New packages for Ubuntu to test out before final release - 3.0.2-1

[fragadelic]

You have been a good help in the past with testing and providing feedback.

These are the 2 facts we are dealing with:

1 - you have autologin set on your system - this is your issue and neither I nor the Ubuntu developers will do anything to change it.

2 - remastersys-gui on 64-bit is not saving the config file properly - this is MY issue and I never said it wasn't but I need more info.

[fragadelic]

ourasi - remove the remastersys-gui package and install the remastersys-gtk I just put up in the first post.  I found an old drive and tried Xubuntu 12.04 and it was a mess with the new gui.  I couldn't even get it to run but the python one from 3.0.1 worked fine so I have just used that one for now.

My apologies for the issues with the gui on your system relating toe the config file issue.  I still recommend you change your autologin choices moving forward to ensure it works as the Ubuntu developers designed it.

[ourasi]

I'm sorry I ever mentioned autologin - you seem to have some strange fixation about that.

1. I made once more clean install from 32-bit Lubuntu 12.04 daily, user "ourasi", no autologin, as you maybe can see here:
http://s20.photobucket.com/albums/b239/ourasi/Remastersys/?action=view&current=-lightdmconf-_002.png

2. Updated system, installed some programs and then latest remastersys.
Made settings for Remastersys like this:
http://s20.photobucket.com/albums/b239/ourasi/Remastersys/?action=view&current=remastersysconf_001.png

3. Build dist-iso with remastersys, copied it to usb-stick and booted - and got this:
http://s20.photobucket.com/albums/b239/ourasi/Remastersys/?action=view&current=this_001.png

So I was user "this" in home folder "this", why??? What I have made possibly wrong again?

4. This is what is in file /etc/casper.conf in my Live system:
http://s20.photobucket.com/albums/b239/ourasi/Remastersys/?action=view&current=-casperconf-_002.png

I reported that error first there: I noticed some other problem:, after which you started insulting me.

[fragadelic]

I reverted the gui back to remastersys-gtk.  Check the first post in this thread for the link to the package.  Uninstall remastersys-gui and install remastersys-gtk and it will be fine.

I don't believe I ever insulted you.  I merely pointed out that part of your autologin issue was self inflicted due to setting autologin on your system you are creating the distro from.  Search the forum and you will mention of that issue.  For some reason the new gui behaves starngely on your system so I have assume it will behave strangely on other ones as well so I reverted back to the tried and true remastersys-gtk frontend.

[ourasi]

I reverted the gui back to remastersys-gtk.  Check the first post in this thread for the link to the package.  Uninstall remastersys-gui and install remastersys-gtk and it will be fine.

Remastersys-gtk seems to work, also with 64-bit versions.

I don't believe I ever insulted you.  I merely pointed out that part of your autologin issue was self inflicted due to setting autologin on your system you are creating the distro from.  Search the forum and you will mention of that issue.  For some reason the new gui behaves starngely on your system so I have assume it will behave strangely on other ones as well so I reverted back to the tried and true remastersys-gtk frontend.

Sorry, I was very tired and maybe should have used better words.

Anyway, I think you still don't understand that error:

Here is again two screen shots (test made Lubuntu 12.04 without autologin):
http://s20.photobucket.com/albums/b239/ourasi/Remastersys/?action=view&current=Remastersys_001.png

and:
http://s20.photobucket.com/albums/b239/ourasi/Remastersys/?action=view&current=fraga_001.png

You should clearly see that Remastersys CD Label's first word (Fraga) becomes Live Users name and home folder - not the name I wanted (ourasi). This happens every time (in 12.04, even without gui) and it has nothing to do with my habit of using autologin.

Difference between autologin on/off is that last one logs automatically with CD Label name, with first one I must give that name or log in as guest (if they are not same as autologin name).

I am not a programmer, and don't know which one (remastersys, casper or lightdm) causes this. Latest remastersys seems to work without this error with older versions of Lubuntu et al. which use LXDM.


Hope this helps you.

[fragadelic]

Ah - I think I finally understand.  It looks like it might be an issue with your locale.  Can you try putting the locale statement you used to fix the  working folder check and see if that fixes it?

Just to be sure, can you show me the /etc/remastersys.conf file?  If it is wrong in the remastersys.conf file then it is an issue with the gui somehow but if the remastersys.conf file is ok then it is an issue with the /usr/bin/remastersys main script.  I have been unable to duplicate your issue so I think it is definitely relating to locale.

Thanks for sticking with this and not giving up.  Now that I have a clearer picture of what is happening I'm sure we can figure it out and get it corrected.

By the way, you will notice in this new version that the check and erroring out of the working folder filesystem type has been removed to just print it out in the log file so I can see it.

[ourasi]

Ah - I think I finally understand.  It looks like it might be an issue with your locale.  Can you try putting the locale statement you used to fix the  working folder check and see if that fixes it?

I dont't quite understand. I tried this I solved Type/Tyyppi error, but noticed no difference.

Just to be sure, can you show me the /etc/remastersys.conf file?  If it is wrong in the remastersys.conf file then it is an issue with the gui somehow but if the remastersys.conf file is ok then it is an issue with the /usr/bin/remastersys main script.  I have been unable to duplicate your issue so I think it is definitely relating to locale.

Remastersys.conf and casper.conf files seem to be OK in source installation:
http://s20.photobucket.com/albums/b239/ourasi/Remastersys/?action=view&current=Valinta_002.png

its only in Live boot system casper.conf has been changed, like this:
http://s20.photobucket.com/albums/b239/ourasi/Remastersys/?action=view&current=Valinta_001.png

[fragadelic]

That looks like a bug with casper especially if the config files for remastersys.conf and casper.conf are normal.

