Lucid Puppy 5.25 Feedback and Bug Reports

[Grandma]

Seamonkey choak and croak running youtube stream - Lupu 5.25
========================================

- Prior post - this thread - similar event with FF3.6 and 4.0

- Platform Lupu 5.25 on Dell Latitude - 2Ghz CPU, 256K Ram, HD 10GB free space, formatted Fat 32

- Run time until seizure: about an hour left alone to play about 20 3 - 4 minute music videos - http://wingsofaaliyah.com - click ENTER - music starts streaming on following page at top. It might be noted that FF4.0 wouldn't even run the javascript that rotates the videos and died almost instantly, whereas FF 3.6 and Seamonkey both survived for a period before croak, with Seamonkey lasting the longest, however Puppy's internal browser doesn't have any problems with the page/stream

- Other aps running: 1 terminal screen looping through a "ping -c1 google.com to verify internet is working. 3 tabs were open - 2 idling at this site.

- A likely cause: insufficient ram for later versions of FF or SM

- Likely solutions (my guess):

a) I could downgrade to prior versions perhaps - but since Lupu is not the likely cause, not sure if that would help since I'd still be running these versions of browsers

b) Increase RAM - naturally a likely method that MIGHT resolve this

c) Switch browsers - as soon as Puppy's browser has cut and paste and a home page this is what I'll probably do.

Am using it now to type this post. Lupu did not croak and gave msg "Window is not responding" and let me kill browser.

Conclusion: much as I love Firefox more than IE, this browser, and another open source made by Internet Surfboard that looks similar to this are both more robust players than the Moz family line.

...my prayers go out to the feature makers of Puppy's browser - that would solve a lot of issues though many might balk at giving up their favorite ap(s) - but it would give the Puppy family a nice benchmark to compare browser performance. Puppy browser has never burped.

[2byte]

Hi Grandma,

I've enjoyed your threads at TC and here. You are helping me to learn a few things too. Firefox...
Are you running Puppy with a save file (frugal) or did you install it to a linux partition?
If it's a frugal, how big is the save file?
Do you have a swap partition or swap file loaded? (256 MB ram).

The reasons for these questions: If a frugal install using a save file, firefox could very well be filling it up with downloaded media streams, causing the crash. Again if a frugal install you could open up two rox filers and drag /root/.mozilla (hidden) to someplace on a mounted linux partition (/mnt/home probably), then drag the folder back to /root/ and choose symlink. That will take it out of the save file and put it on the HD. Then FF won't be filling up the save file when it runs.

Memory will be a problem if you don't have a swap enabled. FF with flash can be a hog, not to mention if you ever use Adobe reader with it.
~

[bigpup]

Grandma,

If your ram is only 256MB, that is the minimum limit for Lucid Puppy 5.2.5
It will run, but it will not be hard to run out of memory.
If you are not using fat32 format, on the partition, for another operating system.
I would do this:
Boot with the Puppy live CD.
At the boot screen click F2 key
Use the boot option puppy pfix=ram
This keeps the hard drive from mounting.
Repartition the hard drive with the program Gparted.
Make two partitions.
One partition formatted ext3 to install Puppy on. (make sure you flag as boot)
one partition 1GB or 2GB in size formatted as a Linux Swap partition.
Do a full install of Puppy.
Re run the boot loader config. I would use Grub4dos bootloader config.
Remove the live Puppy CD.
Reboot.
Because you where running with the live CD it will ask about making save file. Answer no.

This will give you the room you need and the added help, to ram, of a swap partition.

Good guide for setting up a wireless connection.
http://www.murga-linux.com/puppy/viewtopic.php?t=65346

[bigpup]

Booting a computer from USB drive, with no internal hard drive installed, was failing to boot.

Barry has found a fix
http://bkhome.org/blog/?viewDetailed=02272

[PaulBx1]

I am anxious for this fix to be implemented in lupu525.

One other possible thing for grandma: Puppy has in the past had trouble running out of room in /tmp, which is normally in ram. If you mount an additional disk volume on /tmp, it takes care of the problem of long videos that crash in the middle.

I wanted to say that my problem #2, reported earlier, has a good work-around (for me) by turning keypad tapping off.

Has anyone confirmed my earlier reports of two items (#1 and #3) that do not survive reboot?

[Grandma]

right now the install was just "dumped" into a folder on the 2nd partition - as W2K owns part-1 with grub4dos controlling the boot, and part-2 is a Fat32 - probably not the best place for many reasons, not the lease of which are Windows viruses that can get into stuff and screw up things that will probably never propagate in Linux, but still mess up binaries and scripts

Once I get the fdisk or cfdisk part of the install fine tuned (on another machine) - it will dump the Puppy boot files into an EXT2 area and establish the swap partition (part 3), as well as a second EXT2/3/4 etc. (part 4), as well as a fat32 (part 5), area allowing a user to move NTFS files from C: (hda1/part1) to hda5/part 5 for times when hda1 is completely off line but docs, xls etc. could be accessed by both systems.

the goal is to wean users off Windows by providing the aps required to keep them productive - and hopefully at some point they say "Gee - haven't booted Windows for 3 months" - and them give em a little "push" - "Want to shrink or remove that NTFS space since you haven't touched it since January?"

