(old)Puppy Linux Discussion Forum

@seaside
Yes, we need more discussion on the script like pinstall.sh and puninstall.sh of the pet.
At this point of time, we can put a script at /etc/init.d or at /root/Startup, those will be called at load. The script at /etc/init.d is called with parameter 'start' after load, and with 'stop' BEFORE unload. But we also need to consider the compatibility with normal shutdown and reboot...

Tested the --list command. Works great. But i can not use --skip-fixmenus and --list together.

Is this the way it has to be or could this be changed?.

If i could use these two commands together, it would made me able to load the needed sfs's and while working with the wanted program (called by the run-script) all the dependencies of the dependent loaded sfs's would be loaded in background.

Then just calling one time fixmenus at the end of my run-script and everything would be fine. I think...

This would be user and programmer friendliness

----
hi, this is off topic but I am trying to help kocg with localisation but I don't know what I'm talking about.

I know people here have experience with localisation, the problem is here-
http://www.murga-linux.com/puppy/viewtopic.php?t=77253

any help appreciated ! ; -)
---

Hi shinobar,

The problems with sfs-load-on-the-fly 1.3, 'built-in' to Saluki 17 and 18, are described by gcmartin and myself, starting here:
http://www.murga-linux.com/puppy/viewto ... 856#617856.
jemimah requested that we test your 1.4 and 1.9. I decided to test 1.9 first, with the following results:
No menu entry> Solved edit categories argument in .desktop to System
Using sfs-load 1.9, loaded google-chrome-stable_12.0.742.112-r90304_i386.sfs. Attempted to run same: notified of possible permissions error. Attempted to unload: problem reported at above link reoccurred.
As the problem may have related only to the above google-chrome, used sfs-load 1.9 successfully to load and unload picasa3.sfs & blender-2.61, both from Saluki repository. While doing so, notice that that repository now contained google-chrome-stable_12.0.742.sfs.
Downloaded the same using via Saluki PPM's SFS Downloader which was able to load and run it. However, trying to unload it via sfs-load 1.9 generated the
above mentioned problem. After a hard-shut-down, the above google-chrome no longer appeared as loaded in sfs-load dialog.
Next loaded gimp-2.6.11 which was not created specifically for Saluki. It could be loaded, but not unloaded, generating the above described conditions. Actually, I think that sfs-load may, in fact, successfully unload a "problem" sfs, but "hangs" at updating its "loaded list." [Sorry, I don't have the technical expertise to better describe what appears to be happening]. I chose not to wait 10 or so minutes while Saluki shut down, so I just pulled the plug. On reboot, gimp was no longer in sfs-loaded "loaded list."
After the above reboot, I use sfs-load to load xnview, which also was not developed specifically for Saluki. Successfully loaded and unloaded.
I downloaded sfs-load 1.4, and used ppm to uninstall 1.9. Tried to install 1.4 and received a "failed to install" notice. Rebooted and was surprised
to find that Salulki's 'built-in' 1.3 was now active. Again attempted to install 1.4 and again received "failed to install" notice.

Edit: disregard discussion of 1.4. I just noticed that its download was "stopped" rather than "finished."

mikesLr

I think I've located the problem.

With kernel 3.2 the command
is causing kernel panics and oopses sometimes.

I changed it to use busybox umount, and it seems to work better.

Also
was crashing it sometimes. I just commented that out and it seems to work ok. What does that do? Is it important?

jemimah wrote:

Code: Select all

losetup -d "$LOOPDEV"

was crashing it sometimes. I just commented that out and it seems to work ok. What does that do? Is it important?

Great. I have made tests on pemasu's Dpup Exprimo 3.2.11 and 3.2.14 kernel. The loop device, should be automatically free by unmounting, but frequently it fails. There is no way to free them once it fails. And the command 'losetup -d "$LOOPDEV" hangs...

Ok, i removed all the 'losetup -d' command from sfs_load-1.9.1. The loop devices stick on if the freeing fails, and we take another loop device, as the Mad Tea Party in Alice.

Try sfs_load-1.9.1. Hopefully more stable and usable.
# 9 Apr 2012 v1.9.1: clean up whiteout once for load, busybox umount and not do 'losetup -d'(jemimah), where to mount(pup_ro?), always register EXTRASFSLIST, no wide search for full path

RSH wrote:Tested the --list command. Works great. But i can not use --skip-fixmenus and --list together.

Is this the way it has to be or could this be changed?.

If i could use these two commands together, it would made me able to load the needed sfs's and while working with the wanted program (called by the run-script) all the dependencies of the dependent loaded sfs's would be loaded in background.

