.

On Thu, 23 Feb 2006 07:47:34 -0500, David Zeuthen wrote:

Don't be silly. You can still use /etc/fstab and/or mount(1). Even with the new /disk/disk/by-label etc. symlinks you can construct a persistent entry in /etc/fstab.

Nonetheless, what was wrong with the old behaviour (dropping a directory in /media for each plugged in device)? I do not even particulary care about the entries in /etc/fstab, as long as "mount /media/usbdisk" does what it is supposed to do.

Can I restore the old behaviour somehow?

David Zeuthen wrote:

On Thu, 2006-02-23 at 14:19 +0100, Ralf Ertzinger wrote:

...

Hi.

On Thu, 23 Feb 2006 07:47:34 -0500, David Zeuthen wrote:

...

Don't be silly. You can still use /etc/fstab and/or mount(1). Even with the new /disk/disk/by-label etc. symlinks you can construct a persistent entry in /etc/fstab.

Nonetheless, what was wrong with the old behaviour (dropping a directory in /media for each plugged in device)? I do not even particulary care about the entries in /etc/fstab, as long as "mount /media/usbdisk" does what it is supposed to do.

The thing is... we don't know a priori what mount point (and mount options) is to be used. With gnome-mount, this decision now comes from the desktop session. Right now the logic is hard wired into the gnome-mount sources itself but we're working on fixing this to read it from gconf; see

http://freedesktop.org/~david/gnome-mount-properties-0.01.png

for some work in progress.

...

Can I restore the old behaviour somehow?

Not really, no. But nothing prevents you from write an fstab-sync like program yourself and submitting it to Fedora Extras.

David

Unfortunately, the arguments are piling on for re-evaluating our distro choice here. Another thread in this group suggested that the poster get one of the handy linux reference books to solve some mounting problems. I bet they say "mount" and "fstab". (Has anyone checked wheter posix mandates "mount" for mounting?)

On 2/23/06, David Zeuthen david@fubar.dk wrote:

http://freedesktop.org/~david/gnome-mount-properties-0.01.png

for some work in progress.

Is there be a cmdline oriented tool planned to do what the gui gnome-mount-properties does to balance the functionality of gnome-mount on the cmdline? Or will we be asked to use gconftool to edit these keys from the cmdline?

-jef

David Zeuthen (david@fubar.dk) said:

...

Can I restore the old behaviour somehow?

Not really, no. But nothing prevents you from write an fstab-sync like program yourself and submitting it to Fedora Extras.

That's a cheat, though - the introduction is removing functionality from the OS.

This is all documented in the release notes, right?

Bill

On 2/23/06, David Zeuthen david@fubar.dk wrote:

You can also use dbus-send in conjunction with hal-find-by-property... it's not exactly rocket science

If I had a better personal understanding of the capabilities of the dbus commandline tools like dbus-send that would help. Are there existing examples of dbus-send usage that do useful work in the wild? I honestly do not grok dbus communication well enough to be able to parse the example you give usefully from a "get crap done with a shell script" sysadmin point of view. And the errors spawned the what I assume to be non-functional example in the dbus-send on the fc4 manpage leaves my mouth agape and my eyes glassy.

If you know of something better than http://dbus.freedesktop.org/doc/dbus-tutorial.html to help sysadmins learn enough about dbus to use dbus-send effectively please let me know. Especially any existing generally useful scripts that exist to do useful work.

-jef

Hi Jef,

On Thu, 2006-02-23 at 11:40 -0500, Jeff Spaleta wrote:

If I had a better personal understanding of the capabilities of the dbus commandline tools like dbus-send that would help. Are there existing examples of dbus-send usage that do useful work in the wild?

This document is out of date

http://webcvs.freedesktop.org/*checkout*/hal/hal/doc/spec/hal-spec.html

but it is where stuff like this will end of for HAL... but we're pre-1.0 so we're working more on code than documentation. We take patches though :-). Send them to

http://lists.freedesktop.org/mailman/listinfo/hal

We can also answer specific questions about what parameters the Mount() method takes. And I'm normally a lot more reasonable on that list than here :-).... The best advice I can give you right now... is to read the gnome mount source... here

http://cvs.gnome.org/viewcvs/*checkout*/gnome-mount/src/gnome-mount.c

it will show you what the parameters are... and what exceptions you need to catch.. Hope this helps...

Btw /etc/pm/hooks/00NetworkManager is one example of using dbus-send.

I honestly do not grok dbus communication well enough to be able to parse the example you give usefully from a "get crap done with a shell script" sysadmin point of view. And the errors spawned the what I assume to be non-functional example in the dbus-send on the fc4 manpage leaves my mouth agape and my eyes glassy.

