logoPuppy developer news:

from 3.00b to 3.00

left-arrow Older news

arrow-rightLater news

puppy 3.00 released 

The announcement and release notes are here:

Get it from here:
http://distro.ibiblio.org/pub/linux/dis ... uppylinux/

A note for frugal installers -- there is still a little quirk. When do a frugal install to a subdirectory, using the 'psubdir' boot parameter, at first boot Puppy cannot find the 'zdrv_300.sfs' file. It is found after the pup_save file is created.
Another quirk is that the zdrv file cannot be found if, when presented with a list of pup_save files, you choose '0' (none).
I decided that these and some other issues are small enough not to hold up release of 3.00 final.

Thanks to kirk for creating a .sfs of the kernel source. PaulBx1, I've put in your improved partview. Apologies to anyone else who contributed in some way but I haven't explicitly acknowledged in this Blog!

Great, Barry, I will post the news at http://puppylinux.org

Very impressive Barry -upgrade went nicely,

Thanks, i will download and install.
Thanks to Puppy still i can use my old k6-2 400 Mhz like a new pc. ;-)

Barry wrote: "Upgrading from earlier version of Puppy. No problem. If you run Puppy from live-CD, just boot the new CD and Puppy will automatically perform any required upgrading of your personal storage file/partition (usually pup_save.2fs file)."

This is what I keep hoping, but it hasn't happened in any of the ver 3s. All are similar in the problems I see. I have reported them all before, but feel some reports are getting overlooked because they get reported in various places. I hope you will again assign a SINGLE place for reporting problems for this version. That's why I'm writing here.

To briefly summarize: My configuration file has been upgraded from 2.13 through 2.17.1 without problems. But in each iteration of version 3:

Neither Seamonkey 1.08 or 1.12 (or any suite components) will run, although files for both are there. Should not 1.12 have replaced 1.08?

Until I get Seamonkey running again I will of course not know if the problem running Enigmail has neen fixed. (I'm not going into that again here ;-)

Abiword and text editing do not run. Such things are probably easily fixed by someone who knows how, not me.

A small problem (not special to ver 3) that was reportedly fixed twice has not been: In Pdict and in Pfind one cannot use the "enter" key to enter the search term, but has to click on "Pdict" and "Search" respectively.

Like many I look forward to each new step in Puppy's growth and know we will get past this. Thanks in advance.


I am trying this 3.0 on another laptop and it is better than 2.17
Network works finally on it and dhcp could find ip, but 2.17 never could.
New icons look great.
A problem at boot, when loading puppy to the ram with pfix=ram, it takes very long time !
Where can we write feedback and bugs etc for this new version ?
Well done Barry :-)

(ps comments get lost easily :-( )

I forgot to include console responses in my previous post:

# geanyshell
geany: error while loading shared libraries: libstdc++.so.6: cannot open shared object file: No such file or directory
# mozmail
/usr/lib/seamonkey-1.1.2/seamonkey-bin: error while loading shared libraries: libstdc++.so.6: cannot open shared object file: No such file or directory
# seamonkey
/usr/lib/seamonkey-1.1.2/seamonkey-bin: error while loading shared libraries: libstdc++.so.6: cannot open shared object file: No such file or directory
# abiword
abiword: error while loading shared libraries: libstdc++.so.6: cannot open shared object file: No such file or directory


I reported Puppy 3.00 bugs in this tread:


OK! libstdc6++.so.6 was missing.

I used PSI ti install libstdc6++.so.6.0.5.pup.

That did it for everything (except the "enter" problem, which also applies to PSI).


ok i have installed 3.00 (frugal) and not grub will not boot it. It seems to have configured the grub menu.lst wrong

the menu list says that windows is on hda8 fat32 but that is where puppy is and grub is in a separate linux partition hda5 ext2

get a kernel panic no init found

Well, may as well get it all out, so back to Enigmail, aka OpenPgp in Seamonkey mail. A lot of Puppy users don't need this - some don't seem to need printers or scanners either. But I use Gnupg in Eudora windows everyday.

To summarize: It has never worked for me in Seamonkey 1.08 or 1.12 in Puppy. I just tried the Enigmail extension in Puppy 3 (Seamonkey 1.12) and it could not be installed for "unexpected error -201"

Then I installed the new Gslapt into Puppy 3 and thereby installed Seamonkey 1.14. This time the extension was accepted and OpenPgp appeared in the Seamonkey menu, but every time I touch it Seamonkey shuts down immediately, same as on previous attempts.

Note: I previously got it to work _only_ when I downloaded Seamonkey 1.14 directly from the Mozilla site and manually, albeit clumsily installed it.

So I think the Slackware package is leaving something out. On the other hand the PCLinuxOS Synaptic package installs and works great!(in PCLinuxOS). Is there a chance that someone could recompile Seamonkey 1.14 so this worked in Puppy, or would it be possible for me to use the PCLinuxOS rpm with "unrpm," or such. I'm out of my element here.


Yes, Henry, get/install the unrpm dotpup, put the rpm file in a folder then in console "unrpm file.rpm". Copy the extracted files to the proper folders.

Barry, would you spare us some tips in humongous initrd creation in Puppy 3? This Puppy is going to school! :-)

Yes, Henry, get/install the unrpm dotpup, put the rpm file in a folder then in console "unrpm file.rpm". Copy the extracted files to the proper folders.

Barry, would you spare us some tips in humongous initrd creation in Puppy 3? This Puppy is going to school! :-)

Thanks, Raffy,

Guess I'm not too old to learn a bit of Linux. Sorry for taking up so much time here with things that I know are trivial compared with the major system advancements.

Anyway, I simply installed the current Seamonkey 1.1.4 from the Mozilla site into /usr/lib/ and deleted things labeled 1.0.8 and 1.1.4 and all is well seemingly, except that 1.1.4 apparently takes up more space than Barry's 1.1.2, which is probably why it works with the add-on.

Thanks, Barry, for a Linux that's fast and sturdy enough to stick with.


gxine crushes all the time

John Doe 
release notes look great.

can't wait to try it out and mess with it a bit.

the fall season is settling in here for the northern hemisphere and more time will soon be spent on puppy by me. been too busy riding my bicycle and fishing, to keep up.

are the reported gxine crashes due to xorg 7.2 i know that when i installed beryl xorg 7.2 and 3D drivers gxine was not at all happy with that i worked only if you didn't move the window around too much

Is there a torrent to download 3.0? The normal download is [i]dog[/i] slow right now.

Hi Barry,

After some time trying to have 3.0 booting on a virtual VMWare I found something that may be a hint for you and that may help you testing the kernel.

When I use the version with the recent kernel, and loglevel=7 the boot process stops when it identifies the Floppy.

If I use the retro kernel, it boots OK.

My laptop does not have a physical floppy but It doesn't matter if i create a floppy image for the VMWare session.

The new kernel won't boot. the retro kernel does.

luck_gh0st (luck_gh0st<at>yahoo.com) 

Puppy just gets better and better! Very good network wizard! I'm impressed that my WUSB54GC worked out of the box!

3.00 final... when? 

A beta3 made a quiet appearance, printing was fixed.

Today I applied a fix for pcmcia posted by pakt.

I propose to upload the 3.00final approximately 24 hours from now. I'm waiting on reports on a couple of things... the beta3 had dhcpcd rolled back to version 1.3.22, there is pakt's pcmcia fix (see here: http://www.murga-linux.com/puppy/viewtopic.php?t=22162).

