(old)Puppy Linux Discussion Forum

Posted: **Wed 23 Jan 2019, 11:31**

@rijsmier: I am sending you a PM.

Posted: **Wed 23 Jan 2019, 12:25**

@rcrsn51 or somebody else

Please tell me how to access a PM.

Thanx

Edit: I figured it out

Posted: **Sat 26 Jan 2019, 20:02**

Added to custom repos:

- Palemoon web-browser v28.3.1 (currently latest).
Install palemoon from Synaptic or with apt-get.
Can be updated (if available) from Palemoon : Help > About

- run-as-user
Support for applications that cannot run as root:
Installs Menu entries for: Chromium, Vlc and Google-chrome (last is for 64-bit only).
(in /root/.local/share/applications)
First time run it will install the application, once installed it will run as unprivileged user (e.g. puppy or (if not exists) as user guest)
This construction (although modified) I took from Slax by Tomas M.
Also included Menu > System > "Run program as (unprivileged) user" to be able to run any application as unprivileged user
It runs modified "run-as-spot" script from Fatdog by jamesbond.
Install run-as-user from Synaptic or with apt-get.
(see pics below)

Fred

Posted: **Wed 06 Mar 2019, 21:26**

Hello again

I have a rather technical question:
I boot poerteus with EXIT and I load MANY squashfs files.

My initial idea was that linux runs in ram and squashfs files are loaded/accessed when needed from the hard disk.

I have noticed that if I delete a squashfs (while the system is live) I can still access its files.
The example in more detail :
I see RAM usage from conky is 250-350MB. A specific squashfs loaded during boot is 500MB. I can access the files ok.
Now I rename the squashfs and I still can access it's files. Ram usage seems still under 500MB.

So, how can this be?

Thanks in advance

Posted: **Thu 07 Mar 2019, 18:10**

trister wrote:Hello again
I have a rather technical question:
I boot poerteus with EXIT and I load MANY squashfs files.

My initial idea was that linux runs in ram and squashfs files are loaded/accessed when needed from the hard disk.

I have noticed that if I delete a squashfs (while the system is live) I can still access its files.
The example in more detail :
I see RAM usage from conky is 250-350MB. A specific squashfs loaded during boot is 500MB. I can access the files ok.
Now I rename the squashfs and I still can access it's files. Ram usage seems still under 500MB.

So, how can this be?

Thanks in advance

Hi, trister,
A technical explanation I cannot give (maybe someone else can?), but I guess that once it's mounted, it stays mounted no matter what you do with the source file.

Same when you mount an ISO, renaming the ISO afterwards doesn't matter, the mountpoint stays the same.
When you look at the output of "mount" after renaming, you'll see that the name of the ISO has changed (same as you renamed) or in case of deletion, it shows (deleted) after the ISO filename.

Fred

Posted: **Sat 09 Mar 2019, 09:51**

all the software files are stored in the /var directory all the installed software has its own folder, I suppose you can delete the software from there.

Posted: **Sun 10 Mar 2019, 07:19**

ralcocer wrote:all the software files are stored in the /var directory all the installed software has its own folder, I suppose you can delete the software from there.

Better to run "apt-get clean" to let apt remove them for you. While you are at it, "apt-get autoremove" to clean up any unneeded dependencies.

Posted: **Sun 10 Mar 2019, 19:33**

ralcocer wrote:all the software files are stored in the /var directory all the installed software has its own folder, I suppose you can delete the software from there.

To add to dancytron's reply:
To really remove software from the system, use synaptic or (from terminal) e.g. "apt-get remove <package>".
The files in /var/cache/apt/archives/ are just downloaded deb packages to be unpacked by the package manager, once they have been installed, removing these will not uninstall any software.

Fred

Posted: **Sun 10 Mar 2019, 22:44**

fredx181 wrote:
trister wrote:Hello again
I have a rather technical question:
I boot poerteus with EXIT and I load MANY squashfs files.

My initial idea was that linux runs in ram and squashfs files are loaded/accessed when needed from the hard disk.

I have noticed that if I delete a squashfs (while the system is live) I can still access its files.
The example in more detail :
I see RAM usage from conky is 250-350MB. A specific squashfs loaded during boot is 500MB. I can access the files ok.
Now I rename the squashfs and I still can access it's files. Ram usage seems still under 500MB.

So, how can this be?

Thanks in advance
Hi, trister,
A technical explanation I cannot give (maybe someone else can?), but I guess that once it's mounted, it stays mounted no matter what you do with the source file.

