TazPup64 (pre-alpha17) - getting close to alpha quality

[s243a]

Latest ISO
TazPup64-PreAlpha-17.iso (See post)

Latest .vdi file (for virtualbox and other virtualization/emulationtools):
TazPup64-PreAlpha-17.vdi (see post)

Older Info
I think we're getting close to alpha quality with TazPup64-PreAlpha-15

Some notes, the ISO will automatically log you into tux in the lxde desktop. On the bottom left (in the panel) there is a red logout/shutdown/reboot button that looks like the power symbol. If you click on it you can log out and then log in as root in the jwm desktop. The password for root is "root".

To use gparted you must install gtkmm.

Code: Select all

tazpkg get-install gtkmm

I recommend starting gparted from the command line. If for some reason you get an error try it one more time from the command line.

The puppy universal installer works. I verified after installing and creating a save file that tux still works but I think there is a catch. If you install any software (e.g. gtkmm) in the pfx=ram mode then creating a save file might mess up the permissions for non root users. Therefore if you install any software in pfx=ram mode then you should first reboot without creating a save file, and then after rebooting create a save file before installing any packages.

Further notes on why not to install any packages before creating a save file can be found in the following post:
http://murga-linux.com/puppy/viewtopic. ... 79#1026679

And if you disregard this advice, then notes on how to fix the save file after the fact can be found in the following post:
http://murga-linux.com/puppy/viewtopic. ... 27#1026527

A .vdi (virtual disk image) for virtualbox is available for this release (see post) but I will be doing further updates to it before the next release.

There is no recent buildkit but an older version of the buildkit is referenced in the following post:
http://murga-linux.com/puppy/viewtopic. ... 77#1026677

The password for root is root. Both spot and tux have a zero length passwords (almost equivalent to having no password but this gets around restrictions in the pam package).

If you login with the screen=text option (equivalent to pfx=nox) then you can start jwm with the command:

Code: Select all

startx_jwm

or alternatively start lxde with the command:

Code: Select all

startlxde

Note that I'll eventually change the name of the command to start lxde because I'm starting Xorg with this script but the standard version of this script does not do this. Also note that I disabled the autostart for many applications. To have them autostart, rename the files in the folder:

Code: Select all

/etc/xdg/autostart

so that they end in ".desktop", This will be more apparent if you are using the rox file manager rather than pcmanfm. This is because for ".desktop" file pcmanfm displays the "application name" rather than the "file name".

Edit 1 To see the rox desktop in jwm for some reason you have to first log in as root, and then "Exit to Prompt" and restart the desktop with the command:

Code: Select all

startx_jwm

I think the reason is that the pcmanfm desktop overlays the rox desktop, and so far the only way I know to remove this is to kill Xorg. As an alternative to the above way of seeing the rox desktop you can boot with the screen=text option and then start jwm with the command startx_jwm.

[s243a]

The build script doesn't seem to be creating the 64bit version of the iso for me regardless of whether or not I use the online or cloud version.

I think that mistfire moved to using tazpkg to build the ISO. On configureation file used by tazpkg is located in

Code: Select all

/etc/slitaz/slitaz.conf

Some important lines in this file are:

Code: Select all

#05 - SLITAZ_RELEASE="$(cat /etc/slitaz-release)"
#30 - MIRROR_URL="http://mirror.slitaz.org"
#33 - ONLINE_PKGS="$MIRROR_URL/packages/$SLITAZ_RELEASE/"

some other vars in this file that look relavant:

Code: Select all

#20 - STABLE="$SLITAZ_HOME/stable"
#21 - COOKING="$SLITAZ_HOME/cooking"
#22 - UNDIGEST="$SLITAZ_HOME/undigest"

I'm not sure yet where this file is being pulled from but probably not the slitaz iso like it should be.

So urls where 64 bit packages can be found are:
http://mirror.slitaz.org/packages/next64/
http://mirror1.slitaz.org/packages/next64/


The above config file I can't find but from the build script:
Ln#452-468:

Code: Select all

cd $curdir/slitaz-livecd-wkg

rcount=$(ls -1 ./rootfs*.gz | wc -l)

if [ $rcount -eq 1 ]; then
 echo "Extracting rootfs gz..."
 extractfs $curdir/slitaz-livecd-wkg/rootfs.gz
elif [ $rcount -gt 1 ]; then
	for rf1 in $(ls -1 ./rootfs*.gz | sort -r)
	do
	 echo "Extracting $(basename $rf1)..."
	 extractfs $rf1
	done
else
 echo "No rootfs found"
 exit
fi

we see that the extrafs command moves the iso contents to the root-fs folder:
LN#207-215:

Code: Select all

extractfs() {
	
if [ ! -d $curdir/slitaz-rootfs ]; then
mkdir $curdir/slitaz-rootfs
fi

(busybox unlzma -c $1 || cat $1 ) 2>/dev/null | (cd $curdir/slitaz-rootfs; busybox cpio -idmu > /dev/null)

}

[mistfire]

@s243a

Did you ran the tazpuppy build kit on 64-bit puppy?