Can you check /home/remastersys/remastersys/dummysys/etc/casper.conf as well. If that one is the same as your system then it is an issue with casper for sure but I think it might not be.

[ourasi]

Sorry about that, I forgot to say that it's also OK.

[fragadelic]

This isn't good.  That is a casper bug from what I can tell.

Let me look into it further.

[fragadelic]

Ok I figured it out.

This is not a bug with remastersys nor is it a bug with casper.  This is how the Ubuntu developers want it.

Here is the /usr/share/initramfs-tools/scripts/casper file which is responsible for the live boot and setup.  I highlighted in red what they changed and you will see it clearly.  This is what they want it to do.

Code:

#!/bin/sh

# set -e

export PATH=/usr/bin:/usr/sbin:/bin:/sbin

mountpoint=/cdrom
LIVE_MEDIA_PATH=casper
[ -f /conf/uuid.conf ] && UUID="$(cat /conf/uuid.conf)"

root_persistence="casper-rw"
home_persistence="home-rw"
root_snapshot_label="casper-sn"
home_snapshot_label="home-sn"

USERNAME=casper
USERFULLNAME="Live session user"
HOST=live
BUILD_SYSTEM=Custom

mkdir -p $mountpoint
tried=/tmp/tried

[ -f /etc/casper.conf ] && . /etc/casper.conf
export USERNAME USERFULLNAME HOST BUILD_SYSTEM

. /scripts/casper-helpers

if [ ! -f /casper.vars ]; then
    touch /casper.vars
fi

parse_cmdline() {
    for x in $(cat /proc/cmdline); do
        case $x in
            showmounts|show-cow)
                export SHOWMOUNTS='Yes' ;;
            persistent)
                export PERSISTENT="Yes" ;;
            nopersistent)
                export PERSISTENT="" ;;
            persistent-path=*)
                export PERSISTENT_PATH="${x#persistent-path=}" ;;
            union=*)
                export UNIONFS="${x#union=}";;
            ip=*)
                STATICIP=${x#ip=}
                if [ "${STATICIP}" = "" ]; then
                    STATICIP="frommedia"
                fi
                export STATICIP ;;
            uuid=*)
                UUID=${x#uuid=} ;;
            ignore_uuid)
                UUID="" ;;
            live-media-path=*)
                LIVE_MEDIA_PATH="${x#live-media-path=}"
                export LIVE_MEDIA_PATH
                echo "export LIVE_MEDIA_PATH=\"$LIVE_MEDIA_PATH\"" >> /etc/casper.conf ;;
            toram)
                export TORAM="Yes" ;;
            todisk=*)
                export TODISK="${x#todisk=}" ;;
        esac
    done
    if [ "${UNIONFS}" = "" ]; then
        export UNIONFS="DEFAULT"
    fi
}

is_casper_path() {
    path=$1
    if [ -d "$path/$LIVE_MEDIA_PATH" ]; then
        if [ "$(echo $path/$LIVE_MEDIA_PATH/*.squashfs)" != "$path/$LIVE_MEDIA_PATH/*.squashfs" ] ||
            [ "$(echo $path/$LIVE_MEDIA_PATH/*.ext2)" != "$path/$LIVE_MEDIA_PATH/*.ext2" ] ||
            [ "$(echo $path/$LIVE_MEDIA_PATH/*.dir)" != "$path/$LIVE_MEDIA_PATH/*.dir" ]; then
            return 0
        fi
    fi
    return 1
}

matches_uuid() {
    if [ -z "$UUID" ]; then
        return 0
    fi
    path="$1"
    for try_uuid_file in "$path/.disk/casper-uuid"*; do
        [ -e "$try_uuid_file" ] || continue
        try_uuid="$(cat "$try_uuid_file")"
        if [ "$UUID" = "$try_uuid" ]; then
            return 0
        fi
    done
    return 1
}

get_backing_device() {
    case "$1" in
        *.squashfs|*.ext2)
            echo $(setup_loop "$1" "loop" "/sys/block/loop*")
            ;;
        *.dir)
            echo "directory"
            ;;
        *)
            panic "Unrecognized casper filesystem: $1"
            ;;
    esac
}

match_files_in_dir() {
    # Does any files match pattern $1 ?

    local pattern="$1"
    if [ "$(echo $pattern)" != "$pattern" ]; then
        return 0
    fi
    return 1
}

mount_images_in_directory() {
    directory="$1"
    rootmnt="$2"
    if match_files_in_dir "$directory/$LIVE_MEDIA_PATH/*.squashfs" ||
        match_files_in_dir "$directory/$LIVE_MEDIA_PATH/*.ext2" ||
        match_files_in_dir "$directory/$LIVE_MEDIA_PATH/*.dir"; then
        setup_unionfs "$directory/$LIVE_MEDIA_PATH" "$rootmnt"
    else
        :
    fi
}

is_nice_device() {
    sysfs_path="${1#/sys}"
    if /sbin/udevadm test-builtin path_id "${sysfs_path}" | egrep -q "ID_PATH=(usb|pci-[^-]*-(ide|sas|scsi|usb|virtio)|platform-sata_mv|platform-orion-ehci|platform-mmc|platform-mxsdhci|platform-omap_hsmmc|platform-sdhci-tegra)"; then
        return 0
    fi
    if echo ${sysfs_path} | grep -q "^/block/dm-"; then
        return 0
    fi
    return 1
}

copy_live_to() {
    copyfrom="${1}"
    copytodev="${2}"
    copyto="${copyfrom}_swap"

    size=$(fs_size "" ${copyfrom} "used")