The final will have the Gnumeric plugins package restored in the live-cd. The betas did not have it -- but it is a PET package.

I don't think anyone posted a anything about rc.network being "foregrounded" making any difference from "backgrounded". The beta1 and earlier had it backgrounded, beta2 and beta3 had it foregrounded.
So, I'll go back to backgrounded for the final.

Anything else... well, there will be a 3.01 :-)

1. I noticed that forum member Matu mentioned a workaround to solve the problem about frugal install with 'psubdir' parameter. I've done some research and found that my sound and network adapters was properly autodetected after I copied zdrv_222.sfs file to root directory on the same drive where the subdirectory with other Puppy-3.00beta3 files resides.

This is Puppy-3.00beta3 section from my menu.lst:

# Boot Puppy-3.00beta3
title Puppy-3.00beta3, idehd, Bp300b3
rootnoverify (hd0,4)
kernel /Bp300b3/vmlinuz PMEDIA=idehd psubdir=Bp300b3
initrd /Bp300b3/initrd.gz

2. At boot I noticed that Puppy is searching for Puppy files significantly longer then in previous versions when the message 'Searching for Puppy files in computer disk drives' is shown on the screen.

3. After first boot Puppy created and saved a personal file named pup_save-1.2fs to /Bp300b2 subdirectory where Puppy-3.00beta2 instalation files reside although psubdir is set to 'psubdir=Bp300b3' subdirectory.

After I manually moved the pup_save-1.2fs from /Bp300b2 to /Bp300b3 subdirectory and renamed it to pup_save.2fs there are no problem with boot anymore.

4. I still can't established my ADSL internet connection although my D-Link DFE-530TX PCI Fast Ethernet Adapter (rev.C) is autodetected. When trying to start connection Roaring Penguin stops with this message shown on the screen:

Error starting connection: TIMED OUT

The Network Wizard properly detect 1 alive 'eth0' network interface but can't make network connection.

The problem with sound is solved.
CUPS works without any problem with my HP LaserJet 4P printer. Gxine plays videos in FLV format significantly better then in previous versions.
This Puppy seems to be very fast.

Barry Kauler 
Leon, are you sure that you have beta2 or beta3? There was a problem with the link at one time, where a forum thread indicated beta2 but the link was to beta1 download -- which got fixed I think.

Barry, I installed beta2 to /Bp300b2 and beta3 to /Bp300b3 subdirectory from this iso files:



Both subdirectories have pup_222.sfs file in them. I suppose that beta3 found pup_222.sfs in /Bp300b2 prior to /Bp300b3 subdirectory and ignored the value of 'psubdir=Bp300b3' parameter.

Hard Drive Install from Puppy-3.00beta2:

As posted by Crash on Fri Sep 28, 2007 12:24 am on:


"With a full install on a hard drive, the partition is not locked. MUT thinks you can mount it, which is OK, but it also thinks you can unmount it, which leads to really interesting results. GParted doesn't see the partition locked either."

I have also found this on two separate machines with IDE (ATA) internal Hard Drives.

Running the "mount" command shows:

rootfs on / type rootfs (rw)
/dev/root on / type reiserfs (rw)
none on /proc type proc (rw)
shmfs on /dev/shm type tmpfs (rw)
none on /dev/pts type devpts (rw)
none on /sys type sysfs (rw)
none on /proc/bus/usb type usbfs (rw)

On both machines the partition to which Puppy is installed does not show up.

Interestingly, a hard drive install to "/dev/sde2" on an all scsi machine (Compaq Proliant 1600) shows up correctly in Mut and partially correctly in "mount":

rootfs on / type rootfs (rw)
/dev/sde2 on / type reiserfs (rw)
none on /proc type proc (rw)
shmfs on /dev/shm type tmpfs (rw)
none on /dev/pts type devpts (rw)
none on /sys type sysfs (rw)
[no functioning USB on the Proliant]

I had to use "puppyinstaller-217" to install to the scsi Hard Drives. It would be desirable to retain this installer in in the forthcoming Puppy-3.00.

On the Proliant and other scsi Hard Drive machines I use a small cpio boot initramfs file to load the scsi drivers.

The initramfs file is easily generated from the scsi modules contained within the "zdrv_222.sfs" file using a slightly modified mkinitramfs archive from the Pardus Linux distribution.

Full details will follow later.

grndoor has solved the scsi boot problem!
Can this be incorporated easily in 3.0?

Hi Barry, sorry can't do any more testing for a while as my M.B died last night before I could check the Beta3. Good luck.
Tony (AKA veronicathecow)


This is /root/.etc/rc.d/PUPSTATE file from Puppy-3.00beta3 frugal install with 'psubdir=Bp300b3' parameter.

#these directories are unionfs layers in /initrd...
#The partition that has the pup_save file is mounted here...
#(in /initrd) ...note, /mnt/home is a link to it.
#this file has extra kernel drivers and firmware...

1. Zdrv file is set to /zdrv_222.sfs but it should be set to /Bp300b3/zdrv_222.sfs.

2. It's interesting then swap file is set to PSWAPFILE='hda5,vfat,/bpg200a4/pupswap.swp' where Grafpup frugal installation files reside although another /mnt/home/pupswap.swp swap file exists in the root directory.

[blockquote]"With a full install on a hard drive, the partition is not locked. MUT thinks you can mount it, which is OK, but it also thinks you can unmount it, which leads to really interesting results. GParted doesn't see the partition locked either."[/blockquote]
MUT is maintained by Jesse, so that would have to raised with him. I maintain Pmount.

Hi Barry & co,

Theres a few simple config settings for Mut in the top of the mut.tcl file, for example, directories that can't be unmounted.
Maybe add in the one for /mnt/home if thats the offending one?
I don't think I've done a hard-drive install (recently) of puppy so I'm not so sure how thats supposed to work these days.


Testing in vmware, Beta3 is the first version to complete booting in a new vm. Betas & alpha before would hang immediately after detecting floppy drive. (running puppy with loglevel=7).

SouthPaws (jaguar1<at>netzero.net) 
Hey Barry I don't know if this would be of any use or not...(something I came across)


Latest Beta 3 was missing UUID Dbus ID.

It wasn't a problem to gen manually and remove dashes. Don't know if this was addressed.

Looks good otherwise. :-)

Hi, Barry-

There's a new update for Flash 9, which came out on 1st October. Shall we include this in the final?



does puppy 3.0beta have the RTL8187 aka r8187 driver built in? if it does i may have a chance t try it out tomorrow...

any tips on getting that wg111v2 working with that driver would be welcome just PM me

barry the 3.00 you just uploaded doesn't boot in virtualbox either
i'll try it some more out later

doesn't finish booting with or without formated virtual disk attached
with or without a pup_save.sfs

puppy loglevel=7 didn't reveal any problems either it just stops at perform a switch root

seems that 2.17 suffers of this as well...

I must admit it is strange because 2.17 and 3.00 have booted fine on real PC's at least in my experience maybe it is a kernel incompatability but not likely

Barry Kauler 
blockquote]Latest Beta 3 was missing UUID Dbus ID.

It wasn't a problem to gen manually and remove dashes. Don't know if this was addressed.[/blockquote]
I don't know anything about dbus, I just threw it in becaause it is a dependency in slackware 12. Would you mind giving details on exactly what you did, and why?

