Saluki

[jemimah]

I have one sfs that crashes the kernel if I try to unload it. AUFS bug I suppose.

i had it with 2 out of 2 I tried (Firefox, OOo4kids) and indeed looks like is a 3.x kernel and/or aufs issue. I had no problem with the same setup and SFSs but a 2.6.35 kernel I made.
Maybe that's why Slacko also moved to Unionfs for the 3.2.x kernel. Though unionfs appears to have its own issues
Edit: Looks like the 3.2 kernel uses 2.1 aufs utils! Is this the case? The developer strongly counter advises this.

Here is aufsutils3. I'm am unclear about whether the kernel needs to be recompiled or if I can just add this to woof (and remove version 2).

[jemimah]

It would appear that tmount ignores fstab entires.
If you define a mount point in fstab for a device that is different than "/mnt/device_label" then a) the mount/unmount widget can not handle it and calls pmount and b) after mounting shows device -> label as mounted and device -> fstab_entry as unmounted.
Trying to mount device -> fstab_entry, unmounts device -> label.
I would think that fstab entries should be considered first.

You'd need to add new entries to /etc/fstab.d/static_entries. But yes, I think it does pretty much ignore the fstab and mount things how it wants them.

[jemimah]

new pets:
truecrypt - encryption
figaros password manager
tangogps - maps
tucan - download manager
tribler - p2p client

thank jemima

i've been testing tribler and it appears to be working well any pointers you could give me at getting it working on other puppy's

You can probably use my tribbler pet but the dependencies need to be compiled on the target.

• Python
  Pygtk
  wxPython
  wxGTK
  M2Crypto
  apsw

[jemimah]

Can anyone explain why Opera cant see libflashplayer.so but Seamonkey can? I have not had this before and its driving me a bit nuts
If it has been previously covered, please point me to the post. I have set the path within Opera prefs, to no avail.

If you install flashplayer from the ppm it will pull in the dependency you need.

[pemasu]

I'm am unclear about whether the kernel needs to be recompiled or if I can just add this to woof (and remove version 2).

I do not think that kernel will need to be recompiled to use aufs3-util.

But to get logitech device unifying feature compiled in...recompile would be useful.

[mavrothal]

Here is aufsutils3. I'm am unclear about whether the kernel needs to be recompiled or if I can just add this to woof (and remove version 2).

You may want to compile it -shared instead of the default -static. Is 20 times smaller.
You do not need to recompile the kernel for that but for optimal results is suggested to compile aufs-utils against the actual kernel to be used with.

[pemasu]

I am recompiling the 3.2.8 kernel with only that logitech unifying module change...it is now compiled in.

I will upload the whole stuff to the repo this evening. This time I will check manually the symlinks. 01micko has struggled with symlink implementation according to Slacko thread so...I suppose they will be fixed in script soon.

[mavrothal]

It would appear that tmount ignores fstab entires.
If you define a mount point in fstab for a device that is different than "/mnt/device_label" then a) the mount/unmount widget can not handle it and calls pmount and b) after mounting shows device -> label as mounted and device -> fstab_entry as unmounted.
Trying to mount device -> fstab_entry, unmounts device -> label.
I would think that fstab entries should be considered first.

You'd need to add new entries to /etc/fstab.d/static_entries. But yes, I think it does pretty much ignore the fstab and mount things how it wants them.

That's where I put the "fstab" entries in and the attached patch respects it.

[jemimah]

I am recompiling the 3.2.8 kernel with only that logitech unifying module change...it is now compiled in.

I will upload the whole stuff to the repo this evening. This time I will check manually the symlinks. 01micko has struggled with symlink implementation according to Slacko thread so...I suppose they will be fixed in script soon.

Thank you.

[jemimah]

Awesome! I will add this to the next release.

It would appear that tmount ignores fstab entires.
If you define a mount point in fstab for a device that is different than "/mnt/device_label" then a) the mount/unmount widget can not handle it and calls pmount and b) after mounting shows device -> label as mounted and device -> fstab_entry as unmounted.
Trying to mount device -> fstab_entry, unmounts device -> label.
I would think that fstab entries should be considered first.

You'd need to add new entries to /etc/fstab.d/static_entries. But yes, I think it does pretty much ignore the fstab and mount things how it wants them.

That's where I put the "fstab" entries in and the attached patch respects it.

[elroy]

I've gotta admit that I was a bit nervous about removing grub in favor of Grub4Dos in the official Saluki 013 release (I realized I could have aquired grub from the package manager...). I had never used G4D actually, because I don't have Windows on my system at all. Didn't think it applied to my setup. Had I known that it worked as simply as it does I would have switched to it a long time ago. I've always used grub, and in the past grub2 (yuch! convoluted), but Grub4Dos located my frugal installs to HD without even needing to really edit menu.lst. And, as a bonus, there's a menu.lst that I'm familiar with! Again, nice choice, J.

I erased the MBR from the terminal using 'dd if=/dev/null of=/dev/sdX bs=446 count=1' in order to keep my partition table intact, ran Grub4Dos, and saved about 15 minutes of web searching/menu.lst editing that I used to need in order to remember everything I needed to do for booting using grub. Grub was flexible, as is Grub4Dos. Excellent choice.

About the only suggestion I have is creating a new Grub4Dos that allows the user to erase the MBR with or without retaining the partition table. That would make it dummy-proof (nothing is ever dummy-proof, but I hope you get what I'm trying to convey...).