I do not believe completely removing Windows / NTFS is really prudent unless they really want to do that, since they may find - as some do - that some "New Utility they always wanted" only runs in Windows. Just because Bill Gates doesn't play nice and read all file systems, doesn't mean Linux can't accommodate - allowing Windows to ride partition 1 (disgusting as it may be), and hog a lot of space for an OS that really won't get much use - but doesn't have to be vaporized completely with the drive sizes available now.

32 GB flash - under $20
100 TB HD - under $100

so leaving Windows in 5GB to squirm around should be ok

Therefore - yes - I thought the swap might be causing FF to choke as well as cause Puppy to run slower than expected.

I did get 2 new users / friends interested in starting their migration - but need to finish this before I go after that...

[PaulBx1]

I think I found my problem #3, the flSynclient settings not surviving over boot, but I don't understand it.

The settings actually do survive, but are not invoked. The settings are saved in the file /root/.flSynclient. Apparently, these are normally restored when X starts up via the file /root/Startup/load-touchpad-settings, in which there is a command "flsynclient -s" that loads from the settings file. But before that there is the line,

Code: Select all

grep -Ewq '(Elantech|Alps|Synaptics)' /proc/bus/input/devices || exit

...which prevents settings from being loaded for, in my case, the "AlpsPS/2 ALPS GlidePoint". Since the settings are not loaded, I am always stuck with the defaults and I have to renew the settings every boot (or run flsynclient -s in a console window, which seems to work properly).

Why would this line be in that startup file? Could it be that touchpad settings are supposed to be handled in my X config instead? But for some reason they are not? I wonder if my X config was set up in my other machine. But I don't think so; I don't believe I ever got it to boot 525, and it doesn't have a touchpad in any case.

[PaulBx1]

OK, I did find the section for ALPS touchpad in xorg.conf, and added

Code: Select all

Option "TouchpadOff" "2"

to it, which works to turn off tapping. Apparently this is how to handle these touchpads. One wishes though, that either
1) some warning had been generated that flSynclient cannot be used to make permanent settings on this touchpad, and that it is done instead in xorg.conf, or
2) take that code out of load-touchpad-settings, and let flSynclient make the settings anyway (after xorg.conf has provided some reasonable default). Since Puppy is supposed to be easy and generally controls settings via gui apps (that is, it is not Arch), maybe the 2nd solution fits better.

Perhaps some other setting in the same place could reduce the extreme sensitivity I encountered, but since I generally prefer to turn tapping off I am satisfied at this point.

Now I just need to figure out why my .mozilla link does not survive reboot.

[jpeps]

Now I just need to figure out why my .mozilla link does not survive reboot.

I've never had any luck with .mozilla links. Best alternative has been making sure it doesn't grow by comparing to a backup copy. (I've never had any luck with touch pads either....)

[PaulBx1]

This worked fine in Lupu 511.

On that bug about not booting if a hard drive is not in the system, I found a recipe for editing init here:
http://www.murga-linux.com/puppy/viewto ... 82&t=53019
With my meager coding skills I looked in there and found a line something like this (working from memory):
[ ATADRIVES = " " ] && ATADRIVES = ""
I made another copy of that except changing " " to "\n"
Well, it didn't work. Don't know if the code is wrong, or if I interpreted Barry wrong, or what.

[jpeps]

This worked fine in Lupu 511.

On that bug about not booting if a hard drive is not in the system, I found a recipe for editing init here:
http://www.murga-linux.com/puppy/viewto ... 82&t=53019
With my meager coding skills I looked in there and found a line something like this (working from memory):
[ ATADRIVES = " " ] && ATADRIVES = ""
I made another copy of that except changing " " to "\n"
Well, it didn't work. Don't know if the code is wrong, or if I interpreted Barry wrong, or what.

"\n" is newline, so don't know why that would work. The code just eliminates a blank space if there isn't any drive listed.

[PaulBx1]

Well, I thought that original code was changing a variable with only a space into a null. My added code was intented to change the variable with only a line-feed into a null. Barry said the problem occurred because the list, which is blank if there is no hard drive, still contains just a line feed in it; if it was completely empty the code should work. At least that's what I thought he said.

[jpeps]

Well, I thought that original code was changing a variable with only a space into a null. My added code was intented to change the variable with only a line-feed into a null. Barry said the problem occurred because the list, which is blank if there is no hard drive, still contains just a line feed in it; if it was completely empty the code should work. At least that's what I thought he said.

I don't think [ X="\n" ] tests for a newline, but a literal value "\n"

Code: Select all

y="hello\n"
 echo $y

To remove a newline, you could first translate to a space:

A=`echo $X | tr "\n" " "`

edit: The newline shouldn't, however affect the test to begin with (so the original code works)

Code: Select all

A=`printf " \n"`
[ "$A" = " " ] && echo "hello"

[ICPUG]

PaulBx1

Barry never responded to my comment on the blog (after yours).