Barry Kauler 
I think you would have to run 'df' and see that if it is a partition mounted on '/' then it cannot be unmounted. The full install is just a normal Linux hd install, so a partition would be on '/'.

Hi Barry,

Yup, Mut does that, but uses output from 'mount' instead of df.
However the / mount point partition is not listed in the 'mount' output for IDE hard disk installs, as grndoor pointed out above.
Perhaps this is a script not quite processing the mount-FULL output correctly?


Here is the 'after' snapshot 

I took this photo this morning, sometime before the judging:

It's stacked high showcasing the products and attractions of the Midwest. Agriculture is the mainstay, but mining is coming on-stream. In fact the entire northern half of Western Australia is enjoying a mining boom, mostly due to demand from China for raw materials. An iron-ore mine is to be built north-east of Perenjori soon, and other mines further north are already in production. You can't see the details in the photo, but it has a complete model open-cut iron mine, with toy machinery -- the guys who built this are really a bunch of big kids!

I'm quite enjoying this work, but it's highly physical on the go all day which I'm not accustomed to, and the problem is by the time I get to sit down in front of my PC to work on the Puppy project I'm dog-tired. And my fuse is short, as someone on the forum found out yesterday. Anyway, I'll keep at it, and aim to fix any show-stopper bugs in the beta2 and get the final out without too much delay. This job I've got is only temporary, until the 6th of October, although they may want me to help pull everything down on the day after the Show finishes. Then back onto Puppy full-time!

Bit disappointing, I'm afraid, Barry!
That circular orifice - I was expecting to see a steaming cyanide bath for gold extraction (and disposing of cats!). Or even a shearing pit with spectators sitting on the shelves now displaying produce.....

Reminds me of the 4H booth at the state fair when I was a kid!


That mostly dealt with cattle,pigs,and corn though. Next we'll expect a shot of you in your Pavilion Steward regalia!

The Perth Royal Show 

I am working in the Centenary Pavilion, which hosts displays of the ten districts of Western Australia. My district is the "Midwest". Most of the displays are created by volunteers from the districts (and the majority of those are senior citizens), though three of them have been done by commercial contractors. Basically, each district is advertising what it does. My lot started work on their display on the 22nd or 23rd (all volunteers, except for me) and I took this snapshot early on the 26th:

These guys spend a week building it, it goes on display for a week, then they have to pull it down.

Tomorrow is the judging ...yes, they are competing for a trophy. I get there at 7.30am and the judging will be at 9.00am. I'll take another snapshot of the finished product. The public get in the day after, Saturday the 29th, and I get to stand around wearing a white coat.

Find out more: http://www.perthroyalshow.com.au/

Puppy 3.00beta2 available 

Get it from here:
http://distro.ibiblio.org/pub/linux/dis ... 3.00beta2/
It is disguised as version 2.22.

The patched kernel source is here:

Hm, its 4am. I have to leave for work at 6.30am, so better get some sleep! (it's not that bad, I did have a nap between 11pm and 1am while waiting for PET packages to upload)

barry if i remember correctly you just upgraded gcc right? I have noticed that there are a few programs that don't like gcc 4.x. so will the old compilers still work in puppy 3.xx just rename ths sfs or would it be better to make a more generic gcc 3.x that will work ok in all the puppies? is that possible or not even needed?


This Puppy boots smoothly and works fast. It seems promising. Thank you for considering our suggestions.

There are only two major problems left.

1. D-Link DFE-530TX PCI Fast Ethernet Adapter (rev.C)
I can't establish Internet connection using Roaring Penguin PPPoE ADSL when set 'Ethernet Interface' to 'eth0' as in previous versions.

2. Intel(r)82801BA/BAM AC'97 Audio Controller
I can't setup ALSA sound.

I had none of this problem with previous versions including Puppy 2.17.1.

Just downloaded 3.00Beta2 and booted my Panasonic Toughbook CF-28
using "Puppy pfix=ram for a clean boot (we tested the prior 3.00
Betas on this machine and I wanted to avoid any old debris. This
laptop is currently loaded with Win2000Pro.

The prior two versions worked fine. This one failed at the video

I chose xorg then C h31.5-48.5v40-70 LCD Panel 1024-768

It was flagged as the Unconfirmed Maximum for monitor, OK for card.

I chose TEST and since I am using only the touchpad left the mouse
setting at ps/2.

The display shows some weird color patterns near the top then opened
to a Linux/Shark/Seabird background.

Since I was typing my observations here I didn't hit CTRL-ALT-Backspace within 60 seconds so it went to a weird mostly
green pattern across the top of the LCD and then only a Hard Reset
(Power down & back On) would break out.

I walked through the process again but this time hit the
CTRL-ALT-Backspace sequence quickly and despite more of the weird
green lines and other visual debris a couple of pop-up windows
later the Puppy was barking and all appeared well!

I went ahead and did a Frugal install then to set up GRUB.

From inside Geany I could find no BOOT folder anywhere (yes, I
toggled on Show Hidden Files).

I use Home to check the directories and finding no /boot/grub I
created them and then created menu.lst with the correct code.

Then I removed the CD and rebooted - choosing all of the defaults
to create a pup_save.2fs file on hda1 and copying the pup_222.sfs
to HDD.

Reboot to GRUB failed and so I hit Enter to continue as insructed
and was booted to Win2000Pro ... yuck!

I rebooted and allowed the CD to boot "normally".

I can see /boot/grub/menu.lst and all looks OK but my laptop
doesn't see it on reboot.

What do I need to do to force GRUB to be seen?

How do we fix this so that a newbie not be scared off by the levels
of complexity in creating a Frugal install?

Thanks! edoc

edoc, for a frugal installation, GRUB is not installed, and I don't recall anywhere in the installation process where there is any suggestion to the contrary. There is information displayed on how you can add the appropriate menu entry to an existing GRUB installation.
You can add an entry if you already have GRUB installed. If not, you would have to install it, and one way is to use the GRUB-config menu entry in the system menu.
GRUB is not installed, as many people have other boot managers such as LILO and can add menu entries to those.

Linux newbies scared off? Anyone doing a frugal or full Linux hard drive install is expected to know something about one of the boot managers, or use a boot floppy. Otherwise they should stick to running from live-CD or Flash drive. The Flash drive works really well, very simple, in fact I'm running off one right now.

Leon, please run Roaring Penguin from a terminal and report any error messages.

Leon, when you state that you "can't setup alsa sound", what do you mean? Was sound auto-detected in previous puppies? In this beta2 are there sound modules loaded? (run lsmod).

A forum thread for beta2 feedback:

cb88, I don't know if you can use the old gcc 3.x in Puppy 3.00. For a start, it would hve to be compiled to use the new glibc 2.5. If you have such an application (that won't compile), go back to Puppy 2.17 or earlier and compile it there. It will likely still work in pup 3.00. I haven't found any apps that won't compile so far.

Mark S 
From Mark-1st time responder-been using Puppy 2+ years.

On 2 computers I can't establish Internet connection using my normal auto DHCP. It just keeps cycling never setting up a connection. They both say they see connection link, though. Always worked before until 3.00 beta 1, same inop with beta 2.

Gateway 2100 with eepro100 card
Dell GX260 with 3c59x card.

><> Thanks, Mark


I posted my answers to your questions at the forum.