    if [ "${copytodev}" = "ram" ]; then
        # copying to ram:
        freespace=$(awk '/^MemFree:/{f=$2} /^Cached:/{c=$2} END{print f+c}' /proc/meminfo)
        mount_options="-o size=${size}k"
        free_string="memory"
        fstype="tmpfs"
        dev="/dev/shm"
    else
        # it should be a writable block device
        if [ -b "${copytodev}" ]; then
            dev="${copytodev}"
            free_string="space"
            fstype=$(get_fstype "${dev}")
            freespace=$(fs_size "${dev}")
        else
            [ "$quiet" != "y" ] && log_warning_msg "${copytodev} is not a block device."
            return 1
        fi
    fi
    if [ "${freespace}" -lt "${size}" ] ; then
        [ "$quiet" != "y" ] && log_warning_msg "Not enough free ${free_string} (${freespace}k > ${size}k) to copy live media in ${copytodev}."
        return 1
    fi

    # begin copying..
    mkdir "${copyto}"
    echo "mount -t ${fstype} ${mount_options} ${dev} ${copyto}"
    mount -t "${fstype}" ${mount_options} "${dev}" "${copyto}"
    cp -a ${copyfrom}/* ${copyto}
    if [ -e ${copyfrom}/.disk ]; then
        cp -a ${copyfrom}/.disk ${copyto}
    fi
    umount ${copyfrom}
    mount -r -o move ${copyto} ${copyfrom}
    rmdir ${copyto}
    return 0
}

do_netmount() {
    rc=1

    modprobe "${MP_QUIET}" af_packet # For DHCP

    /sbin/udevadm trigger
    /sbin/udevadm settle

    configure_networking
    export DEVICE

    if [ "${NFSROOT}" = "auto" ]; then
        NFSROOT=${ROOTSERVER}:${ROOTPATH}
    fi

    [ "$quiet" != "y" ] && log_begin_msg "Trying netboot from ${NFSROOT}"

    if [ "${NETBOOT}" != "nfs" ] && do_cifsmount ; then
        rc=0
    elif do_nfsmount ; then
        NETBOOT="nfs"
        export NETBOOT
        rc=0
    fi

    [ "$quiet" != "y" ] && log_end_msg
    return ${rc}
}

do_nfsmount() {
    rc=1
    modprobe "${MP_QUIET}" nfs
    if [ -z "${NFSOPTS}" ]; then
        NFSOPTS=""
    fi

    [ "$quiet" != "y" ] && log_begin_msg "Trying nfsmount -o nolock -o ro ${NFSOPTS} ${NFSROOT} ${mountpoint}"
    # FIXME: This while loop is an ugly HACK round an nfs bug
    i=0
    while [ "$i" -lt 60 ]; do
        nfsmount -o nolock -o ro ${NFSOPTS} "${NFSROOT}" "${mountpoint}" && rc=0 && break
        sleep 1
        i="$(($i + 1))"
    done
    return ${rc}
}

do_cifsmount() {
    rc=1
    if [ -x "/sbin/mount.cifs" ]; then
        if [ -z "${NFSOPTS}" ]; then
            CIFSOPTS="-ouser=root,password="
        else
            CIFSOPTS="${NFSOPTS}"
        fi

        [ "$quiet" != "y" ] && log_begin_msg "Trying mount.cifs ${NFSROOT} ${mountpoint} ${CIFSOPTS}"
        modprobe "${MP_QUIET}" cifs

        if mount.cifs "${NFSROOT}" "${mountpoint}" "${CIFSOPTS}" ; then
            rc=0
        fi
    fi
    return ${rc}
}

do_snap_copy ()
{
    fromdev="${1}"
    todir="${2}"
    snap_type="${3}"

    size=$(fs_size "${fromdev}" "" "used")

    if [ -b "${fromdev}" ]; then
        # look for free mem
        if [ -n "${HOMEMOUNTED}" -a "${snap_type}" = "HOME" ]; then
            todev=$(cat /proc/mounts | grep -s " $(base_path ${todir}) " | awk '{print $1}' )
            freespace=$(df -k  | grep -s ${todev} | awk '{print $4}')
        else
            freespace=$(awk '/^MemFree:/{f=$2} /^Cached:/{c=$2} END{print f+c}' /proc/meminfo)
        fi

        tomount="/mnt/tmpsnap"
        if [ ! -d "${tomount}" ] ; then
            mkdir -p "${tomount}"
        fi

        fstype=$(get_fstype "${fromdev}")
        if [ -n "${fstype}" ]; then
            # Copying stuff...
            mount -t "${fstype}" -o ro,noatime "${fromdev}" "${tomount}"
            cp -a "${tomount}"/* ${todir}
            umount "${tomount}"
        else
            log_warning_msg "Unrecognized fstype: ${fstype} on ${fromdev}:${snap_type}"
        fi

        rmdir "${tomount}"
        if echo ${fromdev} | grep -qs loop; then
           losetup -d "${fromdev}"
        fi
        return 0
    else
        return 1
        [ "$quiet" != "y" ] && log_warning_msg "Unable to find the snapshot ${snap_type} medium"
    fi
}

try_snap ()
{
    # Look for $snap_label.* in block devices and copy the contents to $snap_mount
    #   and remember the device and filename for resync on exit in casper.init

    snap_label="${1}"
    snap_mount="${2}"
    snap_type="${3}"