I guess we could now download the latest wary 5.1.1.56 and look how he coded the change.

Let us know if you find out!

[01micko]

ICPUG, PaulBx1...here's the code block direct from woof, starts at line 578, I'll attach the whole file.

Code: Select all

[ "$ATADRIVES" = " " ] && ATADRIVES=""
[ "$ATAOPTICALDRIVES" = " " ] && ATAOPTICALDRIVES=""
ATADRIVES0="`echo -n "$ATADRIVES" | tr ' ' '\n'`"
touch /tmp/ATADRIVES0 #110509 fix for LF only messes up grep later.
[ "$ATADRIVES" != "" ] && echo "$ATADRIVES0" > /tmp/ATADRIVES0 #110205 broken busybox grep -F. 110509
ATAOPTICALDRIVES0="`echo -n "$ATAOPTICALDRIVES" | tr ' ' '\n'`"

#filesystems...
modprobe nls_cp437     #needed by windows filesystems.
modprobe nls_iso8859-1 #needed by linux filesystems.
#modprobe nls_utf8 #w007
#modprobe nls_cp850 #w468 now default for fat f.s. (2.6.29.2 kernel)
modprobe $LAYERFS #unionfs or aufs.
modprobe fuse #for ntfs-3g driver.

#110425 get drive info before usb drives become available...
ALLDRVS="`ls -1 /sys/block | grep -E '^scd|^sd|^mmc|^sr' | tr '\n' ' '`" #110204
[ -e /proc/ide ] && ALLDRVS="${ALLDRVS}`ls -1 /proc/ide | grep '^hd' | tr '\n' ' '`" #110204 ex: sda sdb hda sr0
ALLDRVS0="`echo -n "$ALLDRVS" | tr -s ' ' | tr ' ' '\n'`" #110205
echo "$ALLDRVS0" > /tmp/ALLDRVS0 #110205 warning, busybox grep -F option different from full grep.
PCPARTSALL="`probepart_init -k`"
PCPARTS0="`echo "$PCPARTSALL" | grep '^/dev/' | cut -f 1-2 -d '|'  | grep -E 'iso9660|ext2|ext3|ext4|reiserfs|msdos|vfat|minix|ntfs' | sed -e 's%/dev/%%'`" #ex: sda1|vfat
LESSPARTS0="$PCPARTS0"
OPTICALDRIVES0="`echo "$PCPARTS0" | grep 'iso9660' | cut -f 1 -d '|'`" #ex: sr0
#warning, busybox grep -F option is different from full grep, have to write this to a file...
touch /tmp/OPTICALDRIVES0 #110509 fix for using with grep.
[ "$OPTICALDRIVES0" != "" ] && echo "$OPTICALDRIVES0" > /tmp/OPTICALDRIVES0 #110509 fix
echo "ALLDRVS=$ALLDRVS " >> /tmp/puppy-file-search.log #for debugging.

Looks like it was a busybox grep issue.

[PaulBx1]

01micko, just for grins I tried plugging that new init into my initrd.gz, but no go (still says "lupu_525.sfs not found"). So the fix is going to be more complicated than that. Probably wouldn't have worked in any case, as the init script is a whole lot different than lupu's. Oh, well...

When it dumped me at the prompt I tried looking at the appropriate variables (ATADRIVES, etc.) but they were all empty. Perhaps that is a result of exiting the script.

[Béèm]

It is frustrating that the network wizard (not the SNS) indicates a successful network configuration, while the ip obtained is a APIPA.

[Chriske]

I'm a newbie using 5.2.5 from a USB drive on a Dell Inspiron 531 and Openbox.

There seems to be a bug around the unhide feature for the Fbpanel/tray.

Autohide itself is successful. The problem is that whenever the cursor hovers near to the tray (about double its height it seems) then the tray pops-up and obscures the bottom part of whatever maximised window currently has focus.

That's an issue if the user wants to click anything in that area - say, the statusbar in Firefox. So, ideally the unhide should only kick in when the mouse hovers at the very edge of the bottom of the display.

[Bill_Gates]

OK, I did find the section for ALPS touchpad in xorg.conf, and added

Code: Select all

Option "TouchpadOff" "2"

to it, which works to turn off tapping. Apparently this is how to handle these touchpads. One wishes though, that either
1) some warning had been generated that flSynclient cannot be used to make permanent settings on this touchpad, and that it is done instead in xorg.conf, or
2) take that code out of load-touchpad-settings, and let flSynclient make the settings anyway (after xorg.conf has provided some reasonable default). Since Puppy is supposed to be easy and generally controls settings via gui apps (that is, it is not Arch), maybe the 2nd solution fits better.

I agree with PaulBx1 and don't understand why flSynclient doesn't work properly in 525 (save settings such as Tap off) when it seems to work fine in 511???

[PaulBx1]

See this post about ISOLINUX:
http://www.murga-linux.com/puppy/viewto ... 662#528662

I ran into that one as well.