Idiot spammer 

There's an idiot spammer who has been targetting my blog recently, ever since the blog got mentioned on Distrowatch Weekly. As I have a graphical image security required for each post, I presume this person has to do each post manually. This person keeps trying. At first I tried blocking IP addresses, but that didn't work. A couple of weeks ago this person (perhaps I should not even be dignifying with the designation of "person") spammed almost all my blog posts back to May 2007 -- no further back as they are read-only -- which turns out to be a good precaution. It took me a couple of hours to get rid of them. After that I banned a word that was in all the spam posts, which worked until today -- now I have banned another word.

So, if you try to post, legitimate post, and it fails, it could be due to having a banned word. Unlikely, as there are only two banned words so far.

Hi Barry, I hate these people as well. Perhaps I'm wrong to do this but I usually post them a few dozen copies of some lovely large scenery pictures I have taken :-)

But then again spam smells so good when it's cooking!

Pleasant Puppy review please follow the link.

Barry and gang "Gung Ho"
The funnest distro on the planet.


Thank you barry, I won't open mouth again, no bark.


Sorry about the link mess up!
Pleasant Puppy review please follow the link.


Verbose bootup 

I removed 'pfix=debug' in 3.00, that a few testers have commented on. Instead, the beta2 will support 'loglevel=n' where 'n' is the verbosity level. For lots of output to the screen while booting, use 'loglevel=7'.

There are two factors determining verbosity. One is that the 'init' script redirects most screen output to a logfile. The second is the kernel loglevel, and I have patched the kernel to a value of 3, which is very low verbosity. If you pass the boot parameter 'loglevel=7', not only does that increase the kernel verbosity, but the init script sees that there is a boot 'loglevel' parameter and does not redirect to a logfile, but lets all stdout and stderr go to the screen.

Cool idea, Barry. I'd suggest your comment, "For lots of output to the screen while booting, use 'loglevel=7' for the boot screen where the list of boot parameters is. In one simple sentence it says clearly what that boot parameter does and suggests its range of arguments.

"but the init script sees that there is a boot 'loglevel' parameter and does not redirect to a logfile, but lets all stdout and stderr go to the screen."

Does that mean we can pass loglevel=3 to make it to stop redirecting the output without increasing the actual verbosity? Not that it matters, I'm just an overly curious person ;-)

Actually, it could matter now that I think about it. If someone uses an old menu.lst file that includes loglevel=3 for some reason, that would make it output everything. Let's see... yep, mine has that. I don't remember if I put it there myself or if it got added automatically by the installer. I think I might have copied it from the isolinux.cfg file at some point.

NoteCase 1.6.9pre7 

I have communicated with Miroslav Rajcic about text formatting bugs that I discovered in NoteCase 1.6.5 and 1.6.6, and he supplied me with 1.6.9pre7, which seems to have fixed the problem.
Although this is not the 1.6.9final, I'll put it into pup 3.00beta2.

I think that I've just about got beta2 ready. Puppy-development has slowed down considerably since yesterday, but things are still happening. I think that beta2 will be uploaded about 24 hours from now.


Please check also the problems with Chooselocale country localization.

1. Timezone was not set to 'GMT+1' and country localization was not set to 'sl-SI' after I choosed 'slovene' keyboard at first boot.

2. The locale files can't be fetched from zdrv file. When 'FETCH' button is pressed this message is shown on the screen:
"Sorry, but there does not appear to be a connection to the Internet. This program needs to download locale files off the Internet.
Connect to the Internet then click the OKAY button..."



I copied the isolinux.cfg from 2.15 CE to 3.0 betha (kernel with ISOMaster file editor and I was able to start puppy on my DELL latitude CPi A, I'm running it, right now.

Please change it.

Barry Kauler 
AlfonsoL, change what? isolinux.cfg is just a text file. What is the content of the one that works for you?

Barry Kauler 
3.00 has gone back to the way 2.16 and earlier work. The keyboard layout dialog does not set timezone or locale. You have to click the 'setup' icon on the desktop to do that.

Barry Kauler 
Also, chooselocale now must have an Intenet connection to download locales as they are not in zdrv.


Thank you for the explanation about Chooselocale.

I hope that beta2 will solved my network problem and I'll be able to establish ADSL Internet connection via my D-Link DFE-530TX PCI Fast Ethernet Adapter (rev.C) as in previous versions.


I copied all the file isolinux.cfg, only one line is different.

This don't work "append initrd=initrd.gz pmedia=cd"

this is OK "append root=/dev/ram0 initrd=initrd.gz loglevel=3 PMEDIA=idecd"

Barry Kauler 
'root=/dev/ram0' and 'loglevel=3' make no difference.
The only thing that perhaps could be a factor, though it seems strange, is 'pmedia=idecd' instead of 'pmedia=cd'.

rc.network back to foreground 

The 3.00 alpha and beta1 had the network setup at bootup "backgrounded", which speeds bootup. However, there was at least one report (veronicthecow) of network not working whereas it did in all previous puppies. Sometime ago when investigating backgrounding at bootup, I did experience some problems with 'ifplugstatus', the utility that detects if a network is 'alive'. So, I need to find out if this is the cause of the problem veronicathecow and maybe others have reported. The beta2 will have rc.nework not backgrounded. If it turns out not to be the cause, then I'll go back to backgrounded for the 3.00final.


Veronica missed everthing: sound and network.
So I asume it was the missing zdrv file problem of beta 1 - because I made the same experience. With the correct zdrv file everthing is running very well.

So please give us the speed of the backgrounded network. ;-);-)

Maybe have both, with a boot flag? pfix=slowboot, with fast (backgrounded) being default? Everything is solved by a boot flag. :-)

Tony (Aka Veronicathecow) 
Hi Barry, GeorgR might be right about that although backgrounding in the past did give me a similar problem (Not sure if it was exactly the same.)
Also after not getting any sound or network I did not check further if other drivers were missing.
Boot switches sounds a good idea.
Cheers mate

Me too had none of this problem with previous versions including Puppy 2.17.1 that works perfectly.

1. D-Link DFE-530TX PCI Fast Ethernet Adapter (rev.C)
I can't establish Internet connection using Roaring Penguin PPPoE ADSL when set 'Ethernet Interface' to 'eth0' as in previous versions.

2. Intel(r)82801BA/BAM AC'97 Audio Controller
I can't setup ALSA sound.

I saved the reports made with PupScan when using Puppy 2.17.1 and Puppy-p3.00beta so you can see the differences with TkDiff.
I attached the archive to my post at:


Barry, I thought we had fixed the 'ifplugstatus' problem. Ref:


Barry Kauler 
pakt, yes, I thought it was fixed, but just to be absolutely sure, beta2 will have rc.network 'foregrounded'. If everyone reports no difference then the final will definitely be 'backgrounded'.

The problem is, I think that I did some tests with ifplugstatus, running it in a terminal:
# ifplugstatus /dev/eth0
# ifplugstatus /dev/eth0 &
...on certain hardware they return different results (and I know, it doesn't make sense).
My memory is not quite clear on this though, so I want to do this final check with beta2.

Endless version-upgrades fixed 

My first day as "pavilion steward" is over, and I've got a little bit of energy left to do some work on Puppy...

BlackAdder reported "Tried an upgrade from 2.17 to 3.0B ( alias 2.21) on USB flash drive. Puppy thinks it is an upgrade on every subsequent reboot, and goes through the upgrade motions each time." I fixed it.