    snapdata=$(find_files "${snap_label}.squashfs ${snap_label}.cpio.gz ${snap_label}.ext2")
    if [ ! -z "${snapdata}" ]; then
        snapdev="$(echo ${snapdata} | cut -f1 -d ' ')"
        snapback="$(echo ${snapdata} | cut -f2 -d ' ')"
        snapfile="$(echo ${snapdata} | cut -f3 -d ' ')"
        if echo "${snapfile}" | grep -qs '\(squashfs\|ext2\)'; then
            # squashfs or ext2 snapshot
            dev=$(get_backing_device "${snapback}/${snapfile}")
            if ! do_snap_copy "${dev}" "${snap_mount}" "${snap_type}"; then
                 log_warning_msg "Impossible to include the ${snapfile} Snapshot"
                 return 1
            fi
        else
            # cpio.gz snapshot
            # Unfortunately klibc's cpio is incompatible with the rest of
            # the world; everything else requires -u -d, while klibc doesn't
            # implement them. Try to detect whether it's in use.
            cpiopath="$(which cpio)" || true
            if [ "$cpiopath" ] && grep -aq /lib/klibc "$cpiopath"; then
                cpioargs=
            else
                cpioargs='-u -d'
            fi
            if ! (cd "${snap_mount}" && zcat "${snapback}/${snapfile}" | cpio -i $cpioargs 2>/dev/null) ; then
                log_warning_msg "Impossible to include the ${snapfile} Snapshot"
                return 1
            fi
        fi
        umount "${snapback}"
    else
        dev=$(find_cow_device "${snap_label}")
        if [ -b "${dev}" ]; then
            if echo "${dev}" | grep -qs loop; then
                # strange things happens, user confused?
                snaploop=$( losetup ${dev} | awk '{print $3}' | tr -d '()' )
                snapfile=$(basename ${snaploop})
                snapdev=$(cat /proc/mounts | awk '{print $2,$1}' | grep -es "^$( dirname ${snaploop} )" | cut -f2 -d ' ')
            else
                snapdev="${dev}"
            fi
            if ! do_snap_copy "${dev}" "${snap_mount}" "${snap_type}" ; then
                log_warning_msg "Impossible to include the ${snap_label} Snapshot"
                return 1
            else
                if [ -n "${snapfile}" ]; then
                     # it was a loop device, user confused
                     umount ${snapdev}
                fi
            fi
        else
            log_warning_msg "Impossible to include the ${snap_label} Snapshot"
            return 1
        fi
    fi
    echo "export ${snap_type}SNAP="/cow${snap_mount#$rootmnt}":${snapdev}:${snapfile}" >> /etc/casper.conf # for resync on reboot/halt
    return 0
}

setup_unionfs() {
    image_directory="$1"
    rootmnt="$2"

    if [ "${UNIONFS}" = 'DEFAULT' ]; then
for union in 'overlayfs' 'aufs' 'unionfs'
do
            modprobe "${MP_QUIET}" -b ${union} || true
            if cut -f2 /proc/filesystems | grep -q "^${union}\$"; then
UNIONFS="${union}"
break
    fi
done
    fi
    if [ "${UNIONFS}" = 'DEFAULT' -a -x /bin/unionfs-fuse ]; then
UNIONFS="unionfs-fuse"
    fi
    # If all else fails fall back to aufs.
    if [ "${UNIONFS}" = 'DEFAULT' ]; then
UNIONFS='aufs'
    fi

    # run-init can't deal with images in a subdir, but we're going to
    # move all of these away before it runs anyway.  No, we're not,
    # put them in / since move-mounting them into / breaks mono and
    # some other apps.

    croot="/"

    # Let's just mount the read-only file systems first
    rofsstring=""
    rofslist=""
    if [ "${UNIONFS}" = "aufs" ]; then
        roopt="rr"
    elif [ "${UNIONFS}" = "unionfs-fuse" ]; then
        roopt="RO"
    else
        roopt="ro"
    fi

    mkdir -p "${croot}"
    for image_type in "ext2" "squashfs" "dir" ; do
        for image in "${image_directory}"/*."${image_type}"; do
            imagename=$(basename "${image}")
            if [ -d "${image}" ]; then
                # it is a plain directory: do nothing
                rofsstring="${image}=${roopt}:${rofsstring}"
                rofslist="${image} ${rofslist}"
            elif [ -f "${image}" ]; then
                backdev=$(get_backing_device "$image")
                fstype=$(get_fstype "${backdev}")
                if [ "${fstype}" = "unknown" ]; then
                    panic "Unknown file system type on ${backdev} (${image})"
                fi
                mkdir -p "${croot}/${imagename}"
                mount -t "${fstype}" -o ro,noatime "${backdev}" "${croot}/${imagename}" || panic "Can not mount $backdev ($image) on ${croot}/${imagename}" && rofsstring="${croot}/${imagename}=${roopt}:${rofsstring}" && rofslist="${croot}/${imagename} ${rofslist}"
            fi
        done
    done
    rofsstring=${rofsstring%:}

    mkdir -p /cow
    cowdevice="tmpfs"
    cow_fstype="tmpfs"
    cow_mountopt="rw,noatime,mode=755"

