[Bash-completion-devel] mount, umount completions
Igor Murzov
e-mail at date.by
Sun Nov 20 16:35:28 UTC 2011
On Sun, 20 Nov 2011 12:08:13 +0100
Raphaël Droz <raphael.droz+floss at gmail.com> wrote:
> On Sun, Nov 20, 2011 at 02:08:20AM +0300, Igor Murzov wrote:
> > Hi, list.
> >
> > There are new linux-specific mount and umount completions in the mount
> > branch. Please review and test. The new completions address Alioth
> > bugs #313166, #313214 and #313437.
>
> just gave a quick look and it seems awesome !
>
> some tiny notes:
>
> * smbclient presence may be tested like done for showmount (or is
> 2>/dev/null enough ?)
Yes, I think that 2>/dev/null is ok here.
> * for umount I would rather strip device path from the list except if
> ^/dev/ is in the command line.
This is sane, but printing device paths could be probably useful in some cases.
> (and only print non-system mount point ?)
> * then $(lsblk -o MOUNTPOINT) may be useful
I don't think this is a good idea, we don't do arguments validation. Also lsblk output misses a lot of mount points for me:
----------------------------------
$ lsblk
NAME MAJ:MIN RM SIZE RO MOUNTPOINT
sda 8:0 0 111,8G 0
├─sda1 8:1 0 31,4M 0 /boot
├─sda2 8:2 0 502M 0 [SWAP]
├─sda3 8:3 0 11,7G 0 /
└─sda4 8:4 0 99,6G 0 /home
sr0 11:0 1 1024M 0
sdb 8:16 0 74,5G 0
└─sdb1 8:17 0 74,5G 0
sdc 8:32 1 1,9G 0
└─sdc1 8:33 1 1,9G 0
$ mount
/dev/sda3 on / type reiserfs (rw,relatime)
proc on /proc type proc (rw)
sysfs on /sys type sysfs (rw)
/dev/sda1 on /boot type ext2 (rw)
/dev/sda4 on /home type reiserfs (rw)
/dev/sdb1 on /home/lxuser/Documents/torrent type btrfs (rw,subvol=torrent)
/dev/sdb1 on /usr/src type btrfs (rw,subvol=src)
/dev/sdb1 on /var/mirror type btrfs (rw,subvol=slackware-mirror)
/dev/sdb1 on /tmp/tgz type btrfs (rw,subvol=build)
/home/lxuser/repository on /srv/ftp/slack-soft type none (rw,bind)
----------------------------------
> * There should be a way to test if we are completing a device (_filedir)
> or a mount point (_filedir -d) (at the end of the completion)
This would probably require full reimplementation of mount's arguments parser :) But if someone can do this, that's would be cool.
> * why not dropping some short options from the list (like --fork and
> -F), while still completing the argument for both form.
Ok, that's not a big deal, will do. Though I don't understand why drop short options.
> * is ' sshfs#HOST:RPATH PATH ' syntax in the scope of the device path
> completion within mount ? (so would be obex#, ... and a bunch of other
> fuse FS :) )
Hm... I don't personally use such things. May be someone more experienced in that things would like to write some code to add sshfs-like syntax support.
> thank you !
You are welcome :)
-- Igor
More information about the Bash-completion-devel
mailing list