Also you must study the slitaz documentation very well before you do it

[mistfire]

@s243a

according to slitaz download page. slitaz-core64 was just 32-bit program suite, only 64-bit was their kernel

[mistfire]

@s243a

this link might help you
http://forum.slitaz.org/topic/slitaz-x86_64

[s243a]

@s243a

this link might help you
http://forum.slitaz.org/topic/slitaz-x86_64

I found some links to ISOs in the discussion, unfortunately the links either expired or I needed permission. I did learn though from the thread that I can build an iso using cookutils so I may gave that a try.

In the mean time, I've replaced the 32bit files in your build kit with the 64bit version of the package from the next64 folder.
**I did this manually and didn't verify the checksums.


This modified build kit can be downloaded from:
https://www.dropbox.com/s/vuirof94f8g38 ... alpha?dl=0
**Build using "local mode". Still some syncing issues.

The following packages are missing 64bit versions:

Code: Select all

grub4dos
libtdl
pcmciautis
util-linux-losetup
Xdialog
xorg-renderproto
xorg-xf86-video-geode
xorg-xf86-video-modesetting
xorg-xf86-video-radeonhd
xorg-xf86-video-v4l



Maybe:
aufs-utils-20150122.tazpkg
flsynclient
squashfs
xorgrdp

Some of these like grub4dos I replaced with a pet version from slaco 6.9.9.9. I noticed there was a folder called, "special packages".

Code: Select all

flsynclient_tazpkg

I added a -32 at the end of the file name in this folder so that it wouldn't be included in the build. I also commented out the associted exit statment and replaced it with a "press enter to continue".

around line#244:

Code: Select all

if [ $(find $curdir/special-packages -type f -name "*.tazpkg" | wc -l) -eq 0 ]; then
  echo "Special tazpup package is missing."
  #exithttps://www.dropbox.com/s/vuirof94f8g38e8/tazpup-builder64-pre-alpha?dl=0
  read -p "Press enter to continue" #added by s243a
fi

I added another "press enter to continue" at the point after the SliTaZ iso is extracted so that one could modify it if desired before the script starts installing packages (this would be required for now to get cloud-build mode to work):

Around line#257

Code: Select all

if [ $rcount -eq 1 ]; then
 echo "Extracting rootfs gz..."
 extractfs $curdir/slitaz-livecd-wkg/rootfs.gz
elif [ $rcount -gt 1 ]; then
	for rf1 in $(ls -1 ./rootfs*.gz | sort -r)
	do
	 echo "Extracting $(basename $rf1)..."
	 extractfs $rf1
	done
else
 echo "No rootfs found"
 exit
fi

#exit
#chroot "$curdir/slitaz-rootfs/" tazpkg setup-mirror http://mirror.slitaz.org/packages/next64/ 
"http://mirror.slitaz.org/packages/next64/">$curdir/slitaz-livecd-wkg/var/lib/tazpkg/mirror
read -p "Press enter to continue"

*Note the line to overwriite /var/lib/tazpkg/mirror was added by me
**Note2 I should be writing to the slitaz-rootfs folder and not the slitaz-livcd-wkg folder but this doesn't matter since this isn't enough to change the repo. If one can't change the repo than cloud mode can't be used to download 64bit packages.

Code: Select all

"http://mirror.slitaz.org/packages/next64/">$curdir/slitaz-rootfs/var/lib/tazpkg/mirror 

***I also tried changing the repo with the "tazpkg setup-mirror" command but this also didn't work for me.

On another note; at the top of the folder structure there is a folder called s243a. This does nothing useful. It is just some snippets that I found interesting from the build script. Anyway, the kernal isn't syncing for me (with 64bit packages) at the point of the swhich-root when using the slitaz-core64 iso. You noted that this iso might not be suitable, so I'll look into building an iso via cookutils. I will also play around with what I built using the above build kit in a chroot environment.

