(old)Puppy Linux Discussion Forum

RetroTechGuy wrote:
Mem: 286964 282380 4584 0 22700
Swap: 546200 0 546200
Total: 833164 282380 550784

That 5 M left really cramps my style.

Run "Top" and you'll find that much of the "used" ram is actually in
cache and buffer.

Running "Htop" would show your ram being used as something
like 35 / 286964 at desktop.

rjbrewer wrote:

Run "Top" and you'll find that much of the "used" ram is actually in
cache and buffer.

Running "Htop" would show your ram being used as something
like 35 / 286964 at desktop.

Actually, the usage problem was that it persisted in loading the .sfs file to ram (even when pfix=noram was used, even after much fiddling with the init file). This was eating 100 MB that it didn't need to...

I have kludged a patch in, based on Pizza's post... Not pretty, but it left my memory free (and the computer is much more responsive already)...

Found it: http://www.murga-linux.com/puppy/viewto ... 197#267197

Thanks Pizza, that at least got me to the correct location.

The changes above apparently don't fix the problem. I also added

&& [ "$PNORAM" != "yes" ]

to the line above the one you referenced, (since it apparently also made decisions about "heaps of ram"). Unfortunately, neither of these actually fixed the problem.

I ended up commenting out the entire section from

if [ "$COPY2RAM" = "yes" ];then

through the "else", and rebuilt initrd. That cured it... It's certainly not pretty, and not portable, but is temporarily patched, until I can track the problem further.

Thanks!

BTW, if anyone else is planning to tinker with this -- be aware that you need to run Pizza's commands on an ext2 partition, not a fat32...symbolic links won't correctly transfer to the fat32 (and you'll munge your files)

[/quote]

I found this problem in pyppylinux 3.0X but it is still actual in newest version.
When some program starts in background from xinitrc script it causes one or more "zombi" jwm processes. (possibly bug?)
Suggestion:
Jwm parses the "StartupCommand" tag in it's configuration file. All programs
placed after this tag run automatically when jwm starts.
I remove programs (clean desk_icons,pup_event_frontend,delayedrun,etc...) from xinitrc and place them to jwm configuration file using this tag. Works fine, without "zombies".
Another problem:
When system is fully (no frugal) installed and runs with disabled eventmanager, (I don't use it)pmount can't mount the usb mass-storage drives because the usb-storage module is not loaded.

Attempted to burn multisession CD-R (also tried CD-RW) the boot hangs at loading vmlnuz.

after 10 minutes or so, it gives some sort of error about trying to re-boot then the same issue happens.

Tried several CDs burnt at low speed

used the same procedure to burn multi session CD-R of 4.21 with no issue

tgz2pet says: "chmod: missing operand after +w".
Advise:
The 9-th string of this script "chmod +w $TARBALL" must be changed to
"chmod 755 $TARBALL".

Kernel 2.6.30.5 has dropped support to Acer Travelmate backlight brightness.
Previous Puppies/puplets/derivatives aka 4.12 and 4.21 has backlight brightness support through video module which was loaded automatically.

Now video module does not load backlight folders and files. Even Acer-wmi does not work. It creates folders and files ok but echo... console command does not change values.

I havent found cure to the problem. Brightness is randomly dim or something other or bright with every boot so something affects brighness.

Also other laptop owners has had samekind problems.

So If anybody knows cure without compiling kernel to Acer Travelmate 6592G models I would be very happy Puppy owner

All the best and Merry Christmas

4.3.1 bug?

The Pudd welcome screen says "This program will copy any drive, partition, or image-file to another drive, partition, or file." I tried to copy an image file, and Pudd refused, on the grounds that the file was not a directory.

Also, when copying to a USB stick, it does not show up on the drive/partitions list. So I click on the USB button (bottom left) and get a red screen saying that it can't be unmounted. It is necessary to mount it, so that Pudd can unmount it and stay happy.

gerry

Sorry Pudd- it's not your fault! If I open a terminal, and use dd, then that won't copy image files either. If I put in a file name, it says "Oi! that's not a directory!"

gerry

Minor bugs:
1) /usr/local/apps/Wallpaper/AppRun
Crashes.
Incorrect use a "separator" tag.
Removing this tag solves problem.
(Knows GtkDialog such a tag?)
2)cfdisk.
Fails when starts without parameters. Puppy has "s" prefixed devices whereas
cfdisk tryes to open the "hda" device by default.
3) Monitor gamma correction button in Xorgwizard does nothing. It calls
"tkgamma" tool which is missing in 4.x releases. (Because tcl/tk was dropped).
I make simple replacement which is attached here:

I can confirm the four separate posts in this thread that GParted is indeed broken -- won't create a bootable FAT32 thumbdrive even if my life depends on it....