Another last minute Bug before next beta is coming:
Booting puppy 3 on a computer with AMD 690G chipset ends with message "puppyserialdetect is running".

I am aware of the problems I have to expect due to the "new" chipset but stopping booting already at this point seems to early.
Do I have an option to let puppy show more detailed booting information on the screen?

There is "pfix=debug", but I don't think that has worked for a while. It doesn't do anything in 3.00, that I can tell. Sure would like that fixed; it would make all these boot problems more solvable.

Oberon85 (Oberon85<at>gmail.com) 
Just did a frugal install of 3.0 next to 2.17.1 on my hd. Running great! good sound, networking, appearance, and boots to the highest resolution my laptop can show!
thanks for great work!

See our Boot Parameters wiki page for normal debug boot options and if someone wants to contribute what [pre]pfix=#debug[/pre] actually does or has other updates, please feel to register on the wiki and edit the page appropriately.

Nothing new information with pfix=#debug option while booting an AMD 690G Chipset.

I tried again puppy 220 because I am able to enter terminal commands with pressing ctrl c after puppy is hanging with message "puppyserialdetect is running".
(With 300 there is only a reaction when pressing ctrl alt del)

I found a lot of logfiles but at first look no special hint for this dead loop situation when booting.
Any tip which file to check to come closer to the point?

AlfonsoL (falfonsolb<at>yahoo.com.mx) 
Hi Barry.

I copied the isolinux.cfg from 2.15 CE to 3.0 betha (kernel with ISOMaster file editor and I was able to start puppy on my DELL latitude CPi A, I'm running it, right now.

I hope this help you to fix it.

Great job, Barry.

No wonder I couldn't get pfix=debug to work. :SURPRISED

Strange it was done with a pound sign. I assume there must be a good reason.

Georg, it looks like there is another boot parameter, loglevel=7. Might be worth a try.

Also if you can get 220 out of the hang, the last thing in the logs should tell what boot step it was working on. You'd need to look at the boot script to figure out what was next.

Did we really have a 220? Guess I missed that...

slightly OT, but what is a "Pavilion steward?"

Barry Kauler 
Kind of like a flight steward, except I'm in a pavilion. And I don't serve food or bring pillows.

Barry Kauler 

PRename upgraded 

'plinej' has updated his nice batch file-renamer program from v0.4 to v0.6. This will also be in the beta2.


I PMed you my error messages on the murga forum seems that pppd is now the only thing preventing connection as the modem is now working

any thoughts on how to improve connection speed?

if any of you guys have an lt modem that is not working try

#modprobe ltmodem
#modprobe ltserial

then use pupdial or whatever

Subito Piano 
Don't know where else to put this, but maybe this something to keep an eye on? machboot

Pmirror upgraded 

zigbert, who is one of our most enthusiastic coders (a couple of other wonderful apps written by zigbert are Pbackup and Pfind), has upgraded PuppyMirror v0.1 to Pmirror v0.2. This will be in the beta2.

Jwmconfig volunteer needed 

'thoughtjourney' wrote Jwmconfig package originally, back in 2005, and there have been various contributions since then. 'kwiller' contributed in 2006, and 'MU' has also had some input.

I want to get rid of the 'gtkdialog' executable. Puppy currently has three different (incompatible) versions of this program, gtkdialog, gtkdialog2 and gtkdialog3. I want to retain just the last two.

I have been going through and finding what still uses the old gtkdialog executable, and Jwmconfig does. My request is for someone to to get the 'jwmconfig-1jul07.pet' package off ibiblio and install it in 3.00beta and then fix it to work with gtkdialog3.

Note, 3.00beta has 'jwmconfig-5may07-1.pet', which is also on ibiblio, but awhile back I created the '1jul07' version with some mods to work better with JWM2 -- but I forgot to select that one for the beta.

If anyone wants to volunteer, could you post here to let everyone know? MU? zigbert?

Sorry if it is not the correct place :
proposition for future puppy :

1 Keep only usefull progs : Gtail not so usefull ? Same for Regexpviewer. Gdmap not so usefull. Same for prename .

2 Keep only one program to do one thing : One editor is enough ? uxplor. why 2 file finders ? 2 files editors ? 2 calculators ? 2 cd ripper ? 2 dvd writer ? Keep only one prog to do all theme settings.

3 Many progs in PETget package manager : Maybe they can be classified into categories ?

Barry Kauler 
linuxcbon, this question has been asked many times in the past. Firstly, some of those apps you don't want, others like very much. Scondly, some of the duplicate apps are so small that it makes virtually no difference o the size of the iso file -- so I have opted to give people choice where it doesn't impact the live-cd size.
Note, uXplor has gone in 3.00.

thanks for your answer, I will try to come with better questions :-)

duplicate apps help with hardware issues (at least one will work,usually).
Original topic- still a C newbie here. But learning.

I have already written a Jwmconfig application in murgaLua if you were interested. Just let me know and I will update it and finish it for you.

I'll take a look


Barry, is this going to break dotpups, etc. that use gtkdialog? Maybe better keep it, unless it is really huge...

or write a script that detects if a dotpup needs gtkdialog then prompt the user to install it and give the option to do it automatically would that even save space?

Dialup bugfix, please test 

Those who have been testing Internet dialup with 3.00beta, could please try this. Before starting PupDial or whatever dialup program you wish to use, open a terminal and type this:
# modprobe -vn crc-ccitt
# modprobe -vn zlib-deflate
...then try dialing the Internet.

# modprobe -vn crc-ccitt
# modprobe -vn zlib-deflate
...then try dialing the Internet.

Dialup works now

While you are addressing connectivity issues, could you replace the DHCP daemon (dhcpcd) with an older version (1.3.22 or 2.0.4 from slackware), to avoid the MAC-address misidentification issue I described in the Bugs forum, http://www.murga-linux.com/puppy/viewto?
It is a showstopper for me because I do not feel safe running the new Puppys while a telecommuter is on the same LAN; IT dept. sometimes blames a problem on my router -- the misbehavior exposes Puppy to be an apparent possible "cause".

wolf pup 
i cant test 3,00 beta, because grub gives error 13 unsupported executable format + md5sum of 3.00 beta is wrong. ill try downloading again today.

barry i think it is a problem with my modem getting detected
it is an ltmodem

device id 0x0440
vendor id 0x11c1 (int)

i beleive that the modem is a fujitsu (also googled the id and seems to confirm that) some other parts are also fujitsu made

those are the ids that pupscan reports in 2.16 where dialup is working

i tried the commands above but no change 3.00beta still no dialup for me

is there any other information i could send you to help

it would be nice if pupscan could have an option to output all the data to a gziped file perhaps even include dogal's hardware test in it

cb88, 11c1:0440 should have been recognised as an ltmodem, so you should have the ltserial and ltmodem modules loaded after bootup. Type 'lsmod' in a terminal and find out.

cb88, I looked at the pppd crashing log that you sent me. What I posted at the beginning of this thread fixes that exact problem.

Yes, dhcpcd-3.1.4 which I contributed here
has misbehaved with certain hardware.
This is the problem with updating packages; we optimistically think that the new versions will be better ... then often discover they have some new problem.

I see that version 3.1.5 is out - http://dhcpcd.berlios.de/
Do we trust it?