[s243a]

As I mentioned in another post; In the build-kit I think we have to patch tazpkg before installing files.

To do this I suggest moving:

--around line 554

Code: Select all

if [ "$xmode" == "local" ]; then
 install_local_pkg
else
 install_online_pkg
fi

after

--Around line 632:

Code: Select all

echo "Applying patches..."

cp --remove-destination -arf $curdir/tazpup-core-files/slitaz/* $curdir/slitaz-rootfs/ 2>/dev/null
cp --remove-destination -arf $curdir/tazpup-core-files/puppy/* $curdir/slitaz-rootfs/ 2>/dev/null

make_pupcore

if [ -f $curdir/slitaz-rootfs/usr/sbin/hald ]; then
 mv -f $curdir/slitaz-rootfs/usr/sbin/hald $curdir/slitaz-rootfs/usr/bin/hald
 sed -i -e "s#\/usr\/sbin\/hald#\/usr\/bin\/hald#g" $curdir/slitaz-rootfs/var/lib/tazpkg/installed/hal/files.list
 sed -i -e "s#\/usr\/sbin\/hald#\/usr\/bin\/hald#g" $curdir/slitaz-rootfs/var/lib/tazpkg/installed/hal/md5sum
 sed -i -e "s#\/usr\/sbin\/hald#\/usr\/bin\/hald#g" $curdir/slitaz-rootfs/etc/init.d/hald
fi

if [ -f $curdir/slitaz-rootfs/usr/share/applications/defaults.list ]; then
 rm -f $curdir/slitaz-rootfs/usr/share/applications/defaults.list
fi

echo "Updating system database..."

chroot "$curdir/slitaz-rootfs/" tazpkg tazpkg recharge
chroot "$curdir/slitaz-rootfs/" /usr/bin/update-desktop-database /usr/share/applications
chroot "$curdir/slitaz-rootfs/" /usr/bin/update-mime-database /usr/share/mime
chroot "$curdir/slitaz-rootfs/" /usr/bin/gtk-update-icon-cache /usr/share/icons/hicolor
chroot "$curdir/slitaz-rootfs/" /usr/bin/glib-compile-schemas /usr/share/glib-2.0/schemas
chroot "$curdir/slitaz-rootfs/" /usr/bin/gdk-pixbuf-query-loaders --update-cache
chroot "$curdir/slitaz-rootfs/" /usr/bin/gio-querymodules /usr/lib/gio/modules	
chroot "$curdir/slitaz-rootfs/" /usr/bin/fc-cache -f

if [ -e $curdir/slitaz-rootfs/usr/sbin/iconvconfig ]; then
 chroot "$curdir/slitaz-rootfs/" /usr/sbin/iconvconfig
fi

I'll try this and see what happens.

Edit 1: I notice that if you apply the patches early then packages won't install because "sh" points to "bash" which doesn't exist yet. My solution was to first install bash then install the patches and then install the rest of the packages.

[s243a]

duplicate

[s243a]

Another wierd thing I notice that after applying the patches that the bash executable doesn't work in the chroot environment but ash does. I think the difference is that ash is linked to busybox which is a static executable whereas the bash executable depends on the linker. I'm guessing that I need to add something like glibc to the package list but that wouldn't explain whey the executable worked before the patches but not after.

[mistfire]

@s243a keep exploring, you can do it.

[s243a]

@s243a keep exploring, you can do it.

I'm making some progress. It now boots up to the log-in screen but slim doesn't load:

Code: Select all

slim: waiting for X server to begin accepting connections.
slim: open_session: Unable to open session: The name org.freedesktop.ConsoleKit was not provided by any .service files

Current ISO:
https://www.dropbox.com/s/uw9z6xba8m0en ... o.iso?dl=0
Current Buidkit (unclearn)
https://www.dropbox.com/s/ymgzbn7jijil0 ... ar.gz?dl=0
Current Buildkit (clean)
https://www.dropbox.com/s/w9ustwlmpv8ca ... ar.gz?dl=0

[s243a]

**Just posting a note for myself here.

At startup:
/etc/init.d/rcS
calls
/etc/init.d/network.sh

around line #262 in funciton boot we have

Code: Select all

[ -s /etc/sysctl.conf ] && sysctl -p /etc/sysctl.conf

This throws the following error if one installs procps-ng

Code: Select all

sysctl: error while loading shared libraries: libproc-3.2.8.so: cannot open shared object file: No such file or directory

however if one installs procps -- which I think is a cutdown version -- then one doesn't get this error. The cutdown version (i.e. procps) uses the busybox version of sysctl.