Then just calling one time fixmenus at the end of my run-script and everything would be fine. I think...

This would be user and programmer friendliness

Hello, shinobar.

It is not necessary, to implement possibility of using --skip-fixmenus and --list at the same sfs_load command. I did solve this using a little script, that does exactly what i want.

So again, thank you very much to implement the --skip-fixmenus command/parameter, for having an ear to my suggestions and also for doing this great job - which means mainly your sfs_load in general.

Hi Shinobar,
Clicking sfs_load-1.9.1 on the first post of this thread results in downloading of pet named sfs_load-1.9. Note absence of the suffix ".1".

mikesLr

mikeslr wrote:Clicking sfs_load-1.9.1 on the first post of this thread results in downloading of pet named sfs_load-1.9. Note absence of the suffix ".1".

Thanks. Fixed.

Why does it now say restarting X is necessary after an unload?

I'm mainly just curious.

jemimah wrote:Why does it now say restarting X is necessary after an unload?

The alert "You may need to restart X to reflect changes." is issued if the sfs loaded/unloaded has fonts or startup scripts.
The changes have done by the sfs_load afterwork, but may not be perfect.

The other case, that the sfs_load does not warn but can be a problem, is treating /etc/profile.d.
You need to reboot PC or do 'exit' on console to reflect the contents of /etc/profile.d.

Hi, shinobar.

sfs_load 1.9 did crash.

I used
to load sfs file. sfs file is stored in /mnt/sdd1 which means /mnt/home if i would use a pupsave. This never did happen on any other version of sfs_load. I do every day, several times, load sfs files on the fly.

I had to push the reset button (like power off).

RSH wrote:sfs_load 1.9 did crash.

Will you try sfs_load-1.9.1?
What is the kernel? 2.6.33.2?

EDIT: if it occurs only with the LP2_Internet.sfs specific, it may be caused by the contents like the libraries under /usr/lib.

shinobar wrote: Will you try sfs_load-1.9.1?
What is the kernel? 2.6.33.2?

EDIT: if it occurs only with the LP2_Internet.sfs specific, it may be caused by the contents like the libraries under /usr/lib.

Yes, the kernel is 2.6.33.2.

The LP2_Internet is several times a day used and loaded and never did happen something similar. I will try to use 1.9.1.
Thanks - just didn't see there is 1.9.1.

Small fix and change.
#13 Apr 2012 v1.9.2: fix was missing mountedpart, skip fixmenus for 'start', on RAM with remastered, saluki stuff

Test release: sfs_load-1.9.3.pet (2012-05-01)
NLS (seperated from 1.9.3): sfs_load_NLS-1.9.3.pet
[url]Download: http://shino.pos.to/linux/puppy/[/url]
Small improvement from 1.9.2:
#1 Mar 2012 v1.9.3: fix was clean up whiteout do not see pup_z, incleased RESERVHDD/RAM, version info, jwm -reload without flicker
Messages are refined by jemimah, in the English NLS in the main PET.

Test release: sfs_load-1.9.4.pet (2012-05-07)
Download: http://shino.pos.to/linux/puppy/

#6 Mar 2012 v1.9.4: fix excess(>6) load was messing up with some version of busybox(lupu), fix was menu disappeared for excess sfs, fix was error but already mounted, avoid 'losetup -d' for some kernel
#6 Mar 2012 v1.9.4: fix 'restart X' warning, never use external download_file, 'Back to sfs_load' button, gtkdialog4

Note that this is a test release, but has many important fixes. All users and pupplet makers using version 1.9.x series are recommended to upgrade.

Typing "sfs_load --help" generates:-

usage: sfs_load [--unload] [--cli] [+/-][EXTRA_SFS_FILE]
Load or unload extra sfs files on-the-fly.

Two questions:-

1. What's the "--cli" option for?

2. Do "sfs_load --unload xxx.sfs" and "sfs_load -xxx.sfs" do the same thing? If not, what's the difference?

Many thanks.

puppypeter wrote:Typing "sfs_load --help" generates:-

usage: sfs_load [--unload] [--cli] [+/-][EXTRA_SFS_FILE]
Load or unload extra sfs files on-the-fly.

It depends on the version what option is accepted, but most recent sfs_load explains:

Test release: sfs_load-1.9.5.pet (2012-05-21)
#19 May 2012 v1.9.5: fix was no move but always copy
#18 May 2012 for lupu-528JP: SUPPORTSIG=yellow

Download: http://shino.pos.to/linux/puppy/
NLS (seperated as for 1.9.3 and later): sfs_load_NLS-1.9.3.pet