Same when you mount an ISO, renaming the ISO afterwards doesn't matter, the mountpoint stays the same.
When you look at the output of "mount" after renaming, you'll see that the name of the ISO has changed (same as you renamed) or in case of deletion, it shows (deleted) after the ISO filename.

Fred

Strictly speaking there is no delete in 'nix, only unlink - which in effect marks inodes as being free to be used. Think of it as two elements, the directories structure, that contains the file/folder name and a pointer to the first inode, and inodes - that point to where the data actually resides on disk (if data extends beyond a block another inode is pointed to such that you (system) can chain down all of the blocks to gleam the full data content. Similar in some respect to packets of data that traverse the net in blocks).

When a 'delete' is made, only the directory structure is updated to disassociate the chaining. inodes however are only updated after all open handles (which includes mounts) on that file are closed. So even if deleted (or renamed) you can still continue to use the opened version as its inode links/data are all still there and your opened process on that file remains as-is (you continue to see it). Once all open handles are closed on a deleted file then only then will the inodes will be marked as empty/free/available. At the time of deletion however, the directory entry will be flagged as deleted, so anyone else looking for that file wont see it.

That structure is why moves of large files within the same filesystem are very quick, as all that occurs is that the directory entry is updated (inodes (actual data location) remain unchanged. cp and rm however require chaining though the inodes/data.

This all assumes on the same filesystem. Difference can apply if moving files/folders across different filesystems.

So in your sfs terms, if you boot with main.sfs such that you have it open (mounted), and then rename main.sfs to oldmain.sfs ... your open position will continue to see all of the inodes/data as your process already has the first inode pointer stored (data chain). Others however would only see the availability of oldmain.sfs (as the directory entry will have been updated), which if they opened would link to the exact same first inode as your opened version. Similar with delete, if you have it open and delete it, then that prevents other processes from opening it, but the inodes/data continues to be available to your process in having it open. Close it, try to reopen it however and you'll be in the same boat as others - file not found (and the inodes will have been marked as free).

Posted: **Tue 12 Mar 2019, 05:39**

Thanks for the replies.
I will do some more checks

Posted: **Tue 12 Mar 2019, 18:26**

Hi everyone, new package "obkey" added to Stretch custom repo, install with Synaptic or from terminal with apt-get:

Code: Select all

apt-get update # update the package lists
apt-get install obkey

Run from Menu > Settings > Openbox key bindings

It's a comprehensive program based on python and gtk3 to set keyboard-shortcuts in Openbox, has nice list overview of all keybinds and has many options.
IMHO, it's better (easier to use and more advanced) than the existing (in custom repo) "openbox-keybinder" program.
Although it has some small bugs, e.g. deleting multiple keybinds doesn't always work as expected (also the undo option may not always work as expected), it's very usable once you know about these limitations.
Direct download deb package:
https://fredx181.github.io/StretchDog/i ... -1_all.deb
On github (forked from: https://github.com/luffah/obkey and made some small changes):
https://github.com/fredx181/obkey

Fred

Posted: **Sun 17 Mar 2019, 10:55**

Updated packages in repo:
- googledrivegui2 (Google-drive Filemanager)
- dropboxgui (Dropbox Filemanager)

Recently these stopped working, should be fixed now.
Install with Synaptic or apt-get (from terminal)

More info about googledrivegui2, here (specially see "Important:"):
http://www.murga-linux.com/puppy/viewto ... 85#1022085

Fred

Posted: **Fri 31 May 2019, 23:40**

Fred,

Did you maybe delete https://raw.githubusercontent.com/Debia ... 815.tar.gz and break the script?

Posted: **Sat 01 Jun 2019, 08:26**

dancytron wrote:Fred,

Did you maybe delete https://raw.githubusercontent.com/Debia ... 815.tar.gz and break the script?

Probably you used older mklive-stretch version?
Yes, I did delete dog-boot-stretch-20180815.tar.gz, but now uploaded it again for compatibility with older version of the script.
The latest update was in january http://murga-linux.com/puppy/viewtopic. ... 34#1015534
(dog-boot-stretch-20190115.tar.gz is the newest)
EDIT: Now updated, see next post.