    # Looking for "${root_persistence}" device or file
    if [ -n "${PERSISTENT}" ]; then
        cowprobe=$(find_cow_device "${root_persistence}")
        if [ -b "${cowprobe}" ]; then
            cowdevice=${cowprobe}
            cow_fstype=$(get_fstype "${cowprobe}")
    cow_mountopt="rw,noatime"
        else
            [ "$quiet" != "y" ] && log_warning_msg "Unable to find the persistent medium"
        fi
    fi

    mount -t ${cow_fstype} -o ${cow_mountopt} ${cowdevice} /cow || panic "Can not mount $cowdevice on /cow"

    case ${UNIONFS} in
        unionfs-fuse)
            (ulimit -n 16384; unionfs-fuse -o cow -o noinitgroups -o default_permissions -o allow_other -o use_ino -o suid /cow=RW:$rofsstring "$rootmnt" || panic "${UNIONFS} mount failed")
            mkdir -p /dev/.initramfs/varrun
            pidof unionfs-fuse >> /dev/.initramfs/varrun/sendsigs.omit || true
            ;;
        aufs|unionfs)
            mount -t ${UNIONFS} -o noatime,dirs=/cow=rw:$rofsstring ${UNIONFS} "$rootmnt" || panic "${UNIONFS} mount failed"
            ;;
overlayfs)
    # Mount the layers pairwise from the bottom onto rootmnt,
    # for the second and later layers rootmnt forms the lower layer.
    mounts=""
    for mount in /cow $rofslist
    do
mounts="$mount $mounts"
    done
    lower=""
    for mount in $mounts
    do
if [ "$lower" = "" ]; then
    lower="$mount"
    continue
fi
mount -t overlayfs -o "upperdir=$mount,lowerdir=$lower" \
    "$mount" "$rootmnt"
lower="$rootmnt"
    done
    ;;
    esac

    # Adding other custom mounts
    if [ -n "${PERSISTENT}" ]; then
        # directly mount /home
        # FIXME: add a custom mounts configurable system
        homecow=$(find_cow_device "${home_persistence}" )
        if [ -b "${homecow}" ]; then
            mount -t $(get_fstype "${homecow}") -o rw,noatime "${homecow}" "${rootmnt}/home"
            export HOMEMOUNTED=1 # used to proper calculate free space in do_snap_copy()
        else
            [ "$quiet" != "y" ] && log_warning_msg "Unable to find the persistent home medium"
        fi
        # Look for other snapshots to copy in
        try_snap "${root_snapshot_label}" "${rootmnt}" "ROOT"
        try_snap "${home_snapshot_label}" "${rootmnt}/home" "HOME"
    fi

    if [ -n "${SHOWMOUNTS}" ]; then
        for d in ${rofslist}; do
            mkdir -p "${rootmnt}/${LIVE_MEDIA_PATH}/${d##*/}"
            case d in
                *.dir) # do nothing # mount -o bind "${d}" "${rootmnt}/${LIVE_MEDIA_PATH}/${d##*/}"
                    ;;
                *)
                    if [ "${UNIONFS}" = unionfs-fuse ]; then
                        mount -o bind "${d}" "${rootmnt}/${LIVE_MEDIA_PATH}/${d##*/}"
                    else
                        mount -o move "${d}" "${rootmnt}/${LIVE_MEDIA_PATH}/${d##*/}"
                    fi
                    ;;
            esac
        done
        # shows cow fs on /cow for use by casper-snapshot
        mkdir -p "${rootmnt}/cow"
        mount -o bind /cow "${rootmnt}/cow"
    fi

    # move the first mount; no head in busybox-initramfs
    for d in $(mount -t squashfs | cut -d\  -f 3); do
        mkdir -p "${rootmnt}/rofs"
        if [ "${UNIONFS}" = unionfs-fuse ]; then
            mount -o bind "${d}" "${rootmnt}/rofs"
        else
            mount -o move "${d}" "${rootmnt}/rofs"
        fi
        break
    done
}

check_dev ()
{
    sysdev="${1}"
    devname="${2}"
    skip_uuid_check="${3}"
    if [ -z "${devname}" ]; then
        devname=$(sys2dev "${sysdev}")
    fi

    if [ -d "${devname}" ]; then
        mount -o bind "${devname}" $mountpoint || continue
        if is_casper_path $mountpoint; then
            echo $mountpoint
            return 0
        else
            umount $mountpoint
        fi
    fi
    [ -e "$devname" ] || continue

    if [ -n "${LIVEMEDIA_OFFSET}" ]; then
        loopdevname=$(setup_loop "${devname}" "loop" "/sys/block/loop*" "${LIVEMEDIA_OFFSET}")
        devname="${loopdevname}"
    fi

    fstype=$(get_fstype "${devname}")
    if is_supported_fs ${fstype}; then
        devuid=$(blkid -o value -s UUID "$devname")
        [ -n "$devuid" ] && grep -qs "\<$devuid\>" $tried && continue
        mount -t ${fstype} -o ro,noatime "${devname}" $mountpoint || continue
        [ -n "$devuid" ] && echo "$devuid" >> $tried
        if is_casper_path $mountpoint && \
           ([ "$skip_uuid_check" ] || [ "$UUID" = "$devuid" ] || matches_uuid $mountpoint); then
            echo $mountpoint
            return 0
        else
            umount $mountpoint
        fi
    fi

    if [ -n "${LIVEMEDIA_OFFSET}" ]; then
        losetup -d "${loopdevname}"
    fi
    return 1
}

find_livefs() {
    timeout="${1}"
    # first look at the one specified in the command line
    if [ ! -z "${LIVEMEDIA}" ]; then
        if check_dev "null" "${LIVEMEDIA}" "skip_uuid_check"; then
            return 0
        fi
    fi
    # don't start autodetection before timeout has expired
    if [ -n "${LIVEMEDIA_TIMEOUT}" ]; then
        if [ "${timeout}" -lt "${LIVEMEDIA_TIMEOUT}" ]; then
            return 1
        fi
    fi
    # or do the scan of block devices
    for sysblock in $(echo /sys/block/* | tr ' ' '\n' | egrep -v "/(loop|ram|fd)"); do
        devname=$(sys2dev "${sysblock}")
        [ -e "$devname" ] || continue
        fstype=$(get_fstype "${devname}")
        if /lib/udev/cdrom_id ${devname} > /dev/null; then
            if check_dev "null" "${devname}" ; then
                return 0
            fi
        elif is_nice_device "${sysblock}" ; then
            for dev in $(subdevices "${sysblock}"); do
                if check_dev "${dev}" ; then
                    return 0
                fi
            done
        elif [ "${fstype}" = "squashfs" -o \
                "${fstype}" = "ext4" -o \
                "${fstype}" = "ext3" -o \
                "${fstype}" = "ext2" -o \
                "${fstype}" = "btrfs" ]; then
            # This is an ugly hack situation, the block device has
            # an image directly on it.  It's hopefully
            # casper, so take it and run with it.
            ln -s "${devname}" "${devname}.${fstype}"
            echo "${devname}.${fstype}"
            return 0
        fi
    done
    return 1
}