So, is there a fix for this?

Is there a GParted dot-pet available that will over-write the default / defective program in Puppy 4.31?

I will note, like the four previous posters, that previous Puppy versions did not seem to have this problem -- and add that Quirky 002 does not seem to suffer ill efects in this area, either.

Any help would be appreciated.

-Roy

Serious bug:
I try to explain, step-by-step:
1)When original Puppy boots it (mean initrd's init) copies some own gzipped modules to new mounted root filesystem.
2)rc.sysinit checks directory /lib/modules/$KERNELVER/initrd. If it exists,
gunzips all modules placed there. At next step, if modules.dep file not exists or
has no "initrd" entries, runs depmod. After this operation, we have ungzipped (*.ko)
modules in /lib/modules/$KERNELVER/initrd/ and links to them in modules.dep
file
3)When Remasterpup creates new squashfs filesystem, it excludes /lib/modules/$KERNELVER/initrd/ but leaves modules.dep unchanged (!)
4) When remastered distro boots, it copies some own gzipped modules to new mounted root filesystem.
5)rc.sysinit checks directory /lib/modules/$KERNELVER/initrd/ , checks string
"initrd" in modules.dep file and because such a string is present, does nothing
(!)
6)In result, we have gzipped (*.gz) modules and strings in modules.dep file, linked
to ungzipped (*ko) modules. System can't "modprobe" them.

I was searching through this thread for a couple of bugs that I ran into after remastering the 4.3.1 CD. One bug was the problem with the modules from the initrd not being gunzipped, which I see that PANZERKOPF already reported on December 25th, and filed a better report than I probably would have, so I don't need to report that.

But I'll post my work-around:

I mounted pup-431.sfs from the remastered CD, and copied the filesystem to a new directory, /tmp/sfs/. Then I removed lib/modules/2.6.30.5/modules.* from /tmp/sfs, ran mksquashfs on it, ran isomaster on the remastered iso image, removing its pup-431.sfs and replacing it with the new one, then burnt a new CD with burn iso2cd. (More detailed steps available on request.)

It is probably enough to just remove modules.dep, but since (I think) modules.alias and modules.symbols are rebuilt at the same time that modules.dep is, I removed all three files.

Probably this could also be done by modifying pup-431.sfs before completing the remastering process by replacing the pup-431.sfs while it is still in the puppylivecdbuild directory, but I've not tried that.

Norm Pierce

Because /usr/sbin/remasterpup2 includes the files under /usr/lib/locale/ in the new CDR, /usr/bin/locale prints a locale. Because this matches the value given for LANG in /etc/profile, /etc/rc.d/rc.country never calls /usr/sbin/chooselocale or /usr/sbin/timezone-set, so the user has to set these manually later or, perhaps, learn a new language and move to the US west coast.

This can be worked-around similarly to the work-around mentioned in my previous post, except that the subdirectory under /tmp/sfs/usr/lib/locale/ (e.g., en_US) is removed.

Again, that could probably be done before completing the remastering process by replacing the pup-431.sfs while it is still in the puppylivecdbuild directory. Or, possibly, the subdirectory could be temporarily removed from /usr/lib/locale before even starting the remastering process, then restored later. I've tried neither of those two ideas, so can't vouch for them.

Norm Pierce

Also "chooselocale" dialog sometimes makes UTF locales even corresponding checkbox is unchecked.
Another theme:
I was changed (for my own needs) rc.sysinit script for solving some problems.
Added some procedures:
1) Checking root filesystem when system is installed in full mode.
2) Cleaning temporary files and other garbage.
3) Solving ungzip/depmod problem reported above. (Not perfect but better than nothing because (IMHO) coping modules from initrd to rootfs is bad idea, all modules should be placed in compressed filesystem.
Use at own risk. l labeled my changes as much as possible.

This afternoon my system reset - not sure how or why. It booted into Slax, as I'm using the Ultilex 5.0.0 LiveCD, which boots to Slax by default if no user input.

When I rebooted Puppy, and opened Seamonkey, it went to the SM home page instead of my set preference. When I tried to access my mail, Seamonkey's password manager failed to provide the password. When I looked in the password file, it was wiped clean...

I checked the Bookmarks, and they were still there. I then tried rebooting. On the first reboot, I got the same situation as the previous time. But on the second reboot, things got worse...