If we revert to an older version, be careful; when Puppy 2.17 was released it had a new kernel with newer wireless extensions, and we discovered that the version of dhcpcd carried over from Puppy 2.16 could cause a freeze with certain wifi adaptors.

Older versions are here -
and maybe an older version will be OK if it's compiled in an environment with the correct wireless extensions headers ... but it's difficult to know how many versions older we should use.

Barry, in the middle of this page:
Firefox had trouble with a missing "libcrypto" link, I believe in the context of dialup, although I think he was working with packaged dialers rather than standard Puppy ones. You might want to include this link in your dialup fixes.

Barry Kauler 
PaulBx1, yes, I have already put that libcrypto symlink into puppy, just forgot to mention it on this blog.

# modprobe -vn crc-ccitt
# modprobe -vn zlib-deflate

Fixed Dialup here too!

More multisession bugfixes 
I've fixed a couple of niggly little long-time issues.

There was something strange if decided not to save a session, a message came up stating could remove the CD before shutdown, but then another message came up wanting it back.

There was a problem when the CD/DVD gets full and have to migrate to a new blank CD/DVD. The new CD/DVD wasn't getting created properly. I discovered this problem just by examining the code, and have not yet actually tested it.

There's a message displayed in blue text, "IF STOPPED AT THIS MSG, MAKE SURE TRAY IS CLOSED THEN PRESS THE ENTER KEY:". I introduced this awhile back, for reasons I can't exactly remember. I think there was some issue about the tray opening then closing, which in the case of a laptop has to be done manually. The thing is though, I'm not getting that spurious open-then-close thing anymore. I would like to get feedback after the beta2 is released, whether that message is needed anymore. If you remember anything about that mesage and why you needed it, let me know, and in particular let me know if the beta2 doesn't need it anymore.

Multisession working 

I fixed the remaining bugs and got multisession working, though I've only tested with a CD-R so far. There are a couple of issues to fix at shutdown, I'll look at that tonight.

'Leon' and 'ICPUG' suggested that if a 'psubdir' boot parameter is passed, then Puppy should only look in that directory for a pup_save file. Okay, done.

Yesterday I had a pleasant surprise when I went to the Post Office -- two parcels waiting. I like opening surprise parcels. Ian had read in this blog that I don't have any CD-RWs, so he sent me a ten-pack, plus three compact DVD-Rs. The other parcel was from Chris, who sent me ...wait for it... a book of poetry, that he himself had written! Chris also wrote on my behalf to Senator Helen Coonan, regarding my mediocre Internet access -- she actually replied, said reply forwarded to me in the parcel. Well, it looks like one of those standard letters where the blanks get filled in -- it was actually one of her underlings that replied. Chris also included a few CDs with assorted software.
Guys, you're too kind!

Monday 24th I start work, and I'm not sure if I'll have the beta2 ready. If not, I'll try and put in at least a couple of hours each evening and hopefully get it out by the 25th or 26th. There are still a few things on the to-do list, especially dialup seems non-functional, from user feedback.

mkisofs is kaput 

I'm on the trail of the bugs preventing multisession-Puppy from working. V3.00 alpha and beta testers reported unable to save a session back to the CD/DVD. Well, I found the cause, and it is strange.

3.00alpha/beta has the cdrtools package from Slackware 12, version 2.01.01a23. This has the 'cdrecord' and 'mkisofs' utilities. It was an oversight, I did not upgrade the dvd+rw-tools package, still using the version from Puppy 2.17 -- this has the 'growisofs' utility.

To save a session to a DVD, the shutdown script executes growisofs, which in turn calls mkisofs. What I found is that the mkisofs utility simply doesn't work. It just exits with ":-( mkisofs has failed: 0".
I tried some tests with mkisofs, it just wouldn't work, so I substituted the mkisofs from Puppy 2.17, and voila, we're in business!

Puppy 2.17 has cdrtools v2.01.01a10, so same minor version number, just a difference in the alpha number. Hah hah, I complained recently that one of my scipts was broken as the output from 'cdrecord -scanbus' had changed -- well it changed somewhere between the alpha number 10 and 23, so the author is not beyond making fundamental changes from one alpha release to the next.

I did a google on "mkisofs has failed", and surprise, found my own blog post back in May 4th this year:
http://www.puppylinux.com/news/index.ph ... 505-075947
...so, I had forgotten, I already discovered that mkisofs has, as I commented then "gone down the drain".

I have fixed one bug, but have not yet got multisession working, as although a session gets saved there is still a bug at bootup -- will tackle that next.

Various little fixes and improvements 

'eccentric' reported that Roaring Penguin pppoe did not work. The reason is that it needs the 'id' program, but is hardwired to run it at /bin/id, whereas it is now at /usr/bin/id. I don't want to move it, nor even create a symlink. Really, a script should not rely on utilities being at fixed locations. Anyway, I edited the Roaring Penguin script and fixed it, took out the path.

'8-bit' reported that the Samba utility 'smbspool' is required to access a Windows network printer. Well, no, it isn't small, it's 830K, but perhaps this is an important utility to have so I put it in.

'rarsa' has a directory at the top level in a hard drive partition that is named "2007-07-10 Vacaciones Mexico", which Puppy thinks is a multisession folder, then suffers from a kernel panic as it isn't.
Should Puppy even be looking on a hard drive for a multisession folder? Yes, I want to explore another mode of storage to hd.
A multisession folder has a name like this:, '2007-09-21-06-52', and I have refined the code to look for exactly that format.

'PaulBx1' reported one small issue with a shutdown message, "pup_save.2fs mounted directly top layer, session already saved.", that might cause confusion if the user is using a pup_save with a different name. So, I have modified the message to display the exact name of the pup_save file.

'Making filesystem usable' fail bugfix 

Some 3.00beta testers reported a 'fail' at 'Making filesystem usable' message during the first boot, before a pup_save has been created.
I found the bug in the rc.sysinit script. Actually, it is not really a fail, just a false alarm. Note, this bug only occurs on PCs without any Linux swap partition.

The PCs that I normally use for testing Puppy all have Linux swap partitions, so I didn't get that bug. However today I tested on my oldest PC, which has a Cyrix 8x86 CPU, 128MB RAM, and a single FAT partition on the hard drive. That's when the bug showed up and I was able to track it down.

However, I think a couple of people reported getting a 'fail' at the 'performing switch_root' message. Unfortunately I cannot reproduce that one, but I'll study the code carefully. After all, the bug that I found in the rc.sysinit script was a silly little thing that was just staring me in the face and I didn't see it.

My old Cyrix PC has a quirk, and I don't recall this happening before. Perhaps it is an Xorg 7.2 thing. It uses the 'sis' video driver, and when running the Xorg Wizard and I choose the 'TEST' button, it works but I have to press CTRL-ALT-BACKSPACE twice to terminate X. Pressing once, it stays in graphics mode with just the mouse cursor displayed.
So, I've added this information to the Xorg Wizard, informing that on some old PCs it may be required to press that key combination twice.

This bug was bugging me until I read your post here. I don't usually use a swap file. I made one with GParted and now 3.00 Beta boots to a full install on the hard drive in 20 seconds flat!