[s243a]

I'm planning to release an ISO soon which will work with the window manager and an updated build kit. I'm putting some notes here for my reference.

The following will be excluded from the iso:

libgiomm-2.56.0-x86_64.tazpkg
gtkmm-2.24.5-x86_64.tazpkg
procps-3.2.8-x86_64.tazpkg and/or procps-ng-3.3.12-x86_64.tazpkg
consolekit2-1.2.1-x86_64.tazpkg
udev-182-x86_64.tazpkg #instead use eudev-3.2.7-x86_64.tazpkg

libgiomm seems to cause segmentation faults and isn't required. It is listed as a dependency for gtkmm and gtkmm is listed as a dependency for gparted. However, gparted will work with an alternative version of gtk.

procsps seems to be missing the dependency libproc-3.2.8.so, but the functions supplied by procsps can be replaced with the busybox version. I added some code in my build script to automatically replace these functions in case one installs procsps ( I think the procps-ng package has the same issues)

Code: Select all

#Somd utilities in procps link to missing library libproc-3.2.8.so
while read -r line; do #https://unix.stackexchange.com/questions/85060/getting-relative-links-between-two-paths
  #https://stackoverflow.com/questions/9293887/reading-a-delimited-string-into-an-array-in-bash
  #https://www.tldp.org/LDP/Bash-Beginners-Guide/html/sect_10_02.html
  arr=($line)
  target_f=$curdir/slitaz-rootfs$prefix${arr[0]}
  source_f=${arr[1]}
  if [ -f $target_f ]; then
    
    fname=`basename $target_f`
    target_d=`dirname $target_f`
    mv $target_f $target_d/$fname-old
    cd $target_d
    pwd
    ln -s $source_f $fname
  fi
done <<EOM
/bin/uname busybox
/sbin/sysctl ../bin/busybox
/bin/ps busybox
/bin/kill busybox
/usr/bin/w ../../bin/busybox
/usr/bin/pmap ../../bin/busybox
/usr/bin/free ../../bin/busybox
/usr/bin/uptime ../../bin/busybox
/usr/bin/pkill ../../bin/busybox
/usr/bin/pgrep ../../bin/busybox
/usr/bin/top ../../bin/busybox
/usr/bin/watch ../../bin/busybox
EOM

Note that uname is supplied by several packages. One version depends on some receipt and I don't understand this process enough. However, hard-coding it to a receipt would be problematic if someone changed the kernel.

I think the version of eudev is newer than the udev package but I can't seem to find the missing library error. This error at least doesn't show up in the boot log. To ensure that I'm using eudev I make sure that it is installed after udev. This seems to eliminate a boot error that I noted on another post.

There is a boot message related to consolekit. I don't think it is required and I think it may cause some problems so currently I don't plan to include it.

[s243a]

The latest challenge I've been having was to get the desktop to work. I've been booting into console mode with the boot option:

Code: Select all

screen=text

I was able to get into the desktop with the command:

Code: Select all

startx

https://www.pearltrees.com/s243a/startx ... id24233101
https://pastebin.com/vdrEK8pg

This function looks for a startup script called $HOME/.xsession, to start the Xorg & desktop session and if it isn't there it first creates the startup script via the function tazx. Tazx also does some basic configuration such as the fonts and keyboards.

The ./xsession script is nice and simple. It simply starts the Xorg server in the background and then starts the desktop. The default configuration file is loaded from:

Code: Select all

. /etc/slitaz/applications.conf

which provides the name of the function to start the windows manager (e.g. lxde-session).

Actually both stat-up scripts .xsession and .xinitrc are generated by tazx. They are generated by installing the respective packages slitaz-configs-base and slitaz-configs. There are some errors in tazx that I note below:

Code: Select all

slitaz_apps_conf() {
	apps_conf=/etc/slitaz/applications.conf #

	[ -f $apps_conf ] || inst_pkg slitaz-configs --forced #s243a: $apps_conf comes from slitaz-configs-base so this statement is wrong.  

	if [ "$(id -u)" -eq 0 ]; then
		user_conf="$apps_conf"
		new_user_conf='/etc/skel/.config/slitaz/applications.conf' #Not referenced by either .xsession or .xinitrc
		mkdir -p $(dirname $new_user_conf)
		cp -f $apps_conf $new_user_conf
	else
		# If started by: startx and with boot opts screen=text
		user_conf="${XDG_CONFIG_HOME:-$HOME/.config}/slitaz/applications.conf" #Not referenced by either .xsession or .xinitrc
		if [ ! -f "$user_conf" ]; then
			mkdir -p $(dirname $user_conf)
			cp $apps_conf $user_conf
		fi
	fi
	[ -f "$HOME/.xinitrc" ] || inst_pkg slitaz-configs-base --forced #s243a: $HOME/.xinitrc comes from slitaz-configs  so this statement is wrong.  
}