Now Seamonkey popped up the User Profile Manager window, showing the only profile - default. And when I clicked to start Seamonkey, I got another popup telling me that:

Seamonkey cannot use the profile "default" because it is in use. Please choose another profile or create a new one.

Several further reboots changed nothing in this situation. So now I had lost acccess to my passwords, bookmarks, and all my Seamonkey settings.

I also tried starting seamonkey from the programs menu instead of the desktop icon, and noticed that there are now three Seamonkey browser tabs there, one with an icon, and two without. But they all produce the same result.

Happily (I thought) I had a 2fs backup from only a week ago, so I tried booting with that, thinking I would only lose a couple of passwords and five or six bookmarks. But that was a vain hope too.

The first sign of trouble was that the Xorg configuration screen opened instead of the desktop. I know I haven't changed anything in my video hardware since this save was made. Next, Xorg failed to accept my choice of resolution, and forced me to tweak the refresh rate repeatedly before I could get a gui. I'd chosen 1200x1024, and had to settle for 1600x1200.

And finally, when the desktop loaded, it was with a background image I'd replaced almost two months ago. Firedog, which I removed the first week of November, was back, and of course, I was missing a lot more passwords and bookmarks than I expected.

While looking among my Flash cards for another backup file, weird things also happend with Pmount. It was showing the flash cards on the USB reader as "Generic Storage Devices" with impressive capacities of 102GB, 196 GB, and 209 GB.

I noticed this after trying to save the password file and the bookmark file from the save to a flash card. The save was executed without error messages, but the files never appeared on the target card.

Remembering earlier problems with MUT, I opened Pmount's preferences window and unchecked the MUT box. After that Pmount showed USB storage normally, and I was able to really save the two files.


And then I had one last idea - what if I installed the Seamonkey 2.0 pet. Maybe I could import the settings, bookmarks, and passwords from Seamonkey 1.1.18 even though I couldn't open it?

Well, it was a nice thought...

Seamonkey 2.0 installed ok, opened on the desktop, asked if I wanted to import stuff from SM 1.x. I said yes, but no cigar.

I still have three Seamonkey tabs in my programs menu, plus the desktop icon. But every one of them opens the same brand spanking new Seamonkey 2.0, with no bookmarks, passwords, or settings other than the defaults.

arghhhhh

In Windows, plugging in my flash drive automounts and turns on the flash drive light. Unmounting it turns off the light.

In Puppy, using the same flash drive, the light turns on when the drive is plugged in, and stays on no matter what the mount status is.

I don't like the Windows automount, but I do very much like the flash drive light reflecting the mount status of the drive. It gives a positive indication ON the drive and may save someone from pulling the wrong drive if they have more than one plugged in. I also like the "beep" you get when unmounting in Windows.

Today I tried some DVD-RAM disks (3x Maxell and 5x Verbatim) written under Windows98SE in dla format (disk-like access?) on Windows 7.

I hadn't been able to access these disks since the motherboard on my Win98 system died a year ago (the replacement MB stubbornly refuses to detect any dvd burners, and I haven't been able to figure out how to enable dla support in Win2k)

Windows 7 (on a new Toshiba laptop) read them w/o problems. I then tried a blank Verbatim 5xDVD-RAM disk and a blank Sony DVD-RW. Win7 asked whether I wanted them formatted USB-flash style.

I'd never been able to mount any DVD-RAM disks in Puppy, but thought I'd try with Slax 6.1.2, bundled in the Ultilex 5.0.0 LiveCD, and was pleasantly surprised to find that SLAX will read and write to UDF formatted (its icon for the disks carries the text "UDF disk") 3X and 5X DVD-RAM and DVD-RW disks formatted by Windows 7.

The only thing it apparently can't do is format them as UDF itself.

Pmount in Puppy 4.3.1, OTOH, shows the disk size, but responds "unable to mount SR0" to every attempt to mount any of these disks.


I reported this shortcoming over a year ago, but gave up on it as there was little expression of interest, and I assumed it was just another Linux deficiency (Knoppix 5.1.1 couldn't read UDF either).

But if SLAX can manage this, why can't Puppy?

(BTW - I'm also puzzled by the fact that Pmount can't detect my USB-connected burner while Pburn 3.1.6 can - SLAX will automount, offer to format, read, and write to the same disk on the USB burner as readily as on the IDE cable)

mtpaint wont save from a screen capture in 431. Can someone confirm?

====================
EDIT 1st March 2010
Sorry. mtpaint does save. The lack of labelling of the 2 entry fields is confusing.

sorry duplicate post