Multimedia upgrades 

Forum member 'henry' has been helping testing Puppy 3.00beta and has reported issues with using Gxine, the media player. When trying to play certain media (Real audio or video I think) there was a message that 'cook.so' or 'sipr.so' was required. I have now placed these codecs in Puppy.

henry also mentioned using PSI to install an extra codecs package. However, please note that earlier codecs packages may install to /usr/lib/win32, whereas now it must be /usr/lib/codecs -- all the distros have moved to this new directory name. PETget package manager will install a correct extra-codecs package, but even if you have a "wrong" one, no problem, just rename the directory.

I got Gxine working with my laptop. What happened was, I tried Xine-ui, which is an Xlib-based media player and got the same crash when tried to play a DVD. I experimented and found that the cause is XV mode of playback. When I chose Xshm it worked. XV gives higher performance, but note if you use Xvesa then XV is not available and you have to use Xshm -- which Gxine and Xine-ui both detect automatically. However when running Xorg, they try to use XV.
In both players, run the setup and choose 'Advanced' options then manually set to Xshm, save, then restart the player and all is well. Plays nicely on my laptop, no jerkiness.

For those who would like some choice, I have made Xine-ui into a PET package.

Forum member 'ArnaudN' has also been testing the 3.00beta and has reported the Pause etc controls for playing a video at youtube.com do not work.
This problem is due to the latest Flash player being used in Puppy ...now, this is going to set me off grumbling (another thing that will set me off is mention of how the kernel keeps changing and breaking things). We had a Flash9 player in Puppy 2.17 and I think it was about 6MB. Well, this latest has climbed to 7MB, yet is still buggy -- in fact nonfunctional.
The older Flash7 player is only 2MB, yet works fine with youtube.

I've taken Puppy back to the Flash7 player. If anyone wants/needs a Flash9 player then they can of course install it. There is an open-source Flash plugin under development and hopefully that will mature enough sometime soon.

It's a wild guess... that a different person or team of people wrote the Flash9 player than whoever wrote the Flash7 player. Whoever they are ...well, I'll restrain myself from further comments.

I continue the love affair with Notecase, the superb outliner application. It still isn't 100% bug free -- I made some text bold, also changed text color to green, but next time I looked at the document the formatting was gone. What is really great though is that the document is stored in SGML format, which is a super-set of XML or HTML, so I opened in a text editor and quickly found that NoteCase did not save the formatting tags properly -- it was easy to fix. NoteCase mostly gets text formatting right, but there's a hiccup sometimes when saving. Note, I'm keeping my Puppy project notes in a NoteCase document, and when I say that formatting was lost, I don't mean for all of the document, just occasionally on a particular highlighted text that I applied some bold and color to.
Note, even images get encoded as text and embedded in the one document.

Change of plan! My job at the Perth Royal Show is no longer Gatekeeper for the horse section. Here in Australia there is currently a scare about Equine Influenza -- this is a horse flue that is over most of the world but not in Australia -- until recently. A horse from Japan managed to infect some other horses in Australia, even though the horse was in a government quarantine station and the infected horses weren't. It quickly spread. Anyway, how it affects me is they decided to up the security at the horse gate -- two trained security guards will be on the gate.

I was offered an alternative job, as 'pavilion steward', which basically entails 'minding' one of the display pavilions. I have to start work on the 24th, even though the Show commences on the 29th, as exhibitors will be moving stuff in -- I have to make sure stuff goes in and nothing goes out!

It won't affect the Puppy project, as I'll still have evenings free, just that my input will slow down somewhat.

ok just a little info for ya

3.00beta is not detecting my laptops modem 2.16 does 2.17 doesn't

also have gateway desktop here and a BCM4212 v.90 56k modem is apparently detected in PupScan vendor:14e4 device:4212 kernel module:slamr

pupdial detects the modem in neither of these PCs

while the 2.16 did detect the modem in the laptop(compaq presario 1810) and gets 26kbs

also there seem to be some issues with the jwm configuration wizard
well now it seems to have dissapeared before i was geting a "config file corrupted" error and it was not saving certain changes

and the sound is still configured wrong on the gateway <SB live> must be unmuted in alsawizard before it works then it is fine

too bad flash9 causes issues

also netsurf seems decent not blazing fast but decent but i have yet to get the laptops modem working in 3.00beta to try it

well good night

Puppy 3.00Beta worked on both of the laptops we tested today.

One is a Panasonic Toughbook CF-28 and one a newer HP belong to a
visiting friend.

BTW: He was impressed that it handled his Broadcom internal wireless
nic as well as everything else he tested.

Will try some more tests of 3.00Beta and 3.00Throwback ... errr k2.6.18.1



Your ps script needs a minor change. It does not work correctly with the kp utility as the full command line parameters to ps are not being passed. The line exec ps-FULL should have $@ at the end to correct this hiccup I think.

Note also that ps-FULL is only needed with ver 0.3 of the kp script to display the start time column as requested by vovchik. If you use ver 0.2 it works with bustybox ps but does not have the stime column.


The minimum system requirements for Flash 9 is definitely higher than for Flash 7, so just for that simple fact it would seem best in keeping w/ Puppy's minimalistic tendencies and stick with Macromedia Flash 7 instead of Adobe Flash 9 and let users add the latest if desired.

ALSA drivers upgraded to 1.0.15rc2 

"It speaks, Igor, it speaks!" ...ahem, I'm very pleased, my laptop has been silent since I upgraded to Puppy 2.17. The culprit is ALSA versions after 1.0.12. However I am pleased to report that the developer of the snd-hda-intel ALSA driver has fixed it in the recently-released 1.0.15rc2 -- and from the changelog it looks like sound is fixed for a range of other computers as well.
Note, I have an Acer Aspire 3681WXMi, with ALC883 sound chip.

I have never been able to play DVDs on my laptop -- Gxine crashes with an Xwindows 'BadAlloc' message. I'm hoping that will get fixed one day too. Note, I haven't yet tried any other media players.

Hi Barry er Igor ahem,

My Pidgin is silent no cooing....Gxine has crashed for me for awhile but seems ok now for cd/dvd....but this 3.0Beta with EZ Pup is much faster and will be my work a day OS.

I find rendering of web pages much better and Opera.pet works well...I missed that.

Have a great time with the horses..but watch where you step.


my screensaver does not blank th LCD again.


Barry Kauler 
The screensave works fine for me. Perhaps your problem has something to do with EZ Pup.

As for the "BadAlloc" message ... for the longest time I thought the only solution is to use Xvesa. But kirk found a solution, check it out here: ttp://www.murga-linux.com/puppy/viewtopic.php?t=14690

Glad my ALSA tip helped :-)

Can you make the ALSA upgrade available for us to test?


It would be even better if this ALSA driver were also included:

Barry Kauler 
Pakt, yes, you set me off on the right trail. Regarding getting it to test, well since I'm fixing lots of bugs it will be best to grab 3.00beta2, which I'll upload soon ... I start work on Monday, and I'll probably be dead tired in the evening, but I'll see if I can get beta2 organised and uploaded Monday night (GMT+8).

'pfix=ram' and 'zdrv' bugs fixed 

Not exactly a bug, but with 'pfix=ram' Puppy should not probe through the hard drives at bootup. Fixed.

There was a problem with finding the zdrv file under certain circumstances -- such as first boot after a frugal install. Fixed.

Did you know ?

There is nice improvement concerning frugal install with Puppy-p3.00beta.
I tested frugal install with psubdir=p300b and all Puppy files stored in /p300b directory.