Both .xsession and .xinitrc import settings from /etc/slitaz/applications.conf but with .xinitrc you can over-ride these settings by specifying the window manager as the first argument:

Code: Select all

. /etc/slitaz/applications.conf
tazbox mkmenu

[ "$1" ] && WINDOW_MANAGER="$1"

she startx script starts the Xorg script but the .xintrc script doesn't. I think this is because if one is using .xintrc then the xinit utility is supposed to start the Xorg server. Here is the ussage of xinit:

Code: Select all

xinit [ [ client ] options ... ] [ -- [ server ] [ display ] options ... ]

https://linux.die.net/man/1/xinit

The window manager (e.g. jwm) is the client and Xorg is typically the server. I think that if one wants to pass an input argument to .xinitrc then they would do the following:

Code: Select all

xinit ~/.xinitrc jwm -- Xorg :0

I'll experiment. Options for the Xorg server are documented at:
https://linux.die.net/man/1/xserver

Many are default. For instance

Code: Select all

:0

is the default display. Actually one could leave off:

Code: Select all

 -- Xorg :0

because Xorg is also the default xserver.

[s243a]

More notes to self:

One issue that I had is that when the desktop started (via startx) up the taskbar wasn't showing. I noticed in in the .xinitrc script there is a call to tazbox but there is no related call in the .xsession startup script. Tazbox has the following mkmenu() function:

Code: Select all

mkmenu() {
	# (re)make the menu button with the user log name
	dd="$HOME/.local/share/desktop-directories"; mkdir -p "$dd"
	echo -e "[Desktop Entry]\nType=Directory\nName=$(whoami)" > "$dd/user.directory"

	# check the panel config file existence, copy from default config if needed
	panels=$(find $HOME/.config/lxpanel -type f -name panel 2>/dev/null)
	if [ -z "$panels" ]; then
		mkdir -p "$HOME/.config/lxpanel"
		cp -a /etc/skel/.config/lxpanel/* "$HOME/.config/lxpanel"
	fi

	for panel in $panels; do
		conf=$(tr '\n' '`' < $panel | sed 's|Plugin|\n&|g') # each plugin on its own line
		if fgrep -q 'user.directory' "$panel"; then
			# if menu already exists - shrink it to '@@@'
			newconf=$(echo "$conf" | sed 's|.*user\.directory.*$|@@@|')
		else
			# otherwise do nothing
			newconf="$conf"
		fi
		menu=$(cat <<EOT
Plugin {
    type = menu
    Config {
        name=user.directory
        image=avatar-default
        tintcolor=#33230D
        item {
            image=system-log-out-panel
            name=$(_ 'Close X session')
            action=tazbox logout exit
        }
        item {
            image=system-reboot-panel
            name=$(_ 'Reboot system')
            action=tazbox logout reboot
        }
        item {
            image=system-shutdown-panel
            name=$(_ 'Shutdown system')
            action=tazbox logout halt
        }
    }
}

I think the related puppylinux script is called fixmenus. I tried testing fixmenus in the chroot enviornment. I noticed that it looks in a folder called:

Code: Select all

$HOME/.config/autostart/*.desktop

tazpup doesn't apear to use this mechanism for statup scripts. However, if I include an acount with the puppy windows manager jwm, then maybe I'll copy the *.desktop files into said folder.

[s243a]

I added a function to create the user profiles in the build script.
https://www.pearltrees.com/s243a/add-us ... id24252575
https://pastebin.com/8Q3hn8qR

The build script will only create the user profiles if the variable $users is defined. I added a file called "defaults" to put such settings in so that one doesn't have to keep clicking through the dialog windows every-time that someone runs the script.

My configuration file has the following:

Code: Select all

users='tux,root,spot:jwm

The default window manager is slim but I thought I would try creating a user called spot that has the window manager jwm.

In this script the window managers are configured with the code:

Code: Select all

    if [ $WINDOW_MANAGER = "jwm" ]; then   
      usr_home=`su -c "echo $HOME" $USER` 
      ( export HOME=$usr_home; fixmenus ) #TODO update fixmenus so that it only updates files for a single user. 
    else
      su -c "sh tazbox mkmenu" $USER # 
    fi      

one issue I have is that the fixmenus script updates the settings for all users rather than just the home directory that I specify.