set_usplash_timeout() {
    if [ -x /sbin/usplash_write ]; then
        /sbin/usplash_write "TIMEOUT 120"
    fi
}

start_usplash_pulse() {
    if [ -x /sbin/usplash_write ]; then
        /sbin/usplash_write "PULSELOGO"
    fi
}

mountroot() {
    exec 6>&1
    exec 7>&2
    exec > casper.log
    exec 2>&1
    tail -f casper.log >&7 &
    tailpid="$!"

    parse_cmdline

    set_usplash_timeout
    start_usplash_pulse
    [ "$quiet" != "y" ] && log_begin_msg "Running /scripts/casper-premount"
    run_scripts /scripts/casper-premount
    [ "$quiet" != "y" ] && log_end_msg

    # Needed here too because some things (*cough* udev *cough*)
    # changes the timeout

    set_usplash_timeout

    if [ ! -z "${NETBOOT}" ]; then
        if do_netmount ; then
            livefs_root="${mountpoint}"
        else
            panic "Unable to find a live file system on the network"
        fi
    else
        # Scan local devices for the image
        i=0
        while [ "$i" -lt 60 ]; do
            livefs_root=$(find_livefs $i)
            if [ "${livefs_root}" ]; then
                break
            fi
            sleep 1
            i="$(($i + 1))"
        done
    fi

    if [ -z "${livefs_root}" ]; then
        panic "Unable to find a medium containing a live file system"
    fi

    if [ "${TORAM}" ]; then
        live_dest="ram"
    elif [ "${TODISK}" ]; then
        live_dest="${TODISK}"
    fi
    if [ "${live_dest}" ]; then
        log_begin_msg "Copying live_media to ${live_dest}"
        copy_live_to "${livefs_root}" "${live_dest}"
        log_end_msg
    fi

    mount_images_in_directory "${livefs_root}" "${rootmnt}"

    log_end_msg

 [color=red][b]   # Allow to override USERNAME and HOST based on media information
    # make it skipable by setting FLAVOUR= in casper.conf
    if [ -f /cdrom/.disk/info ] && [ -z "$FLAVOUR" ]; then
        FLAVOUR="$(cut -d' ' -f1 "/cdrom/.disk/info" 2>/dev/null | tr '[A-Z]' '[a-z]')" || FLAVOUR=
        if [ -n "$FLAVOUR" ]; then
            HOST=$FLAVOUR
            USERNAME=$FLAVOUR
            export HOST USERNAME
            sed  -i "s,USERNAME=.*,USERNAME=\"$FLAVOUR\",g; s,HOST=.*,HOST=\"$FLAVOUR\",g" /etc/casper.conf
        fi
    fi
[/b][/color]
    # unionfs-fuse needs /dev to be bind-mounted for the duration of
    # casper-bottom; udev's init script will take care of things after that
    if [ "${UNIONFS}" = unionfs-fuse ]; then
        mount -n -o bind /dev "${rootmnt}/dev"
    fi

    # Open up two fifo's fd's for debconf-communicate to use. Speeds up
    # the Casper process considerably.
    log_begin_msg "Creating debconf-communicate fifo mechanism"
    mkfifo /tmp/debconf-in.fifo
    mkfifo /tmp/debconf-out.fifo

    # Make the template database read-only, so that passthrough debconf
    # instances can write to it directly; otherwise templates are only
    # passed through when necessary.  Use temporary config databases as
    # well; we'll copy their contents back at the end.
    DEBCONF_TMPDIR="$(chroot /root mktemp -dt debconf.XXXXXX)"
    cp -a /root/var/cache/debconf/config.dat "/root$DEBCONF_TMPDIR/"
    cp -a /root/var/cache/debconf/passwords.dat "/root$DEBCONF_TMPDIR/"
    sed "s,^Filename: /var/cache/debconf/\(config\|passwords\).dat$,Filename: $DEBCONF_TMPDIR/\1.dat,; /^Name: templatedb/a\
Readonly: true" /root/etc/debconf.conf >"/root$DEBCONF_TMPDIR/debconf.conf"

    DEBCONF_SYSTEMRC="$DEBCONF_TMPDIR/debconf.conf" chroot /root debconf-communicate -fnoninteractive casper > /tmp/debconf-out.fifo < /tmp/debconf-in.fifo &
    debconfpid="$!"

    if [ ! -p /tmp/debconf-in.fifo ] || [ ! -p /tmp/debconf-out.fifo ]; then
        log_warning_msg "failed to setup debconf-communicate channel"
    fi
    log_end_msg

    # Order matters!
    # These file descriptors must stay open until we're finished with
    # debconf-communicate.
    exec 4</tmp/debconf-out.fifo 3>/tmp/debconf-in.fifo

    maybe_break casper-bottom
    [ "$quiet" != "y" ] && log_begin_msg "Running /scripts/casper-bottom"

    run_scripts /scripts/casper-bottom
    [ "$quiet" != "y" ] && log_end_msg

    if [ "${UNIONFS}" = unionfs-fuse ]; then
        umount "${rootmnt}/dev"
    fi

    # Close the fd's associated with debconf-communicate.
    exec 3>&- 4<&-
    rm -f /tmp/debconf-in.fifo
    rm -f /tmp/debconf-out.fifo
    wait $debconfpid