If you know of something better than http://dbus.freedesktop.org/doc/dbus-tutorial.html to help sysadmins learn enough about dbus to use dbus-send effectively please let me know. Especially any existing generally useful scripts that exist to do useful work.

From another mail: Actually since we are talking about scripting dbus actions with the cmdline tools... how do i get access to dbus's introspection ability via the cmdline tools so that as a sysadmin I can explore the object path and valid messages?

There is not really an easy way to do this yet. But someday we'll have nice introspection and a generic browser UI for these.

David

David Zeuthen (david@fubar.dk) said:

...

That's a cheat, though - the introduction is removing functionality from the OS.

Strong disagreement.

Prior releases automatically mounted filesystems in all situations; now they are only mounted in the desktop.

You can use gnome-mount and/or mount(1) - they work just fine and gnome-mount will (at least in the future) do so much more than we could ever do with the fstab-sync approach. KDE also got some stuff for this but I don't follow KDE development very closely, sorry.

You can also use dbus-send in conjunction with hal-find-by-property... it's not exactly rocket science

    [davidz@daxter ~]$ dbus-send --system --dest=org.freedesktop.Hal --print-reply /org/freedesktop/Hal/devices/volume_uuid_43F1_517C org.freedesktop.Hal.Device.Volume.Unmount array:string:
    method return sender=:1.2 -> dest=:1.62
     uint32 0
    [davidz@daxter ~]$ dbus-send --system --dest=org.freedesktop.Hal --print-reply /org/freedesktop/Hal/devices/volume_uuid_43F1_517C org.freedesktop.Hal.Device.Volume.Mount string: string: array:string:
    method return sender=:1.2 -> dest=:1.65
     uint32 0

This isn't exactly common use currently, though - I suspect more people understand basic perl line noise than this. :)

Is there a way to easily query a d-bus object (from the command line, or even python -c '...') for the supported methods?

...

This is all documented in the release notes, right?

If you think it's necessary.. feel free to add it.

I really think we could use one.

Bill

Bill Nottingham wrote:

...

...

This is all documented in the release notes, right?

If you think it's necessary.. feel free to add it.

I really think we could use one.

Bill

Available at http://fedoraproject.org/wiki/Docs/Beats/OverView. Will be in the overview section in the FC5 release notes in the ISO images.

On Fri, 2006-03-03 at 08:12 +1100, Rodd Clarkson wrote:

On Thu, 2006-02-23 at 08:31 -0500, David Zeuthen wrote:

...

see

http://freedesktop.org/~david/gnome-mount-properties-0.01.png

for some work in progress.

Okay, this screen shot is just plain teasing.

Would you like to explain to the masses (or maybe just me) how you got SD/MMC support working in Fedora.

I notice that 'modprobe mmc_block' now installs the modules needed, but beyond this it's not doing anything.

Doesn't notice the insertion of cards into the reader (which lspci reports as a):

03:01.2 Class 0805: Ricoh Co Ltd R5C822 SD/SDIO/MMC/MS/MSPro Host Adapter (rev 17)

Nothing shows up in /dev

Nothing mounts, that's for certain.

This isn't the plug N play I'm getting used to in Fedora, so what do I need to do. It would be great to have support for SD/MMC in FC5 and it certainly looks like we're close.

Hmmm, could this be because the kernel doesn't have sdhci support?

R.

On Mon, 2006-03-06 at 11:00 +0100, Matthias Saou wrote:

Rodd Clarkson wrote :

...

Would you like to explain to the masses (or maybe just me) how you got SD/MMC support working in Fedora.

I notice that 'modprobe mmc_block' now installs the modules needed, but beyond this it's not doing anything.

Doesn't notice the insertion of cards into the reader (which lspci reports as a):

03:01.2 Class 0805: Ricoh Co Ltd R5C822 SD/SDIO/MMC/MS/MSPro Host Adapter (rev 17)

Nothing shows up in /dev

Nothing mounts, that's for certain.

This isn't the plug N play I'm getting used to in Fedora, so what do I need to do. It would be great to have support for SD/MMC in FC5 and it certainly looks like we're close.

Well, these might be separate issues, as many USB connected SD/MMC readers work by default (only with unencrypted media, I guess), whereas the device you list reminds me a lot of the internal reader I have in my Dell X1 laptop, which simply doesn't work with Linux, from the lack of a device driver... it's a quite different device (PCI connected, not USB).

Yes, I've got one of these in my Dell Inspiron 9300 at the moment.

Fortunately for us, a driver has been developed and has been submitted for inclusion in the kernel. It's currently in -mm and should make it into 2.6.16 or 2.6.17.