Fred

Posted: **Sat 01 Jun 2019, 10:45**

*** Updated mklive-stretch script ***

Make executable:

Code: Select all

chmod +x mklive-stretch

And run on a linux filesystem e.g. ext2, ext3 or ext4

Changed: install newest Debian stable kernel (currently version 4.9.0-9 )

For info: the archive containing boot scripts, aufs and squashfs modules for v4.9.0-9 is now: dog-boot-stretch-20190601.tar.gz

Fred

Posted: **Sat 01 Jun 2019, 20:35**

fredx181 wrote:
dancytron wrote:Fred,

Did you maybe delete https://raw.githubusercontent.com/Debia ... 815.tar.gz and break the script?
Probably you used older mklive-stretch version?
Yes, I did delete dog-boot-stretch-20180815.tar.gz, but now uploaded it again for compatibility with older version of the script.
The latest update was in january http://murga-linux.com/puppy/viewtopic. ... 34#1015534
(dog-boot-stretch-20190115.tar.gz is the newest)
EDIT: Now updated, see next post.

Fred

I could have sworn I was up to date. My bad.

Just ran latest version and all worked (minimal iso 135 meg). Posting from it now running your firefox portable.

Posted: **Sat 01 Jun 2019, 21:06**

dancyron wrote:Just ran latest version and all worked (minimal iso 135 meg). Posting from it now running your firefox portable.

That's nice, I'm curious how you made it so small, can you post your build_setup.conf (or build_setup.txt ) ?

Fred

Posted: **Sat 01 Jun 2019, 23:13**

fredx181 wrote:
dancyron wrote:Just ran latest version and all worked (minimal iso 135 meg). Posting from it now running your firefox portable.
That's nice, I'm curious how you made it so small, can you post your build_setup.conf (or build_setup.txt ) ?

Fred

I basically chose "skip" and then took out more stuff.

Live system built on x86_64 for amd64 on Sat Jun 1 09:58:56 CDT 2019
Preinstalled on top of debootstrap build:
live-boot wget menu dialog apt-utils dbus ca-certificates apt-transport-https xserver-xorg-legacy

Base Install:
wget net-tools ifupdown wireless-tools sysvinit-core xserver-xorg-core xserver-xorg psmisc fuse x11-utils x11-xserver-utils dbus-x11 busybox sudo mawk xinit xterm pciutils usbutils file rsync dosfstools volumeicon-alsa alsa-utils pm-utils xdotool wmctrl desktop-file-utils mime-support cryptsetup-bin squashfs-tools fakeroot xserver-xorg-input-evdev pv xserver-xorg-input-synaptics ntfs-3g

Base Dog Packages Install:
yad gtkdialog obshutdown pup-volume-monitor peasywifi chpupsocket edit-sfs-pcmanfm filemnt-pcmanfm remaster-scripts quick-remaster apt2sfs sfsload fixdepinstall greybird-theme-dd-stretch makedebpackage peasyclock

Applications Install:
leafpad viewnior pfind

Desktop Environment Packages:
openbox obconf pcmanfm lxpanel lxrandr lxinput lxappearance

Firmware:

Extra Dog Packages:

Remove xserver-xorg-video-vmware=TRUE

Posted: **Sun 02 Jun 2019, 08:01**

dancytron wrote:I basically chose "skip" and then took out more stuff.
....

Thanks, I tested, without synaptic and browser probably makes the most difference (and xz compressed .squashfs), not bad though, 135MB for a system with X , the Desktop managed by pcmanfm and some useful applications included.

For who is interested, I added this setup, 'openbox_min.conf' to the configs:
https://debiandog.github.io/MakeLive/co ... x_min.conf
(to download, right-click 'Save link as')
To build, run e.g:

Code: Select all

./mklive-stretch -gui /path/to/openbox_min.conf

Fred

Posted: **Sun 02 Jun 2019, 16:34**

fredx181 wrote:
dancytron wrote:I basically chose "skip" and then took out more stuff.
....

Thanks, I tested, without synaptic and browser probably makes the most difference (and xz compressed .squashfs), not bad though, 135MB for a system with X , the Desktop managed by pcmanfm and some useful applications included.

For who is interested, I added this setup, 'openbox_min.conf' to the configs:
https://debiandog.github.io/MakeLive/co ... x_min.conf
(to download, right-click 'Save link as')
To build, run e.g:
Code: Select all
./mklive-stretch -gui /path/to/openbox_min.conf
Fred

And no drivers, so if your computer needs drivers it may not work.

(old)Puppy Linux Discussion Forum

Create Debian 9 (Stretch) minimal ISO similar to DebianDog

Technical question about squashfs

Re: Technical question about squashfs

Create Debian 9 (Stretch) minimal ISO similar to DebianDog

Re: Create Debian 9 (Stretch) minimal ISO similar to DebianDog

Re: Technical question about squashfs

Updated packages: googledrivegui2 and dropboxgui