    # Copy config database changes back to the master files.
    chroot /root debconf-copydb tmpdb config \
--config=Name:tmpdb --config=Driver:File \
--config="Filename:$DEBCONF_TMPDIR/config.dat"
    chroot /root debconf-copydb tmpdb passwords \
--config=Name:tmpdb --config=Driver:File \
--config="Filename:$DEBCONF_TMPDIR/passwords.dat"
    rm -rf "$DEBCONF_TMPDIR"

    exec 1>&6 6>&-
    exec 2>&7 7>&-
    kill "$tailpid"
    cp casper.log "${rootmnt}/var/log/"
    if [ -f /etc/casper.conf ]; then
        cp /etc/casper.conf "${rootmnt}/etc/"
    fi
}

The good news is that I can add "FLAVOUR=" to the casper.conf file to override this as this is a new feature they put in to set the live username to whatever actual Ubuntu flavour they have.  This makes it easier for them but not for us.

I'm getting a new package ready that will have this in it.  Having this won't do anything for older versions of Ubuntu(10.04-11.10) so it is safe to put in.

I'll post here when I have a new package ready for you to test.

[ourasi]

OK. Hope you can correct that.

I apologize my little bit harsh wording during conversations (English not my native language).
And  I want to thank you again for this great program.
 

EDIT: I saw (during my writing) that you have already found some solution. 

[fragadelic]

Yes I am testing it right now after the new package has been created.  I also have some more tweaks for the remastersys-skelcopy so it copies more config files for window managers like icewm, fvwm and afterstep.

I will post back again if it works and put up the new package in the first post.

[fragadelic]

Ok the new package is up and it is corrected.  Remastersys now adds FLAVOUR="$LIVEUSER" to /etc/casper.conf to override the new Ubuntu casper change where it used the first word of the live cd name as the username for live.

ourasi - thanks for helping me find this.  I never noticed on my home remastersys as I use "rem" as the username and all my home remastersys start with "REM" like my current one "REM LXDE LTS" so as you can see I wouldn't have noticed what the change was.

You only need the new base package - the gtk gui has stayed the same.

Let me know how this one works for you.

[ourasi]

Everything seems to be OK with this first installation I tested.
I have some 3 (hard disk) versions and 2 VBox to test.

I post again tomorrow (it is 05:45 am now here in Finland) after more testing.

Thank you very much.
(and have a nice day  )

[fragadelic]

You need to get to bed - lol.

Just let me know after you are done testing.

[tester]

Hello,

Ubuntu 12.04 + updates + Remastersys 3.0.2-1 using Backup option, restored suscesfully.
One thing, after installation you need to reinstall ubiquity-frontend-debconf before update your system, to avoid download kde stuff.

[fragadelic]

Do you have the latest 3.0.2 package?  The first one I put out did not have the dependencies right.

I finally got the issue with the new gui fixed and will be putting up new gui packages - they are arch dependent now.  The base remastersys is common to both platforms but the gui is specific to the architecture.

[ourasi]

I noticed one more locale related annoying feature in Remastersys.

In Finnish "Desktop" folder has name "Työpöytä". When Remastersys (backup mode) adds ubiquity desktop file it adds also folder "Desktop" to my installed system (Lubuntu 12.04 daily).

http://img6.imagebanana.com/img/v3e5afc5/ourasi_001.png
http://img6.imagebanana.com/img/1oa8q6ja/ourasi_002.png

I can delete that folder afterwards, but would it be better to put that desktop file to home folder? I recall that those names Desktop/Työpöytä caused problems with earlier lxde versions, system get confused and I couldn't remove either.

[fragadelic]

Hmmm..can you try to add "export LC_MESSAGES=en_US.UTF-8" like you did earlier and see if that helps it?  If not, I'll have to try and figure something else out.

[fragadelic]

I have the new gui working on 64-bit and had to create architecture specific packages for it.  Please try them out and see the first post also for those that would like to help translate the gui to their native language.

[ourasi]

I don't think that "export LC_MESSAGES=en_US.UTF-8" helps. It's a complicated situation, because my favorite light distro (Peppermint Two) doesn't translate Desktop to "Työpöytä".

[nessus]

On an Intel mobo with a Pentium D and 2 gB of RAM, under Oneiric, 32 bit, with Mate:
I used GDebi to install both the base and the GUI. No error messages. I started the GUI from the menu entry it put under admin.  Positioning of the window is better than in the previous version for me. Before I always had to drag the window up and left to get it to  usable position. Bottom of the window is too low for me. I didn't notice that until I clicked the customise and configure buttons and there was no visible "next" or "main" or whatever. So I hid my top and bottom panels and drug the window up some and could then get to the button on the bottom.  Did a backup (complete with home, not the dist option) with no error messages. Booted the iso in a VM. I get some inconsistent error messages from my VM but I'm fairly confident that is not a problem with the iso because I'd had it on other good isos. Anyway, it booted. It looks normal, allowing for the fact I can't get the VM to give me a decently large window, but that is a Virtual Box problem, not a problem with the iso. And I am posting this from the VM booted from the iso.  So if my spelling and grammer suck I'll blame it on this tiny little window that is all Virtual Box is willing to give me at the moment so that I have to scroll back and forth to get any idea what I've typed. Bottom line: It seems to work fine with only some minor imperfections in the appearance of the GUI.  I've seen this same problem in lots of other programs and it probably is related to my size 20 fonts.  I'll actually burn a DVD as soon as I can get to town and buy some more blank media. Thumbs up!