[jemimah]

I have one sfs that crashes the kernel if I try to unload it. AUFS bug I suppose.

i had it with 2 out of 2 I tried (Firefox, OOo4kids) and indeed looks like is a 3.x kernel and/or aufs issue. I had no problem with the same setup and SFSs but a 2.6.35 kernel I made.
Maybe that's why Slacko also moved to Unionfs for the 3.2.x kernel. Though unionfs appears to have its own issues
Edit: Looks like the 3.2 kernel uses 2.1 aufs utils! Is this the case? The developer strongly counter advises this.

Here is aufsutils3. I'm am unclear about whether the kernel needs to be recompiled or if I can just add this to woof (and remove version 2).

Just tested this out. Didn't get a kernel panic this time. Looks good.

[pemasu]

jemimah. You could test aufs3-util with this...or even better...recompile it using this kernel:
http://www.smokey01.com/pemasu/Saluki/k ... ve-hid-dj/

I compiled aufs3-util-shared in dpup. I changed just one "static" expression to "shared" in Makefile. The result as pet > 26 k. Impressive. Aufs3-util-static as pet 1058 k. I wonder did I it right. Can the difference be so big ?
At least LibreOffice loads and unloads with sfs_load with no problem.

[aragon]

They are for me. What did you do differently?

tried opening from rox
thunar works

now trying to open an .ncd file
this is created by the notetaker in series 5 puppys
Anyone know which program this needs? Is it available in Saluki?

lobster,

should be notecase.


aragon

[smokey01]

Notecase is in the Saluki repo.

[mavrothal]

I compiled aufs3-util-shared in dpup. I changed just one "static" expression to "shared" in Makefile. The result as pet > 26 k. Impressive. Aufs3-util-static as pet 1058 k. I wonder did I it right. Can the difference be so big ?

Yes. My pet for the 3.1 kernel is 52K unstripped
Static includes a big chunk of kernel stuff.

[jemimah]

jemimah. You could test aufs3-util with this...or even better...recompile it using this kernel:
http://www.smokey01.com/pemasu/Saluki/k ... ve-hid-dj/

I compiled aufs3-util-shared in dpup. I changed just one "static" expression to "shared" in Makefile. The result as pet > 26 k. Impressive. Aufs3-util-static as pet 1058 k. I wonder did I it right. Can the difference be so big ?
At least LibreOffice loads and unloads with sfs_load with no problem.

I assume the headers will be the same. All you changed is the one module right?

Also, I think you put the dbup kernel source sfs in the saluki directory.

[pemasu]

Only change is that one module change:
CONFIG_HID_LOGITECH_DJ=y ....it was earlier =m

I had earlier compiled the bluetooth communication modules...as modules...
which was compatible addition with first preemptive kernel.
CONFIG_BT_L2CAP=y
CONFIG_BT_SCO=y
CONFIG_BT_RFCOMM=m
CONFIG_BT_RFCOMM_TTY=y
CONFIG_BT_BNEP=m
CONFIG_BT_BNEP_MC_FILTER=y
CONFIG_BT_BNEP_PROTO_FILTER=y
CONFIG_BT_HIDP=m

So...this latest kernel pet has those bluetooth modules and the change from one module to compiled in.

And the right kernel sources is uploading now
I have too many kernel-kit folders now...

[Geoffrey]

I'm in contact with a friend who isn't on the forum, he is trying Saluki 012 and 013, he appears to be having a problem with usb flash and sd cards,

he is unable to access the drives and has to reformat the drives before he can reinstall, the error he gets is

Code: Select all

Failed to open directory "home".
Error when getting information for file '/mnt/home/.Trash-0: Input/output error.

does anyone have an idea as to what may be causing this, it has happened on two usb flash drives and one sd card on two different PC's and is random in nature

Some more info would be needed but usually I/O errors is faulty hardware or firmware. That would be strange with 3 different devices so my second guess would be formating.
Linux utilities are doing a very poor job on sticks and cards, while anything above ext2 is very demanding for this media.
Does the problem exists with a factory fresh card or stick without reformatting?
My third guess would be aggressive power management. Some hardware cut power to USB/SD ports and on remount everything is a mess.
Some kernel logs would be useful anyway.
HTH

Hi mavrothal,

My friend has been looking at it a little more and seems to think it may be due to him deleting sfs file's in /mnt/home/ the files will fill /mnt/home/.Trash-0/files so if there isn't enough free space on the drive he gets the drive lockup.

It might pay to either move the files to another partition or permanently remove them when working with small usb flash and SD cards with large save file.

[kimathi]

jemimah wrote:

Code: Select all

I've compiled avahi and mdns. Please let me know if it works. 

Also I uploaded a new mpd. Let me know if it's fixed

Have problems with avahi though it should not be hard to resolve. First thing is that avahi doesnt like to run as root so i have to first create user avahi

Code: Select all

adduser avahi

then input a password. to start avahi by running in terminal

Code: Select all

avahi-daemon

it complains about avahi being unable to run something to do with dbus due to policy. Anyhow my only work around that works is to install the previous version that i had posted and then to clobber it by upgrading to you version of avahi, then everything works. I guess your avahi needs some way to autolaunch either through /etc/init.d script or a way to auto create avahi user or to use another existing user like spot or daemon. If you get avahi to auto start then all is solved.

Mpd works too though i realised you did not include a sample config file. I wonder if someone who has never seen one would be in a position to create one,attached here are my config files. Drop them in /etc and edit where necessary to point to where your music is stored. start them from console/terminal. Also for Mpd create a file in /var/mpd/tag_cache, for forked-daap create a file in /var/cache/forked-daap/songs3.db. Thats it they all work. Your Minidlna too does not have the config file its also attached here.