This is the Puppy-p3.00beta boot section from my menu.lst:
title Puppy-p3.00beta, idehd, p300b
rootnoverify (hd0,4)
kernel /p300b/vmlinuz root=/dev/ram0 PMEDIA=idehd psubdir=p300b
initrd /p300b/initrd.gz

At every boot Puppy offers a menu to choose which personal file to use:
0 none
1 hda5 /p200/pup_save.3fs
2 hda5 /p300b/pup_save.2fs
3 hda5 /p2171/pup_save.2fs

We can choose which personal file to use at every boot.

- When mistakenly choosing personal file from previous version then Puppy would probably upgrade it.
- Searching for personal files at every boot increases boot time.

Possible solution
Additional parameter, something like 'psavefile' to specify which personal file to use. The menu to properly choose personal file can be created using Grub menu.lst file.

kernel /p300b/vmlinuz root=/dev/ram0 PMEDIA=idehd psubdir=p300b psavefile=p300b/pup_save.2fs

Jim Lyons (lyons<at>canada.com) 
Recent versions do not recognize my 3com 905B network card hence I have to revert to Windows to post this message :-(


In re: Puppy 2.17.1 USB stick, here's the domain ownership info: http://http://www.zoneedit.com/whois.ht


That should have been: Guest07/other:

In re: Puppy 2.17.1 USB stick, here's the domain ownership info: http://http://www.zoneedit.com/whois.ht

Sorry 'bout double post.


meh ... I give up. Remove one of the "http://" from the above URL

Leon - I'm with you all the way regarding the disadvantage of searching all the drives for pupsave files when a psubdir is defined.

However I think there is another possible solution.

Surely, the major reason for psubdir is to localise the puppy files used and keep separate subdirs for separate puppy versions. If so, then when we specify a psubdir we only want to use the pupsave in that directory. My solution would be:
When psubdir is specified search ONLY the psubdir directory for the pupsave file. If it is not there than act as if it is a new installation and only allow saving a pupsave file to that directory.


Stupid question: are these bug fixes for the final release or can they be had now?

Talking about probing the HDD at boot:

I found a strange error booting:

I had a folder with a long name with about 30 folders each with many pictures (one folder per day from my Vacations in Mexico)

When booting Puppy I would receive a message that it could not load that folder and then a kernel panic.

I have no idea why Puppy was trying to load that folder.

Thanks alienjeff for the link. SeaMonkey truncated it by itself. I do not know them though, but perhaps someone can tell ?

ICPUG, Yes, I think you are right. If you specify a psubdir, Puppy should be constrained to only looking inside it for the pup_save file. I'll make that change.

Rarsa, what is the name of that folder? I think Puppy is mistaking it for a multisession storage, as Puppy looks for a folder with a name composed of the date, something like 2007-10-11-05-05 ...not exactly like that, but similar.

Leon, note that 'root=/dev/ram0' is now optional. This parameter was required in previous puppies in which the 'initrd.gz' file contained a ext2 filesytem. For pup 3.00alpha/beta, 'initrd.gz' is now a cpio archive, which is just a collection of files stacked together along with directory information, and does not contain a filesystem. So, it cannot be mounted, and the kernel cannot mount it on a ramdisk.

There is no problem with leaving that boot parameter there. The kernel detects that 'initrd.gz' is a cpio archive and simply ignores 'root=/dev/ram0' if it exists. Instead, the kernel creates what is called a 'initramfs' and extracts the contents of initrd.gz to that.
One great advantage of a initramfs is that it is not a fixed size, unlike the ramdisk -- it just uses as much memory as required to hold the files, and has no maximum size limit other that the amount of free RAM.

Although, while I'm explaining this, I should complete the explanation -- although a ramdisk is a fixed size, after the initrd.gz is mounted, any unused space in the ramdisk is freed -- however the expanded initrd.gz filesystem cannot be any bigger than the max size of the ramdisk as compiled into the kernel, unless the 'ramdisk=' boot parameter is used to set the initial size of the ramdisk. Anyway, that's a historical note, as I've moved to the initramfs.

Jim Lyons, how recent? What was the last version that worked?


Thank you for your detailed explanation about 'root=/dev/ram0' parameter and initramfs.

When talking about frugal install, if psubdir is specified and Puppy should only look inside it for the pup_save file then if only one file is found the menu should not be shown to interrupt the boot process. Puppy 2.17.1 already works that way.


I have the same problem as Jim Lyons and I hadn't this problem and none of the other problem that I mention in this post with previous versions including Puppy 2.17.1 that works perfectly.

1. At first boot this message is shown on the screen:
Registering unionfs 2.1.2 (for
fuse init (API version 7.8)
Pausing for 60 s...

After pup_save file is saved and choosed at next boot this problem doesn't appears any more.

2. D-Link DFE-530TX PCI Fast Ethernet Adapter (rev.C) - I can't establish Internet connection using Roaring Penguin PPPoE ADSL when set Ethernet Interface to 'eth0' as in previous versions.

3. Intel(r)82801BA/BAM AC'97 Audio Controller - I can't setup ALSA sound.

4. Timezone was not set to 'GMT+1' and country localization was not set to 'sl-SI' after I choosed 'slovene' keyboard at first boot.

5. The locale files can't be fetched. When 'FETCH' button is pressed this message is shown on the screen:
"Sorry, but there does not appear to be a connection to the Internet. This program needs to download locale files off the Internet.
Connect to the Internet then click the OKAY button..."

6. When Roaring Penguin PPPoE ADSL Start button is pressed this message is shown on the screen:
"Error starting connection:
/usr/sbin/pppoe-start: line 44: /bin/id:
No such file or directory
pppoe-start: You must be root to run this script"

I saved the reports made with PupScan when using Puppy 2.17.1 and Puppy-p3.00beta so you can see the differences with TkDiff.
I attached the archive to my post at:


downloaded the beta(the 2.21.7) last night and it still does not run correctly in virtualbox (just stops if there is no drive or if pfix=ram is used) I had a pup_save on one of the virtual drives and i attached it it detected it and began upgrading so i assume that it was going to work but it was taking a long time so stoped it

I also ran it as a livecd on my laptop P2 300mz 64mb and it seemed pretty fast and netsurf started up ok pidgin also seemed to be working I will try to get on the net with it tonight and i let you know how that goes

also geting the same thing as leon

1. At first boot this message is shown on the screen:
Registering unionfs 2.1.2 (for
fuse init (API version 7.8)
Pausing for 60 s...

i'll have to us it some more to be of any real help though

Barry said:
--> Rarsa, what is the name of that folder? ... something like 2007-10-11-05-05 .

You got it. The folder was called "2007-07-10 Vacaciones Mexico"

Should Puppy be looking for multisession on a HDD?

What would you like to know about Linux CD Shop linuxcdshop.ca?

Guest07 and other

What would you like to know about Linux CD Shop linuxcdshop.ca?

Font rendering bug in Abiword, etc. 

3.00 beta testers have reported documents not displaying in Abiword, and various other font conversion problems. The reason for these problems is that I reduced the size of /usr/lib/gconv. This is a directory that is part of libc, and it has a lot of files -- about 6MB uncompressed. I am sure that some of them are for exotic unlikely font conversions, but deciding which can go is difficult. I chose a subset that seemed to work, but as we are finding, it doesn't for others.

So, I have restored the full gconv directory.

(c) Copyright Barry Kauler 2008, all rights reserved. http://puppylinux.com/