Personally, I'd love to see it included as a patch into the kernel, but I know Dave's perspective on it and respect that position. I also know that occasionally Dave makes exceptions to the rule, so while I hoped that this might be one of them, I'm also not going to start holding my breath.

Rodd

On Thu, 2006-02-23 at 10:39 +0100, Ralf Ertzinger wrote:

Hi.

On Wed, 22 Feb 2006 14:10:42 -0500, Jeff Spaleta wrote:

...

console is unreasonable in your estimation? Using the new gnome-mount command on the cmdline in these situations to mimic what the gnome desktop's automounter does is unreasonable?

gnome-mount --help

Hey, I need the complete Gnome library zoo just to mount a USB stick! This is progress for sure.

Last time I looked, gnome-mount was only using GTK+.

Hi.

On Thu, 23 Feb 2006 11:01:03 -0500, David Zeuthen wrote:

I think gnome-mount (pre 0.4 CVS snapshot) in Rawhide currently only requires GTK+. Upstream (what will be 0.4) currently pulls in libgnomeui and libgnome-keyring for passworded media.

[sun@dhcp05 ~ :) 3]$ rpm -q gnome-mount gnome-mount-0.4-0.cvs20060213.1 [sun@dhcp05 ~ :) 4]$ ldd /usr/bin/gnome-mount linux-gate.so.1 => (0x007f8000) libdbus-glib-1.so.2 => /usr/lib/libdbus-glib-1.so.2 (0x06295000) libhal-storage.so.1 => /usr/lib/libhal-storage.so.1 (0x00dec000) libhal.so.1 => /usr/lib/libhal.so.1 (0x0625e000) libdbus-1.so.2 => /lib/libdbus-1.so.2 (0x00511000) libgtk-x11-2.0.so.0 => /usr/lib/libgtk-x11-2.0.so.0 (0x05e3a000) libgdk-x11-2.0.so.0 => /usr/lib/libgdk-x11-2.0.so.0 (0x00668000) libatk-1.0.so.0 => /usr/lib/libatk-1.0.so.0 (0x0073e000) libgdk_pixbuf-2.0.so.0 => /usr/lib/libgdk_pixbuf-2.0.so.0 (0x0064e000) libm.so.6 => /lib/libm.so.6 (0x009e7000) libpangocairo-1.0.so.0 => /usr/lib/libpangocairo-1.0.so.0 (0x0075c000) libpango-1.0.so.0 => /usr/lib/libpango-1.0.so.0 (0x006f4000) libcairo.so.2 => /usr/lib/libcairo.so.2 (0x00586000) libgobject-2.0.so.0 => /usr/lib/libgobject-2.0.so.0 (0x00c37000) libgmodule-2.0.so.0 => /usr/lib/libgmodule-2.0.so.0 (0x00c7a000) libdl.so.2 => /lib/libdl.so.2 (0x00a0e000) libglib-2.0.so.0 => /usr/lib/libglib-2.0.so.0 (0x00b82000) libc.so.6 => /lib/libc.so.6 (0x008bb000) libcap.so.1 => /lib/libcap.so.1 (0x00648000) libnsl.so.1 => /lib/libnsl.so.1 (0x00c16000) libX11.so.6 => /usr/lib/libX11.so.6 (0x00a36000) libfontconfig.so.1 => /usr/lib/libfontconfig.so.1 (0x00398000) libXext.so.6 => /usr/lib/libXext.so.6 (0x00b35000) libXrender.so.1 => /usr/lib/libXrender.so.1 (0x003d7000) libXinerama.so.1 => /usr/lib/libXinerama.so.1 (0x00dfa000) libXi.so.6 => /usr/lib/libXi.so.6 (0x00767000) libXrandr.so.2 => /usr/lib/libXrandr.so.2 (0x00642000) libXcursor.so.1 => /usr/lib/libXcursor.so.1 (0x00612000) libXfixes.so.3 => /usr/lib/libXfixes.so.3 (0x005da000) /lib/ld-linux.so.2 (0x0089e000) libpangoft2-1.0.so.0 => /usr/lib/libpangoft2-1.0.so.0 (0x005e8000) libfreetype.so.6 => /usr/lib/libfreetype.so.6 (0x00d7c000) libz.so.1 => /usr/lib/libz.so.1 (0x00a14000) libpng12.so.0 => /usr/lib/libpng12.so.0 (0x0027f000) libXau.so.6 => /usr/lib/libXau.so.6 (0x00a31000) libXdmcp.so.6 => /usr/lib/libXdmcp.so.6 (0x00a29000) libexpat.so.0 => /lib/libexpat.so.0 (0x00d59000)