[s243a]

I noticed that when I copped the therme packages: from WoofCE that I was missing the folder:

Code: Select all

/root/.jwm

I figured out that this folder is created by the post install script:

woof-CE/woof-code/rootfs-packages/ptheme/pinstall.sh

I ran this script in my chroot environment, now I'll try rebuilding the ISO to see what happens

p.s. another issue I had was that the full version of ps is required to avoid errors on shutdown. To fix this issue I copied the following two files from slaco6.9.9

Code: Select all

/bin/ps
/bin/ps-FULL

the first file is a script that selects the full version of ps if available and otherwise links to busybox. I had to copy this file from slaco64.6.9.9.9 because the procps package in the next64 branch was causing segmentation faults.

[s243a]

As noted above, I was having trouble with slim so I wrote a script to install the jwm files from woof-CE
https://pastebin.com/xWxfLYez
(see function: add_puppy_jwm_files() )

I also think the version of lxpanel available on the repo is causing segmentation faults. I'm going to replace it with the one from slackware64-14-2-extra repo. Some packages depend on lxpanel such as lxde and lxde-extra. I'll be replacing these also. However, I notice that the slackware repo has more lxde packages so I have to decide which ones I want.

On another note, I was having trouble with ssl certificates so I added the package:

ca-certificates-20181210-noarch-1_slack14.2.txz.tazpkg

and than ran the script:

Code: Select all

/usr/sbin/update-ca-certificates

hopefully this script will address the certificate issues.

[s243a]

Some changes in the coming release:
- Replaced 32 bit version of /usr/local/pup_event_pup_event_frontent_d with the 64bit version
(I copied this from my slacko64-6.9.9.9 system.

Code: Select all

# grep -rn -e 'jwm-xdgmenu'
builtin_files/xdg_puppy_jwm:4: jwm-xdgmenu

- replaced 32 bit version of /usr/local/bin/jwm-xdgmenu with 64 bit version.
- symbolically linked roxfiler to rox-filer

Code: Select all

cd /usr/local/bin
ln -s rox-filer roxfiler

made sure that I exported CURRENTWM=jwm in my startx_jwm script.
https://pastebin.com/dYGEnHvC
which calls a slightly modified version of ~/.xinitrc called ~/.xinitrc_puppy
https://pastebin.com/9F176njn
(I think I only commented out one line)
I copied the following folder over form slacko64-6.9.9.9

Code: Select all

/usr/local/lib/X11/pixmaps

I can't figure out what package it is from.

Code: Select all

# grep -nr . -e '/usr/local/lib/X11/pixmaps'
./Iron-69.0.3600.0-amd64.files:40:/usr/local/lib/X11/pixmaps/
./Iron-69.0.3600.0-amd64.files:41:/usr/local/lib/X11/pixmaps/up_down.png
./builtin_files/wcpufreq:5:/usr/local/lib/X11/pixmaps

**this might have not been necessary

on another note, when I try to start jwm I get the following error:

Code: Select all

/bin/sh: xdg_autostart.sh: command not found

I"m not sure where I would find xdg_autostart

One more thing. I made the following symbolic link:

Code: Select all

cd /usr/share/fox-filer
ln -s ROX-Filer rox-filer

I also installed "icewm-xdgmenu"

which gave me two dependencies
"gnome-menus"
"icewm"

I forget the error that made me decide to try this.

[s243a]

Some more notes:

Code: Select all

# Copy /sbin/losetup-FULL from slacko64-6.9.9.9
gtkwidget.x:9993: widget class 'GtkProgressBar' has no property named 'through-under-steppers'
#Copy /usr/local/apps/ROX-Filer from slacko64-6.9.9.9
#Copy /etc/xdg/lxsession ~/.config/lxsession #Shouldn't be necessary
#Copy /etc/xdg/lxpanel ~/.config/lxpanel #Shouldn't be necessary
#Copy /etc/xdg/templates ~/Templates #Shouldn't be necessary

Midori isn't working for me but firefox is. I fixed the tsl issue by installing some tsl libraries.

I'm able to get lxpanel working on jwm but not on lxde. That said I fixed all the errors that I saw in the run.log for lxde (not counting locale errors). and the above GTK progress error.

I will try to upload something soon to capture these changes and hopfully correctly incorporate them into the build system.