[fragadelic]

nessus - thanks for the feedback.

As for you VM issue, if you are using virtual box, make sure to install the client drivers on the system you are remastering.  That will allow you to resize the window and other things like mouse capture and keyboard capture will behave better.

I can lock the font size down but I'd rather not for those that are visually impaired and need bigger fonts.  Not sure if I can fix the issue but I'll see what I can do for the larger fonts.

[nessus]

Thank YOU. Don't sweat it much on the font size related issues. Anybody who uses large fonts regularly has probably run into problems like that before and has figured out how to cope with it one way or another. I made myself a font toggler button to go between all size 10 and all size 20 which is convenient. Size 10 didn't eliminate this issue now that I think about it, so maybe font size isn't the cause. I've had other weird stuff with windows running off the page and so on. Maybe my graphics card is just retarded. I always manage to cope. I agree locking down the fonts would be a bad idea. I think the fix for issues like this really isn't in your script anyway but more likely in the code in the graphics tools you use to make something like this.

Anyway, an update:

I've made several more isos and burned one DVD. They all boot ok but I've messed up the config file somehow so my timeout is 0 and I get no menu. Details appear in another post in the same thread as before in the help forum. If I can't figure out anything more elegant I'll delete all that sutff and reinstall and be more careful this time.  Pretty sure that isn't a bug - that's just me not understanding the syntax in the config file correctly - I must have screwed up the first line. When I finally get it right, I'll try deleting and restoring the installation on the HD from the DVD.

[ourasi]

There is a way test to check the name of Desktop folder:
http://freedesktop.org/wiki/Software/xdg-user-dirs

I found a solution for my Lubuntu installation this way:
in /etc/xdg/user-dirs.conf I changed enabled=False

then I copied my (empty) Home folders to /etc/skel
eg. Desktop, Asiakirjat, Lataukset, Kuvat, Videot ...

Now folder names don't change with language change and every user gets same folders.

[ourasi]

What is this rem-lxde "daily" version?
I installed it and noticed couple of things.

1. show-language-selector must set to true to be able to change language
2. how do I easily/automatically mount other partitions in my hard disk to /media (as usually in Ubuntu)?
3. how it is made that I must not give root/sudo password to use synaptic etc? Is it secure to do so?

Nice programs you have there, like deb builder. I already used it to change Dukto R5 *.deb file - its dukto.desktop doesn't show automatically in LXDE. I think I loaded some script version of deb builder from lkjoel last year?

[fragadelic]

Those are my personal remasters.

The user is always the weakest link in security.  Asking for the sudo password after the user logs in is just a test of whether they know their password and they should.  It does nothing to improve security by repeatedly having to enter your password.

Deb builder is something I made for myself and for Gmaq of AVLinux so we can easily make all kinds of deb packages.

This was never released unless you find it on the AVLinux ISO or my ISO.

It is updated to the date you see on the iso name.  Ubuntu 12.04 hasn't been released stable so they are still making changes.

[nessus]

Probably not Remastersys related but just in case it is, and is of use, to make long story short:
Drowned my monitor. Hooked up old VGA. Discovered Oneiric understands VGA and Maverick doesn't. Decided to try PC-BSD on the Maverick partition. Didn't like the VGA either. Neither did Dragonfly, an old Sabayon 5.2, an Oneiric mini (which I really expected to work), or a Debian. But somewhere along the line they screwed up GRUB so I couldn't boot the perfectly good Oneiric (well, as perfectly good as Oneiric gets anyway, meh).

But my fresh remastersys live disk of my oneiric in sda7 is working fine (I'm posting from it) so I figured I'd install a copy on my erstwhile Maverick partition, sda2, hoping it would give me a good grub and at least 1 hd based oneiric if not 2. But it didn't work. Tried repeatedly. I have to invoke it with sudo ubiquity --no-migration-assistant because otherwise I can never get past the invalid username error but that has worked fine for me before so I doubt it has anything to do with the problem. It would go ok through the whole graphical installation down to the point where you get a progress bar and then fail after maybe 5 or 10 minutes. Then it gives me one of those cursed gui error message boxes that won't let you copy the message. I believe is said something like "ERRNO9 . . . [a whole lot more stuff suggesting eleventy seven different possibilities]" after which it asked to do an automatic error report and I let it (as I always do unless I happened to have been emailing my co-conspirators about our  secret plot to overthrow the government and install Bessie the Cow as a figurehead for the Darkaniti). After the error report it offers a more specific suggestion on the basis of a log it read in making the error report apparently and indicates a bad DVD as the likely cause.  My DVD drive is fairly new. I burned from the iso with K3b at 8x, the slowest speed it lets me set, and used the verify option. The DVD looked clean, and it has only been a day or so since I made it, but I cleaned it thoroughly anyway after the first time the restoration failed and it continued to fail in the same way. My own suspicion is that this is probably not remastersys related but I report it because this is the first time I've tried to use the restore-to-hd function of the beta. I can still navigate my HD fine when booted from the DVD and the attached log should reflect the making of the DVD I'm using since it is the last iso I made, the one that has "vga" in the name.

I'll try other stuff later and keep you posted. Probably I can hand edit the grub or install LILO maybe.

[sageng]

Hello,

tested this with ubuntu 12.04 + current updates + remastersys 3.0.2 -- backup option

it copies and installs fine till to the point when it gets to "importing user profile data" or something like that, It just pops out a message "Installer crashed" and can't continue from there.

It can boot the live option though.

Remade the iso with the backup option again. Still the same